REEFS AND CARBONATE PLATFORMS IN THE PACIFIC AND INDIAN OCEANS
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
SPECIAL PUBLICATION NUMBER 25 OF THE INTERNATIONAL ASSOCIATION OF SEDIMENTOLOGISTS
Reefs and Carbonate Platforms in the Pacific and Indian Oceans EDITED BY G. F. CAMOIN AND P. 1. DAVIES
b
Blackwell Science
© 1998 International Association of Sedimentologists published by Blackwell Science Ltd Editorial Offices: Osney Mead, Oxford OX2 OEL 25 John Street, London WC IN 2BL 23 Ainslie Place, Edinburgh EH3 6AJ 350 Main Street, Malden MA 02148 5018, USA 54 University Street, Carlton Victoria 3053, Australia 10, rue Casimir Delavigne 75006 Paris, France Other Editorial Offices: Blackwell Wissenschafts-Verlag GmbH Kurftirstendamm 57 I0707 Berlin, Germany Blackwell Science KK MG Kodenmacho Building 7-10 Kodenmacho Nihombashi Chuo-ku, Tokyo 104, Japan The rights of the Authors to be identified as the Authors of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the copyright owner. First published 1998 Set by Semantic Graphics, Singapore Printed and bound in Great Britain at the Alden Press Ltd, Oxford and Northampton The Blackwell Science logo is a trade mark of Blackwell Science Ltd, registered at the United Kingdom Trade Marks Registry
DISTRIBUTORS
Marston Book Services Ltd PO Box 269 Abingdon, Oxon OX14 4YN (Orders: Tel: 01235 465500 Fax: 01235 465555) USA Blackwell Science, Inc. Commerce Place 350 Main Street Malden, MA 02148 5018 (Orders: Tel: 800 759 6102 781 388 8250 Fax: 781 388 8255) Canada Login Brothers Book Company 324 Saulteaux Crescent Winnipeg, Manitoba R3J 3T2 (Orders: Tel: 204 224-4068) Australia Blackwell Science Pty Ltd 54 University Street Carlton, Victoria 3053 (Orders: Tel: 3 9347 0300 Fax: 3 9347 5001) A catalogue record for this title is available from the British Library ISBN 0-632-04778-X Library of Congress Cataloging-in-publication Data Reefs and carbonate platforms in the Pacific and Indian oceans I edited by G.F. Camoin and P.J. Davies. p. em. -(Special publication number 25 of the International Association of Sedimentologists) Includes bibliographical references and index. ISBN 0-632-04778-X l. Coral reefs and islands-Pacific Ocean. 2. Coral reefs and islands-Indian Ocean. 3. Rocks, Carbonate-Pacific Ocean. 4. Rocks, Carbonate-Indian Ocean. I. Camoin, G.F. (Gilbert F.) II. Davies, P.J, III. Series: Special publication ... of the International Association of Sedimentologists; no. 25. QE565.R426 1996 97-28586 551.42'4'09164-dc2l CIP
Contents
vm
Preface
G. F. Camoin & P. J Davies
Processes Operating 3
Exposure, drowning and sequence boundaries on carbonate platforms
W Schlager 23
The origin of the Great Barrier Reef-the impact of Leg 133 drilling
P. J. Davies & F. M. Peerdeman 39
Development and demise of mid-oceanic carbonate platforms, Wodejebato Guyot (NW Pacific)
G. F. Camoin, A. Arnaud- Vanneau, D. D. Bergersen, P. Enos & Ph. Ehren 69
Stable tropics not so stable: climatically driven extinctions of reef-associated molluscan assemblages (Red Sea and western Indian Ocean; last interglaciation to present)
M Taviani 77
Sedimentary cycles in carbonate platform facies: Fourier analysis of geophysical . logs from ODP Sites 865 and 866
P. Cooper
Platform Case Histories 95
Aptian-Albian eustatic sea-levels
U Rohl & J G. Ogg 137
Origin of white sucrosic dolomite within shallow-water limestones, ODP Hole 866A, Resolution Guyot, Mid-Pacific Mountains: strontium isotopic evidence for the role of sea water in dolomitization
P. G. Flood 145
Computer simulation of a Cainozoic carbonate platform, Marion Plateau, north-east Australia
K. Liu, C. J Pigram, L. Paterson & C. G. St C. Kendall v
Contents
VI
163
Quaternary and Tertiary subtropical carbonate platform development on the continental margin of southern Queensland, Australia
J F. Marshall, Y. Tsuji, H Matsuda, P. J. Davies, Y. lryu, N Honda & Y. Satoh 197
Pleistocene reef complex deposits in the Central Ryukyus, south-western Japan
Y. Iryu, T. Nakamori & T. Yamada
Oceanic Reef Case Histories Atolls and Volcanic Islands 219
Morphology and sediments of the fore-slopes of Mayotte, Comoro Islands: direct observations from a submersible
W -Ch. Dullo, G. F. Camoin, D. Blomeier, M. Colonna, A. Eisenhauer, G. Faure,
J Casanova & B. A. Thomassin 237
Tectonic and monsoonal controls on coral atolls in the South China Sea
Wang Guozhong 249
Steady-state interstitial circulations in an idealized atoll reef and tidal transients in a deep borehole by computer simulation
A.-M Leclerc, D. Broc, Ph. Jean-Baptiste & J Rancher
Active Margins 261
Environmental and tectonic influence on growth and internal structure of a fringing reef at Tasmaloum (SW Espiritu Santo, New Hebrides island arc, SW Pacific)
G. Cabioch, F. W Taylor, J Recy, R. Lawrence Edwards, S. C. Gray, G. Faure, G. S. Burr & T. Correge
Passive Margins 281
Lagoonal sedimentation and reef development on Heron Reef, southern Great Barrier Reef Province
B. T. Smith, E. Frankel & J S. Jell 295
Terrigenous sediment accumulation as a regional control on the distribution of reef carbonates
K J Woolfe & P. Larcombe
Contents 3 11
Comparison between subtropical and temperate carbonate elemental composition: examples from the Great Barrier Reef, Shark Bay, Tasmania (Australia) and the Persian Gulf (United Arab Emirates)
C. P. Rao, Z. Z. Amini & J. Ferguson 325
Index Colour plates facing p. 88, p. 160 and p. 304
VII
Preface
The Pacific and Indian Oceans, with their complex
journey and participated in what both promised
and diverse tectonic histories, are naturally fertile
and turned out to be a very lively meeting. To all
ground for the study of carbonate platforms, of
those who participated in the Sydney meeting, we
fering different perspectives in age, scale, architec
extend our thanks and appreciation. We appreci
ture, global position and product from their better
ated your promptness, understanding and good
known Atlantic counterparts.
humour in an increasingly busy daily schedule.
Throughout the late 1980s and early 1990s, there
This Special Publication has been a labour of
fore, platforms of many different types in these
love, in which many have played a part in the
oceans were the subject of important studies by the
consummation: first, the many authors who turned
Ocean Drilling Program and by national organiza
presentations into publications, and second, but
tions such as the Australian Bureau of Mineral
almost as important, our many colleagues who have
Resources and the Japan National Oil Corporation.
acted as referees for the papers in this volume. An
It therefore seemed entirely appropriate to hold a
often thankless, but essential job,
meeting, not to compare the Pacific and Indian
would have been still-born without the help of so
this volume
Ocean platforms with others, but to define their
many, and particularly we thank the following: A.
diverse characteristics and growth response, with
Arnaud, H. Arnaud, M. Aurell, A. Bosellini, T.
the clear objective of expanding the spectrum of
Brachert, C. J. R. Braithwaite, A. Droxler, W.-Ch.
platform types useful as geological analogues. That
Dullo, A. Eisenhauer, D. Feary, R.N. Ginsburg, M.
meeting was held at the University of Sydney,
Grammer, D. Hopley, H. Kayanne, C. Kendall, I. G.
Australia, in July 199 5, and this Special Publication
Macintyre, J. Marshall, J.-P. Masse, L. F. Montag
defines many of the deliberations presented at that
giani, T. M. Quinn, F. Rougerie, W. Schlager, P. K.
meeting. This book is therefore the first to examine
Swart, B. A. Thomassin and J. Wise. Finally, we
the carbonate platforms of two oceans which offer
wish to thank Michael Talbot, Andre Strasser and
such a wide diversity of tectonic and climatological
the staff at lAS for encouragement and help in the
variables, so relevant to platform development.
editing and publication of this volume.
The organization of the Sydney meeting would have been more difficult without the help and
G. F. CAMOIN
sponsorship provided by the Earth Resources Foun
CEREGE, Aix-en-Provence, France
dation of the University of Sydney and the Petro
P.J.
leum Exploration Society of Australia. However,
DAVIES
University of Sydney, Australia
that help and sponsorship would have been fruitless without the support of those who made the long
viii
Processes Operating
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
Spec. Pubis int. Ass. Sediment. ( 1 998) 25, 3-21
Exposure, drowning and sequence boundaries on carbonate platforms W. S C H LA G E R Vrije Un iversiteit !Eart hScien c es, De Boelelaan 1085, 1081 HV Amsterdam, The Net herlands
ABSTRACT Events that reduce or terminate shoa1water carbonate production have a pronounced effect on the anatomy of reefs and carbonate platforms. Both exposure and drowning may cause such disturbances and thus generate bounding surfaces for sequence stratigraphy. The most common scenarios are: (i) exposure followed by shallow flooding, i.e. restoration of shallow-water conditions; (ii) exposure followed by drowning, and (iii) drowning without prior exposure. Sequence boundaries generated by scenarios (i) and (ii) fit into the standard systems-tract model of sequence stratigraphy; scenario (iii) does not, because it implies that a highstand systems tract is overlain by a transgressive tract withqut intervening exposure. As a rule, this contact is unconformable because it represents a profound change in sediment input and dispersal, and because drowning is often followed by extensive submarine erosion as the sharp topography of the drowned platform amplifies ocean currents. A growing number of case studies show that drowning without preceding exposure has produced distinct unconformities that can be used as regionally or even globally correlatable markers. They should be accepted as sequence boundaries.
INTRODUCTION platform top. This production is prolific but it is easily disturbed or terminated by environmental change such as sea-level fluctuations. Sea-level fall and exposure, for instance, will immediately termi nate carbonate production and convert the former carbonate factory into a site of carbonate destruc tion. Sea-level rise and deep flooding can submerge the platform top so deeply that production is re duced or terminated. Since Darwin ( 1842), the term 'drowning' is in use for deep flooding of reefs and carbonate platforms. Modern sedimentology distin guishes between 'incipient drowning', that is flood ing to less-than-optimal conditions yet still in the photic zone, and 'complete drowning' where the platform top becomes submerged below the photic zone and benthic carbonate production ceases for all practical purposes (Kendall & Schlager, 198 1; Read, 1982). Both exposure and drowning entail drastic changes in facies and lithology as well as in sediment input and dispersal. This paper assesses the relative importance of· exposure events and drowning events in controlling the anatomy of reefs and carbonate platforms. It
It is well known that either exposure or flooding with concomitant sediment starvation may generate signifi cant events and bounding surfaces in the stratigraphic record. Opinions differ on which of these contrasting processes dominates the record. Sequence stratigraphy assumes that exposure during the fall of sea-level produces the most pronounced events and consequently sequence boundaries were set at these junctures (Vail et a!., 1977; Posamentier & Vail, 1988;Sarg, 1988a). Flooding events such as transgressive surfaces and maximum flooding sur faces are assumed to be important but clearly less so than lowstand events. Genetic stratigraphy takes the opposite view. It assigns the dominant role to the flooding phase and considers lowstand exposure a less important intermission (Frazier, 1974; Gallo way, 1989a,b; Meckel & Galloway, 1996). The debate still continues and illustrates that the answer is not obvious and that the situation may vary from place to place. For carbonate platforms, this discussion takes on an added dimension because the sediment is pro duced within the depositional environment, at the
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
3
4
W Schlager
relies on observations of seismic data, boreholes and large outcrops. The seismic expression of these events is an important part of this discussion. Furthermore, seismic images commonly show the entire platform edifi c e; outcrops rarely do. On the other hand, seismic images also contain pitfalls, such as the difficulty in distinguishing between gen uine unconformity and rapid lateral facies change (Schlager, 1992; Stafleu & Schlager, 1995) and the suppression of geologically important events if they are not endowed with sufficient impedance contrast. Outcrops and seismic models of outcrops can com pensate for these shortcomings of seismic data and serve as calibration points for seismic interpretation.
Production (%of maximum)
0 ' \
0
50
100
_ Inter-and supratidal
-!'--.;;:;� ;, ��--....---..:... - Meanlow water
E
� Zone of light .
0..
saturation
..<:
., -o
� "' E ·;;; 2
50
a. a.
�
100
-- Approximate base of photic zone
in clear ocean waters
SEDIMENTOLOGICAL CONSIDERATIONS Sediment for the growth of carbonate platforms is produced within the photic zone on the flat top of the platform. Production rates depend on a variety of biological and physico-chemical factors,and vary considerably in time and space. Work on carbonate production has gone far enough to reveal certain basic characteristics of the carbonate production function. Figure 1 shows a tropical carbonate pro duction function in relation to sea-level and the photic zone. The zone of optimum production is narrow and shallow-for modern scleractinian cor als it includes the top 10-20 m of the water column; for other organisms it may be somewhat deeper. It is clear, however,that production decreases rapidly above the low-tide mark and goes to zero in the upper part of the supratidal zone. Carbonate pro duction in the terrestrial environment of the tropi cal and temperate zones is negative, as carbonate dissolution by C02-bearing rain water far exceeds carbonate precipitation there (see Rates of calc rete formation versus karst erosion in Kukal, 1990, pp . 81 & 155; Robbin & Stipp, 1979). The lower limit of optimum production is more gradual. However,at the lower end of the photic zone production is generally very small and the onset of marine dissolution of carbonates characteristically occurs somewhere between 150 and 2000 m water depth at low latitudes. The carbonate production function of Fig. 1 shows that shallow-water carbonate systems are pre cariously perched between Scylla and Charybdis. Ex posure will kill the system almost instantaneously; demise by drowning will be more gradual, yet still
Fig. I. Schematic curve of carbonate production versus water depth, based mainly on corals and green algae (data from Schlager, 1981; Bosscher & Schlager, 1992). It should be noted that maximum production is in the upper few tens of metres, decreasing upward to zero in the upper supratidal zone and to negative values in the terrestrial zone; production also decreases downward to very low values at the limit of the photic zone. The shape of this production function implies that both exposure and submergence to deeper water (drowning) will severely affect carbonate production and thus the anatomy of the platform. Modified after Bosscher & Schlager (1992).
rapid by geological standards. In both instances,ero sion is likely to remove part of the record of change. Erosion associated with exposure is well known under the term 'karst processes'. Erosion and win nowing during platform drowning have received far less attention and merit mention here. The mor phology of many carbonate platforms differs from that of siliciclastic shelves by its raised rim and steep slopes. Reefs or rapidly lithified sand shoals at the platform edge may protect the platform interior against ocean currents. As a platform drowns and becomes submerged to greater depths, waves and currents can sweep across the top and erode what has previously been protected by the platform rim. In this way, the first part of the sediment record of drowning may be destroyed. However,erosion may continue far below the range of surface currents and waves. With their raised rims and steep slopes, drowned platforms often constitute what is known in physical oceanography as 'sharp topographic features'. These are known to intensify currents by
·
Exposure and drown ing of carbonate platforms
topographic trapping of tidal waves and a variety of other effects (e.g. Brink, 1987, 1989; Roden, 1987). On Fieberling Guyot in the North Pacific, current velocities of topographically trapped waves reached maxima of over 40 em s-1 in an ocean setting where the open-ocean velocity of tidal currents is esti mated to be only 1-2 em s-1 (Fig. 2) (Genin et al. 1989; Eriksen, 1991). The observed and calculated magnitudes of these local oceanographic effects are in agreement with the sediment record on top of Pacific guyots (Fig. 3) (Lonsdale et al., 1972; Brink, 1989; Genin etal., 1989). Current intensification by topography provides a plausible alternative to the sea-level fluctuations that have been invoked to ex plain hardgrounds and traction-current deposits in the pelagic caps of drowned platforms (e.g. Seyfried, 1981; Martire, 1992; Zempolich, 1993 ; Clari et al., 1995).
I conclude that exposure or drowning are equally capable of interrupting or terminating carbonate sedimentation and generating an erosional uncon formity. Indeed, in the geological record we find evidence for both processes. Taking into account that either process may act alone or in conjunction with the other, four scenarios for the generation of unconformities in carbonate platforms may be dis tinguished:
5
1 exposure only, with no return to marine sedimen tation; 2 exposure followed by shallow flooding, i.e. return to shallow-water deposition as before the exposure event; 3 exposure followed by drowning (incipient drown ing or complete drowning); or 4 drowning only (incipient drowning or complete drowning). Scenario 1, simple exposure, removes the reef or platform permanently from the sedimentary basin. This option is not very relevant in the context of basin analysis and seismic interpretation, and will not be dealt with here. Scenarios 2 and 3 both fit the standard model of sequence stratigraphy. Scenario 2 represents two highstand tracts, separated by a lowstand with exposure. Scenario 3 may be inter preted as a highstand tract, terminated by exposure during a relative sea-level fall, followed by a trans gressive tract that may ultimately pass into a high stand tract if the system completely recovers. Scenario 4 is not considered by the standard model of sequence stratigraphy. It implies that a highstand tract passes rapidly upward into a transgressive tract. Drastic changes in facies and lithology accompany this transition and make it a prominent marker in outcrop and seismic data. We will evaluate it in de tail after looking at exposure and flooding scenarios for comparison and background information.
50
EVENTS INCLUDING EXPOSURE (SCENARIOS 2 AND 3)
40
'j'
&
"0 �
g_
"'
30
20
10
26
27
28
29
June 1986
Fig. 2. Oscillatory currents on top of Fieberling Guyot, NE Pacific (water depth c. 480 m). Maximum speed exceeds 40 em s-1 in each diurnal cycle. The currents are interpreted as predominantly resonance effect between the diurnal tide (few em s-1) and seamount topography. The intensification of currents by sharp bottom topography on drowned platforms may explain many of the erosional features observed in the deep-water sediment cover of these platforms (see Fig. 3). Reproduced from Genin et al. ( 1 989, Fig. 6), with permission from Elsevier Science Ltd.
The expression of an exposure event in carbonates is karst, calcrete soils and pervasive alteration by freshwater diagenesis. These processes leave a clear petrographic record and there are numerous case studies where major exposure events have been clearly identified and used as sequence boundaries. One example is the Cainozoic of the Great Bahama Bank, where seismic data and core borings were combined to reconstruct the Miocene to Recent his tory of the platform ( Eberli etal., in press). Seismic expression of exposure is good on the platform be cause of the strong contrast of hard karst surfaces and friable marine intervals. On the slope, exposure seems to correlate with periods of submarine erosion and lithification. If confirmed, the extensive erosion and non-deposition of deep-water slopes during lowstands cannot simply be the effect of sea-level lowering . It is well known that at least planktonic
W Schlager
6 (a)
Allison Guyot (pelagics Paleocene-Recent, platform Albian)
5 km
(b)
Limalok Guyot (pelagics Eocene-Recent,
(c) Darwin Guyot (platform Cret.)
platform Paleocene-Eocene)
(d) Blake Plateau (pelagics Late Cret.-Recent, platform Aptian-Albian) W
E
== ==== �========�� 2
�
-
- 3sec
(e)
Campeche Bank (pelagics Late Cretaceous-Recent, platform Early Cretaceous) SE
NW
-2
!Okm
production and concomitant deposition of carbon ate ooze continue during exposure of the platforms. Slope erosion requires an oceanographic amplifier. One possibility is the Gulf Stream, the speed of which may increase as the cross-section of the Florida Straits shrinks during a sea-level fall. Other case studies on exposure unconformities have been complied by Budd et a! . ( 1995); further more, Goldstein et a!. (1990) has provided an inte grated diagenetic approach and Handford (1995) a detailed analysis using sediment geometry and stratigraphy.
- 3sec
Fig. 3. The pelagic cover of
drowned platforms is thin and often absent at the platform margins, attesting to the intensification of currents by the sharp topography. The effect can be observed on isolated guyots (a-c) as well as on drowned platforms attached to land (d,e). Sources: Allison Guyot after Sager, et a/. (1993); Harrie (Limalok) Guyot after Premoli-Silva et a/. (1993); Darwin Guyot after Ladd et a/. (1974); Campeche Bank after Shaub (1983); Blake Plateau after Dillon et a!. (1988).
Despite the many examples of successful use of exposure surfaces as sequence boundaries,there are also problems. The expression of exposure surfaces in seismic data or outcrops is most distinct in young, immature carbonate rocks where the harder, better cemented exposure horizons contrast with the softer, poorly cemented marine intervals (e.g. Beach, 1995). This difference decreases as burial diagenesis progresses and rocks mature to fully cemented calcitic limestones or dolomites. Thus; events combining exposure with shallow flooding are often difficult to see in seismics or from afar in
Exposure and drowning of carbonate pla(forms
outcrops of mature carbonates (see Eberli et a!., 1993, Fig. 10, on a mid-Cretaceous karst in out crop). Bosellini (1989) pointed out that the abun dance of platforms 'grafted' on top of one another may be seriously underestimated because the phe nomenon is so difficult to perceive. The Triassic Schlern Platform in the Southern Alps is a case in point. (Fig. 4) (Bosellini, 1989; Brandner et a!., 1991). Large outcrops of the plat form flank and the adjacent basin fill show that platform growth was interrupted several times by tectono-volcanic events that led to exposure with karst and soil; carbonate deposition was resumed after the disturbances and new generations of plat form were grafted onto the old edific. Despite the excellent ourcrops, the sutures between different platform generations cannot be traced continu ously. This example is typical of many in the area of the Dolomites. The difficulty of spotting exposure followed by shallow flooding is aggravated by another peculiar-
7
ity of the carbonate systems-the strong tendency to put the new highstand margin on top of the old one; in other words, the tendency to put reef on reef and sand shoal on sand shoal. The reason for this is the favourable environment on top of the old margin. The fixed margin also keeps the lagoon and the fore-reef zone anchored in the same position. Be cause facies belts do not shift much at exposure events, differences in sediment dispersal between the two highstands tend to be small and reflection patterns devoid of lap-outs. This unfavourable re flection geometry adds to the unfavourable imped ance configuration and redues the seismic contrast at exposure surfaces in pure carbonates. Exposure and drowning This situation conforms to the standard model of sequence stratigraphy. Shallow-water deposition on the flat platform top (highstand systems tract) is terminated by a relative fall in sea-level that leads to
Fig. 4. Triassic Schlern Platform in the Southern Alps of Italy, an example of exposure and shallow flooding. Platform on the right, basin on the left. Five unconformities (dashed lines marked 'u') could be mapped in this Triassic succession. The unconformity between formations Sdl and Sdll illustrates the difficulty of tracing exposure events within a platform edifice. The Sdl-11 unconformity can be easily identified in the basin where redeposited iron ooids indicate exposure of the hinterland. On the platform, however, the position of the suture remains uncertain, despite intensive mapping efforts and excellent outcrops. From Brandner et a!. (1991 ), reproduced with permission of the authors.
8
W Schlager drowning successions and accentuate the abrupt ness of change. It is important to note that these additional contrasts are contributed by the process of drowning, not by the preceding exposure. Indeed, the seismic expression of exposure and drowning events closely resembles that of simple drowning events devoid of exposure. Two examples may illustrate this point. The Miocene platform of the Marion Plateau shows the typical morphology of a drowned platform with convex top, asymmetric backstepping and well preserved flank deposits. The 7 Myr of exposure that preceded the drowning left only minor vestiges of karst in some seismic records (Fig. 5) (Davies et a!., 1991; Pigram eta!., 1993). A second example is the Miocene E.II platform off north-west Borneo described by Epting (1989). Several exposure and drowning events interrupted platform growth and
exposure of the bank top. The subsequent rise of the sea floods the platform again and either submerges it to less-than-optimal conditions or drowns it completely. The seismic expression is greatly enhanced where drowning follows an exposure event. In these in stances, the relative sea-level rise after exposure is rapid enough to outpace carbonate production and establish deeper water conditions on top of the platform. Pelagic or hemipelagic,often argillaceous, deposits cover the platform. They differ strongly from the underlying platform carbonate both in lithology and acoustic impedance. Furthermore, the change from shallow-water carbonate production to (hemi)pelagic deposition entails a pronounced change in sediment dispersal, commonly visible as changes in bedding geometry and lap-out. Con densed intervals and hiatuses are common in these
(a)
s
N
1200 25i1130
13/JO 257.1140
1400 0
257.1150
1500 °
1600 257.1200
� Q)
0.8
E ::2 Q) >
.�
>-
"'
3: 6 3:
1.0
f-
Fig. 5. Miocene platform on the
l
Site 816
(b)
___ _ _ __
8
N
o ====�=2s=7.�1=o4=o== === 2= 2 o==�====2 =7s .=1=o3 s7=10 == ���- --- ==
o.s�����;;·aT�·rr§��� '"
�
.§
!
,_
"' � I
0
� ,_
1.0
Marion Plateau off eastern Australia with ODP Site 8 1 6. The platform was exposed for c. 7 Myr, then rapidly flooded. Climatic deterioration precluded reinitiation of carbonate production; the platform drowned and was covered by pelagic sediments migrating in from the side. Despite the very long and extensive exposure, the seismic record shows the asymmetric growth morphology and the drowning event of the platform very well (a). Only few lines provide some clues to the exposure event, such as the eroded top left of marker 257.1030 in (b). (Modified after Davies et a!., 199 1 ; Pigram et a!., 1993).
Exposure and drown ing of carbonate platforms
were recognized in cores and also partly on seismic data. The most prominent seismic reflector is the topmost platform surface. It records backstepping and final drowning of the platform. The exposure events that preceded each backstep and the final drowning were recognized in cores but do not betray themselves in the seismic section.
DROWNING WITHOUT EXPOSURE (SCENARIO 4) There is a growing number of examples where reefs and platforms were deeply flooded without first being shut down by exposure. The record is an upward deepening succession that leads from shallow-water deposits to deposits of the lower photic zone (incipient drowning) and commonly on to pelagic deposits of the aphotic zone (complete drowning). In outcrop morphology and seismic expression, these events differ little from exposure and drowning events. In both instances, shallow water deposits are covered by deeper water deposits with different geometry and physical properties. Hiatuses are common and seismic lap-out may occur even where rapid transitions exist in outcrop sections, simply because the transition zone is below seismic resolution. The following sections are descriptions of well-documented examples of drowning events, progressing from young to old. Miocene Liuhua platform, South China Sea (Fig. 6; Erlich eta!., 1990, 1993; Moldovanyi eta!., 1995; Wagner eta!., 1995) Buried by less than 2000 m of homogeneous hemi pelagic mudstones, this early Miocene platform (Zhujiang Limestone) is unusually well docu mented, by high-quality seismic data of up to 80-Hz frequency and several carefully studied core bor ings. The platform displays both exposure and drowning events. A major exposure event, thought to c.orrelate with a eustatic lowstand of -100 m or more, marks the top of the Lower Zhujiang Lime stone. Less prominent exposure surfaces occur within the Upper Zhujiang; their exact number is a matter of debate (Moldovanyi eta!., 1995; Wagner eta!., 1995).
A major drowning event occurred at the base of the Upper Zhujiang, following exposure. The plat form stepped back and became completely reorga nized. Growth of the Upper Zhujiang is marked by
9
repeated drowning and backstepping. The most prominent unconformity of the entire system, and the reason for including this case study here, is the drowning unconformity at the very top. It shows a smooth, high-amplitude reflector with several back steps (Fig. 6). In the boreholes, the final drowning is marked by a rapid transition from shallow-water grainstones to glauconitic packstones and finallly siliciclastic mudstones. The closest exposure surface lies c. 30-40 m deeper in the section and is far less prominent in the seismic data than the drowning unconformity. Similarly, the major exposure surface at the top of the Lower Zhujiang is not prominent seismically except where the Upper Zhu jiang is missing. Oligocene of Kalimantan (Borneo) (Fig. 7; Saller eta!., 1993) Using seismic data, boreholes and river outcrops, four sequences were recognized. Exposure of the shelf can be shown for only one out of five sequence boundaries. (The association of this event with folding and tilting indicates a tectonic cause.) The other sequence boundaries represent rapid deepen ing events with concomitant backstepping of the margins observed in seismic profiles. Saller et a!. (1993) concluded that the stratigraphical turning points that allow one to recognize sequences were flooding and drowning events. This example clearly demonstrates that one can establish viable sequence subdivision, trace it regionally and use it for corre lation even where no major lowstands have been recognized. Cretaceous of southern Pyrenees (Fig. 8; Simo, 1993) The Cretaceous of the southern Pyrenees shows a long history of alternating carbonate and siliciclas tic deposition with both prominent exposure and drowning events. Most sequence boundaries show exposure with soils followed by incipient drowning. However, the top of sequence 5 in Fig. 8 lacks evidence for longer-term exposure and represents drowning only. It serves as a mappable and corre latable marker surface none the less. Furthermore, the first-order trend of sequences 1-5 is one of southward backstepping, probably governed by sub sidence from lithospheric flexure.
10
W Schlager ..
._·�_-_._·_-_ _�-:�- _ :_�_· - �j��i''' �"'===�����l � -:-F= I� � . .
.
·
....
�---- -.. � �--���-
f- -:-:-
�
-
. ..,
...
t��;:.=�t
_____ __ _
:;_:--¥: =}:�7� �����:
.
·=�����_ =�;_ :�":� � .
_
.•
::=:=:::... ··· ·====:: .: ::-:-- ·-- -. ···::..:.:::::....�===··------- - -:: :: --·-·'"·__ ---.. ---··----====--====-::::::::-�:-···-··.:::::=:::··-:. . -:::: --_:: - · :-: · .
--- ----- ..... .---·----
�
-
.
-·- · -
·
.
. ...
__
_______ _ ...
-::::=:: � ===-...:::
,
.
__
.
· ..::::........: .:... :.._., :�:::::.:::::::...:....------------· .
����§i�� ....... ''"""' -�:...
-··
..�_:::---------...::r.;;;r,;.;;;;:-...: :--:::..:.·
.
-
..... ..........
.
.........
(a)
---�·""·. . -·:'::-"- "'"":;.,,.: �.
........
-:; ... .
:
·""' ..
---
·•
SE LH 11-1-4 porosity
dei13c
� -8
+2
LH 11-1-1A porosity
dei13c
� -8
+2
LH 11-1-3 porosity
dell3c
� -8
+2
c::::J zone of freshwater diagenesis, including vadose features (b) Fig. 6. (a) Miocene Liuhua platform in the South China Sea, with borehole L4-l-L The platform backstepped several
times and finally drowned completely when the elevated reef rim on the SW side became submerged. Borehole showed 7-8 m of rapid upward transition from shallow-water coralgal reef to deeper water, glauconitic limestones and finally hemipelagic mudstone. Modified after Erlich et a/. (1990). (b) Distribution of exposure horizons based on examination of rocks and logs, modified after Moldovanyi et a/. (1995). The shaded exposure zones show highly negative carbon isotope ratios ('soil carbon') and low porosity (additional cement). The exact number of exposure events is a matter of debate (see Wagner et a/., 1995, for a different view). There is general agreement that the youngest part of the platform was not exposed and the dominant seismic event in (a) is a drowning unconformity without exposure.
�---------· ---------
11
Exposure and drowning of carbonate platforms
;gEi¥2 !§Ill . �
TST4 (backstepped)
SB 3/4 (drowning unconformity on HST3)
TST3 (empty bucket)
Fig. 7. Oligocene platforms,
Kalimantan (Borneo). Model of growth stages of stratigraphic sequences 2, 3 and 4 recognized by Saller et a/. (1993). Exposure and terrestrial erosion can be demonstrated for the earliest event, sequence boundary 1-2. Subsequent boundaries are drowning events that produce hardgrounds and condensed intervals, followed by backstepping or empty-bucket morphology (raised rim and deep lagoon). The empty lagoon formed during the transgressive phase forces the highstand tracts to prograde in two directions-a common feature of carbonate systems. Despite the lack of exposure, the drowning events were successfully used as correlatable and datable sequence boundaries by Saller et a!. (1993). Bold lines, sequence boundaries; dotted lines, prodelta shales.
SB 2/3 (drowning unconformity onHST)
-----------;::;;.--���.o:::::..:
Early Cretaceous, Helvetic Domain, Swiss Alps (Fig. 9; Funk eta!., 1993; F6llmi et al., 1994) In the Early Cretaceous, the southern continental margin of Europe was intermittently occupied by carbonate platforms, now well exposed in the Hel vetic nappes of Switzerland and western Austria. Building on 150 yr of painstaking field observa tions and stratigraphical analyses, Funk et al. (1993) and F6llmi et al. (1994) added an unusually thorough documentation of facies anatomy and
TST2 (empty bucket)
SB 1/2 (exposure)
stratigraphy, accompanied by sequence-stratigra phical and palaeo-oceanographic interpretation. Platform growth was interrupted by both exposure and drowning events. The significant stratigraphi cal turning points used for subdivision of the Early Cretaceous rocks are five events of incipient or complete drowning. Their onset is often diachro nous, their termination synchronous. F6llmi et a!. ( 1994, p. 738) repeatedly observed highstand systerns tracts overlain by transgressive tracts without intervening exposure or lowstand tracts.
·
12
W Schlager South
North
20km
platform and upper slope
C
nearshore sandstone and platform w. strong clastic admixtures
sequence boundary (drowning only)
-......_ deep shelf and basin (argill. limestone, marl)
0)
sequence boundary (exposure+ drowning) sequence number
Fig. 8. Late Cretaceous, southern Pyrenees. Schematic cross-section of carbonate platforms, concomitant basin facies
'
and sequence boundaries. Most sequence boundaries show exposure of the inner platform and coeval drowning of the platform margin (Simo, 1993, p. 338). The boundary of sequences 5-6 lacks exposure and represents drowning and backstepping only. It should be noted that the long-term trend of platforms 1-5 is one of drowning and backstepping, probably governed by lithospheric flexure. Modified after Simo (1993).
Cretaceous (Comanchean) platforms of southern USA (Tyrrell & Scott, 1989; Scott, 1990a) The Comanchean contains several carbonate plat forms separated by siliciclastics. A variety of com binations of exposure and drowning can be observed in this package. It is listed here because drowning events that put deeper water shales on top of plat forms are particularly prominent in the seismic images of these platforms. One of these boundaries, the top of the Edwards Limestone in Vernon Parish, Louisiana, is a drowning surface used as a sequence boundary by Scott ( 1990a, p. 43) and Tyrrell & Scott ( 1989). Cretaceous Maracaibo platform, Venezuela (Stiteler et al., in press) The Aptian-Albian Cogollo Group consists of shoalwater carbonates that pass laterally into deep water marl and shale. Drowning events repeatedly interrupted and finally terminated platform growth. The drowning events are represented by intercala tions of deep-water marl; the final cover of the platform is the deep-water La Luna Limestone. The drowning surfaces are levels of erosion or non-
deposition and distinct lithological boundaries. Stiteler et al. used a variety of techniques to show that the drowning horizons are the dominant seis mic events in this succession and the natural choice for sequence boundaries. Cretaceous Shuaiba platform (Persian Gulf; Scott; 1990b; Wagner, 1990; Alsharhan & Nairn, 1993; Burchette, 1993, p. 195; Wagner eta!., 1995) The Shuaiba Formation has been well studied in several areas around the Persian Gulf. Its boundary with the overlying Nahr Umr Shale is everywhere a prominent stratigraphic marker but conditions varied. In certain areas, the Shuaiba platform was first exposed and deeply leached, and subsequently drowned and covered by deeper water Nahr Uinr Shale. Elsewhere, the Shuaiba-Nahr Umr contact is an incipient drowning event where no evidence for exposure has been found (Wagner, 1990; Wagner et al., 1995). Scott (1990b) pointed out that the exposure seems to be caused by regional tectonics, whereas the drowning event (which often entails complete drowning) correlates with drowning iri Mexico and the southern USA, indicating a super regional, possibly global, event.
13
Exposure and drown in g of carbonate platforms s
N .t:!.
<
110 (Ma)
phase
115
.�
c.
<
120
125
130
135
140
145
� E.illill � l5:i5l
sboalwater shale & marl
nate,
shoolwater cart>o
incl. slope debris
• IZJ
bemipelagics (siliclastics & carbonates) carbonate platform margin
hiatus
0
lOkm
condensation & reworking phosphate, glauconite, sand
Fig. 9. Early Cretaceous stratigraphy of the Helvetic Domain, Swiss Alps, in a time-distance diagram. The section
shows part of the European continental margin (ocean on the right, i.e. south). Shoalwater carbonates, shoalwater siliciclastics and hemipelagic deposits alternate. Flooding events that generate drowning unconformities are the dominant subdivisions in this succession and are marked by extensive hiatuses generated by submarine erosion and non-deposition. Short hiatuses in the Berriasian-Valanginian probably resulted from exposure. It should be noted that in sequences S7, S8 and S9 the highstand tract with prograding margin is followed immediately by drowning, whereas in S3 highstand progradation is followed by exposure. Drowning hiatuses pass laterally into condensed facies with glauconite, phosphate and hardgrounds. Modified after F6llmi et a/. (1994).
Cretaceous Vercors platform, French Alps (Fig. 1 0; Arnaud-Vanneau & Arnaud, 1990; Jacquin et a!., 199 1; Everts, 1994; Stafleu et a!., 1994; Fouke et a!., 1995; Everts et a!., in press) Tens of kilometres of mountain cliffs show a pro grading and retrograding platform on the scale of a seismic profile. A network of deeply incised valleys provides three-dimensional insight into the anat omy of the platform margin and the interfingering of platform limestones with deeper water marls of the Vocontian Basin. ·The large-scale geometry conforms to the stan dard model of sequence stratigraphy. However, tracing of beds in outcrop, 'fingerprinting' of debris beds by quantitative compositional analysis and detailed diagenetic studies of exposure surfaces
indicate significant deviations from the standard model: what appear in large-scale geometry as basin-restricted wedges are really tongues of basinal marl that interfinger with slope deposits and do not indicate extended lowstands of sea-level ( Everts, 1994; Everts eta!., in press). Exposure events on the platform are scarce and, judging from the minute size of the basin-restricted wedges, of only short duration (Fouke et a!., 199 5). The most significant exposure feature consists of several exposure events that coalesced into one surface. This group of exposure events has only minor expression in the large-scale anatomy of the platform margin (Fig. lOb); the dominant feature in margin anatomy is the alternation of prograding slopes, separated by drowning events that bring open-marine marls up onto the flat platform top. Seismic models demon-
14
W Sc hlager SE
NW
E
8 00
� 6:::::j
800m
platform interior wackest.-packst.
� I:::::::,.,J
platform margin & slope
deeper-water marl,
grainst. - packstone
argillaceous limest.
(a) Archiane (platform-top) Bi &--.1
Fontaine Colombette (flank)
I
�
Rocher
Bi &--2
� Bi &-I
r:,:l----ol!f
..___,_ 80 m
Top Bi5 platform tongue
� 40%
(b)
Open-sea Biota·
0
40% Colds
Fig. 10. Cretaceous platform-basin transition in the Vercors, French Alps. (a) Overview of stratigraphy. Limestones composed of shallow-water material interfinger on various scales with argillaceous limestones and marls of deeper water environments. Both exposure events and incipient drowning events can be demonstrated in the platform section. Large-scale geometries resemble those of the standard model of sequence stratigraphy and suggest extended lowstands with thick basin-restricted wedges of marl. Detailed correlations in the box area (b) show that this impression is false and that stratigraphical turning points in the middle and near the top of the succession are drowning events. Modified after Everts et a/. (in press). (b) Diagenetic studies of proposed sequence boundaries, tracing of beds in outcrop and 'fingerprinting' of beds by quantitative grain composition indicate that: (i) limestones and argillaceous (basin) sediments interfinger down to metre scale, (ii) basin-restricted wedges of lowstands are minute (one of the largest examples is shown as 'wedge with dolomitized clasts') and (iii) the stratigraphic junctures proposed as sequence boundaries represent major incipient drowning events during which the argillaceous sediments encroached onto the platform top. From Everts (1994, Fig. 1.4.9), reproduced with permission from the author and copyright holder.
strate that at standard industry frequencies of 2030 Hz, seismic reflection data show no recognizable difference between the interfingering model of Everts et al. (in press) and the model of onlap and extended lowstands (Jacquin et al., 199 1).
Late Jurassic Haynesville platform, Texas Dravis (1989) carefully studied the diagenetic his tory of the oolitic shoal complexes of the upper Haynesville Formation and the contact to the over lying Bossier Shale. He concluded that the contact
Exposure and drown ing of carbonate platforms
was a drowning event without evidence for sub aerial exposure. Devonian reefs and platforms, Western Canada Basin (Fig. ll; Wendte eta!., 1992) This succession consists of vast and varied basin fill where reefs and platforms played a major role. The growth anatomy reconstructed from cores, logs and seismic data shows that drowning events commonly represent major stratigraphical markers or turning points in stratigraphy. In particular, many isolated build-ups were drowned without prior exposure and the flanks show drowning unconformities as fore reef is onlapped and buried by subsequent progra dation of basin sediments. A particularly well documented example is the Judy Creek reef, the entire growth history of which is segmented by backstepping and incipient drowning; only one cycle boundary has been attributed to sea-level lowering (Wendte & Muir, 1995). Another example of a drowning event is the termination of the Cooking Lake platform (Wendte eta!., 1992, p. 61; see Fig. 11).
DISCUSSION The case studies mentioned above may suffice to demonstrate that the three basic scenarios of growth interruption in carbonates-exposure and shallow flooding, exposure and drowning, and drowning without exposure-are all well documented in the literature. The task that remains here is to compare the expression of exposure events and drowning events in the rocks and in seismic images, and to evaluate the potential of drowning events as se quence boundaries.
Fig. ll. Devonian Cooking Lake platform and Leduc reefs, Western Canada Sedimentary Basin. The top of the Cooking Lake platform is a drowning event with no demonstrable exposure, yet it represents a major stratigraphical turning point where the area of shallow-water deposition was greatly reduced and stratigraphical style changed from large, low-relief platforms to areally limited but high-rising Leduc build-ups. Modified after Wendte et a!. ( 1992).
15
Seismic expression of exposure and drowning events The seismic visibility of exposure events is best in young, diagenetically immature carbonate rocks. In these rocks, the exposure surface may be associated with major impedance contrasts where freshwater cementation and soil formation have produced a hard layer that is sandwiched between friable and diagenetically immature marine deposits. Ansel metti & Eberli (in press) reported variations of sonic velocity from 2000 m s-1 to over 5500 m s-1 related to patchy meteoric cementation. From Ocean Drilling Program (ODP) sites on Mid-Pacific seamounts, Sager et a!. (1993, p. 154 & 234) re ported cyclic variations in velocity between 1600 and 5000 m s-1; the cycles were attributed to epi sodic exposure and meteoric cementation and cal crete formation. This contrast is responsible for the ' good seismic expression of exposure surfaces in the upper part of extant carbonate platforms such as the Great Bahama bank or Enewetak. However, burial diagenesis is likely to reduce or completely elimi nate this contrast. In diagenetically mature lime stone and dolomite, impedance contrasts at exposure surfaces tend to be low because rocks of similar mineralogy and facies occur beneath and above the surface and because the rugged karst surface diffuses seismic energy. The seismic visibility of drowning events is gen erally good because the events juxtapose argilla ceous or chalky pelagic rocks of low impedance and platform rocks of high impedance. The impedance contrast may be enhanced further by extensive submarine cementation (e.g. calcite, phosphate, iron-manganese oxides) during periods of non deposition and erosion. Fundamental changes in the pattern of sediment input add to the seismic
16
W Schlager
visibility of drowning events. The change from shallow water to deeper water pelagic or hemipe lagic deposition goes hand in hand with a funda mental change in sediment supply as the shallow water carbonate source is shut down and pelagic or siliciclastic sources take over. This produces changes in the dip of bedding planes and seismic reflectors. Drowning events as sequence boundaries? Events of incipient and complete drowning have repeatedly been used as sequence boundaries (e.g. Erlich eta! ., 1990; Scott,1990b; Saller eta!., 1993; Simo, l 993;Clarieial. 1995,p. 118). Most of these studies indicate furthermore that there is little difference in the seismic expression of drowning events preceded by exposure and those devoid of exposure. In either case, the dominant effect is the switch from shallow-water to deep-water deposition and the concomitant change in impedance and bedding geometry. I believe that the use of drowning horizons as sequence boundaries agrees with the original defini tion of sequence boundary by Vail et a!. (1977, p. 53): '. . .observable discordances in a given strati graphic section that show evidence of erosion or non-deposition with obvious stratal terminations, but in places they may be traced into less obvious paraconformities recognized by biostratigraphy or other methods'. I suggest we accept this practice rather than insisting on there being an exposure event that precedes it. Carbonate platforms tend to build so close to sea-level that short-term exposure is very common and may not be a significant marker to delimit sequences in the million-year domain. Furthermore, recognition of important exposure events is hampered by the fact that some of the most commonly observed features are not really diagnostic of exposure by sea-level fall. Va dose cement, for one, is not diagnostic of exposure because it may form on storm ridges in the su pratidal zone. Any carbonate platform can build up to this level by depositional processes-no sea-level drop is required. Similarly, selective dissolution and calcite cementation are not diagnostic of meteoric environments. Both have been shown to develop in cold deep ocean water (Schlager & James, 1978) as well as under shallow burial in the marine realm ( Melim et a!., 1995). Only truly terrestrial features, such as subaerial karst, soils or relics of terrestrial plants in situ, may ultimately
serve as diagnostic criteria for relative falls of sea-level. It has been argued that exposure events are superior sequence boundaries because they generate erosional hiatuses whereas drowning events gener ate condensed sections and transitional boundaries even where the seismic image shows a drowning unconformity (Christie-Blick, 1991). I believe that this argument is based on a misunderstanding. Both exposure and drowning events commonly produce hiatuses and unconformable contacts,only the con trolling processes and the location in a shore-to basin profile are different. Exposure of a reef or carbonate platform produces an erosional hiatus that extends from the platform top to the lowstand level; below this level, sedimentation continues, even though the event may be traced further down slope by changes in facies and sediment geometry. Drowning leads first to erosion of formerly pro tected lagoons as the rim goes under. Subsequently, the 'sharp topography' at the platform margin intensifies currents and maintains a regime of at least intermittent erosion that may last for tens of millions of years (Figs 2 and 3). Erosion may extend for tens of kilometres towards the platform interior. On Wodejebato Guyot, the drowned top lay bare for c. 40 Myr; Darwin Guyot, drowned c. 100 Myr ago, remains largely bare to this day. Gaps of up to 40 Myr were also found on drowned Mesozoic platforms of ancient Tethys (e.g. Clari eta!. , 1995). In a seaward direction, the erosional regime at the platform margin may merge with the erosional zone of the slope, where slumping and turbidity currents tend to erode the slope (Schlager & Camber, 1986; Purdy & Bertram, 1993). These erosional slopes commonly extend to abyssal depths and take tens of millions of years to be buried by more gently dipping siliciclastics (Clari eta! . , 1995). Drowning events match exposure events in terms of intensity of erosion and length of hiatuses. Fur thermore, there is a growing number of examples where drowning was not preceded by exposure. Some of them have been described above. These events have been shown to represent excellent seis mic stratigraphical and lithostratigraphical markers that served as sequence boundaries; yet they repre sent a configuration where a highstand systems tract is unconformably overlain by a transgressive tract without intervening lowstand. This situation is not considered by the standard model of systems tracts (Vail, 1987; Posamentier & Vail, 1988), but it is implicit in the original definition of sequence
Exposure and drown in g of carbonate platforms
boundary by Mitchum in Vail eta!. (1977). Sequence stratigraphy originally assumed that both exposure surfaces and condensed intervals represent lap-out horizons that can serve as se quence boundaries (Vail eta!., 1977, p. 66ft'). Sub sequently, downlap onto condensed intervals was excluded because it commonly is a transitional boundary that runs slightly oblique to the time lines (Vail eta!. , 1984; Cross & Lessenger, 1988) . How ever, this argument does not hold for drowning unconformities and their extended erosional hia tuses. The main argument against drowning unconfor mities as sequence boundaries is that they need not be related to subaerial exposure and a fall in relative sea-level. Schlager (1991) argued against restricting the term 'unconformity' to surfaces with 'evidence of subaerial erosional truncation . . . or submarine exposure' (Van Wagoner eta!., 1988) . The recurring use of drowning unconformities as sequence bound aries illustrates the advantages of the earlier, broader definition of sequence boundary by Mitchum in Vail eta!. (1977,p. 211): 'a surface of erosion or non-deposition that separates younger strata from older rocks and represents a significant hiatus'. Sedimentologically, this definition implies that the sequence boundary represents a 'geometri cally manifest change in the pattern of sediment input and dispersal' (Schlager, 1991). Drowning events as global markers The use of drowning events as sequence boundaries may enhance rather than reduce the role of se quence stratigraphy in global correlation. A number of drowning events have been shown to affect reefs and platforms all over the world (e.g . Schlager, 1981). However, global distribution does not neces sarily imply that the drowning is caused by eustasy. It may also be related to a sharp decline in carbon ate production as a result of environmental stress (Schlager, 1991; Pratt & Smewing, 199 3) . The coupling of platform drowning with oceanic anoxic events is a case in point. The correlation has been conjectured by various workers with different data sets (Arthur & Schlanger, 1979; Schlager, 1981, 1991; Scott, 1990a,b). Because deep-ocean circula tion connects all ocean basins, conditions such as oxygen deficiency are likely to be felt world-wide, albeit not everywhere with the same intensity. The distribution of oceanic anoxic events confirms this view.
17
Global extent and (near) synchroneity are neces sary conditions for eustasy but they are not suffi cient to prove a eustatic cause for stratigraphical events. Certain changes in chemical cycling, mass extinctions and other reorganizations in the ocean a lso meet this criterion . These changes may produce effects that are just as widespread as the rise and fall of the sea-level. Seismic evidence for exposure before drowning In the discussion on exposure and drowning, re searchers have repeatedly underlined the impor tance of examining the rocks rather than relying on seismic data and wireline logs (e.g. Dravis, 1989; Schlager, 1989, p. 23; Handford, 1995, p. 3 36). However, both of these remote sensing techniques complement rock studies in important ways and very commonly represent the only information at hand . Therefore, the quest for criteria to recognize exposure and lowstands on carbonate platforms from seismic data is legitimate and necessary. The most reliable diagnostic feature for exposure is a lowstand shelf margin grafted sideways onto a highstand slope . Sarg (1988b), Eberli & Ginsburg (1989) and Kennard et a!. (1992) have successfully applied this technique. The presence and shape of gravity flow deposits is not diagnostic because marine onlap can be generated equally well by sediment bypassing on steep slopes, by slope failure and by lobe switching (e.g. Schlager, 1992). Lowstand wedges and karst landforms are also the best criteria to recognize where exposure had preceded drowning at an unconformity (e.g. Fig. 5b). If these features cannot be identified, the seismic images will show little difference between surfaces of exposure plus drowning and surfaces of drowning only. In both instances, the morphologi cal expression of the event will be dominated by the change from platform morphology to a pelagic cap with the concomitant change in sediment input. Handford (1995) presented an instructive case study illustrating the difficulty of distinguishing between drowning unconformities and lowstand unconformities in seismic images (Fig. 12) . In this example,the lowstand wedge dominates the pattern and the drowning unconformity is the subordinate feature . If the basin had received little sediment during the lowstand (as was the case with the Bahamian basins during the Quaternary low stands), the drowning unconformity would have been a high-relief feature. Handford (1995) rightly
W Schlager
18
(a)
drowning unconformity exposure unconformity
----shale cover
����=
�J
(b)
--lowstand wedge (shale, sandstone)
HST (Burlington)
drowning unconformity exposure unconformity
l �:]'=����-----=--
§J
shale cover
HST (Burlington)
Fig. 12. Close superposition of exposure and drowning unconformity. Burlington platform, Mississippian, Arkansas. (a) Stratigraphy as reconstructed by Handford (199 5) shows from bottom to top: (1) Burlington prograding platform (HST), (2) exposure surface and coeval siliciclastic lowstand wedge in the basin, (3) TST with drowning unconformity, (4) deep-water shale cover. Exposure and drowning unconformity are only 10 m apart at the platform top. Exposure unconformity has greater relief and dominates the pattern. (b) Hypothetical stratigraphy assuming that there was no lateral siliciclastic sediment supply to build the lowstand wedge. In these circumstances, the basin receives very little sediment during lowstands and the drowning unconformity becomes the dominant feature. Bahamian basins in the Quaternary can serve as models for this situation.
sition is best visible in young, diagenetically imma ture limestones with a high impedance contrast between a tightly cemented exposure horizon and friable marine deposits; the contrast decreases as the rocks become diagenetically mature limestones or dolomites. Drowning horizons tend to be reliable and prominent seismic events over a wide range of depths because : (i) they juxtapose different litholo gies in the form of (hemi)pelagic deposits and shoalwater carbonates; (ii) they often represent extensive erosional unconformities with hiatuses that span millions of years and are accentuated by submarine lithification ; this submarine erosion is fuelled by currents that become intensified by the sharp topography of the drowned platform. Drowning unconformities have been used repeat edly,and very successfully, as sequence boundaries. This practice is in agreement with the original definition of sequence boundar y. It implies, how ever, that a sequence boundary does not necessarily represent a relative fall in sea-level. There is a growing number of drowning unconformities where a highstand systems tract is unconformably overlain by a transgressive tract without intervening expo sure or lowstand deposits.
ACKNOWLEDGEMENTS Robert W. Scott provided important information and comments on Cretaceous drowning events. Thomas Roep and John Woodside commented on a draft of this paper, and Chris Kendall and Alfonso Bosellini reviewed it for the editor. To all of them I owe thanks. Finally, Doug Bergersen, Gilbert Cam oin and Peter Davies created the incentive to pursue this topic for the Symposium on West Pacific and Indian Ocean Reefs.
pointed out that seismic information is inadequate to reconstruct the course of events in this instance. This could only be done from the rock record.
REFERENCES A.S. & NAIRN, A.E.M. (1993) Carbonate platform models of Arabian Cretaceous reservoirs. In: Cretaceous Carbonate Platorms (Eds Simo, J.A.T., Scott, R.W. & Masse, J.P.), Mem. Am. Assoc. petrol. Geol., Tulsa, 56, 173-184. ANSELMETTI, F.S. & EBERL!, G.P. (in press) Sonic velocity in carbonates-a combined product of depositional lithology and diagenetic alterations. Spec. Pub!. Soc. sediment. Geol. . ARNAUD-VANNEAU, A. & ARNAUD, H. (1990) Hauterivian to Lower Aptian carbonate shelf sedimentation and sequence stratigraphy in the Jura and northern subalALSHARHAN,
CONCLUSIONS Both exposure and drowning (deep fl o oding) can interrupt the growth of reefs and carbonate plat forms, and numerous examples of exposure and drowning events have been observed in the geolog ical record. The seismic expression of these events differs. Exposure followed by renewed shallow-water depo-
Exposure and drown ing of carbonate platforms
pine chains (SE France and Swiss Jura). In: Carbonate Platforms, Facies, Sequences and Evolution (Eds Tucker, M.E., Wilson, J.L., Crevello, P.D., Sarg, J.R. & Read, J.F.), Spec. Pubis int. Ass. Sediment., No 9, 203-234. Blackwell Scientific Publications, Oxford. ARTHUR, M.A. & SCHLANGER, S.O. ( ! 979) Cretaceous 'oceanic anoxic events' as causal factors in development of reef-reservoired oil fields. Bull. Am. Assoc. petrol. Geol. , 63, 870-885. BEACH, D.K. (199 5) Controls and effects of subaerial exposure on cementation and development of second ary porosity in the subsurface of Great Bahama Bank. In: Unconformities and Porosity in Carbonate Strata (Eds Budd, D.A., Saller, A.H. & Harris, P.M.), Mem. Am. Assoc. petrol. Geol., Tulsa, 63, 1-34. BoSELLINI, A. (1989) Dynamics of Tethyan carbonate platforms. In: Controls on Carbonate Platform to Basin Development (Eds Crevello, P.D., Sarg, J.F., Read, J.F. & Wilson, J.L.), Spec. Pub!. Soc. econ. Paleont. Miner., Tulsa, 44, 3-14. BOSSCHER, H. & SCHLAGER, W. (1992) Computer simula tion of reef growth. Sedimentology, 39 , 503-512. BRANDNER, R., FLOGEL, E., KOCH, R. & YOSE, L. (199 1 ) The northern margin o f the Schlern/Sciliar-Rosen garten/Catinaccio Platform. Dolomieu Conference on Carbonate Platforms and Dolomitization; Guidebook Excursion A. Ortisei Tourist Office, Ortisei, Italy BRINK, K.H. (1987) Coastal ocean physical processes. Rev. Geophys. space Phys. , 25, 204-216.
K.H. ( 1 989) The effect of stratification on seamount-trapped waves. Deep-Sea Res. , 36, 825-844. BUDD, D.A., SALLER, A.H. & HARRIS, P.M. (Eds) (1995) Unconformities and Porosity in Carbonate Strata. Mem. Am. Ass. petrol. Geol., Tulsa, 63, 1-313. BuRCHETTE, T.P. (1993) Mishrif Formation (Cenoman ian-Turonian), Southern Arabian Gulf: carbonate plat form growth along a cratonic basin margin. In: Cretaceous Carbonate Platforms (Eds Simo, J.A.T., Scott, R.W. & Masse, J.P.), Mem. Am. Assoc. petrol. Geol., Tulsa, 56, 1 85-200. CHRISTIE-BLICK, N. (1991) Onlap, offiap, and the origin of unconformity-bounded depositional sequences. Mar. Geol. , 97, 35-56. CLARI, P.A., DELA PIERRE, F. & MARTIRE, L. (1995) Discontinuities in carbonate successions: identification, interpretation and classification of some Italian exam ples. Sediment. Geol. , 100, 97-121. CROSS, T.A. & LESSENGER, M.A. (1988) Seismic stratigra phy. Ann. Rev. Earth planet. Sci., 16, 319-354. DARWIN, C.R. (1842) The Structure and Distribution of Coral Reefs, reprinted 1984. University Arizona Press, Tucson. DAVIES, P.J., McKENZIE, J.A., PALMER-JULSON, A. et a/. (1991) Proceedings ofthe Ocean Drilling Program Initial Reports, 133. Ocean Drilling Program, College Station, TX. DILLON, W.P., TREHU, A.M. & VALENTINE, P.C. ( 1 9 8 8) Eroded carbonate platform margin-the Blake escarp ment off southeastern United States. Am. Assoc. petrol. Geol., Stud. Geol. , 27 (2), 140-148. DRAVIS, J.J. (1989) Deep-burial microporosity in Upper Jurassic Haynesville oolitic grainstones, East Texas. Sediment. Geol. , 63, 325-341. BRINK,
19
G.P. & GINSBURG, R.N. (1989) Cenozoic progra dation of northwestern Great Bahama Bank, a record of lateral platform growth and sea-level fluctuations. In:
EBERL!,
Controls on Carbonate Platform to Basin Development
(Eds Crevello, P.D., Wilson, J.L., Sarg, J.F. & Read, J.F.), Spec. Pub!. Soc. econ. Paleont. Miner., Tulsa, 44, 339-351. EBERL!, G.P., BERNOULLI, D., SANDERS, D. & VECSEI, A. (1993) From aggradation to progradation: the Maiella Platform, Abruzzi, Italy. In: Cretaceous Carbonate Plat forms (Eds Simo, J.A.T., Scott, R.W. & Masse, J.P.), Mem. Am. Assoc. petrol. Geol., Tulsa, 56, 213-232. EBERL!, G.P., KENTER, J.A.M., McNEILL, D.F., GINSBURG, R.N., SWART, P.K. & MELIM, L.A. (in press) Facies, diagenesis and timing of prograding seismic sequences on western Great Bahama Bank. Spec. Pub!. Soc. sediment. Geol. EPTING, M. ( 1 989) Miocene carbonate buildups of central Luconia, offshore Sarawak. In: Atlas of Seismic Stratig raphy (Ed. Bally, A.W.), Am. Assoc. petrol. Geol., Stud. Geol., 27 (3), 168-173. ERIKSEN, C. C. (1991) Observations of amplified flows atop a large seamount. J. geophys. Res. , 96, 15227- 1 5236. ERLICH, R.N., BARRETT, S.F. & Guo, B.J. (1990) Seismic and geologic characteristics of drowning events on carbonate platforms. Bull. Am. Assoc. petrol. Geol. , 74, 1523-1537. ERLICH, R.N., LONGO, A.P., JR & HYARE, S. (1993) Re sponse of carbonate platform margins to drowning: evidence of environmental collapse. In: Carbonate Se quence Stratigraphy-Recent Developments and Applica tions (Eds Loucks, R.G. & Sarg, J.F.), Mem. Am. Assoc.
petrol. Geol., Tulsa, 57, 241-266. EvERTS, A.J. (1994) Carbonate sequence stratgraphy of the Vercors (French Alps) and its bearing on Cretaceous sea level. Pub!. Dissert., Vrije Universiteit, Amsterdam.
EVERTS, A.J.W., SCHLAGER, W. & REHMER, J.J.G. (in press) Carbonate platform-to-basin correlation by means of grain-composition logs: an example from the Vercors (Cretaceous, SE France). Sedimentology. FOLLMI, K.B., WEISSERT, H., BISPING, M. & FUNK, H.P. (1994) Phosphogenesis, carbon-isotope stratigraphy, and carbonate-platform evolution along the Lower Cre taceous northern Tethyan margin. Geol. Soc. Am. Bull. , 106, 729-746. FOUKE, B.W., EVERTS, A.-J. W., ZWART, E., SCHLAGER, W., SMALLEY, P.C. & WEISSERT, H. (1995) Subaerial expo sure unconformities on the Vercors carbonate platform (SE France) and their sequence stratigraphic signifi cance. In: High Resolution Sequence Stratigraphy: Inno vations and Applications (Eds Howell, J.A. & Aitken, J.F.), Spec. Pub!. geol. Soc. London, No 104, pp. 295320. Geol Soc. London, Bath. FRAZIER, E.D. (1974) Depositional episodes: the relation ship to the Quaternary stratigraphic framework in the northwestern portion of the Gulf basin. Univ. Texas, Austin, Bur. econ. Geol., Geol. Circ. , 74 (1), 1 -28. FUNK, H.P., FOLLMI, K.B. & MOHR, H. ( 1 993) Evolution of the Tithonian-Aptian carbonate platform along the Northern Tethyan margin, Eastern Helvetic Alps. In: Cretaceous Carbonate Platforms (Eds Simo, J.A.T. Scott, R.W. & Masse, J.P.), Mem. Am. Assoc. petrol. Geol., Tulsa, 56, 387-408.
20
W Schlager
W.E. ( 1 989a) Genetic stratigraphic sequences in basin analysis 1: Architecture and genesis of flooding surface bounded depositional units. Bull. Am. Assoc. petrol. Geo/. , 73, 125-142. GALLOWAY, W.E. (1989b) Genetic stratigraphic sequences in basin analysis II: Application to Northwest Gulf of Mexico Cenozoic Basin. Bull. Am. Assoc. petrol. Geo/., 73, 143-154. GENIN, A., NoBLE, M. & LoNSDALE, P.F. ( 1 989) Tidal currents and anticyclonic motions on two North Pacific seamounts. Deep-Sea Res. , 36, 1803-1815. GOLDSTEIN, R.H., FRANSEEN, E.K. & MILLS, M.S. (1990) Diagenesis associated with subaerial exposure Miocene strata, southeastern Spain: implications for sea level change and preservation of low-temperature fluid inclu sions in calcite cement. Geochim. Cosmochim. Acta, 54, 699-704. HANDFORD, C.R. ( 1 99 5) Baselap patterns and the recogni tion of lowstand exposure and drowning-a Mississippian-ramp example and its seismic signature. J. sediment. Res., 865 (3), 323-337. JACQUIN, T., ARNAUD-VANNEAU, A. & ARNAUD, H. (1991) Systems tracts and depositional sequences in a carbon ate setting: a study of continuous outcrops from plat form to basin at the scale of seismic lines. Mar. petrol. Geo/. , 8, 122-139. KENDALL, C.G.STC. & ScH LAGER, W. (198 1 ) Carbonates and relative changes in sea level. Mar. Geo/. , 44, 181-212. KENNARD, J.M., SOUTHGATE, P.N., JACKSON, M.J. & O'BRIEN, P.E. (1992) A new sequence perspective on the Devonian reef complex and the Frasnian-Famennian boundary, Canning Basin, Australia. Geology, 20, 1135-1 138. KUKAL, Z. (1990) The Rate of Geological Processes. Aca demia, Praha. LADD, H.S., NEWMAN, W.A. & SoHL, N.F. (1974) Darwin Guyot, the Pacific's oldest atoll. Proceedings of the 2nd
TX. PuRDY, E.G. & BERTRAM, G.T. ( 1 993) Carbonate concepts from the Maldives, Indian Ocean. Am. Assoc. petrol. Geol., Stud. Geo/. , 34, 1-56. READ, J.F. (1982) Carbonate platforms of passive (exten sional) continental margins-types, characteristics and evolution. Tectonophysics, 81, 195-212. ROBBIN, D.M. & STIPP, J.J. ( 1 979) Depositional rate of laminated soilstone crusts, Florida Keys. J. Sediment. Petrol. , 49, 175-180. RODEN, G.I. (1987) Effect of seamounts and seamount chains on ocean circulation and thermohaline structure. In: Seamounts, Islands and A tolls (Eds Keating, B.H., Fryer, P., Batiza, R. & Boehlert, G.W.), Geophys. Monogr., Am. geophys. Union, Washington, DC, 43, 335-354. SAGER, W.W., WINTERER, E.L., FIRTH, J.V. et a/. (1993)
international Coral Reef Symposium, Brisbane, Great Barrier Reef Committee, Part 2, 513-522.
Proceedings of the Ocean Drilling Program, Initial Re ports, 143, pp. 111-180. Ocean Drilling Program, Col
GALLOWAY,
P.F., NORMARK, W.R. & NEWMAN, W.A. (1972) Sedimentation and erosion on Horizon Guyot. Geo/. Soc. Am. Bull. , 83, 289-3 1 5 . MARTIRE, L. (1992) Sequence stratigraphy and condensed pelagic sediments. An example from the Rosso Ammo nitico Veronese, northeastern Italy. Pa/aeogeogr. Pa/ae oclimatol. Palaeoeco/., 94, 169-191. MECKEL, L.D. III & GALLOWAY, W.E. (1996) Formation of high-frequency sequences and their bounding surfaces: case study of the Eocene Yegua Formation, Texas Gulf Coast, USA. Sediment. Geo/. , 102, 155-18 1 . MEUM, L.A., SWART, P.K. & MALIVA, R.G. (1995) Meteoric-like fabrics forming in marine waters: impli cations for the use of petrography to identify diagenetic environments. Geology, 23, 7 5 5-758. MOLDOVANYI, E.P., WAAL, F.M. & ZHANG JUN YAN (1995) Regional exposure events and platform evolution of Zhujiang Formation carbonates. Pearl River Mouth Basin: evidence from primary and diagenetic seismic facies. In: Unconformities and Porosity in Carbonate Strata (Eds Budd, D.A., Saller, A.H. & Harris, P.M.), Mem. Am. Assoc. petrol. Geol., Tulsa, 63, 1 25-140. PIGRAM, C.J., DAVIES, P.J. & CHAPRONIERE, G.C.H. ( 1 993)
LoNSDALE,
Cement stratigraphy and the demise of the early middle Miocene carbonate platform on the Marion Plateau. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 33 (Eds McKenzie, J.A., Davies, P.J. & Palmer-Julson, A.), pp. 499-512. Ocean Drilling Program, College Station, TX. POSAMENTIER, H.W. & VAIL, P.R. (1988) Eustatic controls on clastic deposition 11-sequence and systems tract models. In: Sea-level Changes: an Integrated Approach (Eds Wilgus, C.K. Hastings, B.S., Kendall, C.G.StC., Posamentier, H.W., Ross, C.A. & van Wagoner, J.C.), Spec. Pub!. Soc. econ. Paleont. Miner., Tulsa, 42, 125-154. PRATT, B.R. & SMEWING, J.D. ( 1 993) Early Cretaceous Platform Margin, Oman, Eastern Arabian Peninsula. In: Cretaceous Carbonate Platforms (Eds Simo, J.A.T., Scott, R.W. & Masse, J.P.), Mem. Am. Assoc. petrol. Geol., Tulsa, 56, 201-212. PREMOL!-SILVA, I., HAGGERTY, J., RACK, F. et a/. (1993) Proceedings of the Ocean Drilling Program, Initial Re ports, 1 44: Ocean Drilling Program, College Station,
lege Station, TX. A., ARMIN, R., ICHRAM, L.O. & GLENN-SULLIVAN, C. (1993) Sequence stratigraphy of aggrading and backstepping carbonate shelves, Oligocene, Central Kalimantan, Indonesia. In: Carbonate Sequence
SALLER,
Stratigraphy-Recent Developments and Applications
(Eds Loucks, R.G. & Sarg, J. F.), Mem. Am. Assoc petrol. Geol., Tulsa, 57, 267-290. SARG, J.F. ( 1 98 8a) Carbonate sequence stratigraphy. In: Sea-level Changes: an Integrated Approach (Eds Wilgus, C.K., Hasting, B.S., Kendall, G.C.StC. Posamentier, H.W., Ross, C.A. & van Wagoner, J.C.), Spec. Pub!. Soc. econ. Paleont. Miner., Tulsa, 42, 15 5-181. SARG, J.F. (1988b) Middle-Late Permian depositional sequences, Permian Basin, West Texas and New Mex ico. In: Atlas of Seismic Stratigraphy (Ed. Bally, W.A.), Am. Assoc. petrol. Geol., Stud. Geol., 27 (3), 140-155. SCHLAGER, W. (1981) The paradox of drowned reefs and carbonate platforms. Geo/. Soc. Am. Bull. , 92, 197-211. SCHLAGER, W. (1989) Drowning unconformities on car bonate platforms. In: Controls on Carbonate Platform to Basin Development (Eds Crevello, P.D., Sarg, J.F., Read, J.F. & Wilson, J.L.), Spec. Pub!. Soc. econ.
Exposure and drowning of carbonate platforms
Paleont. Miner., Tulsa, 44, 15-25. W. (1991) Depositional bias and environmen tal change-important factors in sequence stratigraphy. Sediment. Geol. , 70, 109-130. ScHLAGER, W. ( 1 992) Sedimentology and Sequence Stratig raphy of Reefs and Carbonate Platforms. Am. Assoc. petrol. Geol., Cont. Educ. Course Note Series, 34, 71. SCHLAGER, W. & CAMBER, 0. (1986) Submarine slope angles, drowning unconformities and self-erosion of limestone escarpments. Geology, 14, 762-765. SCHLAGER, W. & JAMES, N.P. (1978) Low-magnesian cal cite limestones forming at the deep-sea floor, Tongue of the Ocean, Bahamas. Sedimentology, 25, 675-702. ScoTT, R.W. (1990a) Models and stratigraphy of mid Cretaceous reef communities, Gulf of Mexico. Soc. econ. Paleont. Miner. (Soc. sediment. Geol.), Concepts for Sedimentology and Paleontology, 3, 1-102. ScoTT, R.W. (1990b) Chronostratigraphy of the Creta ceous carbonate shelf, southeastern Arabia. In: The Geology and Tectonics ofthe Oman Region (Eds Robert son, A.H.F., Searle, M.P. & Ries, A. C.), Spec. Pub!. geol. Soc. London, No 49, pp. 89- 1 08. Blackwell Scientific Publications, Oxford. SEYFRIED, H. (198 1 ) Genesis of 'regressive' and 'transgres sive' pelagic sequences in the Tethyan Jurassic. In: Rosso Ammonitico Symposium Prceedings (Eds Fari nacci, A. & Elmi, S.), pp. 547-579. Ediz. Tecnoscienza, Rome. SHAUB, F.J. ( 1 983) Origin of Catoche Tongue. In: Seismic Expression of Structural Styles (Ed. Bally, A.W.), Am. Assoc petrol. Geol., Stud. Geol., 15 (2), 129-140. SIMO, J.A.T. ( 1 993) Cretaceous carbonate platforms and stratigraphic sequences, South-Central Pyrenees. In: Cretaceous Carbonate Platforms (Eds Simo, J.A.T., Scott, R.W. & Masse, J.P.), Mem. Am. Assoc petrol. Geol., Tulsa, 56, 325-342. STAFLEU, J. & SCHLAGER, W (1995) Pseudo-unconformities in seismic models of large outcrops. Geol. Rundsch. , 84, 761-769. STAFLEU, J., EVERTS, A.J. & KENTER, J.A.M. ( 1 994) Seismic models of a prograding carbonate platform: Vercors, south-east France. Mar. petrol. Geol. , 1 1 , 513-527. STITELER, T.C., SCOTT, R.W., CHACARTEGUI, F.J., ORTEGA, O.J., AZPIRITXAGA, I. & TAYLOR, C.K. (in press) Carbon ate platform seismic sequence attributes, Maracaibo Basin, Venezuela. Spec. Pub!. Soc. explor. Geophys. TYRRELL, W.W. & ScoTT, R.W. ( 1 989) Early Cretaceous shelf margins, Vernon Parish, Louisiana. In: Atlas of Seismic Stratigraphy (Ed. Bally, A.W.), Am. Assoc. petrol. Geol., Stud. Geol., 27 (3), 11-17. VAIL, P.R., MITCHUM R.M., TODD R.G. e t a/. (1977) Seismic stratigraphy and global changes of sea level. In: SCHLAGER,
21
Seismic Stratigraphy-Applications to Hydrocarbon Ex ploration (Ed. Payton, C.E.), Mem. Am. Assoc. petrol.
Geol., Tulsa, 26, 49-212. P.R., HARDENBOL, J. & TODD, R.G. (1984) Jurassic unconformities, chronostratigraphy, and sea-level changes from seismic stratigraphy and biostratigraphy. In: Interregional Unconformities and Hydrocarbon Accu mulation (Ed. Schlee, J.S.) Mem. Am. Assoc. petrol. Geol., Tulsa, 36, 129-144. VAIL, P.R. (1987) Seismic stratigraphic interpretation using sequence stratigraphy. Part 1: Seismic stratigra phy interpretation procedure. In: Atlas of Seismic Stratigraphy (Ed. Bally, A.W.), Am. Assoc petrol. Geol., Stud. Geol., 27 (1), 1-10. VAN WAGONER, J.C., POSAMENTIER, H.W., MITCHUM, R.M., et a!. (1988) An overview of the fundamentals of sequence stratigaphy and key definitions. In: Sea-level Changes: an Integrated Approach (Eds Wilgus, C.K. Hastings, B.S., Kendall, C.G.StC., Posamentier, H.W. Ross, C.A. & van Wagoner, J.C.), Spec. Pub!. Soc. econ. Paleont. Miner., Tulsa, 42, 39-45 . WAGNER, P.D. (1990) Geochemical stratigraphy and po rosity controls in Cretaceous carbonates near the Oman mountains. In: The Geology and Tectonics of the Oman Region (Eds Robertson, A.H.F., Searle, M.P. & Ries, A.C.), Spec. Pub!. geol. Soc. London, No. 49, pp. 12713 7. Blackwell Scientific Publications, Oxford. WAGNER, P.D., TASKER, D.R. & WAHLMAN, G.P. (1995) Reservoir degradation and compartmentalization below subaerial unconformities: limestone examples from West Texas, China and Oman. In: Unconformities and Porosity in Carbonate Strata (Eds Budd, D.A., Saller, A.H. & Harris, P.M.) Mem. Am. Assoc. petrol. Geol., Tulsa, 63, 177-196. WENDTE, J.C. & MUIR, I. (1995) Recognition and signifi cance of an interformational unconformity in Late Devonian Swan Hills Reef Complexes, Alberta. In: Unconformities and Porosity in Carbonate Strata (Eds Budd, D.A., Saller, A.H. & Harris, P.M.), Mem. Am. Assoc. petrol. Geol., Tulsa, 63, 259-278. WENDTE, J.C., STOAKES, F.A. & CAMPBELL, C.V. (1992)
VAIL,
Devonian-Early Mississippian Carbonates of the West ern Canada Sedimentary Basin: a Sequence Strati graphic Framework. Soc. econ. Paleont. Miner., Short
Course 28, Calgary. W.G. (1993) The drowning succession in Jurassic carbonates of the Venetian Alps, Italy: a record of supercontinent breakup, gradual eustatic rise, and eutrophication of shallow-water environments. In: Car
ZEMPOLICH,
bonate Sequence Stratigraphy-Recent Developments and Applications (Eds Loucks, R.G. & Sarg, J.F.), Mem.
Am. Assoc. petrol. Geol., Tulsa, 57, 63-105.
Spec. Pubis int. Ass. Sediment. ( 1998) 25, 23-38
The origin of the Great Barrier Reef-the impact of Leg 133 drilling P. J. D A V I E S* and F. M. P E E R D E M ANt
*Ocean Science Institute, University of Sydney, NSW 2006, Australia; and tS hell International, Muscat, Oman
ABSTRACT
Drilling of the shallow fore-reef slope during Leg 133 of the Ocean Drilling Program has allowed the definition of an event history critical for understanding the initiation and evolution of the Great Barrier Reef. Within or immediately before the time period of isotope stages 8 and 9, a fundamental change in climate, driven by a switch from 19 000-yr obliquity to I 00 000-yr precessional orbital cycles, led to raised sea surface temperatures and the initiation of the Great Barrier Reef. It is therefore only 300 000 yr old and an ecosystem response to environmental change. Subsequent development occurred as a series of high sea-level slices effected by four or five sea-level oscillations and growing progressively retrogressively to the west. The subreef section, although unknown, is postulated to be analogous to a mid to outer shelf coralline dominated environment comparable with that growing on the shelf to the south of the Great Barrier Reef today. The vertical and lateral (latitudinal) facies variations obey Walther's Law of Succession as a consequence of a sedimentary response to subsidence, latitude and climate change.
INTRODUCTION
This paper explores a view of the origin of the Great Barrier Reef somewhat different from that normally accepted by others. It is different because it at tempts to cast dogma aside and proposes a new order of things. This new order has arisen from a great tl�al of thought in the time-frame 1990-199 5 but building on studies with submersible and drill ships in the previous decade. It is both a truism and a tautology to say, however, that it has already been received with 'less than enthusiasm' by establish ment Australia, and sceptically by those scientists who have done well under the old ideas. Be that as it may, we believe that the evidence is compelling and likely to have implications in the biological as well as the geological study of reefal systems. Our intention is therefore to ignore Machiavelli's maxim that 'There is nothing more difficult to take in hand, more perflous to conduct, or more uncer tain in its success, than to take the lead in a new order of things'. We propose therefore that the Great Barrier Reef is very young, was catalysed by a major climate change and that the analogue for its subreefal sequence occurs today south of the tropics
on the shelf of southern Queensland. In the next 6 months we expect to have tested these assertions in the field and we will know 'whether we have truly been thinking' or 'just being logical'.
BACKGROUND
The Great Barrier Reef (Fig. 1) is the largest epicon tinental reef system currently existing on this Earth (Davies et a!. , 1989). Its geological and biological relevance is therefore immense and understanding its origin defines a worthy scientific objective. As such, it has been the subject of intense scientific interest (Richard & Hill, 1942; Maxwell, 1968; Tanner, 1969; Palmieri, 1971, 1974; Lloyd, 1973; Ericson, 1976; Hopley, 1982; Davies, 1983) and particularly in the last decade it has been the focus of substantial national and international effort, for example, studies by the Bureau of Mineral Re sources (BMR) throughout the 1980s and the Ocean Drilling Program (ODP) drilling on Leg 133 under taken in 1990 (Davies et a!., 1987, 1988a,b, 1989,
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
23
24
P. J. Davies & F M. Peerdeman A
\
•
•
:.�
\ Ribbon Reef 5
.... . -
•
\
�
I
. J, ,,
-
50
,,
·�
+
•+ torai Sea ·Basin + +
� Great Barrier Reef c=J Late CretaceousfTertiary Rift Basins �Oceanic Basement
0
200 km
+
+
+
+
....
...
• +
• +
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
" I
25
30'
Origin of the Great Barrier Reef
25
GRAFTON PASSAGE SEISMIC GRID
Fig. 2. Site survey seismic tracks and the location of ODP Sites 8 19, 820 and 82 1 in Grafton Passage in the northern Great Barrier Reef, on the upper slope to the east of Cairns. The seismic data through the drill sites are shown in Fig. 3.
1992; Davies, 1991; McKenzie et al., 1993). Fol lowing the detailed seismic studies conducted by BMR in the late 1980s, Davies et at. (1989) postu lated that Cainozoic plate motion to the north was a principal force in defining the timing of reef initia tion off north-east Australia, a postulate proved during Leg 133 drilling on the Queensland Plateau. In addition, one of us (P.J.D.) initiated and orga nized a joint research programme with the Japan National Oil Corporation to the south of the Great Barrier Reef in the belief that such an area repre sented a current model of what may underlie the Great Barrier Reef, the postulate being that Walth er's Law of Succession operates in a lateral as well as a vertical sense as a consequence of the joint effects of northward plate motion and subsidence. Fig. 1. (Opposite) (a) The position of the Great Barrier Reef relative to the bathymetry and main physiographical and geological features of north-east Australia. The position of Grafton Passage where ODP drilled three holes in 1990 is marked. (b) The shelf in the vicinity of Cooktown where an international consortium of five countries will drill three holes through the Reef in late 1995, and where a submersible survey of the outer barrier was conducted in 1985 and gave rise to the data presented in Fig. I 0.
In the present paper therefore, our aims are to reconcile the data from the three sites in Grafton Passage off Cairns (Fig. 2) so as to predict the timing and causes for the formation of the Great Barrier Reef and, through utilization of Walther's Law, to predict by analogy the facies that underlie the Great Barrier Reef. Finally, it is our intention to use these data to define a hypothesis for the evolu tion of the Great Barrier Reef and to discuss the consequences of such an evolution.
LEG 133 DRILLING ON THE SLOPE OF THE GREAT BARRIER REEF
In September 1990, the JOIDES Resolution drilled four holes in the slope of the Great Barrier Reef, in Grafton Passage to the east of Cairns (Fig. 2). Three of these holes (819, 820 and 821) occur on the upper slope immediately in front of the outer barrier: drilling the slope sequences provided evi dence that was compelling, but some found it difficult to believe. However, all agreed that earlier propositions of an age of 3-15 Ma (see Davies et at., 1989) were likely to be wrong and that the Great Barrier Reef appeared to be much younger. The
·
26
P. J. Davies & F. M. Peerdeman
conflict focused on just how young, and there were two schools of thought. On the basis of poorly defined seismic mounding, Feary et al. (1993) pro posed that reefs were present low in the section and with an age greater than 1 Ma. In the same vein but with harder evidence from the distributions of foraminifer (but not corals) in the section at Site 821, Montaggioni & Venec-Peyre (1993) also sug gested an age older than 1 Ma, and went as far as to propose a succession of reef tracts throughout the Pleistocene as precursors to the modern reef system. In a different camp, Davies et al. (1991) and Davies (1991) used the age of seismic reflectors drilled at Site 820 and which were traceable to the west beneath the reefs, to suggest an age of younger than 500 ka, and Davies (1991) defined the geological and biological consequences of such an hypothesis. In this paper we try to reconcile the two schools of thought into the most likely scenario through an analysis of the seismic, sedimentological, strati graphical and isotopic data, and from which we
erect an event stratigraphy most likely to explain the phenomenon which is the Great Barrier Reef. The three drill holes occur on a fore-reef terrace in 280 m of water and less than 3 km from the surface shelf edge reefs that form the outer barrier of the Great Barrier Reef in this sector. The hole at Site 819 occurs in 565 m of water; that at Site 820 occurs in 285 m of water and at Site 821 occurs in 225 m of water. Although the section at Site 821 is probably the most complete sedimento1ogically, it suffers from extensive diagenesis, which has made the development of an isotope stratigraphy difficult. Site 820 therefore represents the most complete high-resolution isotopic section obtained on the north-east Australia margin. However, the results from all three holes are used in the following analysis. The seismic signatures at Sites 819, 820 and 821
Feary et al. (1993) interpreted the relatively high-
Fig. 3. Seismic data through Grafton Passage and ODP Sites 820 and 82 1. (a) Uninterpreted; (b) interpreted. The sequence boundaries are those discussed in the text.
27
Origin of the Great Barrier Reef 2.5km
Fig. 4. Sedimentological and stratigraphical variation across ODP Sites 8 19, 820 and 82 1. Data from Feary eta!. ( 1993) and Davies eta!. ( 1992). The data for Site 820 show the major sedimentological variation in terms of the distribution of wackestones and packstones, whereas for Sites 8 19 and 82 1, only the positions of the major stratigraphical boundaries are shown. The dates are those defined by nanostratigraphy in Leg 133 ODP, Final Results.
resolution seismic profiles available at the time (Fig. 3a & b) and recognized three mega-sequences in the fore-reef section in front of the Great Barrier Reef. In vertical section these are: 1 Seismic sequence 1 (0-490 ms), an essentially aggrading section of high-amplitude parallel and onlapping reflectors but with clear evidence of erosion at several depths, but most clearly seen at the base where the sequence boundary is an offlap surface. 2 Seismic sequence 2 (490- 5 5 5 ms), an essentially aggrading section, but divided into two parts: (i) from 490 to 525 ms, an aggrading sequence down lapping on a strong reflector at 525 ms, and (ii) a prograding section which is onlapping a strong reflector at 5 5 5 ms. 3 Seismic sequence 3 (below 5 5 5 ms), clearly a prograding section composed of many sequences each of which displays discontinuous to continuous reflectors of low to moderate amplitude. The basal sequence boundaries in most cases represent onlap surfaces. Essentially and in turn, the three sequences define a clearly aggradational section above, a middle sequence that is transitional and a lower sequence that is progradational. Peerdeman (1993) inter preted the sections as point source and clastic dominated below and changing vertically into car bonate dominated above. Feary et a!. (1993) de scribed mounds in the lower part of the prograding sequence and proposed that they were reefal mounds. We have re-examined them and con clude that although they may be mounds they are not different from mounds on the present-day slope in deeper water that are not reefal: further,
we fail to see such mounds in higher-resolution seismic data collected since the drilling over the same area. Sediment structure and variation at Sites 819, 820 and 821
The stratigraphical and sedimentological variations at Sites 819, 820 and 821 are shown in Figs 4 & 5 (Davies et a!., 1992). At Site 819 a major hiatus occurs in the upper part of the section (Alexander & Kroon, 1993) and sedimentation rates above the hiatus are also much lower than at Site 820. On the basis of seismic data collected since the cruise, Site 821 contains the most complete sedimentological section, but it has also been affected by the most intense fresh-water diagenesis (K. Konishi, personal communication), so that from a combined sedi mentological and isotopic viewpoint, Site 820 is considered to be by far the most complete section and therefore forms the basis for the analysis defined below. Three sedimentological units are recognized (Fig. 5): Unit A between 0 and 65 m, Unit B between 65 and 145 m, and Unit C between 145 and 220 m. Unit A corresponds to the top 65 m of the section and is composed of wackestones separated by 8 m thick sandy bioclastic packstones composed of planktonic and benthic forams, molluscs, ptero pods, minor echinoids, corals, corallines, bryozo ans, rhodoliths (at 7 and 32 m) and Halimeda. Unit B corresponds to the section between 65 and 145 m and is divided into an upper part (BI) between 65 and II0 m and a lower part (B2) between 110 and 145 m. Unit B l is characterized by wacke-
·
28
P. J. Davies & F. M. Peerdeman
Seismic analysis
Sediment analysis
Om
Aggrading with
A
clear erosional
Wackestones
sedimentation
geometries
separated by
rates
reflecting
8
sandy
Low lowstand
SL variations
bioclastic
sedimentation
packstones
rates
50m
High highstand
81 Aggrading Onlapping
Wackestones and thin persistent sandy
100m
packstones
82 Prograding
Wackestones and thicker packstones
150m Prograding
Low highstand sedimentation rates High lowstand sedimentation rates
Low highstand sedimentation rates
c
Low highstand
Packstones
and High lowstand rates
stones and thin packstone beds, whereas Unit B2 is characterized by wackestones but the sandy pack stones are much thicker than in BI, there again being five of them. Unit C corresponds to the section below 145 m and is dominated by sandy packstone beds with minor wackestones. The variations in sedimentation rates in the three sequences are shown in Fig. 6, from which the following conclusions can be drawn: 1 Sedimentation rates are highest in sequences A and C. 2 The boundary between A and B defines a major change in sedimentation style, with maximum rates
Fig. 5. Summary of seismic and sedimentological variation as seen in the interpreted seismic data (after Feary eta/., 1993) and the data from ODP 820 (after Peerdeman, 1993; Feary eta/., 1993). SS 1-3 are seismic sequences and A-C are the sedimentary units identified in the above-named publications. The rate data are from Peerdeman (1993).
occurring in the highstand in sequence A and generally in the lowstands before sequence A. 3 The boundary between subsequences BI and B2 is also clearly seen in the rate data. The isotope event stratigraphy seen at Sites 820 and 819
The isotope stratigraphy is defined specifically and in detail through reference to Site 820 and compar ison with Site 819. At both sites, the isotope curves are shown in Fig. 7 and summarized in Fig. 8: Three principal isotopic sequences are recognized: I Between 0 and 75 m at 820, the signal is domi-
29
Origin of the Great Barrier Reef Sedimentation Rate - Cm per 1000 years 0
10
20
30
40
�2 9
7 18 �20 -.------, 27
c::J
10 16
60
70
3
5 12
Fig. 6. Variation in sedimentation rates for the top 150 m at ODP Site 820. Numbers refer to isotope stages. The three sequences, A, B and C are those defined through the sedimentological analysis of the core and correspond to identical devisions defined through seismic and isotopic analysis.
so
80
I
Sequence A
7
8
B1
14
15
100 120
90
Sequence B
B2
22 Sequence C
31
Highstand Sedimentation
Depth (mbsf)
28
30
19
liill
Lowstand Sedimentation
w 0
'"' 0
0 0 0
b
� "' 0
"' 0
'"' 0
Depth (mbsf)
w 0
.:... .:...
0 00 'i:l Cl tp (/)
u,
�·
.:.,
N 0
00
.:.,
0
0
.:...
0 00 'i:l Cl tp (/)
@"' '-" .:, Fig. 7. Isotope data for ODP Sites 820 and 8 19 (data from Alexander eta!., 1993; Peerdeman eta!., 1993). Readers should not attempt to correlate, as the depths are of substantially different ages. Interpretations are provided in Fig. 9.
00 "'
P. J Davies & F. M Peerdeman
30
Site 819
Site 820 Low Frequency High Amplitude
I
75m High Frequency Low Amplitude
145m
High Amplitude Low Frequency
I
High Frq/Ampl
2/2
300k
300k 2/1 400-SOOk
V.High Frequency High Amplitude
Low Frequency High Amplitude
Low Frequency High Amplitude
2/2 750k
)UUK
/)UK
3
1.02Ma
3
Fig. 8. Summary of isotope characteristics for the upper parts of the sections at ODP Sites 8 19 and 820 in Grafton Passage in front of the Great Barrier Reef.
1.02Ma
210m 4
nated by low-frequency, high-amplitude variation defined by Peerdeman et a!. (1993) as related to the orbital eccentricity frequency. This isotopic se quence extends from stage 1 to 8 and has a base therefore at about 300 000 yr. This same section, although appreciably thinner, is seen at Site 819 and has been interpreted by Alexander & Kroon (1993) in a similar manner. Further, Alexander & Kroon (1993) interpreted the character of the iso tope signal to indicate the progressive decoupling of surface and bottom water temperatures and raised surface water temperatures. 2 Between 75 and 145 m at Site 820, the signal is characteristically dominated by a much higher fre quency as well as high amplitude. It has been defined by Peerdeman et a!. (1993) as dominated by the obliquity signal, and extending from isotope stage 9 to 19, bottoming at about the Brunhes Matuyama (B-M) boundary. It is proposed here, however, that the section may be divided into two parts: (i) between 7 5 and 110 m (isotope stages 9-13) characterized by a lower-frequency, high amplitude signal, and (ii) between 110 and 145 m (isotope stages 13-19) characterized by a much higher frequency but also high-amplitude signal. At Site 819, isotope stages 9-13 are missing and although the section below is condensed, the high frequency section is still easily identifiable. 3 Between 145 and 220 m, the signal is character istically again much lower frequency (perhaps dom inated by the eccentricity signal) and has high amplitude, and as such is similar to the section above 7 5 m. This same signal is seen at Sites 819 and 820.
Peerdeman (1993) and Peerdeman et a!. (1993) defined the boundary at 75 m as a critical event in the history of the margin (Fig. 8). Using the differ ence between planktonic 8180 in the core and benthic 8180 at the deep-sea site 677 A (Shackleton & Hall, 1989), they defined sea surface temperature (SST) variation for the north-east Australian margin (Fig. 9) that indicated a substantial rise in temper ature of between 3 and soc at around the 75 m boundary and interpreted this as occurring around isotope stage 9. This temperature variation corre sponds to the change in frequency and amplitude of the isotopic signal, which is thought to arise from the change from obliquity to precessional orbital cyclicity. The change in SST may therefore relate to the beginning of a change leading to a warmer Earth. Alexander & Kroon (1993) recognized a similar temperature change separating sequences on either side of the hiatus incorporating isotope stages 9-13. Reconciliation of the seismic, sedimentological and isotopic data
The conclusions from the analysis of seismic, sedi mentological and isotope data are also summarized in Fig. 9. The boundaries defined at 6 5 m, 110 m and 145 m occur in all three data sets. There is little doubt that the seismic, sedimentological and isoto pic sequences are real and the boundaries recogniz able. At a more detailed level, there is a clear correlation of sand units with strong seismic signa tures and with positive 8180 excursions, i.e. low� stands. At a sequence level and especially in the topmost sequence between 0 and 65 m, there is a
31
Origin of the Great Barrier Reef
ol 8poB Site 820 ·I ·2
+fl.�
ol8pos Site 819
_,
�; Fig. 9. Interpreted isotope curves for Sites 820 and 819 together with the stratigraphical record for Site 820. The sea surface temperature curve at Site 820 was calculated using the difference between the planktonic 180 at Site 820 and the benthic values at Site 677A (Shackleton & Hall, 1989).
consistent relationship over three major sea-level oscillations of sedimentary characteristics (Table I). This shows that process and process-response are identical over three glacial cycles. However, the response to interstadial sea-level lowering is dif ferent from that of the last sea-level lowering. Also, in sediments older than 300 000 yr, a marked change in the abundance of sand and sedimentation rates is seen. We have no doubt that the event stratigraphy shown in Figs 4, 5 & 8 which encom-
Isotope stages
-
Wackestones
D
Packstones
passes the isotopic, sedimentological and seismic data is correct and that it can be explained in terms of global and regional causes. Jansen et a!. (1986) recognized a fundamental change in isotope fre quency at the 9-8 boundary defining a transition from lower to higher sea surface temperatures linked to perturbations of the orbital eccentricity cycle and leading to the establishment of more interglacial conditions in the period of isotope stages 1-8. This is clearly the case at Sites 820 and
Table I Change in sediment characters over three sea-level oscillations in top part of hole 820, north-east Australian margin.
Regression Lowstand Early transgression Late transgression Highstand
Sand
Mud
Carbonate
Magnetic susceptibility
High Very high Lower Low Low
Low Low Higher High High
High High Lower Low High
Low Low High High Low
32
P. J. Davies & F. M. Peerdeman
819, and the boundary at 65 m at Site 820 is there fore a globally driven phenomenon dividing colder from warmer environments, with distinct sedimen tological and seismic signatures. This event occurred at around 300 ka and is clearly a fundamental signature. Two groups working in the southern ocean recognized a change in palaeoceanographic conditions at around stage 13 (i.e. 110 m at Site 820), i.e. Kuijpers (1989) defined the period 19-13 as one of less intense circulation of Antarctic Inter mediate Water, and Nelson et a!. (1988) defined the period 13-9 as one of increased circulation actually leading to erosion on the Chatham Rise. The time frame of around 500 000 yr may therefore have coincided with at least a regional palaeoceano graphic reorganization. Why this should be repre sented in the shelf edge section at Site 820 is unclear. However, we do see a fundamental change in sediment style and seismic expression corre sponding to a change in isotope frequency and amplitude. The lower part of the section appears to be dominated by the obliquity signal whereas the upper may be the transitional change to the eccen tricity signal dominating in the period after 300 ka. The change from B2 to B1 may therefore represent both a change in water depth (perhaps shallowing) and a rise in temperature, this corresponding to the effects reported by Kujpers (1989) and Nelson eta!. (1988) all related to increased temperature gradi ents between tropics and poles. Many workers have defined changes at the B-M boundary (Stage 19) as a probably global event (Keany & Kennett, 1972; Prell, 1982; Imbrie et a!., 1984; Imbrie, 1985; Ruddiman et a!., 1986, 1989; Maasch, 1988); how ever, there is some uncertainty as to both the cause and/or the effect. There is, however, no doubting the global nature of the event. At Site 819, Alex ander & Kroon (1993) concluded that the data indicated a significant reorganization of the climate and oceanography at about the B-M boundary, and at Site 820, Peerdeman et a!. (1993) defined a major change in isotope frequency from obliquity domi nant below comparable with that occurring in the period after 300 ka. Summarizing therefore, on the slope immediately in front of the Great Barrier Reef, we can identify three sequences (1, 2 and 3), the boundaries of which occur at 65 m, and 14 5 m, and which are manifestations of globally driven processes. Within sequence 2 (at stage 13, 110 m) an event manifested elsewhere as a change in southern ocean intermedi ate water circulation resulted in the division of the
sequence into two parts. Alexander & Kroon (1993) defined the lower part (13-19) at Site 819 as shallower water because of the separation of plank tonic and benthic 180, whereas Peerdeman (1993) at Site 820 defined the upper part (9-13) as having low sedimentation rates (deeper water) and the lower part (13-19) as having high sedimentation rates (correspondingly shallow water). The data from Sites 820 and 819 are reconcilable, although the upper part of isotope sequence 2 at Site 819 is missing. This stratigraphy has profound implica tions for the formation of the Great Barrier Reef.
IMPLICATIONS FOR THE GREAT BARRIER REEF
The profile of the Great Barrier Reef
Seismic data in Grafton Passage show part of the external profile of the Great Barrier Reef (Fig. 3a&b). These also show a major break in slope at around 100 m, with the section above sloping more gently than that below. In addition, a relatively low angle terrace has developed below 165 m. A more complete profile obtained from submersible exami nation of the upper slope together with echo profiles is shown in Fig. 10. This shows clearly that the outer slope of the Great Barrier Reef can be divided into three parts: A between 0 and 100 m, a terraced slope with major changes of altitude at 25 m, 50 m (Fig. 10-1), 70 m and 100 m (Fig. 10-2); B from 100 to 200 m, a truly vertical wall with caves at around the 125 m and 160 m levels; C deeper than 200 m a talus slope (Fig. 10-4) lead ing to a smooth terrace which merges with the slope leading to the Queensland Trough. The talus slope and lower terrace are clearly depositional and represent the equivalent locations of Sites 819-821. The vertical wall is composed of laterally discontinuous ledges (Fig. 10-3) approxi mately 18 em thick separated by recessive areas half the thickness, the overall appearance of which is that of discontinuous bedding. There are no obvi ous structures in the wall indicating corals, as can be seen in vertical walls associated with raised reefs and as one would expect to see if the wall was composed of corals that had been eroded. We conclude therefore that the wall is not reefal. Above 100 m, there is a sequence of terraces at 2 5 m, 60 m and 85 m (Fig. 10-2) with steep walls extending
33
Origin of the Great Barrier Reef
2
3
4 Distance (km)
0
�
s
.c
15. " Q
0 40 80 120 160 200 240 280 320 360 400 440 480
2
3
4
5
6
7
Unit A equivalent RhodoReef
Unit A Stage 9
Unit C e quivalent
-------t--1 -------+--1
UnitBl
Stage 13 UnitB2 Stage 19
UnitC
Fig. 10. Cross-section of the front of Ribbon Reef No. 5 and the equivalent position of ODP Site 820. The positions of photographs 1-4 are shown on the cross-section. (1) The slope of the 50 m reef. (2) The very rough surface of the terrace above 100 m. (3) The ledges which form the wall. (4) The stand and rubble slope at the foot of the wall.
34
P. J. Davies & F. M. Peerdeman
from 0 to 15 m, from 50 to 60 m, and from 70 to 80 m. Submersible examination of the steep walls between 50 and 60 m (Fig. 10-2), and 70 and 80 m indicates conclusively that they are composed of reefal framework. The steep walls and terraces above 100 m are therefore constructional, repre senting reefs culminating in the present reef above 25 m. If earlier reefs exhibit the same shape as the modern one, then there must be four or perhaps five reefs between 0 and 100 m. The profile indicates clearly that the architecture is recessive to the west, in contradiction to the vertical wall and to the aggradational nature of the lower terrace, and sug gesting that the successive reefs are backstepping to the west. The Great Barrier Reef profile and the results from Sites 819-821
If the above interpretations are correct, then the depth of 100 m represents a major stratigraphical boundary separating reefal from subreefal. Such a boundary, defining the initiation of the main reef, would require a substantial change in environment, and it is proposed therefore that it is equivalent to the boundary of sequence A and sequence B at Sites 819-821, i.e. the time of the major climatic change. The potential correlation of the two data sets is shown in Fig. 10. The consequences of this interpre tation are that the Great Barrier Reef is thin and young, and that it owed its initiation to the climate change at 300 000 yr accompanying the change in orbital parameters. The subreef environment-analogue on the Fraser Island Shelf
As a corollary to the above interpretation, the subreef section would be equivalent to sequence B in Fig. 10 and would have formed in a colder environment, and could therefore be similar to that found today on the mid to outer shelf to the south of the Great Barrier Reef and documented in a cruise jointly organized by BMR and the Japan National Oil Corporation (Davies eta!., 1991). This analogue occupies the 40 km wide shelf to the east of Fraser Island (Fig. 11), which is dominated by Gardner Bank in water depths of 28-110 m on the mid to outer shelf. Although the main part of the bank occurs in water depths of 30-50 m, three morpho sedimentary zones can in fact be identified: 1 Zone 1-the western margin. Mixed terrigenous
and carbonate sediments dominate along the inner western edge of the bank (Fig. 11). The carbonate fraction is composed of benthic Foraminifera, car bonate rock fragments shell fragments and other unidentified bioclastic grains (Fig. 11-1). The sedi ments also include biogenic gravels composed of larger benthic Foraminifera such as Marginopora and Operculina(?), and pelecypods, gastropods, bry ozoans, barnacles and echinoid spines. 2 Zone 2-the Main Bank (Fig. 11). This occurs between 30 and 50 m of water where the faunal composition is dominated by coralline algae and corals (Fig. 11-2), but pelecypods, larger Foramin ifera (Operculina-Cycloclypeus?) Halimeda, bryo zoans, gastropods, echinoid spines, barnacles and branching articulated coralline algae are also very important. Platey and small massive corals, notably Montipora and Goniopora spps, grow in competi tion with both the red algae and extensive areas of fleshy algae. Several alternations of platey coral and coralline algae attest to the severe competition occurring in the environment. Further, dead corals are rapidly attacked by a whole community of borers and reduced to boulder- and pebble-sized fragments, which become coated by coralline algae and transformed into rhodoliths. 3 Zone 3-the eastern margin (Fig. 11). Between 50 and 110 m of water, the coralline algae are particu larly important and form extensive and thick rhod oliths which are pebble to cobble sized and red in living modern forms and white in dead forms (Fig. 11-3). Some living rhodoliths have smooth surfaces, but some have pinnacles on their surface. The nuclei of the rhodoliths are corals or benthic Foraminifera. The rhodoliths show the typical con centric structure and are frequently extensively bored. Other important components of the fauna include larger Foraminifera 2 em in maximum di ameter, Halimeda plates as large as 2 em diameter but in which the calcification is low, and bryozoans in a variety of forms including net, branching and encrusting types. This depth-related community of hard and soft algae, corals and minor players have built structures with relief above the surrounding sea-floor, which form hard rocky outcrops. Such structures, al though dominated by encrusting coralline algae and corals, are not reefs in the sense of coral reefs. They are mounds on which corals grow sparsely and become destroyed after death, although often incor porated in situ into a coral rubble by encrusting coralline algae. The final architectural feature is
(I) Highstand sub-tropical sedimentation
Depth (m)
0 ������������ Highstand
40 Quartz/carbonates
80 120 160 (2) Lowstand sub-tropical sedimentation
Rhodolith/foram facies
Large benthics; Operculina
Montipora and Goniopora plates;
and Marginopora
articulated corallines, halimeda
Encrusting corallines, benthic
Bryozoans, barnacles,
bryozoans, barnacles, echinoids
and planktic forams,
bivalves, gastropods and
encrusting corallines
bryozoans and abundant
echinoids
Intensive boating destruction
rhodoliths
5 km
Depth ( m)
0 ��-------, Lowstand
40
�
s.
80
""
Coral/coralline facies Thin facies periodically destroyed by storms
120 160
Widespread rhodolith/ forams facies-mixed benthic and planktics
J-------� 5 km
(3) Highstand sub-tropical sedimentation
$?
c)Q' s·
� tt)
�
::::
�-
:;._,
""
Depth (m )
0 ��======� 40
�
Highstand
80 120
quartz large benthics, barnacles facies
Coral/coralline/Halimeda bank facies
Distribution related to
Backstepping due to colonization
sea-level position as is
of low sea-level eroded
quartz content
highstand bank
160 �------� 5km
Fig. 11. Subtropical facies development on the shelf of southern Queensland as an analogue for the subreefal facies of the Great Barrier. Insets I, 2 and 3 represent different sea-level scenarios. Picture I defines the sediment facies west of Gardner Bank. Picture 2 shows the corals on Gardner Bank, and Picture 3 shows the rhodoliths forming on the sea-bed in water depths greatet than I00 m.
w Vl
36
P. J. Davies & F. M. Peerdeman
therefore that of a coral-algal rubble bank flanked seawards by a rhodolith facies and landwards by a mixed terrigenous-carbonate facies dominated by large benthic forams, barnacles and molluscs. This lateral facies relation defines growth at a particular high sea-level (Fig. 11-1), but we believe that it could be repeated vertically as a consequence of repeated sea-level oscillations (Fig. 11-2&-3) in a subtropical to temperate environment. Further, we suggest that such vertically superimposed coralline coral-foram-bryozoan dominated facies constitute the subreef facies of the Great Barrier Reef and the shelf equivalent to sequence B in Fig. 10.
CONCLUSIONS 1 The history of the Great Barrier Reef can be
understood from an analysis of its fore-reef sedi ments and from a knowledge of its morphology and architecture. 2 The Great Barrier Reef is thin and young. 3 The subreef is composed of related sedimentary facies which today occupy the mid to outer shelf to the south of the Great Barrier Reef. Such sediments are composed of sparse coral-coralline facies and rhodolith-foram-bryozoan facies. 4 Walthers's Law of Succession is obeyed in a vertical sense because of subsidence and in a lateral sense as a result of plate tectonics and climate change. 5 The initiation of the Great Barrier Reef was the result of a fundamental change in global climate approximately 300 000 yr BP, and as such is a regional expression of a global event. The same event must be seen elsewhere in the world's oceans. For example, an identical reefal profile occurs in the Atlantic Ocean at Belize. 6 The Great Barrier Reef has grown at least four (or perhaps five) times in the past 300 000 yr, respond ing to major sea-level oscillations. Each successive reef has grown in a position further shorewards than its predecessors. In a gross sense, this is an example of the backward growth defined by Davies (1983) for the modern reef, and is the direct consequence of high sea-level growth in a high-energy environ ment, coupled with frontal erosion and retreat in the lowstand and antecedent leeward growth in the following highstand. The Great Barrier Reef is therefore an analogue in shape, size and composi tion to reefs responding to successive sea-level oscillations with 100 000 yr periodicities. We know
from the modern reef that actual reef growth occurs for only a small fraction of that time (probably 10 000-15 000 yr maximum) associated with the highstand. 7 Corals may occur in the subreef section, but if our Gardner Bank analogue is correct, then they are subordinate to algae and indicate at best a marginal reef environment.
DISCUSSION
If our conclusion regarding the origin of the Great Barrier Reef has validity, then it would be likely that the event has expression elsewhere in the world's oceans and that, if global, such an event would have some effect on global ocean biochemis try. Regarding the possibility of reefs elsewhere growing at this time, we point out the almost identical profile exhibited by the Belize barrier reef system (James & Ginsburg, 1979), and also the similar morphology defined for Mayotte (Dullo et a!., this volume, pp. 219-236). It would be interest ing to speculate therefore that modern reef systems owe their origin to sudden initiation in stage 9 or thereabouts. If so, it is worthwhile speculating on the likely eflects such an event may have had on climate and oceanography. Reefs are substantial carbon sinks and it is clear that variations in global calcification rates may influence aquatic and atmo spheric carbon dioxide levels. A prevalent view amongst the uninformed is that calcification will act to depress atmospheric carbon dioxide, but Kinsey & Hopley (1991) and D. Kinsey (personal commu nication) have shown in fact that calcification leads to an increase in oceanic pC02 with a likely increase in atmospheric C02• However, if current produc tion is an indication, then it is worth noting that the sudden effect of global reef turn-on was the removal of 900 x 106 t of C yc 1, which has continued inter mittently (highstands) for the last 300 000 yr or so. Although only a small percentage of the annual oceanic carbon balance (Crossland et a!., 1991), this nevertheless represents a gigantic figure taken on the time-scale of the last 300 000 yr. This occurred comcomitantly with a major temperature rise, the effect of which must have been decreased solubility of C02 in sea water, and the injection of the C02 into the atmosphere. This would have directly effected a further temperature increase. Global warming began with the advent of the global reef systems in or just before isotope stage 9.
Origin of the Great Barrier Reef POSTSCRIPT
This paper was presented as a lecture at the Pacific and Indian Ocean Carbonate Platform Meeting in Sydney, Australia, in July 1995. In a few months, an international consortium organized by P.J.D. will attempt to drill through the Great Barrier Reef in three locations and we will then know whether the predictions made so confidently above are in any way approximations of the truth or are· incorrect. No doubt, all concerned will be kept posted.
REFERENCES
ALEXANDER, I. & KROON, D. (1993) Isotopic and magnetic susceptibility variations in ODP Hole 819, Great Bar rier Reef. In: Proceedings of the Ocean Drilling Program, Scientific Results, 133 (Eds McKenzie, J.A., Davies, P.J. & Palmer-Julson, A.). Ocean Drilling Program, College Station, TX. CROSSLAND, C.J., HATCHER, B.C. & SMITH, S.V. (1991) Role of coral reefs in global ocean production. Coral Reefs, 10, 55-64. DAVIES, P.J. (1983) Reef growth. In: Perspectives on Coral Reefs (Ed. Barnes, D.J.), pp. 69-106. Australian Insti tute of Marine Science, Clouston, Canberra, A CT. DAVIES, P.J. (1991) Origins of the Great Barrier Reef. Search, 23, 193-196. DAVIES, P.J., SYMONDS, P.A., FEARY, D.A. & PIGRAM, C.J. (1987) Horizontal plate motion: a key allocyclic factor in the evolution of the Great Barrier Reef. Science, 238, 1 697-1700. DAVIES, P.J., SYMONDS, P.A., FEARY, D.A. & PIGRAM, C.J. (1988a) Facies models in exploration-the carbonate platforms of northeast Australia. Australian Petroleum Exploration Association J., 28, 123- 143. DAVIES, P.J, SYMONDS, P.A., FEARY, D.A. & PIGRAM, C.J. (1988b) RIG SEISMIC Research Cruises 4 and 5: Northeast Australia post-cruise report. Report No. 281, Bureau of Mineral Resources of Australia, Canberra, ACT. DAVIES, P.J., SYMONDS, P.A., FEARY, D.A. & PIGRAM, C.J (1989) The evolution of the carbonate platforms of Northeast Australia. In: Controls on Carbonate Platform to Basin Development (Eds Crevello, P., Sarg, J.F., Read, J.F. & Wilson, J.L.) Spec. Pub!. Soc. econ. Paleont. Miner., Tulsa, 44, 233-258. DAVIES, P.J., TsuJI, Y., SYMONDS, P. et a/. (1991) Tropical and temperate carbonate environments-the effects of sea level, climate and tectonics on facies development. Aust. Geol. Surv. Org. Record 1991188. DAVIES, P.J., McKENZIE, J.A., PALMER-JULSON, A. et a/.. (1992) Proceedings of the Ocean Drilling Program, Ini tial Reports, 133. Ocean Drilling Program, College Station, TX. DULLO, W.-CH., CAMOIN, G.F., BLOMEIER, D. et a/. (1997) Morphology and sediments of the fore-slopes of May otte, Comoro Islands: direct observations from a sub mersible. This volume, pp. 219-236.
37
ERICSON, E.K. (1976) The Capricorn Basin. In: Economic Geology of Australia and Papua New Guinea, Vol. 3: Petroleum (Eds Leslie, R.B. et a!..), Australas. Inst. Mining Metal!., Monogr. 7, 464-473. FEARY, D.A., SYMONDS, P.A., DAVIES, P.J., PIGRAM, C.J. & JARRARD, R.D. (1993) Geometry of Pleistocene facies on the Great Barrier Reef outer shelf and upper slope seismic stratigraphy of Sites 819, 820 and 821. In: Proceedings of the Ocean Drilling Program, Scientific Results, 133 (Eds McKenzie, J.A., Davies, P.J. & Palmer-Julson, A.), pp. 327-351. Ocean Drilling Pro gram, College Station, TX. HOPLEY, D.H. (1982) The Geomorphology of the Great Barrier Reef Quaternary Development of Coral Reefs. Wiley, New York. IMBRIE, J. (1985) A theoretical framework for Pleistocene ice ages. J. geo/. Soc. London, 142, 417-432. IMBRIE, J., SHACKLETON, N.J., PISIAS, N.G. et a/. (1984) The orbital theory of Pleistocene climate: support from a revised chronology of the marine 018 record. In: Mi lankovitch and Climate (Part 2) (Eds Berger, A.L., Imbrie, J., Hays, J., Kukla, G. & Saltzman, B.), pp. 269-305. D. Reidel, Dordrecht. JAMES, N.P. & GINSBURG, R.N. (Eds) (1979) The Seaward Margin of the Belize Barrier and Atoll Reefs. Spec. Pubis. Int. Ass. Sediment. No. 3. Blackwell Scientific Publications, Oxford. JANSEN, J.H.F., KUYPERS, A. & TROELSTRA, S.R. (1986) A mid Brunhes climatic event: long term climatic changes in global atmosphere and oceanic circulation. Science, 232, 619-622. KEANY, J. & KENNETT, J.P. (1972) Pliocene-early Pleis tocene palaeoclimatic history recorded in Antarctic Sub Antarctic deep-sea cores. Deep-Sea Res., Part A, 19, 529-548. KINSEY, D. & HOPLEY, D. (1991) The significance of coral reefs as global carbon sinks-response to Greenhouse. Palaeogeogr. Palaeoclimatol. Palaeocol., 89, 363-377. KuiJPERS, A. (1989) Southern ocean circulation and global climate in the middle Pleistocene (early Brunhes). Pa/aeogeogr. Pa/aeoclimatol. Palaeoecol., 76, 67-83. LLOYD, A.R. (1973) Foraminifera of the Great Barrier Reef bores. In: Biology and Geology of Coral Reefs, Geology, Vol. I (Eds Jones, O.A. & Endean, R.), pp. 34 7-366. Academic Press, New York. MAASCH, K.A. (1988) Statistical detection of the mid Pleistocene transition. Clim. Dyn., 2, 133-143. MAXWELL, W.G.H. ( !968) Atlas of the Great Barrier Reef Elsevier, Amsterdam. McKENZIE, J.A., DAVIES, P.J. & PALMER-JULSON, A. (Eds) (1993) Proceedings of the Ocean Drilling Program, Sci entific Results, 133. Ocean Drilling Program, College Station, TX. MONTAGGIONI, L.F. & VENEC-PEYRE, M.-T. (1993) Shallow water foraminiferal taphocoenoses at site 821: implica tions for the evolution of the central Great Barrier Reef, northeast Australia. In: Proceedings of the Ocean Drill ing Program, Scientific Results, 133 (Eds McKenzie. J.A., Davies, P.J. & Palmer-Julson, A.), pp. 36 5-378. Ocean Drilling Program, College Station, TX. NELSON, C.S., HENSY, C.H. & DUDLEY, W.C. (1988) Qua ternary isotope stratigraphy of hole 59 3, Challenger Plateau, South Tasman Sea: preliminary observations
38
P. J. Davies & F. M. Peerdeman
based on foraminifers and calcareous nannofossils . In: Initial Reports of the Deep Sea Drilling Project, 90 (Eds Kennett, J.P. & Von der Borch, C.C.), pp. 1413-1424. US Government Printing Office, Washington, D C. PALMIERI, v. (1971) Tertiary subsurf ace biostra tigraphy of the Capricorn Basin. Report No. 52, Geological Survey of Queensland. PALMIERI, V. (1974) Correlationand environmental trends of the subsurf ace Tertiary Capricorn Basin. Report No. 86, Geological Survey of Queensland. PEERDEMAN, F.M. (1993) The Pleistocene climatic and sea-level signature of the northeastern Australian conti nenta l margin. PhD thesis, Australian National Univer sity, Canberra, ACT. PEERDEMAN, F.M., DAVIES, P.J. & C H IV AS A.R. (1993) Isotopic and trace-element indicators of palaeoclimate and sea-level, Site 820. In : Proceedings of the Ocean Drilling Progra m, Scientific Results, 133 (Eds McKen zie, J.A., Davies, P.J. & Palmer-Julson, A .) , pp. 163173. Ocean Drilling Program, College Station, TX. PRELL, WL. (1982) Oxygen and carbon isotope stratigra phy of the Quaternary of Hole 502B: evidence for two modes of isotope variability. In: Initial Reports of the Deep Sea Drilling Project, 68 (Eds Prell, W.L. & Gard-
ner, J.V.), pp. 4 5 5-464. US Government Printing Of fice, Washington, DC. RICHARDS, H.C. & HILL, D . (1942) Great Barrier Reef Cores, 1 926and 1 937-Descriptions, Analysesand Inter pretations. Report No. 25, Great Barrier Reef Commit tee. RUDDIMAN, W.F., MACINTYRE, A. & RAYMO, M. (1986) Matuyama 41 ,000 year cycles: North Atlantic Ocean and northern hemisphere ice sheets. Earth planet. Sci. Lett. , 80, 117-129. RUDDIMAN, W.F., RAYMO, M.E., MARTICNON, D.C., CLEM ENT, B.M. & BACKMAN, J. (1989) Pleistocene evolution: northern hemisphere ice sheets and North Atlantic circulation . Paleoceanography, 4, 353-412. SHACKLETON, N.J. & HALL, M.A. (1989) Stable isotope history of the Pleistocene at ODP Site 677. In: Proceed ings of the Ocean Drilling Program, Scientific Results, I l l (Eds Becker, K. & Sakai, H.}, pp. 295-316. Ocean Drilling Program, College Station, TX. TANNER, J.J. (1969) Theancestral Great Barrier Reefin the GulfofPapua. UN Economic Commission for Asia and the Far East, Mineral Resources Development Series, 41.
Spec. Pubis int. Ass. Sediment. (1998) 25, 39-67
Development and demise of mid-oceanic carbonate platforms, Wodejebato Guyot (NW Pacific) G. F. C A M O I N*, A. A R N A U D -VANNEAU t, D D B E R G E RS E Nt, P. E N O S§ and P h . E B R E N* .
.
*CEREGE, CNRS UMR 6536, Universite Aix-Marseille III, B.P. 80, 13545 Aix-en-Provence cedex 4, France; tinstitut Dolomieu, Universite J Fourier, Rue M Gignoux, 38031 Grenoble cedex, France; tUniversity ofSydney, Department of Geology, NSW 2006, Sydney, Australia; and §University ofKansas, Department of Geology, 120 Lindley Hall, Lawrence, KS 66045-4974, USA
ABSTRACT
A model for the development and the demise of mid-oceanic carbonate platforms is based on a detailed sedimentological, palaeontological, seismic, and geochemical study of Wodejebato Guyot (Marshall Islands, NW Pacific). After a prolonged period of subaerial exposure (a few million years) characterized by the pedogenic alteration of the volcanic basement, the flooding of the platform resulted in the accumulation of skeletal sand piles near the shelf margin, rimming a volcanic island. The overlying sequence is a series of retrogradational and then progradational and aggradational sand shoals that progressively flooded the central part of the guyot. Atoll features at Wodejebato Guyot include well-differentiated depositional elevated platform rims, partly composed of early cemented organic frameworks, enclosing a central lagoon, and flanked by steep talus leading down to volcanic slopes. The subsequent development of the carbonate platform includes two successive periods of emergence: a first limited fall in sea-level and a second deeper fall in sea-level before the final drowning of the carbonate platform during Maastrichtian time. The last platform carbonate sequence corresponds to a wedge deposited during a sea-level lowstand on the outer ridge, whereas the inner part of the platform was subaerially exposed and probably karstified. After its demise, the carbonate platform sank to deep water, between the aragonite and the deeper calcite saturation limits, resulting in the dissolution of aragonite and concomitant precipitation of calcite cements, possibly throughout the Cainozoic to the present. Post-drowning deposits include phosphate-manganese crusts that formed from the late Palaeocene to the middle Eocene. The drilling results of Leg 1 44 and other data indicate that there were at least three major episodes of carbonate platform drowning within the tropical Pacific, i.e. Albian, late Maastrichtian and middle Eocene. We propose that these events were caused by short-term regression-transgression cycles connected to short-term cooling events and deep oceanographic changes.
INTRODUCTION
Guyots are flat-topped volcanic seamounts and drowned atolls submerged to depths of the order of 1 500 m; they are most common on Mesozoic-age Pacific oceanic crust. Dredging of Pacific guyots commonly recovered mid-Cretaceous rudist lime stones (Hamilton, 1 9 56; Matthews et al. , 1 97 4; Konishi, 1 989; Gr6tsch & Flugel, 1 992) but did not allow the reconstruction of a comprehensive model
of carbonate platform development. Key questions remained, such as how a tropical carbonate plat form started to grow on the volcanic substrate, why it should suddenly cease to keep pace with subsid ence after constructing a massive carbonate cap up to 1 000 m thick and tens of kilometres in diameter, or why the relative ages of the underlying volcano and of the shallow-water carbonate cap appeared to
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
39
40
G. F. Camoin et a!.
lack the coherent patterns predicted from hot-spot and subsidence theories (ODP Leg 1 44 Shipboard Scientific Party, 1 993). 0DP Legs 1 43 and 1 44 were devoted to answering these and other questions by systematic drilling of the carbonate platforms and underlying seamount volcanics of representative guyots within different geographical areas of the Mid-Pacific Mountains and Japanese Seamounts (Fig. 1 ) . Wodej ebato Guyot ( 1 2.0° N, 1 64. 9oE), formerly Sylvania Guyot, is by far the best documented guyot drilled during the two legs. This guyot is located c. 74 km north-west of Pikinni Atoll (a.k.a. Bikini) in the northernmost part of the Ralik Chain in the Marshall Islands (Mid-Pacific Mountains; Fig. 1 ). A volcanic ridge of 20 km long and about 1 500 m deep connects Wodej ebato and Pikinni. The first observations on Wodejebato came from Operation Crossroads (Emery et al. , 1 9 5 4); and initial subsurface interpretations were based on limited seismic data (Schlanger et al., 1 9 87; Berg ersen, 1 993). Early dredges on this guyot recovered only basaltic rocks and tuff breccias, the cracks of which were filled with phosphatized, planktonic foraminifers of Eocene age (Hamilton & Rex, 1 9 59; Duennebier & Petersen, 1 982). Four later dredges recovered shallow-water limestones attributed to
the Albian and Campanian-Maastrichtian (Lincoln et al., 1 993). During Leg 1 44, five sites were drilled to determine the facies pattern from the centre of the broad guyot (Site 873) to the outer rims to the north and east (Sites 874 and 877 on the seaward summit of the inner ridge, and Sites 8 7 5 and 876 on the outer ridge). Site 869 was drilled on the slope at 4826-m depth during Leg 1 4 3 .
MORPHOLOGY AND SEISMIC DATA
The summit of Wodejebato Guyot, at 1 3 3 5-m depth, is about 43 km long and increases in width from less than 1 2 km in the south-east to more than 25 km in the north-west. Four spurs project from the edifice and, along with the volcanic spur attach ing Wodej ebato Guyot to Pikinni Atoll, give the guyot a distinct 'starfish' appearance. The flanks of the guyot may be divided into a steep upper slope (20-24o) and a more gently inclined lower slope (about 7 °) below 2500 m depth. On seismic profiles, the acoustic basement shallows toward the centre of the summit plateau to form a basement high; it is c. 50 m higher at the 'lagoon' Site 873 than on the elevated platform rims (Bergersen, 1 995) (Fig. 2). Two concentric elevated platform rims lie along the
Fig. 1. Location of Leg 144 drilling sites and ship track of the JOIDES Resolution.
41
Development and demise of mid-oceanic carbonate platforms " " ""
;::. ""
"-' i-
[f) ""' " 0:
u.;
t
A
[f)
u.; i-
�'J') " 0: "-' i-
[f)
A'
-$ Ql
s
'J:j Ql > � b ;;..., �
� I
0
� E-o
2.6 0 0 '0:1' """
Time
(Z)
�Channcls1
Fig. 2. Migrated seismic profile across the elevated platform rims of Wodejebato Guyot and areal distribution of the carbonate platform.
shelves formed by the northern and north-eastern flank ridge, and along the ridge extending south-east towards Pikinni Atoll (see Bergersen, 1 993, 1 995; Camoin et a!. , 1 99 5). These elevated platform rims broaden across areas where the shelf margin widens (e.g. adjacent to the flank ridges) and are generally separated by a trough about 50 m deep (Fig. 2). As noted by Bergersen ( 1 993), the inner ridge appears at the edge of the summit plateau in all seismic profiles, with the exception of the south flank where faulting has removed portions of the original edifice. The maximum height of this ridge above the synchronous lagoonal deposits is 48 m, although the average height is 36 m (Bergersen, 1 995; Camoin et a!., 1 99 5); the top of the ridge varies in width from less than 1 00 m to a maximum of 800 m. The summit of the outer ridge is about 35 m lower than the inner ridge (Fig. 2). It is 200700 m wide and rises to about 50 m above the trough separating the two ridges. Slopes on the landward side of the inner ridge vary from less than 1 up to 1 0 whereas those on the seaward side are o
o,
slightly steeper (up to 1 5 ° ) and closely comparable with that of the outer ridge. The top of platform carbonates at Site 873 is 1 3 and 3 3 m lower than on the inner ridge (Sites 8 7 4 and 8 77) and 2 1 and 1 1 m higher than on the outer one (Sites 875 and 876) (Fig. 3). At the top of the guyot, a surface with perhaps a metre or more of relief with unconnected depres sions and bio-encrusted knobs was observed over an appreciable area of the sea-floor during video surveys on the perimeter of Wodejebato.
STRATIGRAPHY
Lithostratigraphy
The four major lithostratigraphical units drilled on Wodejebato Guyot are, from the top to the bottom (Premoli Silva et a!., 1 993) (Fig. 3): (i) manganese crust and manganese-phosphate coated limestone conglomerate; (ii) platform carbonates; (iii) ferrugi-
42
G. F.
Camoin et al.
Site 873 Site 877 mbsl
Site 874
1350
Site 876
Site 875
Algal-rudist
1400
II a
Bdst Rudst Fltst Gst 1450
Foraminiferlid
Wckest Pckst Gst
IV v
1550
Foraminifer-
Algal-coral Bdst
rudist-algal lllb
skeletal
1500
Algal-rudist Gst-Adst
VI
lid
Foraminifer
Gst Skeletal coral Pckst
skeletal
Ferrvglnous
Clay & claystone
lie
Gst
Basalt Volcan� breccia
[ill] D.S. 1
CJ D.S. 2
� D.S. 3
CJ D.S.4
..
Mn crust
..
Pelagic cap
Fig. 3. Lithostratigraphy and depositional sequences (DS, see the text) on Wodejebato Guyot (Sites 873-877). Numbers in the stratigraphical columns refer to lithostratigraphical units.
nous clay, claystone, and extremely altered vesicu lar basalt; and (iv) basalt and volcanic breccia. 40Ar/39 Ar ages obtained on the drilled basalts range from 78.4 to 85 Ma (Pringle & Duncan, 1 995). The basalts are reversely magnetized and attributed to Chron 3 3 R of early Campanian age (79 Ma) (Na kanishi & Gee, 1 995). Drilled volcaniclastic breccia at Site 869 (Sager et al., 1 993) and dredged material (Lincoln et al. , 1 993) suggest that Wodejebato expe rienced an earlier, probably Cenomanian, volcanic activity. Biostratigraphy
The oldest marine sediments underlying the plat form carbonates consist of black clays that contain a well-preserved late Campanian calcareous nano flora (CC22 biozone, around 76 Ma; see Erba et al., 1 995). Biostratigraphy of platform carbonates relies mainly on larger benthic foraminifers that constitute successive assemblages including Pseudorbitoides cf. trechmanni, Sulcoperculina sp. and Asterorbis sp., and finally Omphalocyclus macroporus (Fig. 4), rang-
ing in age from the late Campanian to the Maastrich tian (see Premoli Silva et al. , 1 995). The rudist assemblages recorded in these limestones include ra diolitids (Distefanella mooretownensis, Distefanella sp.) and caprinids (Mitrocaprina sp., Coralliochama orcutti, Coralliochama sp. , Plagioptychus aff.fragilis, P. aff. minor, Antillocaprina sp.), which have been reported in Campanian-Maastrichtian strata from Jamaica, Mexico, Cuba and California (see Camoin et al. , 1 995). Sr-isotope data indicated that the major part of the carbonate sequence on the inrter ridge may be attributed to the Maastrichtian, whereas Campanian values were measured for the lowermost part of the sequence (see Wilson et al., 1 995). Solution cavities occurring in platform carbon ates are partly filled by pelagic sediments that contain late Maastrichtian, early late Palaeocene, and early Eocene planktonic foraminifers (Erba et al., 1 995). The manganese crust and manganese phosphate coated conglomerate which cap the shallow-water limestones contain pelagic sediments ranging in age from the late Palaeocene to the middle Eocene (Erba et al., 1 995) (Fig. 4).
Site 873 Depth {mbsl)
Site 877
1350
Pal. Eo
Site 874 Site 876
Site 875
1400
.g ;:s
c
�
.!!! .E 0 ·;:::
�
�� � ;:: l'l...
11450
l} ;:s
:E
c;;· ""
� ;:s
1500
�
�;::
v
r;·
1550
G. gr gansseri +
Occurrence
1<<<<1
8
VI
g. ;::
Su/coperculina
I Distribution
Sequence I
CJ
Pseudorbitoides
I
Distribution
Sequence II
Omphalocyc/us
Orbitoides sp.
+-First occurrence
Occurrence
t<<;:<j
Sequence Ill
Lepidorbitoides •
�
Vida/ina sp. cf. V. hispanica
Distribution
SequenceiV
.
!
&•
Distribution
11ll1111ill
� (i;" � iS"
s., a
�"'
Pelagic cap
Fig. 4. Biostratigraphy, ages and depositional sequences (DS) on Wodejebato Guyot.
""" \.;J
44
G. F Camoin et al. WEATHERING PROFILES OF VOLCANIC SUBSTRATE AND INITIAL MARINE DEPOSITS
The clay units that cap the volcanic basement range in thickness between 7. 3 and 1 4 m on the inner ridge and up to 1 9.2 m at the 'lagoon' Site 873 (Fig. 3). The lack of a distinctive weathered horizon on the outer ridge may suggest either very brief subaerial exposure or complete erosion of the soil before the deposition of the carbonate unit (Enos et a!., 1 99 5a). Clay units consist of dusky red to reddish brown claystone breccias and claystones corresponding to a typical volcanic island weather ing profile of the basalt presumably accumulated in erosional depressions (Holmes, 1 995). The contact between the basalt and the clay units is gradational. In contrast, there is no evidence from core or logs of a gradational contact between the clay and the platform carbonates (Premoli Silva et a!. , 1 993). The degree of alteration of the basalt decreases downward and the texture of the former igneous breccia or altered basalt, including laths of plagio clase and vesicles infilled by white patches of zeo lites, is locally preserved. At Site 873, a gradation from mottled clays downward into clays over printed on basalt textures and into fresh basalts indicates a lack of transport in the soils (Holmes, 1 99 5). On the inner ridge, the formation and preserva tion of grey to black organic-rich clays with calcar eous nanofossils and larger foraminifers represent the first influences of marine waters (e.g. Site 877; see Camoin et a!., 1 995). The high sulphur content of these clays probably reflects bacterial reduction of sulphates provided by marine waters. Preserva tion of kerogen-type woody material suggests the prevalence of low-energy conditions; on the other hand, the preservation of clay and the lack of any reworking at the base of marine deposits suggests either a quiet, shallow-marine environment during initial submergence and/or rapid burial.
Depositional sequences and facies
On small mid-oceanic carbonate platforms, trans gressive systems tracts (TST) and highstand systems tracts (HST) prevail. Lowstand systems tracts (LST) are poorly developed because the steep slopes below the slope break allow little space for carbonate pro duction and accumulation. By integrating seismic, well-log and core data into a sequence framework, the platform carbonates of Wodej ebato Guyot may be divided into four depositional sequences that contain five environmentally controlled foraminifer assemblages (see Arnaud-Vanneau et a!., 1 995) (Fig. 4).
Depositional sequence 1 (DS-1; late Campanian) DS- 1 is the thinnest sequence ( 1 9 m thick at Site 874) and represents the colonization of the volcanic substrate by platform carbonates. It pinches out on the flanks and onlaps the central part of the guyot, where it forms a thin horizon at the base of the carbonate sequence. Seismically, it is a series of stacked reflectors truncating against the emerged volcanic basement and corresponding primarily to a TST (Fig. 5). The facies primarily consist of white, pink and reddish yellow, fine-grained, lightly cemented skel etal packstones and grainstones, locally interlayered with wackestones at Site 873. Major components include red algae (fragments, crusts and rhodoliths of corallinaceans, solenoporaceans and peyssonne liaceans; up to 65% of the skeletal components), benthic foraminifers (Pseudorbitoides pa1aeoeco1o gica1 assemblage V in Fig. 4-and miliolids, lituo lids and textulariids; up to 23% of the skeletal components), fragments of corals, rudists (radio litids) and other bivalves (inoceramids and pycn odonts). There are a few fragments of green algae ( dasycladaceans, codiaceans), gastropods, bryozo ans and echinoderms. Calcisphaerulids and plank tonic foraminifers were noted in one interval which could correspond to a maximum flooding surface (Arnaud-Vanneau et a!., 1 995). -
SHALLOW-WATER PLATFORM CARBONATES
The thickness of platform carbonates varies from 82.2 m at the central Site 873 to a maximum of 1 8 3 m on the inner ridge, whereas 1 26 and 1 4 5 m were measured at the outer ridge sites (Fig. 3).
Fig. 5. (Opposite) Steps in the depositional history of
Wodejebato derived from the migrated seismic profile A'-A (modified after Arnaud-Vanneau et al., 1 995). (a) TST of depositional sequence 1 ; (b) TST of depositional sequence 2; (c) sequence boundary Mal truncating the previous HST of depositional sequence 2; (d) last HST deposits of depositional sequence 3; (e) erosional surface of sequence boundary Ma2.
45
Development and demise of mid-oceanic carbonate platforms
� (a)
E 1.8
Ql
'"'
Qi >
�
> co
� � 2.1
Central basement -----� high (volcanic) Volcanic �� --l o------; -- --------c;= -::-debris flow (7) -:-:-:-----:_,., _ _ """"--/ Volcanic flow T ST
r-- ----
��--,��-;-:-_..,... - �"'----;.:�.;-�· "'··· --·-2·--� �,:.:;r.�::-;:·�.r-.;.;� ·�:"f�":.: :�-::..-i.:: �,:.�
/
'"
___....��:
/
·.
0
� (b)
E 1.8 Ql
·;:;
Qi
>
�
'"' > co
� � 2.1
0
f-
� (c) Ql 1.8
E
'"'
Qi
> co
.::
> co
� � 2.1 0
f-
�(d) Ql 1.8 f-----
E
·.;::; Qi >
lnfill of erosional depression formed during last sea-level fall -
------
Minor aggradation
�
> co
� � 2.1 0
Last stage of 'reef' growth
Sequence boundary (HYP1l
46
G. F. Camoin et a!.
The fossil content and sedimentological criteria imply a shallow-marine depositional environment with moderate- to high-energy conditions inter preted as sand shoals (Camoin et a!., 1 995), both on the inner ridge and in the central part of the guyot. Vague oblique and low angle cross-lamination, as well as the imbrication of large clasts, suggests a sedimentary regime dominated by currents.
Depositional sequence 2 (DS-2; Maastrichtian) DS-2 ranges in thickness from 60 m at Site 873 to 1 4 5 m at Site 8 77; it is absent on the outer ridge (Fig. 3). The sequence boundary at the base of DS-2 is marked by the disappearance of Pseudorbitoides (Arnaud-Vanneau et a!., 1 995) and by a pronounced peak in resistivity related to a sharp increase in cementation (Camoin et a!., 1 99 5). This sequence includes a series of retrogradational reflectors interpreted as representative of an episode of back stepping of the platform margin onto the central basement high (TST), overlain by a series of stacked horizontal reflectors characterizing the aggradation of the carbonate platform (HST) (Fig. 5). The ero sional surface truncating the top of DS-2 pre-dates the first occurrence of Omphalocyclus (Arnaud Vanneau et a!. , 1 995).
Central part of the guyot. Grey, skeletal packstones, grainstones and wackestones contain fragments of bivalves (including rudists, which may represent up to 77% of the skeletal components), echinoderms, sparse corallinacean algae and ostracods (Fig. 6A). The larger benthic foraminifers, Sulcoperculina and Asterorbis, are rare, whereas Dicyclina, miliolids, textulariids and rotaliids are common. A peak in the abundance of green algae (Salpingoporella and Terquemella) and planktonic foraminifers seem ingly coincides with a maximum flooding surface (Arnaud-Vanneau et a!., 1 995). An abundance of pyrite results from sulphate reduction soon after deposition. At the top of DS-2, subaerial exposure is indicated both by sedimentological criteria such as rhizoliths and alveolar texture (Fig. 6D) and by isotopic excursions (see 'Diagenesis' section). This interpretation is strengthened by the occurrence of considerable yellow staining, commonly in solution pores, which may be iron oxides (Enos et a!. , 1 995b). The depositional environment of this platform interior sequence is interpreted as sheltered, locally restricted, but affected by storm deposition, espe-
cially during the early stage of carbonate platform development.
Inner elevated platform rim. DS-2 comprises three successive units that reflect a shallowing-upward sequence: 1 Poorly cemented skeletal grainstones and pack stones rich in rudist, algal, echinoderm and bivalve fragments; the foraminifer assemblage is primarily composed of Asterorbis and Sulcoperculina (palaeo ecological assemblages III and IV in Fig. 4). These bioclastic units are similar to the sand shoal facies of DS- 1 and are interpreted as a TST. 2 Algal-rudist grainstones and rudstones, with local inclined lamination contain a few packstone layers and rudist-coral frameworks (TST). Core recovery and Formation MicroScanner (FMS) imagery sug gest that these facies are interbedded, with a general 'saw-tooth' pattern of resistivity related to differen tial cementation. Major components of skeletal grains include rudists (radiolitids and caprinids), benthic foraminifers (orbitoids: Sulcoperculina and Siderolites; miliolids: Triloculina and Quinquelocu lina; discorbids; lituolids: Ammobaculites; rotaliids; and small textulariids) and red algae (corallinaceans, solenoporaceans, peyssonneliaceans) which repre sent respectively 5 5%, 2 1 % and 1 3% of the skeletal components at Site 877. Rudists are generally frag mented, except the small radiolitids, which are ap parently in life position and form small encrusted clusters. Other grains include fragments of ino ceramids, corals (hexacorals and a few octocorals), echinoids, gastropods and green algae (dasycla daceans). Although the existence of a true framework is dif ficult to ascertain in cores, some recovered intervals seemingly include in situ framebuilders. At Site 8 74, they consist of tabular coral colonies (hexacorals; octocorals: Polytremacis) and a few stromatopor oids, whereas at Site 877, rudists are more abundant and include clusters of small radiolitids replaced up ward by loosely packed caprinid assemblages. Red algae (corallinaceans, peyssonneliaceans and, to a lesser extent, solenoporaceans), may form rhodoliths or encrustations, up to few millimetres thick. The abundance of rhodoliths at the base of the sequence suggests that these organisms may have played a role in the stabilization of a previously mobile substrate, thus promoting larval attachment of framebuilders and initiating framework development. Fossils and sedimentological criteria suggest that these two units were deposited in a shallow-marine ·
tl "'
� .g � "' i:!.
� � l:l..
� �
c;· "'
� �
.�
��
;:::;·
8 ;.
<::> � � ;;; � � s, <::>
�
"'
Fig. 6. Thin-section photomicrographs of skeletal-foraminifer packstone and wackestone occurring in the central part of the guyot (Site 873). Primary pore-spaces and solution cavities are lined by bladed or equant pyramids of limpid calcite with uniform extinction, generally finely crystalline (white arrows). The local occurrence of geopetal internal sediments in solution cavities should be noted (B & C). D shows alternation of spar cements and micrite around a probable mould of plant rootlet.
.j>. -.]
..
48
G. F. Camoin et al.
environment, probably less than 1 0 or 20 m deep. Evidence of turbulent waters includes abrasion of shell fragments and the prevalence of grainstones and rudstones. The abundance of grainstones within the recovered cores and corresponding downhole logging signatures strongly suggests that the initial ridge of the atoll consisted of skeletal sand shoals with patchy organic frameworks. These frameworks were probably not wave-resistant structures because the organic binding was weak, but rigidity and sta bility may have been provided by early marine ce mentation (see 'Diagenesis' section). 3 The top of DS-2 is interpreted as a late HST and comprises very pale brown to white skeletal grain stones, packstones and floatstones, interlayered with sparse grey burrowed, dolomitic wackestones. The contact between these facies is typically an omission surface with borings, local erosion and reworking of intraclasts implying early lithification. Major components of skeletal packstones and floatstones consist of leached micritized fragments ofgastropods, rudists (radiolitids and scarce caprin ids) and other bivalves (inoceramids and probable pycnodonts); in situ radiolitid clusters occur locally. Benthic foraminifers include miliolids (Istrilocu lina), rotaliids, discorbids and, to a lesser extent, orbitoids and lituolids; placopsilinids encrust large bioclasts. Other grains include, in order of decreas ing abundance: fragments of green algae ( dasycla daceans: Terquemella), red algae (corallinaceans and peyssonneliaceans), corals (hexacorals and oc tocorals) and echinoderms. Calcisphaerulids and planktonic foraminifers (Rugoglobigerina sp.) occur in trace abundance. Minor dolomitic wackestone beds are character ized by the abundance of fenestrae that may form irregular networks of vertically elongated tiny tubes reminiscent of fluid escape structures. Skeletal content is low and limited to small gastropods (cerithids), benthic foraminifers (peneroplids, oph thalmiids, discorbids, textulariids, cf. Cuneolina, fragments of orbitoids), smooth-shelled ostracods, green algae (Terquemella) and few abraded frag ments of corals, red algae and rudists. A shallow-marine depositional environment, probably a few metres deep, may be inferred from the fossil content and the sedimentological criteria of the foraminifer-gastropod wackestones. The fine grained matrix and the lack of current indicators imply quiet-water conditions. The abundance of fenestrae is consistent with a shallow subtidal, inter tidal or supratidal environment. Periods of emer-
gence may be deduced from early dissolution of shells and the local reworking of caliche lithoclasts. The local prevalence of depauperate assemblages of very small foraminifers, smooth-shelled ostracods and gastropods (cerithids) may indicate temporary episodes of restriction and inimical ecological conditions (e.g. low illumination or oxygenation, temperature and salinity fluctuations, changes in nutrient content or turbidity). Skeletal packstones and floatstones interlayered with wackestone beds may have been redeposited from the outer debris shoals by currents.
Depositional sequence 3 (DS-3; Maastrichtian) DS-3 has been identified across the entire platform and varies in thickness from a minimum of 1 0 m on the inner ridge to a maximum of 9 5 m on the outer ridge; it is 30 m thick in the central part of the guyot (Fig. 3). The resolution of the seismic data pre cludes identification of the TST and HST stratal patterns in this depositional sequence, but carbon ate deposition was presumably aggradational after the flooding of the central volcanic high. The top of DS-3 corresponds to an unconformable sequence boundary interpreted as a pronounced surface of emergence associated with karstification, as sug gested by the hummocky upper surface of platform carbonates (Fig. 5) and isotopic data (see 'Diagene sis' section).
Central part ofthe guyot. Dominant facies consist of skeletal packstones and grainstones with wacke stones at the top of the carbonate sequence. Major skeletal constituents include fragments of rudists (radiolitids, up to 5 7% of the skeletal components), larger benthic foraminifers (abundant Sulcopercu iina and Asterorbis associated with Idalina and Vida/ina and then Omphalocyclus-respectively as semblages II and I in Fig. 4), echinoderms and gastropods. Red algae are locally abundant. Rela tive to the underlying DS-2, the prevalence of packstones and grainstones suggests a higher-energy environment and increased biotic diversity indi cates open marine conditions. Inner elevated platform rim. Dominant facies in clude skeletal grains tones-rudstones with few algal octocoral-rudist frameworks which are more abundant at Site 877 than at Site 874. Skeletal grainstones and rudstones are generally poorly sorted and medium, coarse or very coarse grained; a
Development and demise of mid-oceanic carbonate platforms few packstone beds occur at Site 877. Major com ponents include rudists (caprinids and radiolitids), corals (hexacorals and octocorals), calcareous sponges (chaetetids) and red algae (corallinaceans, peyssonneliaceans: Polystrata alba, solenopora ceans: Pycnoporidium). Large skeletal fragments are usually micritized and display millimetre-sized borings. Benthic foraminifers (orbitoids, rotaliids, nodosariids, miliolids and ophthalmiids) are mod erately abundant. Other grains are scarce and in clude echinoids, gastropods and green algae (Ter quemella). The local occurrence of calcisphaerulids and small ammonites may characterize a maximum flooding surface (Arnaud-Vanneau et al. , 1 995). Or ganic frameworks are formed by laminar coral colonies ( octocorals: Polytremacis; hexacorals), rud ists (radiolitids and caprinids) and red algal bushes (solenoporaceans) that are heavily encrusted by red algae (corallinaceans and peyssonneliaceans: Poly strata alba) and, to a lesser extent, by foraminifers and Bacinella. Probable Microcodium structures were noted at the top of DS-3 at Site 874. Fossils and sedimentological criteria suggest that the frameworks grew in a shallow-marine environ ment, probably less than 1 0-20 m deep, and characterized by high-energy conditions. Further evidence of turbulent waters includes abrasion of shell fragments and the prevalence of grainstone and rudstone textures.
Outer elevated platform rim. Dominant facies in clude coarse skeletal grainstones and packstones characterized by an extensive porosity (principally mouldic and intergranular). The major skeletal components include larger benthic foraminifers (abundant Sulcoperculina and Asterorbis associated with Omphalocyclus) and fragments of rudists, red algae and echinoderms; minor components are corals, ostracods and planktonic foraminifers. Large leached intraclasts of wackestone bearing moulds of small gastropods, ostracods and discor bid foraminifers probably result from the erosion and the reworking of the upper part of the DS-2 of the inner ridge (Enos et al., 1 995a). Depositional sequence 4 (DS-4; Maastrichtian) DS-4 is the last platform carbonate sequence and is only identified on the outer ridge, where a maxi mum thickness of 85 m is recorded at Site 876 (Fig. 3). Its deposition at the edge of the guyot and at a topographically lower level than the top of the
49
platform suggests that it was deposited during a sea-level lowstand, whereas the major part of the platform was subaerially exposed, thus correspond ing to a lowstand wedge. It is composed of poorly sorted white and highly porous coarse skeletal grainstones rich in benthic foraminifers (abundant Asterorbis assemblage VIa in Fig. 4 ); fragments of rudists and red algae represent 20% of the skeletal components, whereas echinoderm and coral frag ments are less abundant; minor constituents are calcareous sponges, stromatoporoids and green algae. -
Diagenesis
Sediments on guyots generally display a complex diagenetic history involving carbonate cementa tion, transformation of carbonate phases, alteration of morphology and chemistry of the particles, dis solution, cavity infillings, phosphatization, and sec ondary mobilization of Fe-Mn oxides.
Porosity and cementation In contrast to carbonates influenced by burial dia genesis, the sediments on guyots still exhibit a high percentage of porosity. Cementation includes a variety of cement morphologies that represent, presumably, a variety of diagenetic environments. Overburden was of minor importance in Wodeje bato sediments, as the stratigraphy indicates a maximum depth of burial less than 200 m, and pore-water was almost certainly at hydrostatic pres sure throughout the evolution of the guyot. Accord ingly, the definitive parameters in cementation were fluid chemistry and temperature. The Wodejebato platform carbonates may have been exposed to a wide range of environments including shallow marine, deep marine, meteoric phreatic, meteoric vadose and subaerial. No indication was found that hypersaline waters were ever present in Wodejebato (Enos et al. , 1 99 5b). The temperature end-members also spanned a limited range. These carbonates are today bathed in cold sea water, beneath the oceanic thermocline (temperatures about 1 0oC; Premoli Silva et al., 1 993), whereas at the time of deposition near sea-level in equatorial waters, ambient temper atures were probably over 2 5oC (Enos et al. , 1 99 5b ). There is no evidence of hydrothermal alteration within the clays that overlie the volcanic rocks at three sites (873, 874, and 877).
·
50
G. F. Camoin et a!.
Petrography. The major part of Wodej ebato plat form carbonates consists of skeletal sands that display intense leaching of the original grains result ing in high mouldic, vuggy and solution-enlarged interparticle porosity. These sediments are gener ally poorly lithified, except in the upper 70 m of the inner ridge, where sediments are strongly cemented by multiple generations of banded calcitic cements. The two major cement types recognized petro graphically within the platform carbonates of Wodej ebato Guyot include radiaxial (locally fi brous) calcites and prismatic limpid calcites. Other cement types are clearly less abundant and include columnar Mg calcite and syntaxial overgrowth ce ments (Enos et al., 1 99 5b ). 1 When present (upper 70 m of the inner ridge), the first cement phase typically corresponds to isopac hous fringes of amber to brownish, radiaxial calcites that exhibit a typical sweeping extinction and in clude bladed and fibrous morphologies (Fig. 7A-C); a botryoidal habit is evident in large intergranular cavities. These cements are largely confined to primary pore-space. Radiaxial calcites are generally interpreted as resulting from the lateral coalescence of radial fibrous cements (Kendall, 1 977), either aragonite (Bathurst, 1 977; Kendall, 1 985) or high-Mg calcite (Lohmann & Meyers, 1 977), or from direct marine precipitation (Sandberg, 1 98 5). Mag nesium contents and strontium values of Wodeje bato radiaxial calcites suggest a magnesian-calcite precursor (Enos et a!. , 1 99 5b ). 2 The second generation of cement consists of limpid, pyramidal, scalenohedral and equant (syn taxial overgrowths) calcite cements that locally cross-cut the isopachous fringes of radiaxial cal cites. They occur in all types of pores, including solution pores, but are more common in primary pores, where they post-date radiaxial calcites (Figs 7A-D). These cements also fill chambers of some foraminifers and cavities in the overlying Eocene pelagic cap. The rounding of tiny pyramidal crystals is attributed to dissolution, suggesting that the precipitation of these cements and dissolution over lapped or alternated in time. Blocky, pore-filling, low-Mg calcites can be the product of meteoric (Longman, 1 980) or of deep-marine (Schlager & James, 1 978) diagenesis. Stable isotope composition. As petrographical fea tures alone are not conclusive as diagenetic indica tors, depositional and diagenetic history may be tracked using the oxygen and carbon isotope com-
positiOns of depositional and late-stage cements. Stable carbon- and oxygen-isotopic compositions of the successive cement phases and matrix are plotted in Figs 8 and 9 (all values are given in PDB). When present, well-preserved rudist shell fragments were analysed to try to estimate the original shallow marine isotopic composition as a starting point for discriminating progressive diagenetic changes. Carbon and oxygen stable isotope composition of radiaxial calcites ranges from + 1 .22 to + 2. 77o/oo (average +2. 1 1 ± 0.44o/oo) and from -5.2 1 to -0. 1 2o/oo (average - 1 .26 ± 1 . 1 6o/oo) for o 13C and o180, respectively (Camoin et al., 1 99 5 ; Enos et a!. , 1 995b). o180 values are consistent with precipita tion in equilibrium with warm (20-3o·q, shallow marine waters, assuming an ice-free world (o180water 1 o/oo SMOW). These values are close to expectation for cements precipitated from late Cretaceous sea water (-2o/oo < o180 < - 1 . 8o/oo; o13C 3o/oo; Lohmann & Walker, 1 989). They are similar to those reported from Cretaceous radiaxial and fascicular-optic calcite cements in the Bahama Escarpment (Freeman-Lynde et al., 1 986) and from the Miocene of Enewetak Atoll (Saller, 1 986), and slightly higher in o180 than those measured on lower Cretaceous radiaxial calcites from the subsur face of East Texas and Mexico (Moldovanyi & Lohmann, 1 984) (Fig. 8). In contrast, these values are slightly lower than modern marine cements, including high-Mg calcite cements from Pikinni (Gonzalez & Lohmann, 1 98 5), but are within the theoretical field of o13C values for calcite precipi tated from modern Pacific sea water (i.e. +2.0 to +2.5o/oo; Kroopnick et a!., 1 977). Thus, petrograph ical observations and stable isotope data support precipitation in a shallow-marine environment. o13C and o180 values measured in clear, pyrami dal calcite cements fall into two distinct groups: 1 The first group display enriched isotopic ratios where o13C and o180 range respectively from + 1 .2 3 t o +3. 1 3o/oo (average +2. 1 9 ± 0.48o/oo) and from -2. 1 0 to + 1 . 1 1 o/oo (average: - 1 .05 ± 0.69o/oo) (Cam oin et al. , 1 995; Enos et a!., 1 995b) (Fig. 9). Similar isotopic values were obtained for deep marine equant spar cements formed below the thermocline in lower and mid-Cretaceous shallow-water lime stones exposed on the Bahama Escarpment (Freeman-Lynde et a!., 1 986), for Campanian gravity-flow deposits from the Bahamas (McClain et al. , 1 988) and for Campanian-Palaeocene hard-· grounds from the Caribbean (Anderson & Schnei dermann, 1 973) (Fig. 9). On the other hand, =
=
�
"' �
0 �
;:s
� !::> :::s
!::>..
� ;:s
<:;· �
�
;:s �
g
� :::s
r;·
8 �
�!::>
� 'l::l i:S" s, 0
�"'
Fig. 7. Thin-section photomicrographs of rudist-coralgal frameworks and associated skeletal grainstone and rudstone of the inner elevated platform rim
(Sites 874 and 877). Early cements consist of thick isopachous fringes of radiaxial and fascicular-optic calcites with fibrous and bladed morphologies (2 in A and I in B). These calcites appear to be zoned by variations of inclusion density. The remaining pore-spaces are partly filled by internal micritic sediment (A) and/or by limpid, pyramidal and scalenohedral calcite cements (3 in A and 2 in B). In D, the primary pores are partly filled by limpid calcite cements. The local occurrence of clear granular calcite at the contact between grains and early calcite cements should be noted (I in A). m, mouldic pore; po, remaining porosity.
lJl
52
G. F. Camoin et a!.
o"c
Sites 873, 874 and 877 4 ,------,
o"c
Sites S73, 874 and 877 4 ,-------. NEP---
3
3 CE
AIC·C
2
2
0
•
0
Matrix
+ RC "cc o
-1
o
EBS
PLUC
-1
-2 �-r--.--.--,--,,--,-� -8 -7 -6 -5 -4 -3 -2 -1 3 2 4 0
-2�--�----�-r�-.�� -8 -7 -6 -5 -4 -3 -2 -1 0 3 4 2
o'"o
o••o Fig. 9. Cross-plot of stable isotopic composition of late
Fig. 8. Cross-plot of stable isotopic composition of
syndepositional cements (shallow marine and meteoric) and matrix, Sites 873, 874 and 877. RC, Radiaxial cement; CC, columnar, bladed, magnesian calcite; EBS, equant blocky spars; BE, Bahama Escarpment shallow marine cements (Lower-Middle Cretaceous; Freeman-Lynde et a!. , 1 986); STC, shallow-marine cements in Stuart City trend strata (Lower-Middle Cretaceous; Prezbindowski ( 1 985) in McClain et a!. ( 1 988)); AIC-M, average isotopic composition of Maastrichtian calcites (Lohmann & Walker, 1 989); AIC-C, average isotopic composition of Coniacian calcites (Czemiakowski et a!. ( 1 984) in Moldovanyi & Lohmann ( 1 984)).
carbon-isotope values are slightly lower than for most modem shallow-marine cements (see Gonza lez & Lohmann, 1 985), suggesting a precipitation in relatively deep-marine waters where 813C of marine bicarbonate is 1 -2%o lighter than in surface marine waters (Kroopnick et al. , 1 977). This interpretation is strengthened by 8180 values that are consistent with precipitation in equilibrium with colder ( 1 020•C) waters of deep marine composition, assum ing an ice-free world (8180water 1 %o SMOW). 2 The second group of blocky spar cements has carbon- and oxygen-isotopic compositions more typical of meteoric diagenesis, especially depleted oxygen values with respect to radiaxial calcite ce ments and well-preserved rudist shells. 813C and 8180 values range respectively from -2. 3 7 to + l . 5 8%o (average +0. 1 4%o) and from -7.80 to -2.49%o (average: -5%o) (Fig. 9). All these samples come from throughout DS-3, from the top of DS-2 and from skeletal grainstone clasts embedded in the =
deep-marine cements, Sites 874 and 877. PLUC, Prismatic limpid cement with uniform extinction; BE, Lower-Middle Cretaceous from Bahama Escarpment (Freeman-Lynde et a!. 1 986); Leg 1 5, Campanian Palaeocene from Leg 1 5 DSDP (Anderson & Schneidermann, 1 973); NEP, Campanian from NE Providence Channel (McClain et a!. 1 98 8); CE, Aptian-Albian from Campeche Escarpment (Halley et a!. ( 1 984) in McClain et a!. ( 1 9 88)).
manganese crust. They indicate two episodes of subaerial exposure with meteoric-water and possi ble soil-gas influence, which may correspond to the termination of the platform.
Skeletal diagenesis Originally aragonitic skeletons exhibit differential preservation throughout the carbonate sequence, whereas dissolution of benthic foraminifers and red algae is peripheral and rare. Caprinid rudists dis play a gradational sequence of transformation, from pristine aragonitic shells to fragments totally re placed by limpid blocky calcites, through partial neomorphic replacement where relics of aragonitic skeletal structures are preserved in inclusion-rich coarse blocky calcite cements (Fig. 1 0). It is possible to relate the degree of stable isotope depletion to these specific fabrics. Stable carbon and oxygen compositions of well preserved aragonitic rudists ( 60- 1 00% aragonite;· Fig. lOA) range respectively from +0. 76 to +6. 3 5%o (average +2.95 ± l . 37%o) for 813C and
•'
,•\
1/
0
"i?l· ' �. •;:,
(.
I
� � .g ;:;:
�
� ::lO r;:,_
� ;:;: <::;·
� ;:;:
� 6
� ::lO
r;·
8 -. g-
::lO � � 'l:::l iS'" s:;, 0 -.
;:;:
c..,
Fig. 10. Thin-section photomicrographs showing diagenesis of rudists in depositional sequences 2 and 3 of the inner elevated platform rim. (A) Well-preserved
aragonitic shell of caprinid rudist; intraskeletal pores are filled by early calcite cements with a fibrous morphology. (B) Sharp and scalloped contact (arrows) between a rather well-preserved rudist shell (I) and limpid block calcites that exhibit only a few relics of the former skeletal structures; the occurrence of thick isopachous fringes of fibrous calcite cements should be noted (2). (C) Early step of transformation of a rudist shell where limpid blocky calcites cross-cut the original aragonitic structures (I) and associated early calcite cements (2). (D) Nearly complete transformation of rudist shell by limpid blocky calcite cements that display few relics of the former skeletal structure (arrows).
Vl w
G. F. Camoin et al.
54
from -2.99 to - l .2 8o/oo (average - 1 . 34 ± 1 . 1 6o/oo) for
8180 (Fig. 1 1 ) (Camoin et al., 1 99 5 ; Enos et al. , 1 995b). Remarkably, five values from Site 877 plot with less than 1 o/oo scatter at 4. 8o/oo o 1 3C and - 1 .2o/oo 8180 (Camoin et al., 1 995). This cluster could be the best value for Maastrichtian near-surface sea water. It compares reasonably well with two values that average -2.2o/oo 8180 and + 3 . 5 5o/oo o 1 3 C from aragonitic rudists (Antillocaprina and Plagiopty chus) in the Upper Cretaceous of Texas (AI Aasm & Veizer, 1 986) (Fig. 1 1 ). Isotopic values obtained for calcitic layers of rudists in Lower and Middle Cretaceous strata exposed on the Bahama Escarp ment (Freeman-Lynde et al. , 1 986) and from the subsurface of Texas (Moldovanyi & Lohmann, 1 984) are depleted in 180 with respect to aragonitic rudists from Wodejebato. The stable isotopic composition of inclusion-rich calcites exhibiting relics of aragonitic skeletal struc tures (Fig. 1 OC & D) range respectively from +2.40 to +3. 87o/oo (average +2. 9 1 ± 0.47o/oo) for o13C and from -2.87 to - l .99o/oo (average -2. 8 3 ± 0.2 5o/oo) for 8180 (Fig. 1 1 ). These values are very close to those obtained on similar neomorphic calcites by AI
Sites 873, 874 and 877
o13C 7
+
6 5
Dissolution and cavity infillings
+
4 3
NEPC
2 '
0 -1 • N eomorphosed rudists + Aragon itic rudists
-2 -3
o 6
-4
Rudist calcitic layers Partly calcitized shells
-5 -6 -8
Aasm & Veizer ( 1 986). With respect to values measured on aragonitic shells or fragments of shells, there is a slight decrease in 8180 values, whereas the decrease in o 1 3C values is negligible. Neomorphic alteration of marine carbonates in meteoric fluids at low water/rock ratios could have lowered the 8180 values while leaving the o 1 3C values virtually un changed (Wyatt et at. , 1 99 5). This suggests that the 8180 and the temperature of diagenetic water did not vary much from that of the coeval sea water during an early stage of diagenesis, implying a wet replacement of aragonitic components. Stable isotopic compositions of limpid calcites with few or no relics of aragonitic skeletal structures (Fig. l OB) range respectively from -2.22 to +2.3o/oo (average + l . l 9 ± 0.84o/oo) for o 1 3C and from -4. 97 to - l .20o/oo (average -3. 5 1 ± 0.97o/oo) for 8180 (Fig. 1 1 ); lower values (between -5 and -8o/oo for 8180) were reported by Enos et al. ( 1 99 5b). Such a shift towards a lower value of both isotopes was also recorded in other skeletal grains (benthic foramini fers and red algae) at the top of DS-2 and DS-3 both on the inner ridge and in the central part of the guyot. Such an isotopic shift is considered as an indicator of meteoric alteration and confirms ·solution-reprecipitation processes, implying alter ation in a later diagenetic stage.
-7
-6
-5
-4
-3
-2
-1
0
2
o'80 Fig. 1 1 . Cross-plot of stable isotopic composition of
rudist shells with different degrees of preservation. Sites 873, 874 and 877. TEX, Aragonitic rudists from the Upper Cretaceous of Texas (AI Aasm & Veizer, 1 98 6); SL-CUP, rudists from the Sligo and Cupido Formations (East Texas; Moldovanyi & Lohmann, 1 984); NEPC, rudists from the Campanian of NE Providence Channel (McClain et a/. 1 988).
The platform limestone ofWodejebato is 40% voids in some intervals (determined visually), and per haps more where recovery was nil. Dissolution affected formerly aragonitic shells (e.g. gastropods, caprinid rudists, corals) and encasing sediments, thus producing extensive secondary porosity in the form of moulds and vugs which are either still open or infilled by cements and/or internal sediments. Solution cavities occur in the upper 50 m of the carbonate sequence. Fluids responsible for dissolu tion could have been meteoric water or undersatu rated sea water. Submergence of the guyot should demarcate these possibilities, suggesting that the distinction could be readily made. Several lines of circumstantial evidence support considerable dissolution in a deep-marine environ ment. The occurrence of the late phase, prismatic limpid calcites, in dissolution pores, its coarser crystallinity in primary pores and the rounding of small crystals are taken to indicate that precipita tion of prismatic limpid calcites coincided with, or alternated with, leaching in deep-marine waters
Development and demise of mid-oceanic carbonate platforms (Enos et a!. , 1 995b). The same waters probably dis solved much of the aragonite and magnesian calcite that originally formed the platform-carbonate cap of Wodejebato Guyot. Simultaneous dissolution of a metastable carbonate phase and precipitation of a more stable phase is feasible in marine waters be tween the aragonite and the deeper calcite saturation levels (Saller, 1 986; Melim et a!., 1 99 5). Syndepositional episodes of exposure to meteoric waters can be evidenced and the relative timing of dissolution processes can be fixed with respect to the ages of sediment infillings. Yellowish, reddish or rusty brown internal sediments with scarce smooth shelled ostracods and very small benthic foramini fers partly fill mouldic pores or irregular solution cavities, where they locally post-date banded isopa chous crusts of radiaxial calcite cements. These cements line mouldic pores at several levels, espe cially at the top of DS-2 and DS-3 (e.g. Site 877; Camoin et a!., 1 995; Enos et a!. , 1 995b). This implies a dissolution during an early stage of di agenesis (i.e. before submergence of the platform) in near-surface sea water undersaturated with arago nite, but more probably in meteoric water. The occurrence of mouldic pores in platform limestone clasts reworked in the Mn crust strengthens this interpretation. Late infillings of centimetre-sized solution cavi ties in the platform carbonates, as deep as 50 m beneath the top of the carbonate sequence at Site 874 (Camoin et a!., 1 99 5}, consist of white to orange-brown micritic sediments (mudstones and wackestones) rich in pelagic organisms associated with micritized and phosphatized fragments of shallow-water organisms (e.g. benthic foraminifers, red algae and rudists). These sediments range in age from the late Maastrichtian to the middle Eocene (Premoli Silva et a!. , 1 993). The time restrictions introduced by biostratigraphical data narrow the timing of carbonate dissolution and imply that: (i) the drowning of the carbonate platform occurred during the Maastrichtian; and (ii) the platform limestones were lithified, eroded and leached to form moulds as early as mid to late Maastrichtian. The mechanism involved in the formation of these cavities cannot be fully deciphered because of the uncertainties concerning their morphology, their width and their relationships with encasing shallow water carbonates. They could be related either to subaerial karstification processes, or to enlargement of fractures by dissolution in undersaturated ma rine waters. However, it seems unlikely that the
55
guyot was deep enough by Maastrichtian to Palae ocene time to have produced extensive dissolution processes as deep as 50 m in the carbonate cap. Considering an average rate of subsidence of 30 m Myc1 (according to the empirical method of Parsons & Sclater ( 1 977) modified by Crough ( 1 978) and Heestand & Crough ( 1 98 1 }}, the summit of the guyot was probably shallower than 300 m before the end of the Palaeocene and, based on the aragonite compensation depth in the modern Pa cific Ocean (c. 300-400 m; Scholle et a!. , 1 983}, dissolution of aragonite components did not start before that time. Thus, it seems likely that subaerial dissolution could be the culprit.
Paragenetic sequence Figure 1 2 summarizes the relative timing of para genetic events. Cementation, neomorphism and dissolution were important factors in forming the platform limestones of Wodejebato Guyot. Radiax ial calcites formed in shallow-marine waters, gener ally during deposition. Syndepositional episodes of exposure are indicated by petrographical evidence (rhizoliths, alveolar textures, Microcodium struc tures, early dissolution processes) and by low 8 1 80 values at several levels. All sites from which stable isotope data are available show covariant lowering of the isotopic values within the fossils and matrix toward the top of the carbonate sequence, which appears to be the signature of at least partial stabilization in meteoric waters. The reduction in 8 1 3C is often slight and suggests that relatively short periods of exposure or unfavourable climate pre vented formation of a soil profile. Alternatively, erosion may have removed the soil zone without removing evidence of exposure within the meteoric phreatic zone. Slight or no reduction in 8 1 3C is not inconsistent with diagenesis within the near-surface freshwater phreatic environment. Lower 8 1 3C val ues are generally found only in narrow zones near subsurface exposure surfaces, whereas limestones sampled from within the deeper vadose and fresh water phreatic environments at low water/rock ratios do not have such low 8 1 3C compositions (Allan & Matthews, 1 977; Moldovanyi & Lohm ann, 1 984; Melim et a!. , 1 995). Petrographical and isotopic evidence of subaerial exposure is concentrated at the top of DS-2 and DS-3 and extends into the veneer of phosphatized and manganese-encrusted limestone, indicating that exposure to meteoric water was necessarily
56
G. F. Camoin e t a!.
Paragenetic events
lime -
Marine sea-floor
Deposition of DS-1 and DS-2 Micritization processes Cementation by radiaxial calcites
Meteoric-subaerial
Partial dissolution of aragonite shells Local cementation by eq u a nt blocky spars Development of rhizoliths a n d associated alveolar textures J uvenile karstification of the platform top
Marine sea-floor
Local infillings of moldic pores Deposition of DS-3 Micritization processes Cementation by radiaxial calcites
-
Relative sea-level +
)
-
-
Sub. exposure I
-
-
)
-
Sub. exposure I I
Meteoric-subaerial
Parti a l dissolution of aragonitic shells a n d neomorphism Local cementation by e q u a nt blocky spars Karsification processes
-
-
Marine sea-floor
-
Deposition of DS-4 with erosion from DS-2 and DS-3 micritization processes
-
Deep marine
-
lnfillin gs of solution cavities by pelagic sediments Extensive dissolution processes Cementation by limpid prismatic calcites · Formation of ma nganese-phosphate crusts Deposition of pelagic cap
-
Drowning
-
Fig. 12. Paragenetic sequence and interpreted relative sea-level changes.
early in the diagenetic history before the atoll was submerged. It seems likely that the formation of solution cavities as deep as 50 m beneath the top of the carbonate sequence may be related to a sub aerial exposure at the end of the platform develop ment. Late diagenetic processes include dissolution and precipitation processes. The fabric selective leach ing of moulds and their enlargement to vugs could
have begun when Wodejebato sank below the level of aragonite saturation, probably during late Palae ocene to early Eocene time. These processes appar ently overlapped precipitation of prismatic limpid calcites in deeper, cooler marine waters, which bathed the guyot after it was submerged below the photic zone. Dissolution and coeval precipitation of prismatic limpid calcites probably continue to the present day.
Development and demise of mid-oceanic carbonate platforms POST-DROWNING FAC IES
The platform carbonates at all sites are topped by a complex succession of phosphate-manganese dense black crusts, ranging in thickness from 3 em (Site 877) to 1 4 em (Sites 873 and 875). Pores between the ferro-manganese crusts are filled with several generations of pelagic sediments ranging in age from the late Maastrichtian to the late Eocene (Erba et a!. , 1 99 5) and similar to those which infill the solution cavities within shallow-water carbonates, implying a deep-marine depositional environment (i.e. tens to hundreds of metres) and very slow accre tion rates. These crusts seal well-cemented foramini fer-rudist grainstone and manganese-phosphate coated limestone conglomerate embedded in a planktonic foraminifer matrix. At Sites 8 74 and 8 77, the contact surface between the platform carbonates and the manganese-phosphate crust is scalloped and displays multiple generations of borings; at Site 876, the contact between platform carbonates and man ganese crust is vertical and was interpreted as a microkarst formed under subaerial conditions be fore the drowning of the platform (Enos et a!. , 1 99 Sa). These observations demonstrate that the Maastrichtian limestones were lithified and eroded before deposition of the pelagic ooze. The crusts include multiple generations of stacked digitate and laminated accretions embed ding clasts of foraminifer rudist grainstone and late Palaeocene pelagic limestone; manganese dendrites and threads grow inward, suggesting a secondary mobilization of the Fe-Mn oxides. The crusts grew probably during periods of slow sedimentation and have been mechanically reworked several times. The production of clasts included in the crusts is seemingly related to the intense boring activity and to erosion and redeposition during Eocene time, probably as a result of bottom-water activity. Be cause the amount of phosphate in the Central Pacific photic zone is almost nil, upwelling deep water is the only available source for phosphate (Grotsch & Flugel, 1 992).
DEPOSITIONAL HISTORY OF WODEJEBATO GUYOT Flooding of the volcanic substrate
The depositional history of Wodejebato Guyot be gan with the end of volcanism before or during Campanian time (78.4-85 Ma; Pringle & Duncan,
57
1 99 5). Like other guyots, it consisted initially of a broad shield volcano. Erosion, weathering and veg etative growth progressively modified the volcanic cone, transforming the summit into a gently sloping plateau with a central basement high and topo graphical depressions filled by the clay-rich sedi ments produced by pedogenic alteration (Figs 5 & 1 3). According to radiometric ages obtained on lavas and biostratigraphical data from the overlying marine deposits, the weathering of the volcanic basement must have occurred during a time-span of 2.4-3 Myr. During this time, slope fans and pro grading slope wedges were not developed, probably because the flanks of the volcano were sufficiently steep to bypass sediment directly to the adjacent deep ocean basin. The initial marine deposits recovered during Leg 1 44 and dated as late Campanian correspond to black organic-rich clays deposited in a quiet, reduc ing, shallow-marine environment. The basal lime stone recovered from above the clay consists of a grainstone that includes fauna and flora (larger foraminifers, corals and red algae) typical of an open shallow-marine environment. The partial preservation of the weathering profile and of the overlying marine clays, and the lack of any clay reworking at the base of the carbonate sequence suggest a slow sea-level rise and a rapid burial. Development of the carbonate platform
The development of the carbonate platform was clearly governed by the growth potential of the carbonate systems, accommodation space and rela tive sea-level fluctuations, which are a product of the interplay between thermal subsidence rates and superimposed eustatic sea-level changes. Early carbonate production on the carbonate platform resulted in the accumulation of skeletal sand piles near the shelf margin (DS-1 ), rimming an emerged volcanic island and probably correspond ing to the early stages of a TST (Figs 5 & 1 3). The sands accumulated in a shallow-marine deposi tional environment characterized by moderate- to high-energy conditions. The abundance of rhodo liths at the base of the carbonate sequence is a common feature in the initial steps of development of isolated Campanian-Maastrichtian carbonate platforms (Camoin et a!. , 1 988) and Cainozoic atolls (Collot et a!., 1 992), where they probably played a significant role in the stabilization of previously mobile substrates.
G. F Camoin et a!.
58
67 Ma
_)
Erosional surface, reworking and turbidites in the basin Ca2 - 73-74 Ma
=-S==eq=·=1=======� =D=e=� ==�=0=H=s=� ssBe:�::��: a =�
����====�==:
Soil and vegetation Erosion Ma
(___n_]:_ d of volcanism ) Fig. 13. Mid-oceanic carbonate platform development based on Wodejebato Guyot. TST, Transgressive system tract; HST, highstand system tract; LST, lowstand system tract. The sea-level curve is from Haq et a!. ( 1 987).
The base of the overlying sequence (DS-2) corre sponds to a series of stacked retrogradational sand shoal units (TST) with reflectors onlapping the volcanic substrate that was still emergent at this time (Fig. 1 3). A subsequent rise in sea-level is responsible for the backstepping of sand shoals followed by skeletal sands (rapidly cemented) asso ciated with rudist-coralgal frameworks on the inner ridge (Fig. 5). Simultaneously, the central volcanic high was flooded and covered by skeletal sand shoals. This flooding provided accommodation on the platform top, resulting in changes in the domi nant stratal pattern from TST to HST, therefore from retrogradational to aggradational and progra dational. As sea-level stabilized, the rudist-coralgal shoal began to aggrade (early HST) and then pro grade (late HST). At this stage of development, the relief of the inner ridge was accentuated, probably on the order of 36 m above the coeval lagoonal sediments ('empty bucket morphology'; Arnaud Vanneau et al. , 1 99 3). The prevalence of a quiet,
shallow-water depositional environment in the cen tral part of the guyot and the differences in biota indicate a decrease in energy from the inner ridge toward the central part of the guyot. The wide spread deposition of quiet, shallow-water facies over much of the guyot in the upper part of DS-2 (late HST; Fig. 1 3) was a result of progradation of reef units. A temporary subaerial exposure at the top of DS-2 is indicated both by petrographical features and stable isotope data (see 'Diagenesis' section) and may be related to a slight lowering of sea-level. During this short-term subaerial expo sure, the platform edge was partly eroded and clasts of platform limestones were reworked downslope, forming the initial deposits on the outer ridge (Fig. 5). The next transgression caused the backstepping of the carbonate factory, forming a TST (DS-3). This rapid transgression and coeval lagoonwari:l shift of depositional environments initially caused the widespread accumulation of skeletal sands,
Development and demise of mid-oceanic carbonate platforms followed by the brief development of rudist coralgal frameworks on the inner ridge. At that time, normal conditions of salinity and oxygenation existed across the entire platform. The demise and the drowning of the carbonate platform
The solution cavities infilled by pelagic sediments that occur as deep as 50 m within the platform carbonates may have formed by solution enlarge ment of fractures in a deep-marine environment or, more probably, by subaerial karstification as sug gested by the timing of their formation (i.e. as early as mid to late Maastrichtian) and by petrographical and isotopic data recorded at the top of DS-3 (see 'Diagenesis' section). The creation of such cavities during exposure recorded at the top of DS-2 is unlikely because of its apparent brevity. There is clear evidence that the surface of the guyot, includ ing the outer elevated platform rim, was eroded substantially before subsidence into the pelagic realm. On the outer ridge, the vertical contact between platform carbonates and the manganese crust may represent a microkarst formed under subaerial conditions before drowning of the plat form (Enos et al., 1 995a). Karst formation at these sites, the deepest part of the guyot cored, would imply exposure of the entire limestone cap of the guyot and a corresponding drop in sea-level of at least 54 m (i.e. the difference in elevation of the top of platform limestones from the perimeter of the guyot to the interior). The last platform carbonate sequence, DS-4, corresponds to a wedge deposited during a sea-level lowstand on the outer ridge, whereas the inner part of the platform was subaerially exposed and proba bly karstified (Figs 5 and 1 3). The trough that separates the outer and inner elevated platform rims (Fig. 5) may be interpreted either as a larger scale of karst formation (i.e. large doline parallel to adjacent slope), implying a minimum sea-level low ering of 50 m, or more likely as a depositional trough in front of the inner elevated platform rim (Enos et al., 1 995a). The ages of pelagic cavity infillings, ranging from the Maastrichtian to the Eocene (see Erba et al. , 1 99 5), imply that the drowning of the carbonate plat form occurred during the Maastrichtian. This im plies a rapid sea-level rise of at least tens of metres during Maastrichtian time, just after exposure. The association of planktonic foraminifers with Gans-
59
serina wiedenmayeri (see Erba et al., 1 995) confines the age of the formation and the infilling of cavities to a time-span of 2-3 Myr, ranging from the Gans serinagansseri Zone to the lower part oftheAbathom phalus mayaorensis Zone. The inferred short-term regressive-transgressive cycle is consistent with sea level changes with an amplitude of 1 00 m at the top of the Supercycle UZA-4 ofHaq et al. ( 1 987) (Figs 3 & 1 4). Post-drowning evolution
After its demise, the carbonate platform sank to deep water, between the aragonite and the deeper calcite saturation depths, resulting in the dissolu tion of aragonite and concomitant precipitation of calcite cements (Fig. 1 4 ). Based on the aragonite compensation depth in the modern Pacific Ocean (300-400 m; Scholle et al. , 1 983) and considering an average rate ofsubsidence of30 m Myc1 (accord ing to the empirical method of Parsons & Sclater ( 1 977) modified by Crough ( 1 978) and Heestand & Crough ( 1 98 1 )), it seems likely that the guyot reached this depth not before late Palaeocene to early Eocene time. Strontium-isotopic composition of pyramidal limpid calcites from the base of Hole 877 A indicates precipitation during the Tertiary, well after the drowning of Wodejebato (Wilson et al. , 1 99 5). Deep-marine waters which were respon sible for these processes probably invaded the car bonate cap of the guyot continuously during its subsidence, through the very permeable sand facies constituting the bulk of the carbonate sequence. Although dissolution rates in deep waters are much slower than in the meteoric environment, it is probable that these processes continued throughout the Cainozoic to the present day at the margins of the guyots. The interior was effectively isolated from seawater circulation by accumulation of pe lagic ooze, beginning in the early Miocene (Erba et al. , 1 995). The formation of phosphate-Mn crusts may have started when the guyot dropped below the photic zone of intensive carbonate production.
INSIGHTS INTO THE DEVELOPMENT AND THE DEMISE OF MID-OCEANIC CARBONATE PLATFORMS
The sedimentary evolution of carbonate platforms on the guyots drilled during Leg 1 44 yields similar scenarios, irrespective of their age (Arnaud-
·
G. F. Camoin et a!.
60 85
80
I
70
75
I
I
I
CAMPANIAN
MAASTRICHTIAN
"
< 1;2 � n &1 >
I
'
60
65
I
I
DANIAN
� z 0> iii t; v;
55
50
I
THANETIAN
I
YPRESIAN
45 I
LUTETIAN
40 ( Time in M. years) _l_
IIAifi'O.
I
Stages
l
�
�>
�.� �6
J
z�
�>
250 m
I�
200 m 150 m lOO m
t ., 0 0 c.
:;· ""
!2.,
50 m
t "
;;
� r;·
�
�
I
::. 0 "
[
i5" "
200 m
;>; !'; 0
300 m
tl "
() c: "' "' """ "' "'
l OOm
tt O
::: rs
�
Om
5"
< 0
� ...,
OQ
400 m
CARBONATE PLATFORM DEVELOPMENT
600 m
- Phosphate-Mn crust
D I
Platform carbonates Volcanic basement
FORMATION OF PHOSPHATE-Mn CRUST
BOO m
1000 m
Fig. 14. Late Cretaceous and early Tertiary sedimentary evolution of Wodejebato. Subsidence path of volcanic
basement is calculated from the age-depth equation for oceanic crust proposed by Parsons & Sclater ( l 977) and modified by Crough ( 1 978) and Heestand & Crough ( 1 9 8 1 ); eustatic curves are from Haq et at. ( 1 987).
Vanneau et a!. , 1 993): late Campanian-Maastricht ian (Wodejebato) to latest Palaeocene-early middle Eocene (Limalok) in the Marshall Islands and Aptian-Albian (MIT, Takuyo-Daisan) in the Japa nese Seamounts. Consistent features in their devel opment include: (i) a prolonged initial period of subaerial exposure of the volcanic pedestal (gener ally of a few million years) characterized by the formation of soil horizons; (ii) the progressive flooding of the volcanic basement by organic-rich
clays, calcareous sandstones bearing plant remains and/or bioclastic limestones deposited in a reducing shallow-marine environment; and (iii) a burial of the carbonate platform by pelagic deposits, some times preceded by subaerial exposure. Subsidence and accumulation rates
The carbonate platforms that developed on these emergent seamounts kept pace with subsidence
Development and demise of mid-oceanic carbonate platforms rates, but none of the drilled platforms had a lifetime longer than 20 Myr. The MIT and Takuyo Daisan platforms developed for about 1 9 Myr and 1 5 Myr, respectively, whereas a lifetime of 1 0 Myr is estimated for the platform on Wodejebato and Limalok. Carbonate accumulation rates (not cor rected for compaction), range generally from 8.2 m Myr- ' (e.g. Takuyo-Daisan) to a maximum of 40 m Myc 1 (upper carbonate sequence of MIT). On Wodejebato, accumulation rates range from 1 3 . 7 m Myc ' in the central part of the guyot up to 25-45 m Myr- ' on the outer elevated platform rim; on the inner ridge, these rates are 1 7 m Myc 1 and 1 8. 3 m Myc1 at Sites 8 74 and 877, respectively. Accumulation rates of 30 m Myc ' are reported on Limalok. The lowest accumulation rates are related to the presence of hiatus(es) in the carbonate se quences (e.g. Takuyo-Daisan; central part and inner ridge of Wodejebato). Origin o f morphological features
A number of NW Pacific seamounts display atoll like features, including raised rims enclosing a central lagoon, flanked by steep talus leading down to volcanic slopes. Early-cemented organic frame works form thin intervals on the inner ridge of Wodej ebato but they cannot be compared to coral gal reefs that characterize modern Pacific atolls. None of the elevated platform rims drilled on early Cretaceous guyots (e.g. Takuyo-Daisan, Resolution) revealed the existence of organic frameworks. It seems likely that these systems kept pace with subsidence and sea-level changes not by construct ing a wave-resistant bulwark, but by producing vast quantities of loose carbonate sediment that formed intermittent skeletal and oolitic sand shoals. On the other guyots (e.g. MIT, Limalok), the perimeter rim was not drilled so its nature is unknown. The elevated platform rims appear to be depositional features rather than relict karst features related to the dissolution of low-energy limestones in the central part of the guyots and the concomitant lithification of high-energy rim deposits as sug gested by Van Waasbergen & Winterer ( 1 993). Another characteristic ofNW Pacific guyots is the occurrence of closed depressions or sinkholes, lo cally 100-200 m deep (e.g. MIT), stream channels and terraces on their summit. These features have been interpreted as the result of subaerial exposure, on the basis of their remarkable similarity to mod ern subaerial karst morphology (Van Waasbergen &
61
Winterer, 1 993). Other mechanisms for generating features that mimic the appearance of sinkholes on a submerged carbonate platform have been consid ered. They may include dissolution of carbonates related to the aragonite compensation depth and/or dissolution of carbonates associated with sulphur rich fluids (Haggerty & Van Waasbergen, 1 99 5), but none of them alone may account for the formation of such large-scale collapse features. It seems more likely that the surface topography of the guyots has been generated by a combination of several mecha nisms involving subaerial karstification and deep marine dissolution (Haggerty & Van Waasbergen, 1 99 5). It has been shown earlier that, on Wodeje bato, solution cavities affecting the upper 50 m of the carbonate cap were formed necessarily before the guyot sank below the aragonite compensation depth, thus implying early subaerial dissolution processes. Similar histories involving a subaerial exposure preceding drowning are recorded in the Mid-Pacific Mountains (Sager et al. , 1 993) and other dredged guyots along the Japanese Seamount chain (Grotsch & Flugel, 1 992). Subaerial karst sinkholes and troughs as much as 75 m deep are recorded in the upper part of Allison and Resolu tion guyots; furthermore, on Resolution Guyot, speleothems occur at least as deep as 60 m, imply ing that the guyot stood at least to this elevation above sea-level during karstification (Sager et al. , 1 993). The sedimentary evolution of many guyots is characterized by one or several substantial periods of exposure before rapid submergence. On Wodeje bato, the late development of the carbonate plat form is characterized by two successive periods of emergence: a first limited fall of sea-level and the water table may have induced juvenile karstifica tion, and then a deeper lowering of sea-level led to karstification and formation of a lowstand wedge, before a rapid rise of greater amplitude (Fig. 1 3). Meteoric diagenesis at the top of the carbonate sequence of Wodejebato is typified by sedimento logical evidence and stable isotope data (see 'Diagenesis' section). On Limalok Guyot, three intervals, including the top of the carbonate plat form, have been interpreted as indicative of mete oric diagenesis with low water/rock ratios (Wyatt et al., 1 99 5). Indications of subaerial exposure are also reported on MIT and Takuyo-Daisan guyots (Hag gerty & Van Waasbergen, 1 99 5 ; Jansa & Arnaud Vanneau, 1 99 5). With few exceptions, the depletion in 8 1 80 generally coincides with an increase in
62
G. F. Camoin et al.
porosity in the limestone, and meteoric cements are scarce in the Leg 1 44 platform limestones. A similar situation has been reported in modern atolls located in low rainfall areas, where meteoric cementation is reduced and the chief diagenetic processes consist of dissolution of the metastable carbonates and the development of karst terrane (see Wheeler & Aha ron, 1 99 1 ). The problem of the drowning
The drilling results of Leg 1 44 indicate that the formation of guyots was not synchronous. There were at least three major episodes of carbonate platform drowning within the tropical Pacific: Al bian, late Maastrichtian and middle Eocene (Pre moli Silva et a!., 1 993). The first episode of drowning was previously placed at 1 05- 1 1 6 Ma, after the early Aptian but before the middle Albian (Winterer & Metzler, 1 984), but it occurred during late Albian time (Sager et a!., 1 993; Premoli Silva et al. , 1 993; Erba et a!., 1 995). A number of mecha nisms, alone or in combination, may cause platform drowning. Among these are rapid relative sea-level changes, tectonics, volcanism and environmental deterioration causing reduced benthic growth rates or death (Schlager, 1 98 1 ; Hallock & Schlager, 1 986; Erlich et a!. , 1 990). Palaeolatitude data (Nakanishi & Gee, 1 99 5) do not support plate motion carrying reefs to latitudes beyond the Darwin Point as it was suggested by Winterer & Metzler ( 1 984). Because of restricted areal extent and low summit relief, small isolated platforms are probably more susceptible to drowning, which can place them out of the photic zone and effectively shut down the shallow-water 'carbonate factory'.
Sea-level changes Rapid drowning is in evidence on three guyots drilled during Leg 1 44: MIT, where late Albian pelagic deposits overlie mid- to late Albian platform carbonates; Wodejebato, where late Maastrichtian pelagic sediments fill cavities at the top of the Maastrichtian carbonate sequence; and Limalok, where mid-Eocene pelagic deposits overlie platform carbonates of nearly the same age (Erba et al. , 1 99 5). In all these situations, the time-span between the age of the youngest shallow-water limestones and the oldest pelagic sediments ranges from 1 to 3 Myr. This implies regression-transgression cycles of considerable amplitude and a duration of a few
million years. Such cycles are well known from continental margins and may have also been re corded in pelagic sequences (e.g. Albian; Sliter, 1 99 5). The drowning appears to be geologically instan taneous, as there is no transitional sequence be tween platform carbonates and pelagic deposits, as a result either of non-deposition or subsequent erosion during reflooding. The complete shut-down of platform carbonate production during reflooding cannot be explained solely by a rise in sea-level. Subsidence rates and long-term sea-level changes are at least one order of magnitude lower than the potential growth rates of carbonate platforms (Schlager, 1 9 8 1 ), even though the carbonate accu mulation rates reported on the Pacific guyots are generally lower than those reported on continent attached platforms �nd large isolated platforms. Accordingly, this suggest_�; that the cause of abrupt platform drowning may be related to deterioration of environmental conditions (e.g. changes in salin ity, temperature, ocean chemistry, circulation, or nutrient supply).
Nutrient excess Nutrient content is thought to control the balance between carbonate production and bioerosion on carbonate platforms (Hallock & Schlager, 1 986). Several major drowning events occurred during episodes of oxygen deficiency in the oceans, espe cially in the middle Cretaceous (Schlager, 1 99 1 ) The stratification in Cretaceous oceans was proba bly less stable and more prone to overturn than is temperature-dominated stratification in modern oceans, especially when overturn coincided with sea-level pulses (Hallock & Schlager, 1 986). Al though continent-attached platforms are more prone to nutrient excess than the mid-oceanic plat forms, several workers have attributed the demise of numerous mid-Cretaceous Pacific atolls to an excess of nutrient-rich waters. The mechanisms involved by these workers include volcanogenic upwelling (Vogt, 1 989), geothermal endo-upwelling (Rougerie & Fagerstrom, 1 994) or equatorial up welling (Larson et al. , 1 99 5). Because mid-plate submarine volcanism was widespread and intense in the middle Cretaceous Pacific, volcanogenic upwelling was considered by Vogt ( 1 989) as one possible mechanism for plat form demise. He pointed out an apparent temporal correlation between volcanism, anoxia and extinc.
Development and demise of mid-oceanic carbonate platforms tions during Aptian time. However, although the Aptian mid-plate volcanic episode overlaps, within dating errors, the 'Oceanic Anoxic Event 1 ' (Jen kyns, 1 980), new stratigraphical data show that most of the NW Pacific guyots drowned much later, during late Albian time (Winterer, 1 9 9 1 ; Grotsch & Hugel, 1 992; Premoli Siva et a!., 1 993; Sager et a!., 1 993) when volcanic activity had measurably de creased (see Fig. 3 of Vogt, 1 989). As pointed out by Rougerie & Fagerstrom ( 1 994), there is a scale problem for generalizing volcanogenic upwelling to the Pacific Ocean as a whole. Volcanogenic up welling cannot account for such a widespread, possibly global, demise of carbonate platforms (Grotsch & Flugel, 1 992) and the subsequent lack of recognized reef development in the Pacific between Cainomanian and Santonian time. The apparent temporal and spatial coincidence between the timing of guyot drowning and when these seamounts were carried across a palaeolati tude range of 0' to 1 2 ' S led some researchers to suggest that the equatorial zone may have played a role in the drowning equation. Larson et a!. ( 1 995) suggested that as these guyots entered this zone of higher nutrient availability and higher temperature, populations ofbioeroders increased and net carbon ate accumulation dropped. However, the strict ap plication of this model to Cretaceous carbonate platform communities may be questionable, as many, if not most, of the major Cretaceous shallow water carbonate producers (e.g. rudists) may have had ecological requirements different from those of modern building organisms (e.g. corals) (Camoin, 1 989). Furthermore, the equatorial upwelling is driven by trade winds and characterizes icehouse oceans that are governed by a thermohaline oceanic circulation (Hay, 1 98 8). The combination of high sea-surface temperatures, possible ice-free poles, and high sea-levels during a greenhouse climatic supercycle resulted in oceanic halothermal circula tion distinctly different from the thermohaline cir culation that characterizes icehouse periods. As a consequence, it has been suggested that no equato rial upwelling acted in the 'greenhouse' oceans, during Cretaceous and early Tertiary time (Hay, 1 988). Furthermore, extensive carbonate platforms are reported in many areas of the Maastrichtian palaeo-equatorial zone (e.g. Indian plate, East Afri can margin; Camoin et a!. , 1 993), suggesting that this zone was not so inimical to carbonate platform development.
63
Climatic and palaeoceanographic changes Short but strong perturbations of climate may explain both rapid sea-level changes and environ mental deterioration which could have removed the possibility of reef colonization and caused the complete drowning of carbonate platforms. Our drilling results, coupled with dredging data from other guyots, indicate that Pacific guyots preferen tially drowned during specific periods: late Albian, latest Maastrichtian and middle Eocene. Strong regression-transgression cycles connected to short term cooling events have been postulated for the late Albian (Grotsch & Fhigel, 1 992; Sager et a!. , 1 993), the late Maastrichtian (Camoin et a!. , 1 993) and the early middle Eocene (Haq et a!. , 1 987; Butterlin et a!., 1 993), and seemingly coincide with a drastic collapse of carbonate platforms through out the Tethyan realm. Widespread hiatuses and/or changes in calcare ous planktonic communities in the Pacific Ocean support the occurrence of strong palaeoceano graphic changes during late Albian time at shallow and deeper sites (Sliter, 1 99 5). Drowning of several carbonate platforms in the Tethys and Atlantic oceans supports a time-dependent event, as op posed to a geographical cause (i.e. palaeoequatorial zone), for the late Albian crisis (Larson et a!., 1 99 5). A plausible explanation is a cooling event, charac terized by a sharp shift of oxygen isotopes in upper Albian pelagic sediments (see Grotsch & Flugel, 1 992), that resulted in more vigorous circulation and deposition of more oxygenated sediments dur ing the Cenomanian (Larson et a!. , 1 99 5 ; Sliter, 1 99 5). During Maastrichtian time, a general cooling trend has been documented in many areas (Saltz man & Barron, 1 982; Frakes & Francis, 1 990; Huber & Watkins, 1 992; Camoin et a!., 1 993). A sharp decrease in sea-surface temperatures to about 2 1 ' C in the Pacific Ocean has been documented through oxygen-isotope fluctuations by Barrera et a!. ( 1 987). This time of moderately cool conditions began in the Maastrichtian and is thought to have continued, with fluctuations, into the late Palae ocene, when considerable warming occurred in bottom or polar water masses {see Camoin et a!. , 1 993). Maastrichtian time corresponds to a transi tional period between two modes of deep oceanic circulation (i.e. from a Cretaceous to a more Tertiary-type pattern), inducing a sharp decrease in organic productivity (see Camoin et a!., 1 993).
64
G. F. Camoin et a!.
Shifts of planktonic foraminiferal assemblages from high to lower latitudes and oxygen-isotopic data also suggest a cooling event in the early middle Eocene (see Shackleton, 1 986; Butterlin et a/. , 1 993; Diester-Haass & Zahn, 1 996). As a conclusion, the three episodes of drowning (i.e. late Albian, late Maastrichtian and mid Eocene) recorded in the Pacific coincided with rapid and high-amplitude sea-level fluctuations that acted in combination with environmental stress on the carbonate-platform ecosystems through climatic and palaeoceanographic changes in circulation and nutrient cycling.
SUMMARY AND CONCLUS IONS
The sedimentary evolution of NW Pacific guyots apparently yields a consistent scenario, irrespective of their age (Aptian-Albian, Campanian-Maastri chtian or Palaeocene-Eocene). A model for the development and the demise of mid-oceanic carbonate platforms is based on a detailed sedimen tological, seismic and geochemical study of Wode jebato Guyot (Marshall Islands, NW Pacific): 1 After a prolonged period of subaerial exposure (a few million years), the volcanic basement was pro gressively flooded and covered by organic-rich clays and pyrite-rich limestones deposited in a quiet and reducing shallow-marine environment. 2 The flooding of the volcanic platform resulted in the accumulation of carbonate sand shoals near the shelf margin, rimming an emerged volcanic island. The overlying sequence is a series of retrograda tional and then progradational sand shoals as the central part of the guyot was progressively flooded. 3 A number of Pacific guyots display atoll-like features including well-differentiated elevated plat form rims that enclose a central lagoon and are flanked by steep talus leading down to volcanic slopes. Early cemented organic frameworks were reported only as thin intervals on the inner ridge of Wodejebato. None of the drilled elevated platform rims on early Cretaceous guyots revealed the exist ence of reefal masses; rather they are composed of skeletal and oolitic sands. On the other guyots (e.g. MIT, Limalok}, the perimeter rim was not drilled, and its nature is unknown. 4 Later development of the carbonate platform is characterized by two successive periods of emer gence: a first, limited fall in sea-level, and a second, deeper fall in sea-level before the final drowning of
the carbonate platform during Maastrichtian time. The last platform carbonate sequence is a wedge deposited during a sea-level lowstand on the outer ridge, whereas the inner part of the platform was subaerially exposed and probably karstified. 5 After its demise, the carbonate platform sank rapidly into deep water, between the aragonite and the deeper calcite saturation depths, resulting in the dissolution of aragonite and concomitant precipita tion of calcite cements, possibly throughout the Cainozoic. The post-drowning succession includes a phosphate-manganese crust that formed from the late Palaeocene to the middle Eocene. 6 The drilling results of Leg 1 44 indicate that there were at least three major episodes of carbonate platform drowning within the tropical Pacific (late Albian, late Maastrichtian and middle Eocene) that apparently occurred during short-term regression transgression cycles connected to short-term cli matic and palaeoceanographic changes.
ACKNOWLEDGEMENTS
The authors wish to warmly thank the ODP Leg 1 44 Scientific Party and the shipboard personnel of the JOIDES Resolution. An early version of this manuscript benefited greatly from substantial re views by Wolfgang Schlager and Bruce Fouke.
REFERENCES AL AASM, I. & VEIZER, J. ( 1 986) Diagenetic stabilization of
aragonite and low-Mg calcite, II. Stable isotopes in rudists. J. sediment Petrol. , 56, 763-770. ALLAN, J.R. & MATTHEWS, R.K. ( 1 977) Carbon and oxygen isotopes as diagenetic and stratigraphic tools: surface and subsurface data, Barbados, West Indies. Geology, 5, 1 6-20. ANDERSON, T. & SCHNEIDERMANN, N. ( 1 973) Stable isotope relationships in pelagic limestones from the central Caribbean: Leg 1 5, Deep Sea Drilling Project. Initial Reports of the Deep Sea Drilling Project, 1 5 (Eds Edgar, N.T. & Saunders, J.B.), pp. 795-803. US Government Printing Office, Washington, DC. ARNAUD-VANNEAU, A., BERGERSEN, D., CAMOIN, G. et a/. ( 1 99 5) A model for the depositional sequences and systems tracts on small mid-ocean carbonate platforms: examples of Wodejebato Guyot (Sites 873-877) and Limalok Guyot (Site 8 7 1 ) . In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J.A., Premoli Silva, 1., Rack, F. & McNutt, M.K.), pp. 8 1 9-840. Ocean Drilling Program, College Station, TX. ARNAUD-VANNEAU, A., CAMOIN, G. & SHIPBOARD SCIEN TIFIC PARTY ( 1 993) Les edifices carbonates des atolls et
Development and demise of mid-oceanic carbonate platforms guyots d u Pacifique nord-occidental: resultats prelim inaires du Leg ODP 1 44. C.R. Acad. Sci. Paris, 317, 947-954. BARRERA, E., HUBERT, B.T., SAVIN, S.M. & WEBB, P.N. ( 1 987) Antarctic marine temperatures: late Campanian through early Palaeocene. Paleoceanography, 2, 2 1 -4 7. BATHURST, R. ( 1 977) Ordovician Meiklejohn bioherm, Nevada. Geol. Mag., 1 14, 308-3 1 1 . BERGERSEN, D. ( 1 993) Geology and geomorphology of Wodejebato (Sylvania) Guyot, Marshall Islands. In: The Mesozoic Pacific (Eds Pringle, M., Sager, W., Sliter, W. & Stein, S.), pp. 367-38 5 . American Geophysical Union, Wasington, DC. BERGERSEN, D. ( 1 99 5) Morphology of Marshall Islands guyots. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J.A., Premoli Silva, 1., Rack, F. & McNutt, M.), pp. 5 6 1 -584. Ocean Drilling Program, College Station, TX. BUTTERLIN, J., VRIELYNCK, B., BIGNOT, G. et a/. ( 1 993) Lutetian (46 to 40 Ma). Explanatory notes. In: Atlas Tethys Paleoenvironmental Maps (Eds Dercourt, J., Ricou, L.E. & Vrielynck, B. ) , pp. 1 97-209. Gauthier Villars, Paris. CAMOIN, G. ( 1 989) Les plates-formes carbonatees du Tur onien et du Senonien de Mediterranee centrale (Algerie, Tunisie, Sicile). These Doct. es-Sci., Universite de Prov ence, Marseille. CAMOIN, G., ARNAUD-VANNEAU, A., BERGERSEN, D. et a/. ( 1 995) Anatomy and evolution of the inner perimeter ridge (sites 874 and 877) of a Campanian-Maastricht ian atoll-like structure (Wodejebato Guyot; Marshall Islands). In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J.A., Premoli Silva, 1., Rack, F. & McNutt, M.), pp. 27 1 -294. Ocean Drilling Program, College Station, TX. CAMOIN, G., BELLION, Y.J.C., DERCOURT, J.G. et a/. ( 1 993) Late Maastrichtian (69.5-65 Ma). Explanatory notes. In: Atlas Tethys Paleoenvironmental Maps (Eds Der court, J., Ricou, L.E. & Vrielynck, B. ) , pp. 1 79- 1 96. Gauthier Villars, Paris. CAMOIN, G., BERNET-ROLLANDE, M.C. & PHILIP, J. ( 1 988) Rudist-coral frameworks associated with submarine volcanism in the Maastrichtian of the Pachino area (Sicily). Sedimentology, 35, 1 23- 1 28 . COLLOT, J.Y., GREENE, H.G., STOKKING, L.B. et a/. , ( 1 992) Proceedings of the Ocean Drilling Program, Initial Re ports, 1 34. Ocean Drilling Program, College Station, TX. CROUGH, S.T. ( 1 978) Thermal origin of mid-plate hot-spot swells. Geophys. J.R. astron. Soc., 55, 4 5 1 -469. DIESTER-HAASS, L. & ZAHN, R. ( 1 996) Eocene-Oligocene transition in the Southern Ocean: history of water mass circulation and biological productivity. Geology, 24, 1 63 - 1 66. DUENNEBIER, F.K. & PETERSEN, L.K. ( 1 982) Summary report, !POD Site surveys in the western Pacific, R/V Kana Keoki cruise KK81 0626, Leg 4. Hawaii Institute of Geophysics, Honolulu. EMERY, K.O., TRACEY, J.l., Jr & LADD, H.S. ( 1 954) Geology of Bikini and nearby atolls. US geol. Surv. Prof Pap. , 260-A, 265 pp. ENOS, P., CAMOIN, G.F. & EBREN, PH. ( 1 995a) Sedimentary sequence from Sites 875 and 876, outer perimeter ridge,
65
Wodejebato Guyot. In: Proceedings of the Ocean Drill ing Program, Scientific Results, 1 44 (Eds Haggerty, J.A., Premoli Silva, 1., Rack, F. & McNutt, M.), pp. 295-3 1 0. Ocean Drilling Program, College Station, TX. ENOS, P., CAMOIN, G.F., WILSON, P.A. et a/. ( 1 995b) Diagenesis of platform carbonates, Wodejebato Guyot, Marshall Islands. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J.A., Premoli Silva, 1., Rack, F. & McNutt, M.), pp. 789-8 1 8. Ocean Drilling Program, College Station, TX. ERBA, E., PREMOLI SILVA, 1., WILSON, P.A. et a/. ( 1 995) Synthesis of stratigraphies from shallow-water se quences at Sites 8 7 1 through 879 in the Western Pacific Ocean (Leg 1 44). In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J.A., Premoli Silva, I. Rack, F. & McNutt, M.), pp. 873886.0cean Drilling Program, College Station, TX. ERLICH, R.N., BARRETT, S.F. & Guo, B.J. ( 1 990) Seismic and geologic characteristics of drowning events on carbonate platforms. Bull. Am. Assoc. petrol. Geol. , 74, 1 523- 1 537. FRAKES, L.A. & FRANCIS, J.E. ( 1 990) Cretaceous paleocli mates. In: Cretaceous Resources, Events and Rhythms (Eds Ginsburg, R.N. & Beaudoin, B.), pp. 273-287. Kluwer Academic, Dordrecht. FREEMAN-LYNDE, R., FULKER WHITLEY, K. & LOHMANN, K.C. ( 1 986) Deep-marine origin of equant spar cements in Bahama Escarpment limestones. J. sediment Petrol., 56, 799-8 1 1 . GONZALEZ, L.A. & LOHMANN, K.C. ( 1 985) Carbon and oxygen isotopic composition of Holocene reef carbon ates. Geology, 13, 8 1 1-8 1 4. GROTSCH, J. & FLOGEL, E. ( 1 992) Facies of sunken early Cretaceous atoll reefs and their capping late Albian drowning succession (Northwestern Pacific). Facies, 27, 1 5 3- 1 74. HAGGERTY, J.A. &. VAN WAASBERGEN, R.J. ( 1 995) Geochemistry of MIT and Takuyo-Daisan carbonate platforms compared with other northwest Pacific guyots drilled during Leg 1 44: implications for their diagenetic history. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J.A., Premoli Silva, 1., Rack, F. & McNutt, M.), pp. 84 1 -872. Ocean Drilling Program, College Station, TX. HALLOCK, P. & SCHLAGER, W. ( 1 986) Nutrient excess and the demise of coral reefs and carbonate platforms. Palaios, 1, 389-398. HAMILTON, E.L. ( 1 956) Sunken islands of the Mid-Pacific Mountains. Geol. Soc. Am. Mem. , 64, 1 -9 1 . HAMILTON, E.L. & REx , R.W. ( 1 959) Lower Eocene phos phatized Globigerina ooze from Sylvania Guyot. US geol. Surv. Prof Pap. , 260-W, 785-797. HAQ, B.U., HARDENBOL, J. & VAIL, P.R. ( 1 987) Chronology of fluctuating sea levels since the Triassic. Science, 235, 1 1 56- 1 1 67. HAY, W.W. ( 1 988) Paleoceanography: a review for the GSA centennial. Geol. Soc. Am. Bull., 100, 1 934- 1 956. HEESTAND, R.L. & CROUGH, S.T. ( 1 9 8 1 ) The effect of hot spots on the oceanic age-depth relation. J. geophys. Res., 86, 6 1 07-6 1 1 4 . HoLMES, M.A. ( 1 995) Pedogenic alteration o f basalts recovered during Leg 1 44. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty,
66
G. F. Camoin et a!.
J.A., Premoli Silva, I., Rack, F. & McNutt, M.), pp. 3 8 1 -400. Ocean Drilling Program, College Station, TX. HUBER, B.T. & WATKINS, D.K. (1 992) Biogeography of Campanian-Maastrichtian calcareous plankton in the region of the Southern Ocean: paleogeographic and paleoclimatic implications. In: The Antarctic Paleoenvi ronment: a Perspective on Global Change (Eds Kennett, J.P. & Warnke, D.A.). Am. geophys. Union, Antarct. Res. Ser., Washington, DC, 56, 3 1 -60. JANSA, L.F. & ARNAUD-VANNEAU, A. (1 995) Carbonate buildup and sea-level changes at MIT Guyot, western Pacific. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J.A., Premoli Silva, I., Rack, F. & McNutt, M.), pp. 3 1 1 -336. Ocean Drilling Program, College Station, TX. JENKYNS, H. C. (I 980) Cretaceous anoxic events: conti nents to ocean. J. geol. Soc. London, 137, 1 7 1 - 1 88 . KENDALL, A . ( 1 977) Fascicular-optic calcite. A replace ment of bundled acicular carbonate cements. J. sedi ment Petrol. , 47, 1 056-1 062. KENDALL, A. (1 985) Radiaxal fibrous calcite: a reappraisal. In: Carbonate Cements (Eds Schneidermann, N. & Harris, P.M.), Spec. Pub!. Soc. econ. Paleont. Miner., Tulsa, 36, 59-77. KONISHI, K. ( 1 989) Limestone of the Daiichi Kashima Seamount and the fate of a subducting guyot: fact and speculation from the Kaiko-"Nautile" dives. Tectono physics, 160, 249-265. KROOPNICK, P.M., MARGOLIS, S.V. & WONG, C.S. (I 977) 1i 1 3 C variations in marine carbonate sediments as indi cators of the C02 balance between the atmosphere and oceans. In: The Fate of Fossil Fuel C02 in the Oceans (Eds Anderson, N.R. & Malahoff, A.), pp. 295-32 1 . Plenum, New York. LARSON, R., ERBA, E., NAKANISHI, M., BERGERSEN, D.D. & LINCOLN, J.M. (1 995) Stratigraphic, vertical subsidence and paleolatitude histories of Leg 1 44 guyots. In: Pro ceedings of the Ocean Drilling Program, Scientific Re sults, 1 44. (Eds Haggerty, J.A., Premoli Silva, I., Rack, F. & McNutt, M.), pp. 9 1 5-934. Ocean Drilling Pro gram, College Station, TX. LINCOLN, J., PRINGLE, M. & PREMOLI SILVA, I. (1 993) Early and late Cretaceous volcanism and reef-building in the Marshall Islands. In: The Mesozoic Pacific (Eds Pringle, M., Sager, W., Sliter, W. & Stein, S.), Geophys. Monogr. Ser., Am. geophys. Union, Washington, DC, 77, 279305. LoHMANN, K.C. & MEYERS, W. ( 1 977) Microdolomite inclusions in cloudy prismatic calcites. J. sediment Petrol. , 47, 1 07 8- 1 088. LOHMANN, K.C. & WALKER, J. ( 1 989) The 0 1 8 0 record of Phanerozoic abiotic marine calcite cements. Geophys. Res. Lett. , 16, 3 1 9-322. LONGMAN, M.W. (1 980) Carbonate diagenetic textures from nearsurface diagenetic environments. Bull. Am. Assoc. petrol. Geol., 64, 46 1 -487. MATTHEWS, J.L., HEEZEN, B.Z., CATALANO, R. et a/. (1 974) Cretaceous drowning of reefs on Mid-Pacific and Japa nese guyots. Science, 184, 462-464. McCLAIN, W.R., FREEMAN-LYNDE, R.P. & LOHMANN, K.C. ( 1 988) Petrography and stable oxygen and carbon iso tope compositions of Campanian grainstones and rud stones, Northeast Providence Channel, Bahamas: ODP
Leg 1 0 1 , Hole 634A. In: Proceedings of the Ocean Drilling Program, Scientific Results, I 0 I (Eds Austin, J.A., Jr & Schlager, W.), 245-25 3 . Ocean Drilling Program, College Station, TX. MELIM, L.A., SWART, P.K. & MALIVA, R.G. (1 995) Meteoric-like fabrics forming in marine waters: impli cations for the use of petrography to identify diagenetic environments. Geology, 23, 755-758. MOLDOVANYI, E. & LOHMANN, K. (1 984) Isotopic and petrographic record of phreatic diagenesis: Lower Cre taceous Sligo and Cupido formations. J. sediment Petrol. , 54, 972-985. NAKANISHI, M. & GEE, J.S. (I 995) Paleomagnetic investi gations of volcanic rocks: paleolatitudes of the north western Pacific guyots. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J.A., Premoli Silva, I., Rack, F. & McNutt, M.), pp. 5 85-604. Ocean Drilling Program, College Station, TX. ODP LEG 1 44 SHIPBOARD SCIENTIFIC PARTY (I 993) Insight on the formation of Pacific guyots from Leg 1 44. EOS, 74, 3 5 8 . PARSONS, B. & SCLATER, J.G. ( I 9 7 7 ) Analysis o f the variation of the ocean floor bathymetry and heat flow with age. J. geophys. Res. , 93, 1 1 753- 1 1 7 7 1 . PREMOLI SILVA, I., HAGGERTY, J., RACK, F. et a/. (Eds) (1 993) Proceedings of the Ocean Drilling Program, Ini tial Reports, 1 44. Ocean Drilling Program, College Station, TX. PREMOLI SILVA, I., NICORA, A., ARNAND-VANNEARU, A. et a!. (I 995) Paleobiogeographic evolution of shallow water organisms from Aptian to Eocene in western Pacific (Leg 1 44). In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J.A. Premoli Silva, I., Rack, F. & McNutt, M.), pp. 887-894. Ocean Drilling Program, College Station, TX. PRINGLE, M.S. & DuNCAN, R.A. (1 995) Radiometric ages of basement lavas recovered at Loen, Wodejebato, MIT and Takuyo-Daisan guyots, Leg 1 44, northwestern Pa cific Ocean. In: Proceedings of the Ocean Drilling Pro gram, Scientific Results, 1 44 (Eds Haggerty, J.A., Premoli Silva, I., Rack, F. & McNutt, M.), pp. 547-560. Ocean Drilling Program, College Station, TX. ROUGERIE, F. & FAGERSTROM, J.A. (1 994) Cretaceous history of Pacific basin guyot reefs: a reappraisal based on geothermal endo-upwelling. Palaeogeogr. Palaeocli matol. Palaeoecol., 112, 239-260. SAGER, W.W., WINTERER, E.L., FIRTH, J.V. et a/. (Eds) (1 993) Proceedings of the Ocean Drilling Program, Ini tial Reports, 1 43. Ocean Drilling Program, College Station, TX. SALLER, A. H. ( I 986) Radiaxial calcite in Lower Miocene strata, subsurface Enewetak atoll. J. sediment Petrol., 56, 743-762. SALTZMAN, E. & BARRON, E.J. (I 982) Deep circulation in the Late Cretaceous: oxygen isotope paleotemperatures from Inoceramus remains in DSDP cores. Palaeogeogr. Palaeoclimatol. Palaeoecol., 40, 1 6 7- 1 8 1 . SANDBERG, P.A. (I 985) Aragonite cements and their occur rence in ancient limestones . . In: Carbonate Cements (Eds Schneidermann, N. & Harris, P.N.). Spec. Puql. Soc. econ. Paleont. Miner., Tulsa, 36, 33-57. ScHLAGER, W. (I 9 8 1 ) The paradox of drowned reefs and carbonate platforms. Geol. Soc. Am. Bull. , 92, 1 97-2 1 1 .
Development and demise of mid-oceanic carbonate platforms ScHLAGER, W. ( 1 99 1 ) Depositional bias and environmen tal change-important factors in sequence stratigraphy. Sediment Geol. , 70, 1 09- 1 30. SCHLAGER, W. & JAMES, N.P. ( 1 978) Low-magnesium calcite limestones forming at the deep-sea floor, Tongue of the Ocean, Bahamas. Sedimentology, 25, 675-702. SCHLANGER, S., CAMPBELL, J. & JACKSON, M. ( 1 987) Post Eocene subsidence of the Marshall Islands recorded by drowned atolls on Harrie and Sylvania guyots. In: Seamounts, Islands and Atolls (Eds Keating, B., Fryer, P., Batiza, R. & Boehlert, G.), Geophys. Monogr. Ser., Am. geophys. Union, Washington, DC, 43, 1 65 - 1 74. SCHOLLE, P.A., ARTHUR, M.A. & EKDALE, A.A. ( 1 983) Pe lagic environment. In: Carbonate Depositional Environ ments (Eds Scholle, P.A., Bebout, D.G. & Moore, C. H.). Mem. Am. Assoc. petrol. Geol., Tulsa, 33, 620-69 1 . SHACKLETON, N.J. ( 1 986) Paleogene stable isotope events. Palaeogeogr. Palaeoclimatol. Palaeoecol. 57, 9 1 - 1 02. SLITER, W.V. ( 1 995) Cretaceous planktonic foraminifers from Sites 865, 866 and 869. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 43 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. et a/.), pp. 1 5-30. Ocean Drilling Program, College Station, TX. VAN WAASBERGEN, R.J. & WINTERER, E.L. ( 1 993) Summit geomorphology of western Pacific guyots. In: The Meso zoic Pacific: Geology, Tectonics and Volcanism (Eds Pringle, M.S., Sager, W.W., Sliter, W.V. & Stein. S.). Geophys. Monogr., Am. geophys. Union, Washington, DC, 77, 3 3 5-366.
67
VoGT, P.R. ( 1 989) Volcanogenic upwelling of anoxic, nutrient-rich water: a possible factor in carbonate-bank/ reef demise and benthic faunal extinctions? Geol. Soc. Am. Bull., 101, 1 225- 1 245. WHEELER, C.W. & AHARON, P. ( 1 99 1 ) Mid-oceanic carbon ate platforms as oceanic dipsticks: examples from the Pacific. Coral Reefs, 10, 1 0 1 - 1 1 4. WILSON, P.A., OPDYKE, B.N. & ELDERFIELD, H. ( 1 995) Strontium isotope geochemistry of carbonates from Pacific guyots. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J.A., Premoli Silva, 1., Rack, F. & McNutt, M.), pp. 447-458. Ocean Drilling Program, College Station, TX. WINTERER, E.L. ( 1 99 1 ) The Tethyan Pacific during late Jurassic and Cretaceous time. Palaeogeogr. Palaeocli matol. Palaeoecol. 87, 2 53-265. WINTERER, E.L. & METZLER, C.V. ( 1 984) Origin and subsidence of guyots in Mid-Pacific Mountains. J. geo phys. Res., 89, 9969-9979. WYATT, J.L., QuiNN, T.M. & DAVIES, G.R. ( 1 995) Prelim inary investigation of the petrography and geochemistry of limestones at Limalok and Wodejebato guyots (Sites 87 1 and 874), Republic of the Marshall Islands. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J.A. Premoli Silva, 1., Rack, F. & McNutt, M.), pp. 429-437. Ocean Drilling Pro gram, College Station, TX.
Spec. Pubis int. Ass. Sediment. ( 1998) 25, 69-76
Stable tropics not so stable: climatically driven extinctions of reef-associated molluscan assemblages (Red Sea and western Indian Ocean; last interglaciation to present) M. TA V I AN I Instituto di Geologia Marina del CNR, via Gobetti 101, 40129 Bologna, Italy
ABSTRACT
The Indo-West Pacific coral reefs have experienced dramatic faunal turnovers since the last interglacial period, as documented by local extinctions and contractions of associated faunas in the western Indian and Pacific oceans. Although refrigeration of sectors of the Indo-Pacific region is a possible concomitant limiting factor, rate of sea-level change is considered the most important constraint in controlling the ultimate fate of Quaternary coral reefs. Disruption of internal organization at times of rapid sea-level fluctuations may lead to progressive depauperation of coral reef biota through local extinctions. In the case of the shallow-silled Red Sea basin, effects of sea-level changes were dramatically amplified and the entire basin underwent massive destruction of its stenoecious biota as a result of the onset of high-salinity conditions. Similar disturbances punctuated the entire Quaternary ice age, possibly causing many, high-frequency faunal rearrangements of variable intensity in coral reef ecosystems. Habitat fragmentation within the Indo-West Pacific region is seen as a major mechanism to account for the high level of species-level biodiversity witnessed throughout the Cainozoic. Speciation is apparently promoted through gene-flow disruption among populations within the Indo-West Pacific during times of relative lowstands. Endemics seem to be preserved rather than lost, when populations reconnect at the re-establishment of highstand conditions. Thus, the Indo-West Pacific tropical region acts as a vast refuge at peaks of glacial difficulties. However, this is not the general rule for the tropics, as indicated by a decrease in coral diversity in the Caribbean. During the last glaciation, the capacity of the Red Sea and Persian Gulf to sequester C02 through calcification in both the pelagic and neritic domains practically reached zero. This fact, and the concomitant reduction of coral reefs in other areas of the Indo-Pacific region, should be taken into account in the evaluation of the global C02 cycle.
INTRODUCTION
short-term (101-102 yr: e.g. Loya, 1976, 1990; Mc Glade, 1990; Smith & Buddemeier, 1992) or very long-term disturbances (105-106 yr: e.g. Fager strom, 1987; Kauffman & Fagerstrom, 1993). Re sponse of reefs to stress at a time-scale ranging from 103-104 yr has only seldom been investigated, al though this is the scale of Pleistocene glacial cycles, which would allow testing of coral reef reaction to both temperature and sea-level fluctuations. The main factors controlling the fate of coral reefs during ice ages have been the topic of many debates, with suggestions ranging from sea-level fluctuations to changes in temperature regime (e.g. Stoddart, 1976; Taylor, 1978; Stanley, 1984; Crame, 1986; Paulay, 1990, 1991, 1996b).
The highly diversified coral reefs, corresponding in the oceans to what tropical rain-forests represent on the continents (Jackson, 1991), have long embodied the common notion of tropical stability (Newell, 1971). This assumption, however, is being chal lenged as our understanding of Pleistocene glacial age dynamics and its impact on marine biota grows. Significant faunal turnovers have punctuated the Cainozoic history, greatly affecting marine benthic communities, including tropical ones (see review by Stanley & Ruddiman, 199 5). The vulnerability of coral reef ecosystems to climatic disturbances should be explored at dif ferent time-scales. Most previous studies have fo cused on the response of coral reefs to either very
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
69
70
M. Taviani
Much concern exists about the recovery of coral reefs from habitat destruction. Proposed scenarios for reefs in the future are grim (McGlade, 1990; Stone, 1995), but coral reef resilience is still largely unknown (e.g. Loya, 1990; Kaufman, 1993). There fore, a better understanding of how these relatively fragile ecosystems respond to climatic disturbances and issues such as extinction and evolution rates of their faunal associates may ultimately provide guidelines for conservation policies (e.g. Wilkinson, 1991; Huston, 1994). The goal of this paper is to review the available evidence of reef-associated molluscan turnovers during the last glacial cycle in the western Indian Ocean, particularly the Red Sea, and to discuss their evolutionary and global-climatic meaning.
LAST GLACIAL EXTINCTIONS IN THE INDIAN OCEAN CORAL REEFS
General
Stoddart (1973, 1976) presented the concept of massive local extinctions in Indo-West Pacific reef communities caused by Pleistocene sea-level fluctu ations. The comparative work on Quaternary mol luscan assemblages from Aldabra atoll by Taylor (1978) provided documentation of significant changes in the structure of pre- and post-glacial Indian Ocean coral reefs. The notion of Pleistocene to Recent coral reef faunal turnovers was reinforced by the study of outcrops in Kenya, on the East African mainland (Crame, 1986). Emblematic of these two studies is the discovery that giant clams, i.e. Tridacna gigas (Linnaeus, 1758), T crocea Lamarck 1819 and Hippopus hippopus (Linnaeus, 1758) inhabited the reefs of the westernmost Indo West Pacific region during the last interglaciation. Today, the geographical range of these clams is considerably restricted (Rosewater, 1965) as a con sequence of climatically driven extinctions which affected the western Indian Ocean during the last glacial period (Crame, 1986). Whether these extinc tions were selective or more widespread is not known for sure. However, it is clear that the local extinction ofT gigas, other giant clams and gastro pods (Crame, 1986) should not be considered iso lated phenomena. In fact, the same fate was shared by other reef-associated molluscs in the Red Sea and Gulf of Aden. Moreover, similar trends have
been reported for Pacific Ocean coral reefs (Kohn, 1980; Vermeij, 1986; Paulay, 1996b). Extinctions of Red Sea marine biota during the last glacial age
By integrating deep-sea and coral reef data, Taviani (1996a,b) concluded that the Red Sea stenoecious fauna and flora were almost completely annihilated during the last glacial period by a combination of high salinity (c. 50o/oo at least) and sea-level changes. Present-day biota merely represent a recolonization which began c. 10 000 yr ago. This eradication of stenoecious biota from the Red Sea was previously suggested by Sewell (1948) and Gvirtzman et a!. (1977) among others. Taviani (1982, 1997b) has ob served that last interglacial (Eemian, isotopic sub stage 5e) coral reefs of the Red Sea-Gulf of Aden region house molluscan faunas which are qualita tively different from their modern counterparts. This author listed examples of molluscan extinctions, i.e. Diodora impedimentum (Cooke, 1885); geographi cal contractions, i.e. Rhinoclavis vertagus (Linnaeus, 1767), Cerithium madreporicolum (Jousseaume, 1930), Columbella turturina (Lamarck, 1822), Co nus litteratus (Linnaeus, 1758), Cucullaea cucullata (Roeding, 1798); and rarefactions, i.e. Cypraea mon eta (Linnaeus, 17 58), Oliva bulbosa (Roeding, 1798), Corbula taitensis (Lamarck, 1818). Analogous con tractions are observed among the corals Turbinaria peltata (Esper, 1794), Cycloseris vaughani (Bosch rna, 1923) and Pavona minuta (Wells, 1954): (see Dullo, 1987). Por (1989) has offered a different view on this last glacial basinal stress of the Red Sea. First, he argued that some metahaline-tolerant organisms (e.g. molluscs, echinoderms, corals) could stand salinities even greater than 48o/oo, as is the case today in the Persian Gulf, thus making a point against the complete destruction of the Red Sea fauna during the last glacial. He also reports that claimed endemism among certain fish species implies permanent viable conditions in the Red Sea, possibly since the Pliocene (Klausewitz, 1983). In practice, the hypothesis that diverse coral reefs and their associated stenoecious faunas could toler ate protracted periods (thousands of years) of salin ities in excess of 50o/oo is difficult to accept. Moreover, the significance of Red Sea fish ende mism is tempered by the admission of Klausewitz (1983), who reported these endemics in the Gulf of Aden, and therefore present in the Indian Ocean.
Stable tropics not so stable
Some Red Sea-Gulf of Aden gastropod endemics have sibling species in the Indo-West Pacific region, as exemplified by the pairs Cypraea tigris (Linnaeus, 1758) (Indo-Pacific)-C. pantherina (Lightfoot, 1786) (Red Sea-Gulf of Aden), Cypraea vitellus (Linnaeus, 1758) (Indo-Pacific)-C. camelopardalis (Perry, 1811) (Red Sea-Gulf of Aden), Cypraea talpa (Linnaeus, 1758) (Indo-Pacific)-C. exusta (Sowerby, 1832) (Red Sea-Gulf of Aden). It is im portant to note that these endemics have been ob served from last interglacial (5e) coral reefs of the Red Sea (Borri et a!., 1982, and personal data from Zabargad island, Egypt). Therefore, the origination of these endemics predates the proposed biological sterilization of the Red Sea during the last glacial period. In fact, these Red Sea endemics had to survive outside the basin. The Gulf of Aden has been proposed as a glacial refuge by Por (197 5). New evidence seems to indicate that Gulf of Aden coral reefs were seriously affected by the last glacial period. The extinction of the limpet Diodora imped imentum, and the contraction of the geographical range of the cerithiid Rhinoclavis vertagus, known from last interglacial coral reefs of both the Red Sea proper and the Gulf of Aden, and now absent from this region, support this view (Taviani, 1997b ). Studies by Taylor (1978) and Crame (1986) show that significant disturbances of coral reef biota of both western Indian Ocean islands and mainland at equatorial latitudes have occurred. To account for these pre-5e endemics, it seems probable that one or more geographical refuges did exist in the western Indian Ocean. Crame (1986) suggested that a pos sible refuge was Oman. It has always been tempting to look at the Red Sea as an ideal place where new species originate, given its relative isolation and 'abnormal' hydrolog ical attributes (e.g. Head, 1987; Por, 1989). The presence of a relatively high number of endemics seemed to support this view. However, the scenario of their eradication from the basin and their succes sive reintroduction from the Indian Ocean after the last glaciation seems to falsify this hypothesis. If these endemics were only confined to the Red Sea, how did they survive outside the basin, under conditions that did not mimic their Red Sea ones? The frequently emphasized biological and hydro logical uniqueness of the Red Sea (e.g. Head, 1987; Por, 1989) is possibly misleading for evaluation of its true role in the economics of the Indo-Pacific species-rich region. In fact, the Red Sea is simply a
71
satellite basin of the Indo-West Pacific, under the recurrent threat of major hydrological disturbances, linked to the very short-term glacial cycles. The Red Sea: a small-scale abrupt mass extinction
During the last glaciation, it is very likely that the Red Sea experienced the extinction events affecting vast areas of the Indian Ocean (Taylor, 1978; Crame, 1986). The basin, however, deviated from equilibrium as predicted by the island biogeography hypothesis, by the unexpected mass destruction, a result of intolerable hydrological conditions. It should be noted that the last glacial hydrological turmoil affecting the Red Sea, with its profound repercussions, may be described as a disturbance regime in island ecology (Whittaker, 1995). Unfor tunately, the critical steps of progressive reactions of Red Sea biota to increasingly harsh environmen tal conditions (disturbance) are virtually unknown. The signature of step-by-step disruption of the 'last interglacial equilibrium conditions' by increasing extinction rates not balanced by immigrations until its final collapse, is probably sealed within drowned relative lowstand reefs, at present unavailable to detailed palaeobiological analyses. The peculiar morphology of the basin amplified to apocalyptic levels what was a general disturbance of the tropical marine biota during the last glaciation because of progressive area reduction as a result of sea-level lowering. The final results were disproportionate only for the Red Sea, because there it ended in a true mass extinction. The Indo-West Pacific as a whole escaped massive glacial extinctions almost unharmed, thus helping to replenish the Red Sea as well as other peripheral areas where coral reef biota had been more or less completely stripped off. An extreme case is the shallow Persian Gulf, which becomes completely exposed and its biota des troyed at times of glacial maxima (Lambeck, 1996). Were it not for its location adjacent to a large marine tropical reservoir, almost all of the Red Sea stenoecious fauna and flora would have disap peared forever. For this reason, the Red Sea may serve as a scenario to understand better the pro cesses of mass extinction and successive recovery (Jablonski, 1986; McLaren, 1986). In my view the Red Sea faunal extinctions and reinvasions offer a clue to understanding speciation in tropical marine environments. Before describing
M. Taviani
72
the possible patterns of the origination and perpet uation of species in the Indo-West Pacific coral reef ecosystems, we must first analyse the two main factors thought to control the evolution of coral reefs, i.e. temperature and sea-level.
TEMPERATURE AND SEA-LEVEL CHANGES AS CORAL REEF LIMITING FACTORS
negatively affecting coral reef ecosystems of the Red Sea, which are located in the north-westernmost comer of the Indian Ocean. It is, however, difficult to believe that temperature was equally significant in causing important faunal rearrangements and extinctions in Aldabra and Kenya (Taylor, 1978; Crame, 1986) as well as in the Pacific (Kohn, 1980; Vermeji, 1986; Paulay, 1990, 1996b). Salinity, thought to have been responsible for mass extinction in the Red Sea proper, is intimately linked to the second disturbance, sea-level change.
Role of temperature
Modem coral reefs are constrained by winter min imum temperatures and are generally restricted within the 18·c isotherm (Newell, 1971; Belasky, 1996). Coral reefs can tolerate much lower winter temperatures, as documented in the Arabian Gulf where reefs routinely survive exposures to winter minima around Ire (Coles & Fadlallah, 1991). However, long-term temperature depression below the 18 • C isotherm will ultimately cause coral reef regression and extinction (e.g. Belasky, 1996). We do not have precise data about significant temperature changes of surficial waters in the trop ical region under scrutiny to be tied to the fate of individual coral reefs during the last glacial period. Stoddart (1973) drew a map showing a significant contraction of the 2o·c isotherm throughout the tropical belt during the glacial Pleistocene. This reconstruction, however, is biased by the absence of ground evidence. Most palaeotemperature data from the region are indirectly derived from stable oxygen isotope data on benthic and plankton Fora minifera and transfer functions using microfossils. Re-evaluation of the latter is now showing that temperature depression of the tropical Atlantic and eastern Pacific oceans during the ice ages was larger than previously thought by CLIMAP reconstruc tions (Mix, 1996). Micropalaeontological and isotopic data from offshore Red Sea indicate glacial temperatures 5 • C lower than at present in the central Red Sea, and 3.5·c lower than at present in the Gulf of Aden (Ivanova, 1985). Fluctuations of such amplitude are considered to be influential on tropical organisms (Barron, 1995). Long-term low temperatures dra matically affect tropical biota diversity in the Car ibbean region (Edinger & Risk, 1994). Before conditions of high salinity became the leading cause of their annihilation, lower temperatures during the last glacial period, may have been indeed a co-factor
Role of sea-level
Sea-level changes control coral reef growth and have been described as the dominant force influencing faunal diversity by affecting area-size (e.g. Wise & Schopf, 1981). Conversely, sea-level changes may also induce extinctions (Hallam, 1989), although probably only at the species level during the Cain ozoic (Jablonski, 1985). Paulay (1990, 1996b) examined the role of sea level fluctuations and concluded that it is the prin cipal mechanism driving oceanic island extinctions in the tropics. Fundamentally I agree with this view, although I believe that the rate of sea-level change is the most important factor. My contention is that fast rates of sea-level fluctuations (both rise and fall) impede the setting of reasonably diverse coral reefs, as they give no time for a fully topographical expansion and, therefore, formation of exploitable habitats where biotic interactions can be fostered at incremental rates. Back-reef (lagoonal) environ ments are more exposed to disruption and require more time for their full re-establishment, and this would explain why Pleistocene local extinctions affected mostly inner-reef specialists (Paulay, 1996b). By using the present-day situation as our stan dard, we can observe that recolonization of entire basins (Red Sea and Persian Gulf ), and wide sectors of continental and oceanic islands was achieved in the last few thousands year. The Red Sea is a particularly striking example, as it re-acquired its pre-glacial morphological and biological complexity from Bab-el-Mandab at 12 •N up to the northern reaches of Aqaba, at 29.27'N, in less than 10 kyr. Stable (still-stand) sea-level conditions were achieved in the last 5000-6000 yr, so setting a maximum age for the present-day reef complexes.
Stable tropics not so stable DISCUSSION
Tropical speciation by sea-level habitat-area fragmentation
Trends of local extinctions between the last and present interglacial periods are commonplace throughout the Indo-West Pacific region, as shown by their occurrence also in Pacific islands (Kohn, 1980; Vermeij 1986; Paulay, 1996b). Although geographically distant, such faunal turnovers basi cally represent similar responses of the Indo-West Pacific region to disturbances triggered by glacial ages. As discussed by Vermeij (1986) and Crame ( 1986), the tropical Indo-West Pacific served as a refuge throughout the Cainozoic. Briggs (199 5) proposed that the East Indies operates as a radia tion centre but also accumulates over time older genera and species generated elsewhere in the re gion. The tremendous advantage of having a refuge of such a size is dramatically shown by the modest rate of extinctions with respect to originations. According to available evidence, only one taxon (Diodora impedimentum) was apparently lost be cause of the severe disturbances to coral reefs in the western Indian Ocean during the last glacial, and it belonged to the most affected area, i.e. the Red Sea-Gulf of Aden. Many other species contracted their geographical range but did survive through the crisis. It is very important to observe that even 'endemics', whose area is by definition relatively small and whose survival is easily jeopardized by area-reduction, survived the biotic crisis. On the other hand, overall biodiversity of this region seems to have increased steadily throughout the Caino zoic. Is it that Neogene ice ages are beneficial to biodiversity in the tropical marine realm? It seems that the answer is yes, but why? As discussed, temperature drops have a negative effect on coral reefs, therefore there must be another link. Sea-level glacio-eustatic fluctuations are the most logical ex planation, as they caused habitat fragmentation within the continuum of the Indo-West Pacific region (e.g. Paulay, 1996b). In this scenario of enhanced 'provincialism', such disruptions elimi nated gene flow between mother and some periph eral populations, ultimately promoting speciation (e.g. Mayr, 1982; Fagerstrom, 1983; Stanley, 1986; Briggs, 1995; Johnson et al., 1995). From a single faunal core (the Indo-West Pacific), a number of original situations may arise at each significant sea-level fluctuation, when some viable, relict pop-
73
ulations may find the conditions suitable for genetic drift, so that a net gain in biodiversity is achieved by accumulation through time (Crame, 1986; Hus ton, 1994; Briggs, 1995). Furthermore, investiga tion of the Red Sea shows that, once originated, new species are not easily lost, and thus, biodiversity continuously increases. This increase probably di rectly correlates with the size of the refuge. The function played by refugia in the Indo-West Pacific, not only to preserve species but also to enhance speciation during the Pleistocene, appears, therefore, conceptually similar to the role embodied by refugia of the Amazonian rain-forests (Haffer, 1969; Lynch, 1988). The Pleistocene alone is punctuated by some 20 major ice cycles; such conditions accommodate multiple separations and re-unifications of marine biota at the scale of thousands or tens of thousands of years. Advances in taxonomy show that Indo West Pacific tropical ecosystems support a large number of sibling species, indicating a level of provinciality among reef organisms unknown until recently, (Knowlton, 1993; Knowlton & Jackson, 1994; Paulay, 1996a, and references therein). Sib ling species may be the product of isolation at times of relative sea-level lowerings, originating in some glacial refuges. Although not necessarily alternative, this model does not require changes in ocean circu lation to explain some patterns of coral reef evolu tion and biogeography (Veron, 199 5). The sea-level hypothesis can be tested by examining the parallel story of Caribbean tropical ecosystems. Caribbean coral reefs show a trend of reduction in biodiversity during the Cainozoic (e.g. Edinger & Risk, 1994). The difference in behaviour between the two tropical regions may be due largely to area differences. The lack of a sizeable area (and related faunal-reservoir) in the Caribbean makes the region's coral reef biota more vulnerable to unfavourable environmental changes (Edinger & Risk, 1994), and may be related to possible extinctions at the time of regressions (Boecklen & Simberloff, 1986). Sea-level changes are not necessarily the only explanation for extinction and origination rates of past tropical reef ecosystems. In fact, before the onset of Quaternary ice ages, with their paroxysmal sea-level fluctuations, temperature may have been the most important single factor in controlling the fate of tropical marine biodiversity (e.g. Stanley, 1986). It is also important to point out that this· model may possibly explain some patterns of short time-scale speciation, and is largely related to the
M. Taviani
74
nentic domain. For instance, global catastrophic events, such as those linked to bolid-impacts or anoxic oceans (McLaren, 1986), offer totally dif ferent avenues to extinctions and originations also at supraspecific level. Extinction and contraction of coral reefs during the last glaciation: global carbon implications
The hypothesis that coral reefs are important C02 reservoirs and sinks, by means of carbonate precip itation and dissolution, is supported by investiga tors concerned with 'greenhouse' scenarios (Berger, 1982; Kinsey & Hopley, 1991; Opdyke & Walker, 1992, Smith & Buddemeier, 1992; Milliman, 1993). Opdyke & Walker (1992) have recently suggested that the locus of global carbonate produc tion is modulated by sea-level changes and shifts from the deep sea during glacials to the shelves, which include coral reefs, during interglacials. Mil liman & Droxler (1996) have estimated that, during sea-level highstands, the neritic environments (which include coral reefs) produce an amount of carbonate comparable with carbonate sequestered by pelagic calcification at times of lower sea-level. Within this frame, the Red Sea is only incom pletely fulfilling what was expected of the simple CaCOrpump proposed by Opdyke & Walker (1992). In fact, calcification on the shelf is active during interglacial highstands (Kinsey & Hopley, 1991; Milliman, 1993; Milliman & Droxler, 1996) and the Red Sea contributes to reducing the oceans' capacity to hold C02• However, basinal high salinity conditions during the last glacial period prevented the establishment of biogenic carbonate factories in both the Red Sea deep-sea (pelagic) and shelf domains (Taviani, 1997a). Thus, carbonate production of the Red Sea did not contribute to sequester any significant amount of carbonate dur ing a part at least of the last glacial period, and this has been the case also for the Persian Gulf and other reefs of the Indo-West Pacific. The total number of missing calcification-reef factories is possibly signif icant enough to be taken into consideration in modelling global carbonate fluxes during the last glacial cycle.
CONCLUSIONS 1 During the last glaciation, the Indo-West Pacific coral reef belt suffered significant faunal turnovers
as documented by comparative studies of last inter glacial and modern reef-associated molluscan fau nas in the Red Sea, Gulf of Aden, Persian Gulf, western Indian Ocean and Pacific islands. 2 The rate of sea-level change is singled out as the most important factor in controlling the fate of Quaternary coral reefs. As these disturbances oc curred regularly throughout the late Cainozoic, one may hypothesize that they are responsible for mul tiple high-frequency faunal rearrangements in coral reef ecosystems through time (Paulay, 1991). 3 The habitat fragmentation within the Indo-West Pacific region during the Quaternary ice ages may promote speciation by severing gene flow during times of relative lowstands. The core of the Caino zoic Indo-West Pacific is probably acting as a refuge preserving older species during glacial times, thus enhancing an increase in biodiversity. 4 The significant reduction of coral reefs during the last glaciation should be considered when modelling the global C02 cycle.
ACKNOWLEDGEMENTS
CNR (Italian National Research Council) and EC grants SCI *CT91-0719, ERB SCI*CT92-0814 (RED SED) and EV 5V-CT94-0447 (TESTREEF) provided funding to investigate the Red Sea and the Western Indian Ocean. A. Crame, W.C. Dullo, L. Montaggioni, G. Paulay and J. Taylor are gratefully acknowledged for useful discussions on this topic and for providing references. P. Bart, G. Camoin, A. Droxler, B. Thomassin and J. Wise critically re viewed the paper and offered useful comments. Drawings were produced by G. Zini. This is IGM Scientific Contribution 1058 and TESTREEF Con tribution 18.
REFERENCES
BARRON, E.J. ( 1995) Tropical climate stability and impli cations for the distribution of life. In: Effects of Past Global Change on Life. Stud. Geophys., National Acad emy Press, Washington, DC, 108- 117. BELASKY, P. ( 1996) Biogeography of Indo-Pacific larger foraminifera and scleractinian corals: a probabilistic approach to estimating taxonomic diversity, faunal similarity, and sampling bias. Palaeogeogr. Palaeocli matol. Palaeoecol., 122, 119- 141. BERGER, W.H. (1982) Increase in carbon dioxide in the atmosphere during deglaciation: the coral reef hypothe sis. Naturwissenschaften, 69, 87-88.
Stable tropics not so stable BOECKLEN, W.J. & SIMBERLOFF, D. (1986) Area-based
extinction models in conservation. In: Dynamics of Extinction (Ed. Elliott, D.K.), pp. 247-276. John Wiley, New York. BORRI, M., SABELLI, B. & TAVIANI, M. (1982) Cypraea (Talparia) exusta Sowerby, 1832, from the central Red Sea: new data. Lav. Soc. Malacol. It., 103-107. BRIGGS, J.C. (1995) Global Biogeography. Developments in Paleontology and Stratigraphy, 14. Elsevier, Amster dam. COLES, S.L. & FADLALLAH, Y.H. ( 1991) Reef coral survival and mortality at low temperatures in the Arabian Gulf: new species-specific lower temperature limits. Coral Reefs, 9, 231-237. CRAME, J.A. (1986) Late Pleistocene molluscan assem blages from the coral reefs of the Kenya coast. Coral Reefs, 4, 183-196. DULLO, W.C. (1987) Fazies und Fossile Ueberlieferung der
Pleistozaen Riffterrassen and der Ostkueste des Roten Meeres. Habilitation Thesis, University of Erlangen, Germany.
EDINGER, E.N. & RISK, M.J. (1994) Oligocene-Miocene
extinction and geographic restriction of Caribbean cor als: roles of turbidity, temperature, and nutrients. Palaios, 9, 576-598. FAGERSTROM, J.A. ( 1983) Diversity, speciation, endemism and extinction in Devonian reef and level-bottom com munities, eastern North America. Coral Reefs, 2, 65-70. FAGERSTROM, J.A. (!987) The Evolution of Reef Commu nities. John Wiley, New York. GVIRTZMAN, G., BUCHBINDER, B., SNEH, A., NIR, Y. & FRIEDMAN, G.M. (1977) Morphology of the Red Sea fringing reefs: a result of erosional pattern of the last-glacial low-stand sea level and the following Ho locene recolonization. Mem. BRGM, 89, 480-491. HAFFER, J. (1969) Speciation in Amazonian forest birds. Science, 165, 131-137. HALLAM, A. (1989) The case for sea-level change as a dominant causal factor in mass extinction of marine invertebrates. Phil. Trans. R. Soc. London, Ser. B, 325, 437-455. HEAD, S.M. (1987) Introduction. In: Key Environments: the Red Sea (Eds Edwards, A.J. & Head, S.M.), pp. 121. Pergamon Press, Oxford. HusTON, M.A. (1994) Biological Diversity. Cambridge University Press, Cambridge. IVANOVA, E.V. (!985) Late Quaternary biostratigraphy and paleotemperatures of the Red Sea and the Gulf of Aden based on planktonic foraminifera and pteropods. Mar. Micropaleontol., 9, 335-364. JABLONSKI, D. (1985) Marine regressions and mass extinc tions: a test using the modern biota. In: Phanerozoic Diversity Patterns (Ed. Valentine, J.W.}, pp. 335-354. Princeton University Press, Princeton, NJ. JABLONSKI, D. ( 1986) Causes and consequences of mass extinctions: a comparative approach. In: Dynamics of Extinction (Ed. Elliott, D.K.), pp. 183-229. John Wiley, New York. JACKSON, J.B.C. (1991) Adaptation and diversity of reef corals. BioScience, 41, 475-482. JOHNSON, K.G., BUDD, A.F. & STEMANN, T.A. (1995) Extinction selectivity and ecology of Neogene Carib bean reef corals. Paleobiology, 21(1), 52-73.
75
KAUFFMANN, E.G. & FAGERSTROM, J.A. (1993) The Phan
erozoic evolution of reef diversity. In: Biological Diver & Schluter, D.), pp. 315-329. University of Chicago Press, Chicago. KAUFMAN, L. (1993) Why the Ark is sinking. In: The Last Extinction (Eds Kaufman, L. & Malory, K.), pp. 1-46. MIT Press, Cambridge, MA. KINSEY, D.W. & HOPLEY, D. (1991) The significance of coral reefs as global carbon sinks. Response to Green house. Palaeogeogr. Palaeoclimatol. Palaeoecol., 89, 363-377. KLAUSEWITZ, W. (1983) Die Entwicklung des Roten Meeres und seiner Kustenfische. I. Evolutionszentrum. Natur. u. Museum, 113 (4), 103-111; II. Palaeogeogra phie, Palaeoekologie und Endemitenentwicklung. Natur u. Museum, 113(12}, 349-368. KNOWLTON, N. (1993) Sibling species in the sea? Ann. Rev. ecol. Syst., 24, 189-216. KNOWLTON, N. & JACKSON, J.B.C. (1994) New taxonomy and niche partitioning on coral reefs: jack of all trades or master of some? TREE, 9, 7-9. KOHN, A.J. (1980) Conus kahiko, a new Pleistocene gastropod from Oahu, Hawaii. J. Paleontol., 54, 534541. LAMBECK, K. (1996) Shoreline reconstructions for the Persian Gulf since the last glacial maximum. Earth planet. Sci. Lett., 142, 43-57. LOYA, Y. (1976) Recolonization of the Red Sea corals affected by natural catastrophes and man-made pertur bations. Ecology, 57, 278-289. LOYA, Y. (1990) Changes in a Red Sea coral community structure: a long-term case history study. In: The Earth in Transition (Ed. Woodwell, G.M.), pp. 369-384. Cam bridge University Press, Cambridge. LYNCH, J.D. (1988) Refugia. In: Analytical Biogeography (Eds Myers, A.A. & Giller, P.S.), pp. 311-342. McGLADE, J.M. (1990) The impact of global climate change on marine ecosystems. In: Global Climate and Ecosystem Change (Eds MacDonald, G.J. & Sertorio, L.}, pp. 165-184. Plenum, New York. McLAREN, D.J. (1986) Abrupt extinctions. In: Dynamics of Extinction (Ed. Elliott, O.K.}, pp. 37-46. John Wiley, New York. MAYR, E. (1982) Processes of speciation in animals. In: Mechanisms of Speciation (Ed. Barigozzi, C.), pp. 1-19. A.R. Liss, New York. MILLIMAN, J.D. (1993). Production and accumulation of calcium carbonate in the ocean: budget of a nonsteady state. Global biogeochem. Cycles, 7, 927-957. MILLIMAN, J.D. & DROXLER, A. (!996) Neritic and pelagic carbonate sedimentation in the marine environment: ignorance is not bliss. Geol. Rundsch., 85, 495-504. Mix, A.C. ( 1996) CLIMAP-style faunal transfer-function paleotemperatures re-assessed: cooler Ice-Age Tropics? Am. geophys. Union 1996 Spring Meeting, Program with abstract, 31A-3, S!57. NEWELL, N.D. (1971) An outline history of tropical or ganic reefs. Am. Mus. Novit., 2465, 1-37. OPDYKE, B.N. & WALKER, J.C.G. ( 1992) Return to the coral reef hypothesis: basin to shelf partitioning of CaC03 and its effects on atmospheric C02. Geology, 20, 733-736. PAULAY, G. (1990) Effects of late Cenozoic sea-level
sity (Eds Ricklefs, R.E.
76
M. Taviani
fluctuations on the bivalve faunas of tropical oceanic islands. Paleobiology, 16, 415-434. PAULAY, G. (1991) Late Cenozoic sea level fluctuations and the diversity and species composition of insular shallow water marine faunas. In: The Unity of Evolu tionary Biology (Ed. Dudley, E.C.), pp. 184-193. Dio scoroides Press, Portland. PAULAY, G. (1996a) Circulating theories of coral biogeog raphy. J. Biogeogr., 23, 279-282. PAULAY, G. (1996b) Dynamic clams: changes in the bivalve fauna of Pacific islands as a result of sea level fluctuations. Am. malacol. Bull., 12, 45-57. PoR, F.D. (1975) Pleistocene pulsation and preadaptation of biotas in mediterranean seas: consequences for Lessepsian migration. System. Zoo!., 24, 72-78. PoR, F.D. (1989) The Legacy of Tethys. Kluwer Academic, Dordrecht. RosEWATER, J. (1965) The family Tridacnidae in the Indo-Pacific. Indo-Pacific Mollusca, 1, pp. 347-396. SEWELL, S.R.B. ( 1948) The free-swimming planktonic Copepoda: geographical distribution. John Murray Exp. sci. Reps., 8, 317-595. SMITH, S.V. & BUDDEMEIER, R.W. (1992) Global change and coral reef ecosystems. Ann. Rev. ecol. Syst., 23, 89-118. STANLEY, S.M. (1984) Temperature and biotic crises in the marine realm. Geology, 12, 205-208. STANLEY, S.M. (1986) Population size, extinction, and speciation: the fission effect in Neogene bivalves. Paleo biology, 12, 89-110. STANLEY, S.M. & RuDDIMAN, W.F. (1995) Neogene ice age in the North Atlantic region: climatic changes, biotic effects, and forcing factors. In: Effects of Past Global Change on Life. Stud. Geophys., National Academy Press, Washington, DC, 118-133. STODDART, D.R. (1973) Coral reefs: the last two million years. Geography, 58, 313-323. STODDART, D.R. (1976) Continuity and crisis in the reef
community. Micronesica, 12, 1-9.
STONE, L. (1995) Biodiversity and habitat destruction: a
comparative study of model forest and coral reef eco systems. Proc. R. Soc. London, Ser. B, 261, 38 1-388. TAVIANI, M. (1982) Paleontological markers in late Pleis tocene raised coral reefs in the Red Sea. XI INQUA Congress, Moscow, Abstracts, 1, 307. TAVIANI, M (1997a) Axial sedimentation of the Red Sea transitional region (22 -25 "N): pelagic, gravity flow and sapropel deposition during the late Quaternary. In: •
Sedimentation and Tectonics of Rift Basins: Red Sea Gulf of Aden (Eds Purser, B.H. & Bosence, D.W.J.), pp. 47 1-482. Chapman and Hall, London.
TAVIANI, M. ( 1997b) Post-Miocene coral reefs of the Red
Sea: Glacio-eustatic controls. In: Sedimentation and Tectonics of Rift Basins: Red Sea-Gulf of Aden (Eds
Purser, B.H. & Bosence, D.W.J.), pp. 580-588. Chap man and Hall, London. TAYLOR, J.D. (1978) Faunal response to the instability of reef habitats: Pleistocene molluscan assemblages of Aldabra atoll. Palaeontology, 21, 1-30. VERMEIJ, G.J. ( 1986) Survival during biotic crises: the properties and evolutionary significance of refuges. In: Dynamics of Extinction (Ed. Elliott, D.K.), pp. 231246. John Wiley, New York. VERON, J.E.N. (1995) Corals in Space and Time. The Biogeography and Evolution of the Scleractinia. Univer sity of New South Wales Press, Sydney. WHITTAKER, R.J. ( 1995) Disturbed island ecology. TREE, 10, 42 1-425. WILKINSON, B.H. (1991) Coral reefs of the world are facing widespread devastation: can we prevent this through sustainable management practices? Proceedings of the 7th International Coral Reef Symposium 1, 11-21. WISE, K.P. & SCHOPF, T.J.M. ( 1981) Was marine faunal diversity in the Pleistocene affected by changes in sea level? Paleobiology, 7, 394-399.
Spec. Pubis int. Ass. Sediment. (1998) 25, 77-92
Sedimentary cycles in carbonate platform facies: Fourier analysis of geophysical logs from ODP Sites
865 and 866
P. COOPER
D epartment ofG eology andG eophysics, Univ ersity of Hawaii at Manoa, Bachman Hall lOS, 2444 Dol e Str e et, Honolulu, HI 96822, USA
ABSTRACT Shipboard analysis of recovered core materials at Ocean Drilling Program Sites 8 6 5 (Allison Guyot) and 866 (Resolution Guyot) suggested that rhythmic repetition of shallowing-upwards facies on several scales occurred within certain depth intervals. Poor recovery at these sites (averaging less than 16%) prevented visual confirmation of what were suspected to be Milankovitch cycles. In a previous study, Formation MicroScanner images from the sites were integrated with conventional log data and core descriptions to provide detailed stratigraphical columns. These detailed data confirmed shipboard conclusions regarding the presence of sedimentary cycles and provided details of facies changes that could be related to small-scale fluctuations in sea-level. Spectral analysis of the geophysical logs, which provide a continuous downhole record of lithological changes, was performed in an attempt to detect the presence of Milankovitch periodicities. For the initial analysis, a sample window was moved down the data at 3-m depth increments resulting in a spectrogram-like image. Imaging the entire length of the logged interval in this way revealed the presence of periodicities in the log data for the depth interval 250 mbsf (metres below sea-floor) to 490 mbsf in Hole 865 and for the intervals from 430 to 670 and from 9 3 5 to 1135 mbsf in Hole 866. Individual spectra within these selected intervals were then used to identify Milankovitch periodicities. At Site 8 6 5 , age controls were insufficient to provide time constraints for the logged interval ( 1 02 . 5-867.0 mbsf). However, the entire depth interval is probably of late Albian age. The strong cyclicity in the depth interval from 250 to 490 mbsf is mainly controlled by variations in porosity. Sedimentation rate and clay content increase slightly downhole. Fourier analysis of the gamma-ray and resistivity logs revealed high-amplitude spectral peaks corresponding to Milankovitch periodicities attributed to eccentricity and obliquity. The 4 1 3-kyr peak dominates all spectra; the spectral peak corresponding to the 123-95 kyr Milankovitch period has high amplitude for the interval from 330 to 490 mbsf. Vertical resolution was insufficient to resolve Milankovitch periodicities related to preces sion. At Site 866, the logged interval (78 .0-16 79.4 mbsf) corresponds to a time interval of about 3 3 Myr. Geophysical logs from two depth intervals showed strong cyclicity: 430-670 and 9 3 5-11 3 5 mbsf. Sedimentation rate is highly variable and increases downhole from about 30 m Myr-1 in the middle-to-late Albian section to about 80 m Myc1 in the Barremian section. Vertical resolution was insufficient to resolve frequencies corresponding to precession; the 41-kyr obliquity peaks are resolvable only in the lower portion of Hole 866A. The 413-kyr eccentricity cycle dominates all spectra. At both sites, the 123-9 5-kyr cycles could be related to alternations between dense wackestones and more porous packstones.
BACKGROUND
Early analyses of time series of climate-sensitive indicators in sediments and sedimentary rocks (e.g. Hays et a!., 1976) provided evidence that a major portion of climate variation is driven by changes in
insolation in response to perturbations in the Earth's orbital variables (Milankovitch, 1941) . Al ternative explanations for observed quasi-periodic climate changes have been presented in the litera-
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
77
78
P. Cooper
ture (e.g. Muller & MacDonald, 1995). Neverthe less, the quasi-periodicities of variations in orbital eccentricity (413 and 123-95 kyr), variations in obliquity or tilt ( 41 kyr), and variations in preces sion (23-19 kyr) are commonly observed in analy ses of cyclical sedimentation in a variety of sediment types and in many stratigraphical inter vals. Exactly how changes in the insolation patterns are translated into climatic changes remains the subject of extensive research. Because the magnitude of the change in insolation caused by the precession-obliquity-eccentricity cycles is so small (Berger, 1977), models that seek to explain climate forcing in terms of variations in in solation include atmospheric-oceanic feedback mechanisms that amplify the magnitude of the re sultant climate changes (Berger et a/., 1984; Barron et al., 1985; Sundquist & Broecker, 1985; Prell & Kutzbach, 1987; Sarnthein et al., 1988). For exam ple, Berger et al. ( 1984), Start & Prell ( 1984), & Ruddiman and Mcintyre ( 1984) have placed great emphasis on ice-sheet amplification of the relatively weak Milankovitch signals. Although the volume of polar ice caps during the Cretaceous-the time frame of this study-is not known accurately (Kemper, 1987), estimates sug gest that the growth patterns of ice caps with even greatly reduced volume would cause eustatic fluctu ations of the order of several metres, sufficient to cause emergence cycles at carbonate platforms (Hardie et al., 1986). The persistence of Milanko vitch periodicities through the Cretaceous, a time of warm seas and possibly minimal ice cover (Barron & Washington, 1985), implies that, whereas the presence of ice sheets may be an important factor in the modulation of global climate, their presence is not required for such modulation to occur (Fisher et al., 1985; Herbert et al., 1986). A variety of indicators is used to detect the presence of Milankovitch periodicities. Early stud ies of Quaternary pelagic sediments utilized varia tions in the percentage of calcium carbonate content and oxygen isotope ratios (e.g. Briskin & Harrell, 1980) for detecting Milankovitch periodic ities of the order of I 00 kyr. Oxygen isotope ratios for benthic marine foraminifers, as well, reflect global ice volume; numerous published studies have shown that Quaternary glacial regimes are related to orbital periodicities (Imbrie et al., 1984). Confirmation of the effects of climatic forcing in sediments of pre-Quaternary age extends over a range of sedimentary systems and depositional en-
vironments. Mean periodicities of 20, 40 and I 00 kyr have been detected for carbonate-marl deposits in the Cretaceous and lower Tertiary of the Apennines (e.g. Arthur & Fischer, 1977), in pelagic sediments drilled in the open ocean (e.g. Dean et a/. , 1977; Arthur, 1979; McCave, 1979), in Triassic shallow marine carbonates in the northern Alps (e.g. Fischer, 1964; Schwarzacher, 1964) and in Eocene (Bradley 1929), and Triassic lacustrine sediments (e.g. Van Houten, 1964; Olsen et al., 1978) in the USA. Climate indicators most useful for these sediment types include variations in porosity or grain size, the relative abundance and mineralogy of clays, and variations in one or more of the bioge nous components. Previous studies of geological time series from Cretaceous deep-ocean sediments (e.g. McCave, 1979; Arthur et al., 1984; Cotilion & Rio, 1984; de Graciansky & Gillot, 1985; Herbert & Fischer, 1986) revealed similar strong power in the eccentricity and precession peaks, regardless of the ocean of origin, indicating that this is truly a global signature that persists through geological time. Spectral analysis of logging data
Spectral analysis has proved to be a powerful method for extracting the dominant periodicities from climatic signals preserved in the sediments. The spectrum is calculated by taking the Fourier transform of a geological time series-some mea sure of the amplitude of climate change such as porosity plotted versus depth, which is used as a proxy for time. Such a spectral analysis can separate the dominant frequencies present in the data, giving them as cycles per unit depth. If sedimentation rates are well known, the frequencies can be converted to cycle periods. Spectral analysis works best in geo logical settings that contain a relatively continuous record of sedimentation. Jarrard & Arthur ( 1989) first explored the feasi bility of using spectral analysis of downhole geo physical logs to detect cyclic changes in mineralogy and porosity in Pleistocene sediments from Ocean Drilling Program (ODP) Sites 645 in the Labrador Sea and 646 in Baffin Bay. Their analysis revealed periodicities of roughly 20, 40 and 100 kyr in the sonic, resistivity and U/Th logs. They reasoned that fluctuations in bottom-water currents in response to Milankovitch forcing caused the observed varia tions in clay content and porosity. Amplitude spec tra of gamma-ray, sonic, and resistivity logs in upper Tertiary sediments at Site 693 on the Antarc-
S edimentary cycl es in carbonat e platformfaci es tic continental margin yielded obliquity and eccen tricity cycles (Golovchenko et a!., 1990). Obliquity and possibly eccentricity cycles also were observed in amplitude spectra of calcium and silica from Site 704 on Meteor Rise (Mwenifumbo & Blangy, 1991; Nobes et a!., 199 1). The results of an analysis of the natural gamma-ray log from Site 798 on the Oki Ridge suggested obliquity modulation of the Pliocene-Pleistocene aeolian dust influx to the Sea of Japan (DeMenocal et a!., 1992). Molinie & Ogg ( 1992) reported cycles of variable concentrations of radiolarians and clay, and a degree of silicification in upper Middle Jurassic to Lower Cretaceous radiolarites from spectral analysis of gamma-ray logs from Site 801. They used the wavelength of the eccentricity-modulated signals to determine sedi mentation rates. Glenn et a!. ( 1993) investigated the translation of climate modulation by Milankovitch like forcing into sedimentary cycles in mixed car bonate and siliciclastic sediments using sonic logs from Site 82 1 off the Great Barrier Reef, Austra lia. Climate modulation at that site incorporates such factors as lags in ocean-climate response to changes in insolation related to variations in orbital parameters, as well as independent changes in sediment supply (e.g. turbidites) and tectonic sub sidence. ODP site settings
This study uses two sets of geophysical logs obtained during Leg 143 at Allison Guyot (Site 865) in the central Mid-Pacific Mountains and at Reso lution Guyot (Site 866) in the western Mid-Pacific Mountains. Deep holes were drilled into the Creta ceous lagoonal facies of Allison and Resolution guyots to address fundamental problems concern ing guyot development (Sager et a!., 1993a). Hole 865A (Fig. 1), atop Allison Guyot, was drilled between the summit of the pelagic cap and the south rim of the guyot. The hole penetrated 140 m of pelagic cap and 698 m of late Albian shallow-water limestone, and bottomed at 870.9 mbsf (metres below sea-floor) in basaltic sills in truded into limestone. Clay, organic matter and the presence of pyrite in sediments from the lower part of the hole indicate a marsh-like setting with restricted-lagoonal environments in the upper part (Sager et a!., 1993b). Hole 866A (Fig. 1), atop Resolution Guyot, 7 16 km to the north-west of Site 865, was located about 1 km inboard, within a trough behind the perimeter mound. This deep hole
79
penetrated 1743.6 mbsf, through 25 m of pelagic sediments and 1620 m of Albian to Hauterivian shallow-water limestones overlying about 124 m of basalt. Dolomitized oolitic and oncolitic grain stones cover the volcanic basement and give way at 1400 mbsf to peritidal facies containing minor coral and rudist reef debris and beach sediments. Clusters of calcrete horizons are present in the lagoonal carbonates of the upper part of the platform (from about 680 mbsf ) (Sager et al., 1993c). Although average recovery was very low ( 15. 1o/o for Hole 865A, 15.4% for Hole 866A), cores and downhole logs suggested cyclic sedimentation on a metre-scale, particularly in the lagoonal facies of the upper part of Hole 866A (430-670 mbsf). Within this depth range, recurring metre-scale cycles typi cally began as laminated organic-rich mudstones and graded upward into bioturbated, less organic rich packstones and grainstones, and finally into wackestone with mouldic porosity. Facies from the lower part of Hole 866A (935-1 165 mbsf ) sug gested cyclical changes of the depositional environ ment from subtidal to intertidal-supratidal and returning to subtidal. Desiccation cracks and calci fied algal mat at the base of a typical sequence indicated emergence of the platform during low relative sea-level. A relative rise in sea-level resulted in reworking of the mudstone and algal mat as a flat-pebble conglomerate followed by deposition of a shallowing-upwards facies as the carbonate accu mulation rate outpaced the relative rise in sea-level. Many sequences are incomplete. In general, few peritidal sequences contain a complete record of climatic fluctuations because only high-amplitude sea-level changes will flood the peritidal platforms. Similar multimetre-scale peritidal sequences were identified by James ( 1977) and Shinn ( 1983) in both ancient and recent carbonate deposits. The analysis by Goldhammer et al. ( 1987) of peritidal facies from the Italian Triassic revealed a very complete record of the c. 20-kyr signal in the form of metre-scale peritidal sequences with vadose dia genetic, dolomitic caps. Many investigators (e.g. Gretzinger, 1986; Hardie et a!. 1986; Strasser, 199 1) have attributed such sequences to sea-level fluctuations related to climatic forcing as described by Milankovitch ( 1941). Much longer cycles (50-100 m) were revealed in laboratory analyses of MnO, Zn and Cu in core samples from Site 866 in the interval from 680 to 1400 mbsf (Rohl et al., 1995). Although these geochemical cycles are based on from four to six
80
P. Cooper
• .
.;
.&,
•
•• •
. ..
0
. . •.
".;jr , ,. ,;-
-
0
··- .
..
-
. .
.I� .... .
Sites 866, 867,and 868
..
...
• 0
..
•
Marshall Islands Magellan Rise
• .
••
•
fJ
' ' 0
.. .
·-
:Line Islands . ..• •
Fig. 1. The region visited by JOIDES Resolution during Leg 143. Sea-floor shallower than 4 km is stippled. The locations of Leg 143 drill sites and the track of the ship are shown.
samples per core, the trends are clear, and addi tional samples would not change the cycle lengths significantly. They are thought to correlate with packets of cycles seen in the lithology. A strong cyclicity was noted in the openhole log responses for Hole 865A from 250 to 490 mbsf (Fig. 2) and for Hole 866A from 430 to 664 and from 935 to 1 165 mbsf (Fig. 3). Further, the por tions of the geophysical logs displaying this cyclical character at both sites corresponded almost exactly to cored intervals displaying cyclic variations in lithologies. The high degree of correlation between the sonic and resistivity logs (Fig. 4) suggested that porosity variations were the dominant effect; how ever, it was difficult to confirm porosity variations and determine cycle length in the core material because of the poor recovery rate. A post-cruise study of the core, conventional log and Formation MicroScanner (FMS) log yielded a detailed litholog-
ical column that clearly revealed the nature of the cyclical sedimentation (Cooper et a!., 1995). Spectral analysis of the resistivity and gamma-ray logs from Sites 865 and 866 and the sonic log from Site 865 was undertaken (i) to investigate a possible relationship between the observed sedimentary rhythms and climatic forcing and (ii) to speculate on a possible cause.
DATA A standard suite of logging runs was made within the open borehole in Hole 865A between 100.5 and 867.0 mbsf and in Hole 866A between 74.5 andl 1679.4 mbsf or less (at Hole 866A the bottom of the: hole was not logged because of cave-ins at or above: 1680 mbsf; Sager et a!., 1993a). The sonic, medium-induction resistivity and natural gamma-
Sedimentary cycles in carbonate platform facies Hole 865A Gamma (API units) 100
0
Log resis.
Sonic (km/s)
% Recov.
70
200
81
the natural radioactivity of the formation and pro vides a qualitative evaluation of the clay or shale content, as radioactive elements tend to be concen trated in the crystal lattice of clay minerals. Cycle skipping and other noise caused by lack of centring of the tool compromised the quality of the sonic logs at both holes. The sonic log for Hole 866A was not used at all; the reprocessed sonic log for Hole 865A was used for the 4 10-490-m interval because no resistivity data were available for that interval. Resolution
300
400 :;:"' .0
E
.r:.
500
0. Q)
0
600
700
800
900L------L--�--�� Fig. 2. Gamma-ray, resistivity and sonic logs for Hole 865A. Fourier analysis was performed on the stippled portions of the data. The resistivity and sonic logs suffered data losses from mechanical failure and noise contamination, respectively. (Note resistivity log scale is logarithmic.)
ray logs were used in this study. The sonic log responds to lithology and porosity; formation veloc ity is calculated by taking the inverse of the sonic log. The resistivity log responds primarily to poros ity variations. The natural gamma-ray log measures
Vertical resolution is the major limitation to using the downhole logging data for this purpose (Jarrard & Arthur, 1989; deMenocal et al., 1992); that is, the resolution and sampling interval of the tool, com bined with a particular sedimentation rate, deter mines the minimum detectable cycle. For example, the resistivity log has a 1-m vertical resolution (Schlumberger, 1987) and requires a minimum sedimentation rate of 20 m Myc 1 to detect the 95-, 123-, and 413-kyr eccentricity signals (because Fou rier analysis requires at least two points per cycle). Similarly, for 1-m resolution a sedimentation rate of 100 m Myc1 is required to detect precessional signals (23 and 19 kyr). Thus, the sedimentation rate and vertical tool resolution combine to act as a high-cut filter. The natural gamma-ray intensity logging tool has a resolution of 0.3 m and the sonic log has a resolution of 0.6 m (Schlumberger, 1987). Slight spectral contamination could exist near the Nyquist frequency because the sampling interval (0. 15 m) is smaller than the vertical resolution of the logs; however, that is outside the frequency range of interest here. Although the gamma-ray spectra possess more power at higher frequencies, the gamma-ray tool also is affected by residual noise that degrades the low-frequency resolution (de Menocal et al., 1992). As mentioned above, the response of the deposi tional environment to astronomical climate forcing is quasi-periodic in time but the data recorded by the logging tool are periodic in depth. As long as the sediment accumulation rate is constant over long time periods, then the cyclic variation of physical properties with depth (depth series) will approxi mate a variation with time (time series). Spectral analysis of a depth series at constant accumulation rate yields temporal cycle frequencies measured in cycles per Fourier window length. If the sedimenta tion rate and, therefore, the spatial wavelength, of
·
82
P. Cooper (a)
(b)
Hole 866A
% Log Gamma (API units) resistivity Recov. 0 70 0 3.5 0 60 Orrrr��TT����
en .c
800
0
Log % resistivity Recov.
100 0
4
8 0 100
en .c
....
....
E
E
.<:
£
a. Q) Cl
c.
Q) Cl
Hole 866A Gamma (API units)
500
800
1 600
900L-----J-----��
1700 L-----'----...1...-.....1
the cycles is not constant over the entire depth section, the resolution may be degraded. The peaks seen in the amplitude spectra are increasingly smeared, or widened, at high frequencies by slight variations in sedimentation rate (e.g. deMenocalet a!., l992; Mayeretal., 1993).
METHOD
All data segments selected for Fourier analysis were detrended and standardized. Sequentially sampled
Fig. 3. Gamma-ray and resistivity logs for Hole 866A. Fourier analysis was performed on the stippled portions of the log. (Note resistivity log scale is logarithmic.)
discrete data, whether digital waveforms or down hole logs, possess three basic components: a trend and both periodic (signal) and random (noise) components. Trends are often seen in the logging data as linear increases in velocity or resistivity with depth, mainly as a result of compaction. A linear regression analysis was performed on all data seg ments to detect the presence of and to remove trends. That is, deviations of the dependent . variable-resistivity, velocity, or gamma-ray-from a straight line fitted to the data are minimized. A unique line thereby can be defined about which
Sedimentary cycles in carbonate platformfacies 4.0
�
3.0
E
-;,
3.2
� ·r::;
2.8
..Q
0 Qi >
Calculation of spectrogram
2.5
3.6
� ·;;
·� ·;:;;
� Cl 0
--'
2.4
2.0 275
0.5
279
283
287
291
83
0 295
Depth (mbsf)
Fig. 4. Expanded plot of the resistivity and sonic logs for a portion of Hole 865A, illustrating the close correspondence between the two porosity logs. (Note resistivity log scale is logarithmic.)
variance is a minimum. If the values along this line are subtracted from the corresponding data points, i.e. detrended, the resulting data set has a mean of zero; removing this trend removes the zero-order term that would otherwise dominate the spectrum (Schickendanz & Bowen, 1977). The data set then was standardized by dividing by the standard devia tion. Standardization ensures numerical stability in calculations and reduces the calculated amplitudes to percentage of total variance. High-frequency noise in the natural gamma-ray logs was not a serious problem, hence all spectra were calculated using unfiltered logs. Complex waveforms are constructed by the addi tion of successive harmonics, and the relative im portance of a particular harmonic is a function of its amplitude, which will be zero if the harmonic is not present. The purpose of Fourier analysis is to extract the dominant harmonic components. In a finite data sequence of N equally spaced points, N/2 harmonics can be computed. By choosing a data segment length (e.g. 525 samples or 80 m) the fundamental wavelength is arbitrarily set. Thus, the windowing process-the selection of a data seg ment-acts as a low-cut filter in that it removes very long-period cycles. Successive harmonics are com puted in terms of this length (e.g. cycles per 80 m for Hole 865A; Plate 1, facing p. 88). Because the windowing process can introduce spurious peaks ('sidelobes') into the spectrum, a Hanning filter (Jenkins & Watts, 1988) was applied to each spec trum. Smoothing in this manner tends to reduce amplitudes slightly and to merge nearby maxima.
A sliding 530-sample (80-m) window was moved down the detrended, standardized natural gamma ray log data at 3.0-m intervals. An interval of 80 m was chosen because the sedimentation rates derived from the various biostratigraphical (Sliter, 1995) and isotopic (Jenkyns, 1995) sources were approxi mate and not always in agreement. A rate of 80 m Myr-1 is equal to or greater than the maxi mum sedimentation rate reported for both holes. The depth at the middle of the data window was taken to be the approximate depth of the Fourier transform. The calculated amplitudes were con verted to a colour scale and plotted versus this approximate depth and frequency in Plates 1 & 2, facing p. 88. This three-dimensional presentation of the entire data set for each hole was compiled to select portions of the data for a more detailed analysis of the spectral content. Sections of the data with no apparent periodicities present, e.g. Hole 866, rows 25-75 (Plate 2), could be eliminated from further analysis. Some adjustments are necessary when making these selections or when comparing the spectrogram with the lithological units. A major discontinuity or change in sedimentation rate will be contained within several overlapping data win dows and will be smeared depthwise. Data segments were chosen for spectral analysis: (i) to coincide with the intervals of cyclic sedimen tation identified in the cores and on the spectro gram, (ii) to avoid sedimentary lacunae, e.g. at the Hole 866A Lithologic Unit V-VI boundary, 791.8 mbsf, and (iii) to avoid major diagenetic events such as the phosphatization in the upper sediments at Hole 865A and the dolomitization in the lower portion of Hole 866A. Identification of Milankovitch periodicities
Two methods are commonly used to demonstrate that spectral peaks are equivalent to Milankovitch orbital periodicities. Preferably, sedimentation rates that are obtained independently using bio stratigraphical data may be used to estimate the frequencies of the sediment cycles. In the absence of reliable sedimentation rates, one may compare the ratios of the frequencies of the observed spectral maxima with the ratios of Milankovitch periodici ties (e.g. Park & Herbert, 1987). No sedimentation rates have been calculated for the Albian section of
84
P. Coop er Gamma (API units)
Hole 865A: Allison Guyot
1 E.=,2 5 : -.,...--r-�------, E;:,__: .
Resistivity (ohm-m)
35 1
al1
Velocity (km/s)
5
E3 250-330 mbsf
330-410 mbsf
410-490 mbsf
0
10
20
30
40 50
60
70
80
Frequency (cycles per 80 m)
Hole 865A (Fig. 5) because no definitive biostrati graphical zonal boundaries could be defined. At Hole 866A, biostratigraphical (Sliter 1995) and isotope data (Jenkyns, 1995) combine to produce the following age controls that were used to estimate accumulation rates: the Albian-Aptian boundary is at 500 mbsf (112 Ma); the Aptian-Barremian boundary is at 900 mbsf ( 125 Ma); and the Barremian-Hauterivian boundary is at 1500 m ( 132 Ma); the geological time-scale used is that of Harland et a!. ( 1990). For Figs 6 & 7, depths were converted to ages using a linear interpolation be tween the ages at these boundaries. Because sedimentation rates were unknown for Site 865 and poorly known for Site 866, amplitude spectra for the resistivity and gamma-ray logs were computed first for 200-m data windows and then the window size was gradually reduced. It was assumed that the sedimentation rate was constant throughout each window. The ratios of the domi nant frequencies contained within these long data windows were consistent with those of the Milank ovitch eccentricity (3.3 and 4.3) and obliquity periods ( 10. 1). Assuming correct identification of the Milankovitch periods present in the data, a sedimentation rate then was estimated from their spatial wavelengths. Figure 8 shows a portion of just such an exercise for Hole 865A: the ratios of the dominant eccentricity peaks are similar for window lengths that ranged in length from 78 to 122 m. Based on the observed spatial wavelengths, the average sedimentation rate at Hole 865A was about
Fig. 5. Gamma-ray, resistivity and sonic log amplitude spectra for the interval from 250 to 490 mbsf are shown in three 80-m intervals. The vertical scale of spectra is percentage of total variance. Continuous curve-gamma-ray log; dotted line-resistivity log in upper two panels, sonic log in bottom panel. Assuming an average sedimentation rate of 65 m Myr-1, predicted locations of orbital periods El (413 kyr), E2 (123 kyr), E3 (95 kyr) and 0 ( 4 1 kyr) are at 3.0, I 0.1, 13.0 and 30.0 cycles per 80 m, respectively. The logs have been plotted to the right of the spectra at a vertical depth scale that corresponds to the summed lengths of the windows.
65 m Mye1 (250-490 mbsf). To allow for varia tions in sedimentation rate, a window length of 80 m was used for all further calculations for Hole 865A. Average sedimentation rates for Hole 866A estimated from observed spatial wavelengths were 37 m Mye1 (431-664 mbsf) and 75 m Myr-1 (935-1 165 mbsf), which compared favourably with sedimentation rates calculated from a linear inter polation between age horizons: 33 m Myr-1 and 77 m Myr-1• Amplitude spectra of the natural gamma-ray, sonic and resistivity logs were calcu lated for Hole 866A using depth windows of 33 m (Fig. 8) and 77 m (Fig. 9), each equivalent to a period of about 1 Myr.
RESULTS
Hole 86SA
Amplitude spectra for three depth windows of the gamma-ray and resistivity logs (top two panels) and gamma-ray and velocity logs (bottom panel) from Hole 865A are shown in Fig. 5. These logs show similar frequency contents in all windows, although the relative positions and magnitudes of maxima change with depth. In all of the spectra, most of the variance is in frequencies of less than 30 cycles per 80 m; therefore, the amplitude spectrum is trun cated at 80 cycles per 80 m without any significant loss of information. To compare the observed frequencies with the
85
Sedimentary cycles in carbonate platform facies Site 866: Resolution Guyot E1 E2
Log of resistivity 0.5
0
Gamma-ray (API units)
2.5 0
80
431-464 mbsf (109.9-110.9 m.y.)
464-497 mbsf (110.9-111.9 m.y.)
497-531 mbsf (111.9-112.9 m.y.)
531-564 mbsf
Fig. 6. Gamma-ray and resistivity spectra for the interval from 431 to 664 mbsf (109.9-116.9 Ma), shown in 3 3-m (1-Myr) intervals. Vertical scale of spectra is percentage of total variance. Continuous line-resistivity log; dotted line-gamma-ray log. Assuming an average sedimentation rate of 3 3 m Myc1, predicted locations of orbital periods El (413 kyr), E2 (123 kyr), E3 ( 9 5 kyr), and 0 (41 kyr) are at 2.4, 8.0, I 0.5 and 24.4 cycles per 3 3 m, respectively. The corresponding segments of the resistivity and gamma-ray logs have been plotted to the right of the spectra at a vertical depth scale that corresponds to the summed lengths of the windows.
·
(112.9-113.9 m.y.)
564-597 mbsf (113.9-114.9 m.y.)
597-631 mbsf (114.9-115.9 m.y.)
631-664 mbsf (115.9-116.9 m.y.)
0
10
20
40
50
60
70
80
Frequency (cycles per 33 m)
predicted frequencies of Milankovitch periodici ties, the following relation was used (Golovchenko et a!., 1990): predicted frequency= [window length (m)/ sedimentation rate (m Myr-1 )]/orbital periodicity (Myr per cycle). The continuous vertical lines in Fig. 5 indicate the predicted positions of the 4 13-kyr (E1), 123-kyr (E2), 95-kyr (E3) and 4 1-kyr (0) Milankovitch periodicities for a sedimentation rate of 65 m Myr-1• In most intervals, the highest amplitude peaks correspond to the predicted peak locations
for variations related to eccentricity-413, 123 and 95 kyr. This sedimentation rate, derived from a reconnaissance spectral analysis of a 200-m data window, is approximately correct only for this top interval (250-330 mbsf). The E2 and E3 peaks are merged. Gamma-ray log spectra show a trend to ward increasing energy with depth in the E2 and E3 peaks. In the interval from 330 to 4 10 mbsf, a spectral peak lies near the predicted location of the 0 spectral peak, about 30 cycles per 80 m. This frequency is well within the resolution of both gamma-ray and resistivity logs at this sedimenta tion rate. Its very low amplitude in the other two
86
P. Cooper Log of Gamma-ray resistivity (API units) 2.5 0 80 0.5
Site 866: Resolution Guyot E1 E2
0 E3
950
935-1012 mbsf (124.5-125.5 m.y.)
1000
.....
VI .0
1012-1088 mbsf (125.5-126.5 m.y.)
1050
E
.r:.
a.
1100
1088-1165 mbsf
Ill Cl
(126.5-127.5 m.y.) 1150 0
10
20
30
40
50
60
70
80
Frequency (cycles per 77 m)
8
35
Hole 865A Gamma-ray Log
7
30
78-m window E2/E1 3.52
E1
�c
=
::J
1l:
� "' E E "' (.9
91-m window E2/E1 3.44 =
6
25
5 20 . 15
\- / 'I
330
=
0
20
40
336
a:
'
2 338
1 340
Fig. 9. Expanded plot of the resistivity log and the gamma-ray log for a portion of Hole 866A, illustrating the close correlation between the two logs. (Note resistivity log scale is logarithmic.)
122- m window E2/E1 3.35 =
60
334
�
· ;;: ·�
Dept h (mbsf)
107-m window E2/E1 3.46
\
332
.£
'3 'iii Ill
,..., _,
10
E E .<:
4
,-, '
5
\
Fig. 7. Gamma-ray and resistivity spectra for the interval from 9 3 5 t o 1 1 65 mbsf ( 1 24.5- 1 27 . 5 Ma), shown in 77-m ( 1 -Myr) intervals. Vertical scale of spectra is percentage of total variance. Continuous line-resistivity log; dotted line-gamma-ray log. Assuming an average sedimentation rate of 77 m Myc1, predicted locations of orbital periods E l (4 1 3 kyr), E2 ( 1 2 3 kyr), E3 (95 kyr), and 0 (4 1 kyr) are at 4 1 kyr at 24.4; 95 kyr at 2.4, 8.0, 1 0 . 5 and 24.4 cycles per 7 7 m, respectively. The corresponding segments of the resistivity and gamma-ray logs have been plotted to the right of the spectra at a vertical depth scale that corresponds to the summed lengths of the windows.
80
Harmonic (cycles per window length) Fig. 8. Progressive spectral analysis of gamma-ray logs from Hole 865A. The size of the window for the Fourier transform was incremented to allow for confirmation of the locations of spectral peaks corresponding to Milankovitch periodicities. Ratios of E I to E2 are shown.
intervals may result from sedimentary processes, such as erosion and redeposition, or diagenesis acting to obliterate the signal by causing the appar ent sedimentation rate to vary significantly within the depth interval and, thus, weaken the high frequency signal. Assuming that the observed spectral peaks repre sent the true periodicities of the orbital parameters, then slight errors in the assumption of constant sedimentation rate are evident as shifts in the positions of observed peaks with respect to pre dicted peaks within a spectrum. For the interval
Sedimentary cycles in carbonate platform facies from 2 SO to 330 mbsf, observed peaks are shifted slightly to the right of the predicted peaks, suggest ing that the sedimentation rate is less than 6 S m Mye1 • Similarly, peaks in the intervals from 330 to 410 and from 410 to 490 mbsf have shifted to the left, suggesting that the sedimentation rate is greater than 6S m Myr-1. Upper portion of Hole 866A
Resistivity and gamma-ray logs for Hole 866A are displayed in Figs 6 & 7 at a vertical depth scale that corresponds to the summed lengths of the windows. The sonic log at this site is of poor quality and was not used. However, the sonic and resistivity logs typically show similar character because both rock properties are controlled by porosity. Vertical lines in the figures indicate the predicted frequencies of Milankovitch periodicities for an average sedimen tation rate of 33 m Myr-1 . The El peak dominates the spectra of all depth intervals. A strong cyclicity is evident in the logs, although the frequency content varies with depth, the upper 33 m (431-464 mbsf ) having longer-period varia tions than the sections below that level. Very little energy is present in that spectrum at frequencies greater than about l S cycles Myr-1 ; further, the gamma-ray (dotted-line spectrum) and resistivity (continuous-line spectrum) spectra and logs show little correlation in this depth interval. The E2 and E3 signals are not well resolved in this interval. The spectrum for the depth interval from 464 to 497 mbsf shows good correlation between the gamma-ray and resistivity logs, but the sedimenta tion rate clearly is much less than the 33 m Mye1• indicated by the continuous vertical lines. A drastic change in sedimentation rate or period of non deposition is probably the cause of the change in character of the spectra from depths greater than S3 l mbsf, perhaps related to the Albian-Aptian boundary at SOO mbsf ( 112 Ma; Sager et a!., 1993c). Below this level, the gamma-ray and resistivity spectra show similar gross frequency contents. The peaks are located near the predicted Milankovitch frequencies and correlate reasonably well between depth intervals. The sedimentation rate is lowest in the interval from 497 to S3 l mbsf and increases downhole to 33 m Myr-1 through the interval 631 to 664 mbsf. Lower portion of Hole 866A
Sedimentation rates continue to increase downhole
87
to about 80 m Mye1 in the lower section of Hole 866A. Gamma-ray and resistivity spectra for the lower portion of Hole 866A, 93S- l l 6 S mbsf ( l 2 4. S- l 27. S Ma) correlate reasonably well (Fig. 7). Usually the gamma-ray log anticorrelates with the resistivity log. In this instance, however, the gamma-ray log closely mimics the resistivity log (Fig. 9), reflecting the association of the biogenic carbon content with high porosity. The El peak either is absent or has very low power in the resistivity spectra, but is present in the gamma-ray spectra; its amplitude increases downhole. The dominant E2 peak reaches maximum amplitude in the interval from l 0 12 to 116 S mbsf. The 0 peak is broad, but clearly present in all three intervals. 'Tuning' the sedimentation rate at Hole 866A
The apparent sedimentation rate estimated from the observed spatial wavelength of the spectral peaks depends on the amplitude variance of the log spectra. Assuming correct identification of the Mi lankovitch periodicities, it should be possible to refine the sedimentation rates using the spatial wavelengths of the spectral peaks. An 80-m sliding window was passed down the resistivity log from 9 SO to 1200 mbsf; spectra were calculated at l 0-m intervals. The locations of the 9 S- l23-kyr spectral peaks with respect to upper and lower bounds on the sedimentation rate are shown in Fig. l 0. Each point is plotted at the window centre depth (i.e. the spectrum calculated at 9 SO mbsf contains data from 20 m above and below that depth). The sedimenta tion rate is seen to vary between 7 S and 83 m Myr-1 over the depth interval from 9 SO to 11 40 mbsf, with a major excursion to a much lower sedimentation rate at about l 060 mbsf. Minor grainstones recov ered over the depth range from l OSO to 1070 mbsf may represent a short-lived highstand stage. The E2- E3 spectral peak is difficult to identify at depths greater than 1130 mbsf, possibly a result of erosion because of numerous episodes of local emergence as evidenced by the presence of laminated limonite mudstone and desiccation cracks from 1 13 S to l l 6 S mbsf (Cores l 2 1R-124R).
D ISCUSSION
The chronology was not at all well constrained for Hole 86 SA; however, age control at Hole 866A was sufficient to demonstrate with a good degree of confidence that the time-scales of the observed
88
P. Cooper Hole 866A: 125 k.y. frequency 5.5
5
E 4.5
0 -.t 'C1) c. "' C1)
4
u > u 3.5
. �I\r-:/l'--r-1'-r-r---
/\
I
I
- X
"/
-�:
.. . '- - ·
\
�
75 m/m.y. 85 m/m.y.
3
2.5 950
1 000
1 050
1 100
1150
Dept h (mbsf)
rhythmic sedimentation have Milankovitch peri· ods. Application of alternative methods of identifi cation of Milankovitch periodicities in the absence of suitable biostratigraphical control provided con· vincing evidence that the peaks observed in spectra of log data from Hole 865 were Milankovitch periodicities. Several questions arise as to what is the mechanism or origin of the cycles, how are forced climate changes translated into lithologies, and what is the significance of the sedimentary cycles of various thicknesses, great and small. Evidence of subaerial exposure strongly indicates sea-level oscillations of either eustatic or tectonic origin or some combination of the two as the mechanism for these Cretaceous carbonate plat· form cycles (Fischer, 1964; Grotzinger, 1986; Har die et a!., 1986; Strasser, 1991). Deposition of carbonates in a platform environment is primarily a function of space. The sedimentation rate is highly dependent on eustatic sea-level and the rate of tectonic subsidence. Variation in productivity may be a small factor, but is probably not an important cause of the observed cyclicity. In the lower part (> 600 m) of Hole 866A weath ering of the volcanic edifice was the main source of clay for the surrounding lagoons (Sager et a!., 1993c). Variations in clay supply may have been climatically controlled and a quiet, shallow lagoonal environment would have encouraged the near-source deposition of the clays derived from weathering of the volcanic edifice. With increasing water depth, circulation within the lagoonal envi·
1200
Fig. 10. Plot of spatial frequency of 123-kyr spectral peaks versus depth for a 40-m sliding window. The data window for Fourier analysis was centred first on 9 5 0 mbsf and moved downwards in increments of I 0 m. The shaded portion indicates a range in sedimentation rates of 758 7 m Myr-1 that brackets most of the time-depth period covered by the calculations. Excursions to lower sedimentation rates are seen at I 060-1070 mbsf.
ronment is more energetic, removing the fine sedi ment fraction and concentrating the coarse-grained components. A probable cause of the cyclic fluctu· ations in resistivity (porosity) and gamma-ray val· ues observed at these sites is increased deposition of clays during lowstands, and winnowing or transport of clays during highstands. The porosity-sensitive resistivity logs usually an ticorrelate with the gamma-ray logs. Clay content increases downhole and, in general, increased clay content, indicated by elevated gamma-ray intensi· ties, is associated with lower velocity and resistivity. Because porosity appears to be the main effect controlling the resistivity and sonic log responses, a likely explanation of these cycles is that they reflect climatic controls on sea-level, which, in turn, con· trois clay supply and sorting. Deposition in shallow lagoons is affected by (i) storms, (ii) changes in current patterns and migra· tion of tidal channels and (iii) short-term eustatic and tectonic sea-level fluctuations. Thus, superim posed on the weak Milankovitch 'signals' is some 'noise' related to depositional conditions, and both signal and noise may be affected by differential compaction and diagenetic overprinting. Small, multimetre-scale shallowing-upwards cycles are es· pecially well developed in the protected platform interior, lagoonal-peritidal settings, where typical cycles consist of mudstone grading up to grainstone or packstone-wackestone. Boundaries of small cy· cles from Hole 866A are commonly marked by calcified algal mat, bird's eye vugs and desiccation
Sedimentary cycl es in carbonate pl atform facies features indicating repeated exposure to subaerial diagenesis. Integration of core, FMS and conven tional logging data (Cooper et a!., 1995) indicate thicknesses for these shallowing-upwards cycles that range from 0.8 to 28.4 m at Hole 865 and from 0.3 to 1 1.2 m at Hole 866A (upper portion). Not all cycles are complete; migrating tidal channels, cur rents and storms inevitably erode the previously deposited sediments. Core materials recovered from nearby Sites 867 and 868 show evidence of storm deposition. Mean thicknesses of 5.6 and 3.4 m were deter mined for the shallowing-upwards sequences of Hole 865A and the upper portion of Hole 866A, respectively, from an analysis of the FMS logs (Cooper et a!.; 1995). These thicknesses are compa rable with the wavelengths of the E2-E3 cycles for Hole 865A and upper portions of Hole 866A-7.9 and 6.2 m, and 4. 1 and 3. 1 m, respectively. The mean thickness of shallowing-upwards sequences, as determined from the FMS log in the lower portion of Hole 866A (Cooper et a!., 1995), is 3.4 m, comparable with the observed 3.2-m spatial wavelength of the 4 1-kyr obliquity cycle. This sug gests that the 123-95-kyr eccentricity cycle is rep resented lithologically by packets of rhythmic packstone-wackestone alternations, whereas the 4 1-kyr obliquity cycle consists of some smaller portion of the packet. Three possible reasons for the absence of the 4 1-kyr signal in data from Hole 865A and from the upper portion of Hole 866A are (i) lack of spatial resolution, (ii) highly variable sedimentation rate, or (iii) very low amplitude. Sedimentation rates at Hole 865A (c. 65 m Mye1) and in the lower portion of Hole 866A (c. 77 m Myr-1 ) are high, so spatial resolution should be comparable at both sites. Deviations from the mean apparent sedimentation rate are also comparable at both sites. The major difference between the lower portion of Hole 866A and the upper portion of Hole 866A, and Hole 865A, is the greater abundance of clay below 600 mbsf. It is speculative, but the additional clay and organic matter from the eroding Resolution volcanic edifice may have amplified the obliquity signal. The distribution of clay seams and packets of clay seams throughout the core in the bottom portion of Hole 866A provides evidence of higher frequency climatic fluctuations; however, the sedi mentation rate is not high enough at this site to allow for resolution of spectral peaks shorter than the 4 1-kyr cycle.
89
The small cycles are grouped into larger sequences that indicate a more long-term cyclic deepening and shallowing of the depositional environment (Arnaud et a!., 1995). The distribution of MnO, Zn and Cu (Rohl & Strasser, 1995) displays cycles of the order of 50-100 m in length in the lower portion of Hole 866A. Peaks in the MnO content appear to be linked with facies deposited under the most restricted conditions and probably correspond to periods of maximum flooding (e.g. sequence boundaries 5, 7, 10, and 15 of Arnaud et a!. ( 1995)). The spatial wave lengths of these cycles are longer than the predicted wavelength for the 4 13-kyr eccentricity cycles (about 3 1 m). Cycles having periods much longer than 413 kyr are not resolved in the spectra because of the dominance of the 4 13-kyr peak, the low-cut filtering effect of the windowing process, and standardiza tion of the signal. Very long, irregularly spaced cycles of lowstand to highstand with distinctive dia genetic horizons are obvious in the core and logging data (Arnaud et a!., 1995; Cooper et a!., 1995), and, although their lengths are highly variable, they represent a likely explanation of the geochemical cycles observed by Rohl & Strasser ( 1995). A good deal of variation in the thickness of cycles exists between Site 865 and the Albian-age portion of Site 866, so there remains the possibility that non-periodic processes influenced cycle develop ment at both the local and regional scales. Spectral maxima that are not near the predicted locations of peaks may have a non-orbital source for the cyclic ity or may reflect a peak shift owing to change in sedimentation rate. Most likely, occasional collapse of sections of beach mound or fringing reef that normally protect the lagoon from wave energy can expose the lagoon to open-ocean conditions. The cycling of such events is not entirely random, because they involve lateral transport and build-up of carbonate debris, slope destabilization and col lapse. Variations in tectonic subsidence rate probably represent a major non-periodic contribution to cycle development. Rapid sediment accumulation rates of about 80 m Myr-1 during the Barremian slowed to about 30 m Myr-1 during the Albian Aptian at Resolution Guyot (Site 866). Accumula tion rates at the much younger Allison Guyot (Site 865) were about 65 m Mye1 during the Albian. We may infer from this that regional differences in tectonic subsidence rates exist that depend on the age of the volcanic edifice, and that the local
90
P. Cooper
subsidence rate may not be uniform throughout the growth of an individual platform and contains pulses of uplift as well. The relatively high frequency Milankovitch climate signal is, therefore, superimposed on a non-periodic low-frequency tec tonic signal. Because the carbonate platforms of the Mid-Pacific guyots are far removed from terrige nous influences, the Milankovitch cycling repre sents as 'pure' a eustatic signal as can be obtained. Therefore, it may be possible to extract the tectonic signal given excellent biostratigraphical or radio metric age control.
sediments as simultaneous vanat10ns in porosity and clay content. Comparison of the resistivity and gamma-ray spectra shows that all the dominant frequencies are common to the two logs. Clay content increases downhole and, in general, in creased clay content, indicated by elevated gamma ray intensities, is associated with low resistivity. Because porosity appears to be the main effect controlling the resistivity response, an explanation of these cycles is that they reflect climatic controls on sea-level that, in turn, control clay content and sorting.
CONCLUSIONS
ACKNOWLEDGEMENTS
Geophysical logs in the Cretaceous lagoonal carbon ate facies of Holes 865A and 866A show pro nounced cyclic variations in porosity and clay content. Spectral analysis of the logs revealed dominant peaks having spatial wavelength ratios that matched the ratios of Milankovitch eccentri city cycles. Mean sedimentation rates calculated from recon naissance spectral analysis of long (200-m) data windows range from 65 m Myc1 for Hole 865A to 33 m Myc ' for the upper portion of Hole 866A, and to 77 m Myc1 for the lower portion of Hole 866A. No sedimentation rates were estimated from shipboard stratigraphy, but these values are consis tent with the notion that Allison (Site 865) was a younger volcanic edifice subsiding at roughly twice the rate of Resolution (Site 866) throughout the late Albian. A sliding window spectral analysis of the lower portion of Hole 866A reveals details of fluctuations in sedimentation rates from 75 to 83 m Myc ' over the depth interval from 950 to 1 150 mbsf. A short episode of very low sedimentation rates from 1050 to 1070 mbsf corresponds to a cored interval con taining minor grainstones and is interpreted as a short-lived highstand stage. For both Hole 865A and the upper portion of Hole 866A, the average thickness of the mudstone wackestone-packstone shallowing-upwards cycles as measured from the interpreted Formation Micro Scanner images is approximately equal to the pre dicted wavelength of the E2-E3 cycles. No similar match was found for the El cycle. The 0 signal has high amplitude only in the lower portion of Hole 866A, perhaps because of the higher clay content. The Milankovitch periodicities are evident in the
This work was funded by the US Science Advisory Committee and the Joint Oceanographic Institu tions. This paper benefited from discussions with Hubert Arnaud, Peter Flood, Ursula R6hl, Will Sager, Annie Vanneau, and Jerry Winterer. SOEST Contribution no. 4532.
REFERENCES ARNAUD, H.M., FLOOD, P.G. & STRASSER, A. ( 1 995) Reso lution Guyot (Hole 866A, Mid-Pacific Mountains): facies evolution and sequence stratigraphy. In: Proceed
ings of the Ocean Drilling Scientific Program, Results, 1 4 3 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 1 3 3- 1 60. Ocean Drilling Program, College Station, TX. ARTHUR, M.A. ( 1 979) North Atlantic Cretaceous black shales: the record at Site 3 9 8 and a brief comparison with some other occurrences. In: Initial Reports of the Deep Sea Drilling Project, 47-2 (Eds Sibuet, J.C. & Ryan, W.B. et a/.), pp. 7 1 9- 7 5 1 . US Goverment Print ing Office, Washington, DC. ARTHUR, M.A. & FISCHER, A. G. ( 1 977) Upper Cretaceous Paleocene magnetic stratigraphy at Gubbio, Italy: 1 . Lithostratigraphy and sedimentology. Geol. Soc. Am. Bull. , 88, 367-389. ARTHUR, M.A., DEAN, W.E., BOTTJER, D. & SCHOLLE, P.A. ( 1 984) Rhythmic bedding in Mesozoic-Cenozoic pe lagic carbonate sequences: the primary and diagenetic origin of Milankovitch-like cycles. In: Milankovitch and Climate (Eds Berger, A., Imbrie, J., Hays, J., Kukla, G. & Salzman, B.), pp. 1 9 1 -222. D. Reidel, Dordrecht. BARRON, E.J. & WASHINGTON, W.M. ( 1 98 5 ) Warm Creta ceous climates: high atmospheric C02 as a plausible mechanism. In: The Carbon Cycle and Atmospheric C02: Natural Variations Archean to Present (Eds Sundquist, E.T. & Broecker, W.S.), Geophys. Monogr. Am. geophys. Union, Washington, DC, 32, 546-5 5 3 . BARRON, E.J., ARTHUR, M.A. & KAUFFMAN, E.G. ( 1 98 5 ) Cretaceous rhythmic bedding sequences: a plausible
Sedimentary cycles in carbonate platform facies link between orbital variations and climate. Earth
planet. Sci. Lett., 72, 327-340. BERGER, A.L. ( 1 9 77) Power and limitations of energy balance climate model as applied to the astronomical theory of paleoclimates. Paleogeogr. Paleoclimatol. Pa leoecol. 2 1 , 227-2 3 5 . BERGER, A., IMBRIE, J., HAYS, J., KUKLA, G. & SALTZMAN, B. (Eds) ( 1 984) Milankovitch and Climate, Parts l and 2. D. Reidel, Dordrecht. BRADLEY, W.H. ( 1 929) The varves and climate of the Green River. US geol. Surv. Prof Pap. 158-E, 8 7- 1 1 0. BRISKIN, M. & HARRELL, J. ( 1 980) Time-series analysis of the Pleistocene deep-sea paleoclimate record. Mar. Geol. , 36, l -22. COOPER, P., ARNAUD, H.M. & FLOOD, P.C. ( 1 995) Forma tion MicroScanner log responses to lithology in guyot carbonate platforms and their implications: Sites 865 and 866. In: Proceedings of The Ocean Drilling Pro gram, Scientific Results, 1 43 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 329-372. Ocean Drilling Program, College Station, TX. COTILLON, P.H. & RIO, M. ( 1 984) Cyclic sedimentation in the Cretaceous at DSDP Sites 5 3 5 and 540 (Gulf of Mexico), 534 (central Atlantic), and in the Vicontian Basin (France). In: Initial Reports of the Deep Sea Drilling Project, 7 7 (Eds Buffer, R.T., Schlager, W. et al. ), pp. 3 3 9-376. US Government Printing Office, Washington, DC. DEAN, W.E., GARDNER, J.V., JANSA, L.F., CEPEK, P. & SEIBOLD, E. ( 1 9 77) Cyclic sedimentation along the con tinental margin of northwest Africa. In: Initial Reports of The Deep Sea Drilling Project, 4 1 (Eds Lancelot, Y., Seibold, E. et a/. ), pp. 965-989. US Government Print ing Office, Washington, DC. DE GRACIANSKY, P.C. & GILLOT, E. ( 1 9 85) Sedimentologic study of mid-Cretaceous carbonaceous limestones at Sites 549 and 5 50, northeast Atlantic. In: Initial Reports of The Deep Sea Drilling Project, 80 (Eds Graciansky, P.C. de & Poag, C.W. et al.), pp. 8 8 5-897. US Govern ment Printing Office, Washington, DC. DEMENOCAL, P., BRISTOW, J. & STEIN, R. ( 1 992) Paleocli mate applications of downhole logs: Pliocene Pleistocene results from Hole 798, Sea of Japan. In:
Proceedings of The Ocean Drilling Program, Scientific Results, 1 2 7/ 1 2 8 (Eds Pisciotta, K.A., Ingle, J.C., Jr, von Breyman, M.T., Barron, J. et al.), pp. 39 3-407. Ocean Drilling Program, College Station, TX. FISCHER, A.G. ( 1 964) The Lofer cyclothems of the Alpine Triassic. Kansas geol. Surv. Bull. , 169, 1 0 7- 1 49. FISCHER, A.G., H ERBERT, T.D. & PREMOU-SILVA, l. ( 1 9 85) Carbonate bedding cycles in Cretaceous pelagic and hemipelagic sediments. In: Fine-grained Deposits and
Biofacies of the Cretaceous Western Interior Seaway: Evidence of Cyclic Sedimentary Processes (Eds Pratt, L.M., Kauffman, E.G. & Zeit, F.B.), Soc. econ. Paleont. Miner., Tulsa, SEPM Guidebook, 9: 1 - 1 0 . GLENN, C.R., KROON, D. & WUCHANG, W . ( 1 993) Sedi mentary rhythms and climate forcing of Pleistocene Holocene mixed carbonate/siliciclastic sediments off the Great Barrier Reef. In: Proceedings of The Ocean Drilling Program, Scientific Results, 1 3 3 (Eds McKenzie, J.A., Davies, P.J. & Palmer-Julson, A., et al. ), pp. 1 89202. Ocean Drilling Program, College Station, TX.
91
GOLDHAMMER, R.K., DuNN, D.A. & HARDIE, L.A. ( 1 98 7 ) High-frequency glacio-eustatic sea-level oscillations with Milankovitch characteristics recorded in Middle Triassic platform carbonates in northern Italy. Am. J. Sci. , 287, 8 5 3-892. GOLOVCHENKO, X., O'CONNELL, S.B. & JARRARD, R. ( 1 990) Sedimentary response to paleoclimate from downhole logs at Site 6 9 3 , Antarctic continental margin. In: Pro
ceedings of The Ocean Drilling Program, Scientific Results, 1 1 3 (Eds Barker, P.F., Kennett, J.P. et al.), pp. 2 3 9-25 1 . Ocean Drilling Program, College Station, TX. GROTZINGER, J. P. ( 1 986) Upward shallowing platform cycles: a response to 2.2 Billion years of low-amplitude, high-frequency (Milankovitch band) sea level oscilla tions. Paleoceanography, 1, 403-4 1 6. HARDIE, L.A., BOSELUNI, A. & GOLDHAMMER, R.H. ( 1 986) Repeated subaerial exposure of subtidal carbonate plat forms, Triassic, northern Italy: evidence for high fre quency sea level oscillations on a 1 04 year scale. Paleoceanography, 1, 447-457. HARLAND, W.B., ARMSTRONG, R.L., Cox, A.V., CRAIG, L.E., SMITH, A. G. & SMITH, D.G. ( 1 990) A Geologic Time Scale 1 989. Cambridge University Press, Cambridge. HAYS, J.D., IMBRIE, J. & SHACKLETON, N.J. ( 1 976) Varia tions in the Earth's orbit: pacemaker of the Ice Ages. Science, 1 94, 1 1 2 1 - 1 1 32 . HERBERT, T.D. & FISCHER, A.G. ( 1 986) Milankovitch climatic origin of mid-Cretaceous black shale rhythms in central Italy. Nature, 321, 7 3 9-793. HERBERT, T.D., STALLARD, R.F. & FISCHER, A. C. ( 1 986) Anoxic events, productivity rhythms and the orbital signature in a mid-Cretaceous deep-sea sequence from central Italy. Paleoceanography, 1, 495-506. IMBRIE, J.M., HAYS, J., MARTINSON, D. G., et al. ( 1 984). The orbital theory of Pleistocene climate: support from a revised chronology of the marine 1 80 record. In: Mi lankovitch and Climate (Eds Berger, A., Imbrie, J., Hays, J., Kukla, G. & Salzman, B.), pp. 269-3 05. D. Reidel, Dordrecht. JAMES, P. N. ( 1 97 7 ) Shallowing upward sequences in carbonates. Geosci. Can. , 4, 1 26- 1 36. JARRARD, R. & ARTHUR, M.A. ( 1 989) Milankovitch paleo ceanographic cycles in geophysical logs from ODP Leg 1 05 , Labrador Sea and Baffin Bay. In: Proceedings of The Ocean Drilling Program, Scientific Results, I 05 (Eds Srivastava, S.P, Arthur, M., Clement, B. et al.), pp. 7 5 77 72. Ocean Drilling Program, College Station, TX. JENKINS, G.M. & WATTS, D.C. ( 1 9 88) Spectra/Analysis and Its Applications. Holden-Day, San Francisco, CA. JENKYNS, H.C. ( 1 995) Carbon-isotope stratigraphy and paleoceanographic significance of the Lower Cretaceous shallow-water carbonates of Resolution Guyot, Mid Pacific Mountains. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 4 3 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 77-88. Ocean Drilling Program, College Station, TX. JENKYNS, H.C., PAULL, C.K., CUMMINS, D.l. & FULLAGAR, P.D. ( 1 995) Strontium isotope stratigraphy of Lower Cretaceous atoll carbonates in the Mid-Pacific Moun tains. In: Proceedings of The Ocean Drilling Program, Scientific Results, 1 4 3 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 89-97 . Ocean Drilling Program, College Station, TX.
92
P. Cooper
KEMPER, E. ( 1 9 8 7) Das Klimat der Kreide-Zeit. Geol. Jb. Reihe A, 96, 402 pp. MAYER, L.A., JANSEN, E., BACKMAN, J. & TAKAGAMA, T. ( 1 993) Climate cyclicity at Site 806: the GRAPE record. In: Proceedings of The Ocean Drilling Program, Scien tific Results, 1 30 (Eds Berger, W.H., Kroenke, L.A., Mayer, L.A., et a!.), pp. 623-639. Ocean Drilling Pro gram, College Station, TX. McCAVE, I.N. ( 1 9 7 9) Depositional features of organic carbon-rich black and green mudstones at Sites 3 8 6 and 3 8 7 , western North Atlantic. In: Initial Reports of The Deep Sea Drilling Project, 43 (Eds Tucholke, B., Vogt, P. et a!. ), pp. 4 1 1 -4 1 6. US Government Printing Office, Washington, DC. MILANKOVITCH, M. ( 1 94 1 ) Konon der Erdbestrah!ung und seine A nwendung aufdas Eiszeitprob!em. Royal Serbian Academy, Spec. Pub!. 133, Section of Mathematical and Natural Sciences, 33 (published in English by the Israel Program for Scientific Translation, for the US Depart ment of Commerce and the National Science Founda tion, Washington, DC, 1 969). MOLINIE, A.J. & 0GG, J.G. ( 1 992) Milankovitch cycles in Upper Jurassic and Lower Cretaceous radiolarites of the equatorial Pacific: spectral analysis and sedimentation rate curves. In: Proceedings of The Deep Sea Drilling Program, Scientific Results, 1 29 (Eds Larson, R., Lance lot, Y. et a!. ), pp. 529-547. Ocean Drilling Program, College Station, TX. MULLER, R.A. & MACDONALD, G.J. ( 1 995) Glacial cycles and orbital inclination. Nature, 377, 1 07. MWENIFUMBO, C.J. & BLANGY, J.P. ( 1 99 1 ) Short-term spectral analysis of downhole logging measurements from Site 704. In: Proceedings of The Ocean Drilling Program, Scientific Results, 1 1 4 (Eds Ciesielski, P.F., Kristoffersen, Y. , et a!. ), pp. 5 7 7 - 5 8 5 . Ocean Drilling Program, College Station, TX. NOBES, D.C., BLOOMER, S.F., MIENERT, J. & WESTALL, F. ( 1 99 1 ) Milankovitch cycles and nonlinear response in the Quaternary record in the Atlantic sector of the Southern Oceans. In: Proceedings of The Ocean Drilling Program, Scientific Results, 1 1 4 (Eds Ciesielski, P.F., Kristoffersen, Y. et a!. ), pp. 5 5 1 -5 7 6 . Ocean Drilling Program, College Station, TX. OLSEN, P.E., REMINGTON, C.L., CORNET, B. & THOMPSON, K.S. ( 1 978) Cyclic change in Late Triassic lacustrine communities. Science, 201 , 729-7 3 3 . PARK, J. & HERBERT, T.D. ( 1 987) Hunting for paleoclimate periodicities in a geologic time series with an uncertain time scale. J. geophys. Res., 92, 1 402 7- 1 4040. PRELL, W.L. & KUTZBACH, J.E. ( 1 9 87) Monsoon variability over the past 1 50,000 years. J. geophys. Res. , 92, 84 1 1 -8425. RbHL, U. & STRASSER, A. ( 1 995) Diagenetic alterations and geochemical trends in Early Cretaceous shallow-water limestones of Allison and Resolution guyots (Sites 865 to 868). In: Proceedings of The Ocean Drilling Program, Scientific Results, 1 4 3 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 1 97-230. Ocean Drilling Program, College Station, TX.
RUDDIMAN, W.F. & MciNTYRE, A. ( 1 984) Ice-age thermal response and climatic role of the surface Atlantic Ocean, 46 ' N to 63 'N. Geo!. Soc. Am. Bull. , 95, 3 8 1 -396. SAGER, W.W., WINTERER, E.L., FIRTH, J.V. et a/. (Eds) ( 1 993a). Proceedings of The Ocean Drilling Program, Initial Reports, 1 43 . Ocean Drilling Program, College Station, TX. SAGER, W.W., WINTERER, E.L., FIRTH, J.V., et a/. (Eds) ( 1 993b). Site 8 6 5 . Proceedings of The Ocean Drilling Program Initial Reports, 1 4 3 . Ocean Drilling Program, College Station, TX. SAGER, W.W., WINTERER, E.L., FIRTH, J.V., et al. (Eds) ( 1 993c). Site 866. Proceedings of The Ocean Drilling Program, Reports, 1 43 . Ocean Drilling Program, Col lege Station, TX. SARNTHEIN, M., WINN, K, DUPLESSY, J.C. & FONTUGNE, M.R. ( 1 9 88) Global variations of surface ocean produc tivity in low and mid latitudes: influences on C02 reservoirs of the deep ocean and atmosphere during the last 2 1 ,000 years. Paleoceanography, 3, 3 6 1 -399. SCHICKENDANZ, P.T. & BOWEN, E.C. ( 1 977) The computa tion of climatological power spectra. J. appl. Meteorol., 16, 3 5 9-367. ScHLUMBERGER ( 1 987) Log Interpretation Principles! Applications. Schlumberger Education Services, Hous ton, TX. SCHWARZACHER, W. ( 1 964) An application of statistical time series analysis to a limestone-shale sequence. J. Geol. , 72 , 1 9 5-2 1 3 . SHINN, E.A. ( 1 9 83) Tidal flat environment. In: Carbonate Depositional Environments (Eds Scholle, P.A., Bebout, D.C. & Moore, C.H.), Mem. Am. Assoc. petrol. Geol., Tulsa, 33, 1 7 1 -2 1 0. SLITER, W.V. ( 1 995) Cretaceous planktonic foraminifers from Sites 8 6 5 , 866, and 869: a synthesis of Cretaceous pelagic sedimentation in the Central Pacific Ocean basin. In: Proceedings of The Ocean Drilling Program, Scientific Results, 1 43 (Eds Winterer, E.L, Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 1 5-30. Ocean Drilling Program, College Station, TX. START, G.G. & PRELL, W.L. ( 1 984) Evidence for two Pleistocene climatic modes: data from DSDP Site 502. In: New Perspectives in Climatic Modeling, Berger, A. & Nicolis, C.), Developments Atmosphere Sciences, 1 6, pp. 3-22. Elsevier, Amsterdam. STRASSER, A. ( 1 99 1 ) Lagoonal-peritidal sequences in car bonate environments: autocyclic and allocyclic pro cesses. In: Cycles and Events in Stratigraphy (Eds Einsele, G., Rieken, W. & Seilacher, A.) Springer-Verlag, Berlin. SUNDQUIST, E.J. & BROECKER, W.S. (Eds) ( 1 9 85). The
Carbon Cycle and Atmospheric C02: Natural Variations Archean to Present, Geophys. Monogr., Am. geophys. Union, Washington, DC, 32. VAN HouTEN, F. B. ( 1 964) Cyclic lacustrine sedimentation, Upper Triassic Lockatong formation, central New Jer sey and adjacent Pennsylvania. Kansas geol. Surv. Bull. , 169, 497-5 3 1 .
Platform Case Histories
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
Spec. Pubis int. Ass. Sediment. (1998) 25, 95-136
Aptian-Albian eustatic sea-levels
U. ROHL* and J . G. OGGt *Geosciences Department, Bremen University, PO Box 33 04 40, D-28334 Bremen, Germany; and tDepartment ofE arth and Atmospheric Sciences, Purdue University, West Lafayette, IN 4 7907, USA
ABSTRACT Carbonate banks record rapid falls of relative sea-level. as emergent surfaces and rapid rises as deepening or drowning events. During the combined Ocean Drilling Program Legs -143 and 144, four . carbonate banks of Aptian-Albian age were drilled on top, of seamounts ('guyots') that span a large region of the north-western Pacific. Simultaneous episodes of emergence ('sequence boundary') or deepening at these guyots must be the result of major eustatic sea-level events. From a combination of cored lithologies and downhole geophysical and geochemical logs we identified depositional sequences. A general geological age framework was assigned from biostratigraphical datum levels and chemo stratigraphical (carbon and strontium isotope) curves. Compensation for thermal subsidence rates allowed assignment of relative durations of the array of sequences within each stage. The number of upward-shallowing cycles or parasequences was also used 'to compare relative durations of sequences among sites. These Pacific carbonate banks record 12 Aptian and 1 2 Albian significant shallowing events, of which a third were associated with major episodes of emergence. The major events; on the guyots can be correlated easily with Aptian-Albian relative sea-level changes observed in European shelf successions, and both regions display the same number of minor events. Therefore, we can apply the relative timing of these events from the thermal subsidence compensation and parasequence counts within the Pacific banks to construct an improved scaling of the associated ammonite zones and biostratigraphical datum levels within the Aptian-Albian interval.
PACIFIC GUYOTS AS RECORDERS OF CRETACEOUS EUSTATIC SEA-LEVEL CHANGES
applicability was not documented. In addition, the derivation of a eustatic sea-level curve from the se quence stratigraphy of a continental margin is con troversial because shifts of deposition patterns can also be caused by irregular tectonic subsidence or uplift (e.g. Cloetingh, 1 9 88) or changes in sediment influx. Frequent high-amplitude swings of eustatic sea level during the late Cainozoic are caused by fluc tuations in continental ice sheets, but the proposed rapid major sequences in the Mesozoic (e.g. Haq et al., 1 98 7) create a perplexing dilemna-where to store massive amounts of water during a presumed globally warm climate. In particular, the mid Cretaceous is generally considered to be a period of elevated carbon dioxide levels and associated green-
Sequence stratigraphy has the basic premise that marginal-marine successions can be subdivided into distinct depositional packets bounded by unconfor mities. These sequence boundaries are generally in terpreted to be the product of rapid drops or regres sions of relative sea-level. If these sea-level falls are caused by eustatic falls of water levels in the world ocean, then the pattern of depositional sequences provides a means to achieve high-resolution corre lations among all continental margins. The assump tion of a master eustatic signal was implicit in the sequence stratigraphical charts and associated sea level curves published by the Exxon research group (e.g. Vail et a/., 1 977; Haq et a/., 1 987, 1 988; Vail, 1 987). These curves were mainly derived from out crops of European shelf sediments, and their global
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
95
96
U Rohl & J. G. Ogg
house warming caused by accelerated submarine volcanism (e.g. Larson, 1 99 1 ; Berner, 1 994 ); there fore, a demonstrated existence of major short-term eustatic fluctuations would present a challenge to our models of the hydrological and climate systems of a 'hothouse' world. A test of eustatic sea-level excursions during the mid-Cretaceous requires depositional settings that are sensitive recorders of sequence boundaries and are isolated from postulated flexural oscillations of continental margins. A suite of shallow-water car bonate platforms on mid-ocean seamounts distant from the tectonic complications of ridges or trenches provides the desired traits. During the Aptian and Albian stages of the mid-Cretaceous, the tropical Pacific contained a thriving population of carbonate banks and atolls. Over millions of years, these isolated carbonate platforms had maintained a depositional environ ment close to sea-level as the underlying volcanic edifices underwent progressive thermal subsidence at average rates of about 30-50 m Myc1• However, rapid rises or falls in eustatic sea-level produced changes in accommodation space and associated deepening of facies or created episodes of tempo rary emergence. At the end of the Albian, these long-lived carbon ate banks were mysteriously terminated simulta neously with the demise of carbonate platforms from Albania to Venezuela (Arnaud et a!., 1 995). The Aptian-Albian carbonate caps drowned and
became part of the population of enigmatic flat topped seamounts, 'guyots', that punctuate the submarine bathymetry of the western Pacific. (The submerged flat-topped seamounts were named 'guyots' after the nineteenth-century geographer Arnold Guyot by Hess ( 1 946).) Not until tens of millions of years later, during the Campanian and Maastrichtian stages of latest Cretaceous, were car bonate platforms again a component of the tropical Pacific (Winterer et a!., 1 993). The terminal drown ing of these guyots is examined in a later section. Four Aptian-Albian carbonate platforms, span ning a region 40' in longitude and 20' in latitude (Fig. I ), were a main focus of deep drilling during Ocean Drilling Program (ODP) Legs 1 43 and 1 44 (Premoli-Silva et a!., 1 993; Sager et a!., 1 993). Leg 1 43 drilled two guyots in the area of the Mid-Pacific Mountains (Sites 865-868, Allison and Resolution guyots). Of the five guyots drilled during Leg 1 44, MIT and Takuyo-Daisan (Seiko) guyots contain Aptian-Albian shallow-water carbonate caps. The successions of up to 1 000 m thickness also recorded short- and long-term oscillations of relative sea level. A prime objective of these ODP legs was to decipher the sea-level history recorded by each guyot, then extract the major events common to all guyots. Any major relative sea-level fluctuation recorded simultaneously by the four dispersed guyots would require a tectonic uplift over an area equivalent to the entire continental USA; in which case, the
180' 160'E 160'W 140'E 40'N n-......�======�........========........�======�......-n
)
. ·
�·
•879
"&
...
. . .
30'
Shatsky 'Rise
·.
878'..
20'
•
.;
.
.\(. '· : ··"
.
•
·..
.•
.
::.::
• .
,
•.
�"
•
& -
··· •
�'... 0. .
'%,. '···
� ·•··
., �� , 4o
· _•
l·' ?.c:.�· IQ:• : ) :; . J!} b"" Q • •
15'N
·
"•
Hessf"1'\ -':�:�Rise .
(}l
..
{
9.'
'·
•
• •
.,""
·•• . :
0 mbsl
••
..
'!, . �
.
•
·':
.
�··o. � -·
.: ·�-
.
�!�·.£t_.: \t_:y. s? ' ,_,.. •• , . -co • ,'1� "\
":'. ,
•
ta1n •
..
..
·���:�ai�an Ridge
.8 �� 6' M. :d-Pacitic Mou���A . / n
•.
...o.r�s:;.·'��
3000 mbsl
�.
tV
•\ '
H
.
�--
�
.-,· I' ..
Fig. I. Location of ODP Sites 865 and 866 (Leg 1 43), and 878 and 879 (Leg 1 44) in the central and north-west Pacific, and principal seamount and volcanic island chains; continuous line indicates shallower than 3000 m; dotted line indicates the coastline.
Aptian-Albian eustatic sea-levels
displaced sea water would cause a eustatic event to be recorded on all continental margins. Such a major event, a 'Darwin Rise' superswell of the central Pacific, has been theorized to occur once during the history of these guyots (see Schlanger et al., 198 1 ; McNutt et al., 1 990; Winterer et al., 1 993), but this was probably a single long-term anomaly. Variation in sediment influx is not a major factor. As a first approximation, the facies patterns of these carbonate banks remain fixed for millions of years because outward progradation was precluded by the steep outer slopes of the guyot. An added attraction in calibrating the eustatic curve using guyot carbonate caps is the ability to estimate relative durations oftime between sea-level excursions by using two independent techniques. First, the underlying thermal subsidence curve for these seamounts is consistent with a simple theoret ical model (e.g. Crough, 1 97 8, 1 984; Winterer et al., 1 993; Larson et al., 1 994). Therefore, an approxi mate time-scale can be applied for relative ages within the 500- 1 000-m thickness of the Aptian Albian succession. Second, these carbonate facies generally display a characteristic cyclicity of shoaling-upward trends (Cooper, 1 995; Ogg et al., 1 995b; Strasser et al., 1 995). The frequency and magnitude of these shoaling-upward 'parase quences' are consistent with Milankovitch climatic cycles at the 1 00 000- and 400 000-yr period of ec centricity modulation of precession (Cooper, 1 99 5). The magnitude of these parasequences may be par tially related to the thermal expansion and contrac tion of the ocean water column in response to the associated high-latitude warming and cooling in a non-glacial world (e.g. Revelle, 1 990; Ogg et al., 1 995b; E.L. Winterer, personal communication). A similar periodicity has been observed in the surface productivity of the tropical Pacific during the Creta ceous (Molinie & Ogg, 1 992). Therefore, the number of parasequences contained within a major deposi tional sequence may provide an estimate of its du ration, although some parasequences may not be re corded during emergent episodes associated with sequence boundaries. In this paper, we present the procedures and impli cit assumptions to interpret depositional sequences in these carbonate banks, then summarize the iden tification of each Aptian-Albian depositional se quence at the various guyots. The correlation and relative duration of these episodes at these guyots provide a scale that can be tentatively correlated with the Aptian-Albian stratigraphy of Europe.
97
INTERPRETATION OF SEA-LEVEL CHANGES IN GUYOT CARBONATE SYSTEMS-THEORY
Nature, coupled with the constraints of scientific drilling, has conspired to present fascinating chal lenges both in the interpretation of sea-level fluctu ations for each individual guyot cap and in the correlation of sea-level events among guyots. In the ideal world, a combination of extensive outcrop exposures with precision biostratigraphy and asso ciated seismic reflection profiles would provide the data base for interpretation of depositional se quences. Unfortunately, the array of guyot observa tions lack nearly all of these features. Seismic reflection profiles across the guyots are generally unable to resolve shifting of depositional facies or unambiguously distinguish surfaces of emergence. The central lagoonal deposits of other guyots display layered reflectors (Van Waasbergen & Winterer, 1 993), but these commonly do not display onlap or offiap relationships, nor would such features be expected after the central peak has drowned. Only on the Campanian-Maastrichtian carbonate bank surrounding the central volcanic peak of Wodejebato Guyot could seismic stratigra phy be applied to identify major sequences (Berg ersen, 1 99 5). Therefore, one must use the patterns of carbonate facies as a monitor of changes in relative sea-level. The Cretaceous guyot carbonate complexes gen erally evolved along a trend similar to that proposed by Darwin ( 1 842) for the formation of modern atolls, but without a reefal bulwark. Seismic reflec tion profiling across these guyot platforms generally indicated an outer rim surrounding a central de pression (e.g. Van Waasbergen & Winterer, 1 993). However, none of the Cretaceous guyots drilled by ODP Legs 143 and 1 44 displayed any indication of a significant biohermal construction similar to mod ern coral-algal reefal boundstones. Instead, the ODP corings of the facies of the outer rims and central depressions are consistent with a model of a lagoon that was partially protected by wave reworked shoals of bioclastic debris (e.g. Premoli Silva et al., 1 993; Sager et al., 1 993). Indeed, rimmed platforms with an internal lagoon domi nate the geological record in tropical environments, because carbonate sand shoals are also able to build stable barriers through the interplay of storm deposition and rapid lithification (Schlager, 1 992). The maximum rate of carbonate production by
·
98
U. Rohl & J. G. Ogg
organisms is between c. I 0 m depth and the surface. On these Cretaceous guyots, the active carbonate production is further concentrated in a band bounded on the seaward side by the plunging slope of the volcanic edifice and on the inner side by shoals of storm-deposited and wave-reworked accu mulations of bioclastic fragments (Fig. 2). Behind these bioclastic shoals was either a semi-protected lagoon or a shallow peritidal flat, depending upon the degree of lagoonal infilling and the relative importance of an internal volcanic island substrate. As the former volcanic edifice subsided, the carbon ate complex of barrier shoals continued to grow upwards, stabilized by the rapid partial lithification of the immense pile of bioclastic rubble. When the volcanic island finally drowned, the depth of the internal lagoon was governed by the balance be tween subsidence rate plus eustasy versus the rate of
sedimentation of bioclastic debris derived from the outer carbonate factory or formed in situ by the restricted assemblage of lagoonal organisms. Sequence stratigraphy models developed for car bonate ramps, which may exhibit responses to sea-level changes similar to siliciclastic ramps, gen erally do not apply to rimmed carbonate banks (Schlager, 1 992). For example, back-stepping onto higher ground during a relative rise in sea-level is possible only for banks in which the former lagoon has been infilled by elevated beach ridges and dunes. Similarly, seaward progradation during ex tended highstands will be severely limited by the steep angle of repose of the outer volcanic slopes. A minor amount of differential subsidence is recorded across some guyots, but such effects appear to be primarily post-depositional long-term adjustment to the loading of the carbonate cap (Winterer et al.,
Model of Typical Carbonate Platform of a Guyot Barrier Shoals
(bioclastic sand; storm rubble); Prograde into
lagoon during sea-level stillstand s Lagoon carbonates. Cemented horizons during exposures. Debris slope
i5ii��
Thermal subsidence during Guyot cap formation
Volcanic seamount
Fig. 2. Model of a typical carbonate platform of a guyot. As these seamounts underwent progressive thermal subsidence, their riding carbonate caps maintained a depositional environment close to sea-level. Rapid rises or falls in eustatic sea-level produced changes in accommodation space and associated deepening of facies or created episodes of temporary emergence. The active carbonate production was concentrated in a band bounded on the seaward side by the slope of the volcanic edifice and on the inner side by bioclastic shoals.
Aptian-Albian eustatic sea-levels
1993). Therefore, the ability to recognize major changes in relative sea-level on these isolated car bonate banks is mainly constrained to emergence episodes ('sequence boundaries') and surges in rates of accommodation space (early portion of 'trans gressive systems tracts'). The ODP Leg 1 43- 1 44 drill sites were preferentially located in the interior of the guyot platforms, therefore we will focus on these depositional settings. During eustatic sea-level falls, subsidence is par tially or completely offset. The internal lagoon may become partially infilled with carbonate mud and bioclastic debris, and may be converted into a complex of peritidal algal-rich flats, shoals of storm redeposited coarse debris, or supratidal areas. In all these situations, the calcareous sediments undergo rapid enhanced lithification and/or partial dissolu tion aided by the influence of enhanced precipita tion from warmed or evaporating sea water and meteoric water percolation. The rate and type of lithification depend on latitude, on exposure to meteoritic waters and on the proportion of metasta ble Mg-calcite and aragonite (Schlager, 1 992). Sur face denudation during emergence occurs mainly through chemical dissolution, and tends to be rela tively slow unless the emergence continues for an extended period under humid conditions (e.g. Sarg, 1 988). Therefore, relative to the underlying carbon ate facies, the sequence boundary interval in such carbonate systems is commonly associated with restricted facies and anomalous cementation below a drowning surface. During the episode of submergence following this stillstand to lowstand condensation and/or emer gence, the carbonate system may experience a tem porary lag in attaining maximum production to keep pace with the deepening. Holocene reefs ex hibited a lag period of only 3000-5000 yr before the reefs had again caught up with sea-level (Schlager, 1 989), but the non-reefal Cretaceous systems may have taken a considerably longer time. Indeed, the interior lagoon may never catch up, but rather be 'caught down' during the next episode of stillstand of relative sea-level (terminology of Soreghan & Dickinson, 1 994). The transgressive 'catch-up' phase is characterized by lime-mud-rich skeletal wackestones, whereas the highstand 'keep-up' phase (when carbonate sedimentation keeps pace with formation of new accommodation space) is charac terized by peloid-skeletal packstone to grainstone or oolitic grainstones (e.g. Kendall & Schlager, 1 9 8 1 ; Sarg, 1 988). However, if the magnitude of the transgressive pulse of sea-level rise produces drown-
99
ing below the depth of efficient carbonate produc tivity, which is c. 20-30 m, then the bank may never recover. However, the average rates of car bonate accumulation during the Cretaceous exceed most estimates of rates and magnitudes of major transgressive submergences, which suggests that drowning of platforms mainly occurs when other environmental conditions have stressed the carbon ate system (Schlager, 1 9 89, 1 99 1 ). The widespread termination of carbonate platforms during the latest Albian (including three of the studied Pacific guyots) was probably associated with the combined impact of a major transgression following a pro nounced sea-level lowstand (e.g. Gr6tsch et al., 1 993; R6hl & Ogg, 1 996) and a coincident, and perhaps associated, global environmental stress on shallow-water carbonate productivity (e.g. Johnson et al. , 1 996). This typical sequence of facies-sequence bound ary and lowstand associated with enhanced cemen tation and partial dissolution of shoaling to emergent deposits, transgressive systems tract of 'catch-up' lime-mud-rich facies, and highstand systems tract of 'keep-up' grain-rich to peritidal facies-applies to both the large-scale trends of third-order ( 1-2 Myr) sequences and to the smaller-scale 'shoaling-upward' fourth- and fifth-order ( 400 and 1 00 kyr) parase quence sets. As originally defined, a 'sequence' must contain conformable strata, therefore Van Wagoner et al. ( 1 990) proposed the term 'composite sequence' when the component parasequences exhibit small scale emergences. However, the guyot successions are observed to exhibit another intermediate hier archy, in which groups of three to four parase quences constitute sets of progressively more shallow-water ('keep-up' to peritidal) character, ter minated by a rapid deepening to another relatively 'deep-water' type of parasequence. In the strict terminology published by the Exxon group and others, these groups are a 'progradational parase quence set', which is generally considered to com prise the highstand systems tract of the depositional sequence. However, in many cases, there were two or three such progradational parasequence sets formed before a major horizon of emergence. In addition, parasequences superimposed on the equivalent of 'late highstand' within the third-order depositional sequence are commonly capped by minor surfaces of emergence. In the ideal case, it appeared that 1 00-kyr parase quences were grouped into '400-kyr' parasequence sets, and two to three parasequence sets were
100
U Rohl & J. G. Ogg
packaged into the main depositional sequence of about 1 -Myr duration. Milankovitch cycles re corded in deposits as varied as tropical lake to deep-sea environments may also exhibit the longer 413-kyr eccentricity period (e.g. Olsen, 1986; Mo linie & Ogg, 1992), so we suggest that some of these parasequence sets on guyots correspond to this type of climatic forcing. A longer 1.5-1.9-Myr cyclicity has also been observed in extended records of the lacustrine Newark Group of Late Triassic age and of Mediterranean marine sediments of Neogene age, and may be associated with long-period fluctu ations in the amplitude of the precession cycle (e.g. Olsen, 1986; Olsen et a!., 1996; F. Hilgen, personal communication), and many of the 'third-order' sequences observed on these Cretaceous guyots (this study) and in the Tertiary of the Gulf of Mexico (Van Wagoner et a!., 1990) have this ap proximate periodicity. To avoid excessive terminology, we will simply refer to all simple shallowing-upward successions as 'parasequences' and to apparent progradational groups of these as 'parasequence sets', regardless of whether such successions culminate in temporary emer,gent episodes, and we will reserve 'sequence' for the larger-scale (c. 1-Myr) features bounded by major surfaces of extended emergence or by con densed packets of peritidal parasequences. The onset of the next 'sequence' is generally assigned to the overlying base of the transgressive deepening above the surface of emergence. In some cases, the sequence boundary of rapid shoaling may actually correspond to a sudden initiation of peritidal depos its, rather than to the later transgressive deepening, but this type of sequence boundary is difficult to resolve from the 'late highstand' depositional sys tems tract within these successions. Genetic strati graphical sequences (sensu Galloway, 1 9 89) are difficult to apply to carbonate settings, because their bounding maximum flooding surfaces are difficult to recognize in the relatively homogeneous lime mud-rich 'catch-up' facies of the transgressive to early highstand systems tract. INTERPRETATION OF SEA-LEVEL CHANGES IN GUYOT CARBONATE SYSTEMS-PRACTICE
At all four guyots, only a single deep site penetrated the entire Aptian-Albian carbonate succession. These sites were generally slightly inboard of the
outer rim, and the recovered facies indicated that this position was dominated by lagoonal sedimen tation of gastropod-foraminifer-peloid wacke stones to packstones. Our sequence stratigraphy interpretations must be based on a single 'outcrop' of 1 0-cm diameter at each guyot. An implicit assumption is that the major surfaces of emergence and parasequences are indicative of widespread facies trends, rather than anomalies caused by local shifting of depositional facies. In the lagoonal set ting, such facies shifting was probably less signifi cant than major sea-level fluctuations. A drillsite into the active carbonate production belt or into the bioclastic barrier-shoal accumulation belt would probably have been more susceptible to local shift ing of features such as storm debris accumulations and tidal channels. Continuous coring was undertaken at all guyot drilling sites; however, the cores generally had less than I Oo/o recovery of the associated intervals. Comparison of the recovered lithologies with For mation MicroScanner (FMS) imagery of the corre sponding cored intervals suggests that core recovery was preferentially from relatively homogeneous, partially cemented wackestone-packstone, espe cially near the top of each 10-m cored interval (e.g. Cooper, 1995; Cooper et a!., 1995, Ogg, 1995). Intervals of very low-resistivity facies, such as poorly cemented to friable wackestone that charac terized transgressive 'deepening', tended to have negligible recovery; and very high-resistivity facies, such as well-cemented horizons that capped 'se quences', probably shattered into unrecoverable pieces during drilling. Drilling operations at sea did not allow recovery of cuttings. Downhole logging was essential to recognize and delimit depositional facies and surfaces of cementation. The cored facies provided an independent monitor of the interpreta tion of facies from the array of downhole logs. A large portion of the sedimentological and micro facies observations were previously made by the shipboard scientific parties (Premoli-Silva et a!., 1993; Winterer et a!., 1 993). The FMS imagery allowed high resolution of resistivity variations. Resistivity of these carbon ates is primarily a function of differential cementa tion, with grain size playing a subordinate role. The density log mirrored the resistivity log in these carbonates (Figs 3 & 4). Burial lithification was not significant, and isotope and microfacies observa tions indicate that most of the diagenesis was dominated by dissolution of aragonitic molluscs
Aptian-Albian eustatic sea-levels
101
Albian sequences 7/8: Carbonate bank
Fig. 3. Reconstruction of emersion horizons from resistivity data (Spherically Focused Resistivity Log (SFLU), mirrored by the density curve Neutron Porosity Log (NPHI)) and Formation MicroScanner images (FMS, light greys) and interpreted shallowing upward cycles (parasequences) that are grouped into sets of higher order (parasequence sets); pure carbonate environment of the Resolution Guyot carbonate bank (Hole 866A) in 200-250 mbsf. NGT, Natural gamma tool; THOR, thorium (modified from Rohl & Ogg, 1996).
and marine-phreatic cements (Rohl & Strasser, 1 995). Therefore, intervals of increased cementa tion and associated increased resistivity generally corresponded to relatively shallow-water to emer gent facies (Fig. 3). Transgressive deposits were char acterized by low resistivity (Fig. 3). The shallow focused resistivity log, coupled with the borehole imagery from the Formation MicroScanner, allowed distinction of the hierarchy of parasequences, parasequence sets and sequences. The natural gamma-ray log in the carbonates was primarily a monitor of uranium content, which seemed to correlate primarily with the presence of cyanobacteria-algal sedimentary features, such as rhodoliths, oncolites and peritidal laminations. Probably the main process acting to concentrate
...,..,.,., very strong emersion wvv strong emersion
uranium in such facies was the scavenging of ura nium from sea water by the micro-redox conditions within bacterial-algal films. The assembly of the Aptian-Albian sequence stratigraphy framework had two phases. The first phase was to integrate the various downhole logs, FMS features, cored sediment facies and biostrati graphy of each guyot. This first phase is similar to the procedure employed in other sequence stratig raphy or genetic stratigraphy studies (e.g. Home wood et al., 1 992; Robaszynski et al., 1 993). Our conceptual model of the carbonate system was applied to this data array to interpret depositional processes, facies associations, shallowing-upward successions or parasequences, sequence boundaries and rapid deepening episodes.
V.
102
Rohl & J. G. Ogg
Site 865 Albian Sequence 1 Volcanic Island Stage
Fig. 4. Reconstruction of emersion horizons from resistivity data (SFLU) and Formation MicroScanner images (FMS, light greys) and interpreted shallowing upward cycles (parasequences) that are grouped into sets of higher order (parasequence sets); Allison Guyot carbonate bank (Hole 866A) in 72 7-794 mbsf: the central volcanic island still exists, therefore calcareous, clayey and marly horizons (gamma-ray and thorium peaks) occur. NGT, Natural gamma tool; THOR, thorium.
Formation MucroScannerImages • mtXed carbonate/clay •
�
wvv
very strong emersion strong emersion
Aptian-Albian eustatic sea-levels
Sequence boundaries were recognized by a com bination of facies and logging signatures. An unusual lithological contrast of shallow to emergent carbonate facies sharply overlain by deeper facies is the key marker. The shallow deposits may include intertidal algal-bacterial mats, beach sands and 'bird's-eyes'. An emergence is indicated by down ward penetration of meteoric-phreatic cementation, solution cavities, pervasive red or yellow staining, blackened pebbles and other features (e.g. Strasser et al., 1 995). These sequence boundaries are recog nized in high-resolution FMS imagery as a sharp undulatory surface terminating an interval of high resistivity (implying well-cemented material) with possible enlarged voids (representing solution fea tures). The underlying sediment shows a progressive upward increase in resistivity, reflecting increasing cementation, which was also apparent in the lower resolution resistivity and porosity logs. Natural gamma-ray logs may display an elevated uranium background associated with algal-bacterial mats, with a rapid upward transition to low levels of ura nium. The degree of development and thickness of these anomalous cemented layers were used to assign 'major' and 'minor' sequence boundaries. In the lower portions of the carbonate succession, terrigenous influx from the emergent volcanic is land was concentrated in distinct layers. These episodes of clay-rich influx could be identified by their very low resistivity coupled with a thorium signature on the natural gamma-ray log, and gener ally were concentrated in the uppermost portion of a shallowing-upward parasequence trend of the carbonates. The clay layers are sharply terminated by a transgressive deepening, therefore these epi sodes are interpreted to occur in the 'late highstand' portion of a sequence when the lagoonal depression between the volcanic island and the drilled site has been finally infilled. Therefore, the tops of such clay layers were interpreted as sequence boundaries. The second phase was to apply the regional datum points and adjusted subsidence rates to determine which major and minor sequence bound aries are coeval among the various guyots. Nearly all sequences could be correlated within the array of guyots, therefore these are assumed to be eustatic in origin. The succession was compared with pub lished interpretations of European margin deposi tional sequences, and all major events could be correlated. The relative duration of each sequence from the subsidence rates of guyots and internal
10 3
number of parasequences was then used to deter mine how the ammonite-zone chronostratigraphy standard for the Aptian and Albian stages could be rescaled. SUMMARY OF GUYOT SEQUENCES
The identification of the sequences for each site is summarized in the Appendix. Detailed discussions of biostratigraphy, carbon and strontium isotopic stratigraphy of the guyot sites are given in a com panion paper (Rohl & Ogg, 1 996). Based on our proposed inter-guyot correlations and composite stratigraphy, we have introduced a common numbering for the sequence boundary interpretations of 'Apt l -Apt 12' for the Aptian and 'Albl-Alb l 2' for the Albian. The corresponding depositional sequences are numbered according to the underlying sequence boundary. Resolution Guyot (Site 866)
Resolution Guyot (formerly 'Huevo' Guyot) is lo cated in the western Mid-Pacific Mountains at 2J.3•N, 1 74.3·w (Fig. 1). Site 866 was drilled near the northern edge of the circular carbonate bank at a depth of 1 362 mbsl (metres below sea-level). Hole 866A recovered a 1 600-m succession of Hauteriv ian to upper Albian shallow-water carbonates over lain by 24 m of pelagic sediments (Shipboard Scientific Party, 1 993b; Arnaud-Vanneau & Sliter, 1 995; Arnaud et al., 1 995; Rohl & Strasser, 1 995; Strasser et al., 1 995). The lower portion of the carbonate platform contains intervals of dolomiti zation and oolite beds, whereas the upper portion has an abundance of peritidal to lagoonal facies that generally yielded less than 1 Oo/o core recovery. Re covered facies allowed subdivision of the carbonate series into 1 4 lithological subunits (Shipboard Sci entific Party, 1 993b,c). We analysed the Aptian-Albian carbonate suc cession. Our study overlaps the independent se quence stratigraphy analysis within the Hauterivian to lower Aptian succession by Arnaud et al. ( 1 99 5), which largely relied on facies patterns. Their inter pretations of sequences at Resolution Guyot were correlated with coeval episodes recorded by car bonate platforms in south-eastern France and in Venezuela. Our analyses of the combined downhole logs, especially the high-resolution FMS, and facies
U. Rohl & J. G. Ogg
1 04 (a)
I Hole 866A
-
Resolution-Guyot
Thorium Gamma ray (SGR, API units)
Shallowfocused resistivity (Ohm · m )
Neutron porosity (ppm)
'-!�t�'-o�==-=r�i! g==���gl LEGEND Rock and sediment types
... � 5:53 hlrJ E:j � .. CJ
Foraminiferal nannofossil ooze Limestone Clayey limestone
-
Clay
..
Conglomerate Basalt, volcaniclastic breccia Sandstone
Fossils
6
Gastropods
•
Corals
fi'
Caprinid rudists
� '\.
Calcisponges Sponge spicules
�
.I"G
.I" R
Orbitolinid foraminifers Green algae Red algae
Structures
®
Ooids Oncoids
�
11"'11'" • �
patterns allowed identification of additional minor sequences that subdivide their major Aptian se quences, and the revision of placement of some sequence boundaries (Fig. 5). For purposes of no-
Algal mat
-
e
Peloids Black pebbles
® Bird's eyes � Keystone
Calcrete vugs Bioturbation ,.-, Desiccation
menclature of sequences, we assume that the base of the Aptian is at about 9 1 0 mbsf (metres below sea-floor), and the base of the Albian is at about 480 mbsf. Detailed biostratigraphical control on
Aptian-Albian eustatic sea-levels
105
(b) 8
Neutron porosity (ppm) 12
16
ALBIAN APTIAN
Fig. 5. (Continued).
each sequence has been given by Rohl ( 1 996).
&
Ogg
Takuyo-Daisan ('Seiko') Guyot (Site 879)
Takuyo-Daisan Guyot (formerly 'Seiko' Guyot) is the easternmost guyot of the Seiko cluster in the Japanese Seamount province at 34.2°N, 1 44.3oW (Fig. 1). Site 879, at a depth of 1 50 1 mbsl, is on the southern perimeter rim of the guyot platform, which rises c. 70-80 m above the interior platform
surface. The Aptian and lower Aptian carbonate bank at Site 879 is only 1 64 m thick (Shipboard Scientific Party, 1 993e). Core recovery at Site 879 averaged only 4% within the platform carbonates. The identification of detailed facies trends and surfaces of exposure is based on the geophysical logs and FMS high resolution imagery (Ogg et a!., 1 995b). The carbon ate succession at Site 879 was initiated in the middle Aptian; the accumulation rate is approximately half of the Aptian accumulation rate on
·
1 06
U. Rohl & J. G. Ogg
680
APTIAN
BARREMfAH
*Sequence
boundary of
Amaudet al.
either Resolution or MIT Guyot. The slow subsid ence rate is caused by the greater age of the underlying oceanic crust (marine magnetic anomaly M 1 6 of Berriasian age; Nakanishi et a!., 1 992) compared with the construction age of Takuyo-
(19�5)
Fig. 5. (Continued).
Daisan seamount (e.g. thermal reset model of Det rick & Crough, 1 978). We have examined the details of the downhole logs of Hole 879A, ih addition to Ogg et a!. ( 1 995b), and we have sub divided the succession into nine sequences. The
Aptian-Albian eustatic sea-levels
nomenclature of these sequences reflects our postu lated correlations with the Aptian and early Albian sequences of Resolution Guyot (see Rohl & Ogg, 1 996). The major mid-Aptian sequence boundary and exposure episode in the Takuyo-Daisan succes sion is assumed to correlate with the major episode 'Apt?' at Resolution, and the major shoaling pre ceding the terminal drowning of the guyot in the early Albian is assumed to correlate with the major sequence boundary and following dramatic trans gression of 'Alb2' at Resolution. Correlation of these two most significant sequence boundaries in each guyot resulted in an almost perfect linear fit of all other minor sequence boundaries. MIT Guyot (Site 878)
Massachusetts Institute of Technology (MIT) Guyot is an isolated feature close to the Wake Seamount Group at 27TN, 1 5 1 .8·w (Fig. 1). The guyot has an irregular upper surface at about 1 300-m depth with dish-like pits up to 1 80 m deep and 500 m wide (Shipboard Scientific Party, 1 993d; Winterer eta!., 1 993, 1 995). This pockmarked surface may represent episodes of karst solution (e.g. Van Waas bergen & Winterer, 1 993) and/or long-term subma rine dissolution associated with the nearly complete removal of all pre-Pliocene pelagic sedimentation. Site 878 was drilled about 1 km inboard of the platform edge. Only about 2% of the carbonate platform series was recovered during coring, there fore downhole logs and FMS imagery play an important role in identifying the sequences (Jansa & Arnaud-Vanneau, 1 995; Ogg, 1 995). The Aptian-Albian carbonate platform of MIT Guyot was interrupted by a late Aptian episode of explosive volcanism that deposited a 200-m breccia of volcaniclastic material mixed with blocks of shattered Aptian shallow-water carbonates at Site 878 (Shipboard Scientific Party, 1 993d). The under lying 1 20-m-thick Lower Aptian platform has a succession from high-energy shoals of skeletal and oolitic grainstone followed by fluctuating lagoonal conditions with at least one episode of open-marine flooding. The upper carbonate platform consists of 400 m of peloid grainstone to gastropod-rich wacke stone of late Aptian and Albian age representing lagoonal to storm-reworked 'island' accumulations. A metre-by-metre summary and interpretation of the FMS imagery and other downhole logs and associated core recovery has been presented by Ogg ( 1 995), and a description of possible 'keep-up' and
1 07
'give-up' facies patterns has been compiled by Jansa Arnaud-Vanneau ( 1 995) from a combined exam ination of microfacies and downhole logs (other than FMS). The downhole logs and recovered lithologies from MIT Guyot were reanalysed by Rohl & Ogg ( 1 996) using the improved calibration of logging signatures of carbonate facies and epi sodes of emergence.
&
Allison Guyot (Site 865)
Allison Guyot is located in the western Mid-Pacific Mountains (Fig. 1) at 1 8SN, 1 79SW. Site 865 in the former interior 'lagoon' penetrated a 700-m thick shallow-water carbonate platform overlain by 1 40 m of pelagic carbonate ooze with an upper surface at a depth of 1 5 1 8 mbsl (Shipboard Scien tific Party, 1 99 3a). The shallow-water carbonate succession begins in the late Aptian to early Albian and was terminated in the late Albian (Arnaud Vanneau & Sliter, 1 995; Arnaud eta!., 1 995; Rohl & Strasser, 1 995; Strasser eta!., 1 995). The Albian lagoonal sediments of Allison Guyot consist in the lower part of the site of wackestones and packstones with varying amounts of benthic foraminifers, mol luscs, calcareous algae and minor clay, whereas the upper part of the site is characterized by wacke stones and packstones with gastropods, calcareous sponges and large siliceous sponge spicules (Sager et a!., 1 993; Arnaudetal., 1 995). The formation of Allison seamount in the latest Aptian implies that the young edifice had a rapid subsidence during the Albian. As a result of the rapid subsidence, the Albian succession is nearly twice as thick as on the older edifices of Resolution and MIT guyots, and some sequence boundaries which are marked by emergent surfaces on other guyots are only indicated by a slowing rate of accommodation space on Allison. CORRELATION OF SEQUENCES BETWEEN GUYOTS
Pacific guyots were generally isolated from tectonic disruptions of thermal subsidence. The rate of thermal subsidence can be modelled as a function of the relative age difference between the volcanic edifice and the underlying older oceanic crustal basement. After an initial thermal reset of the relative basement age, the seamount subsidence follows an exponential curve (e.g. Detrick &
1 08
U. Rohl & J. G. Ogg
Crough, 1 978; Crough; 1 984; Epp, 1 984). There fore, once a common age-calibration horizon is established among the guyots, the relative accumu lation rates can be adjusted for the computed differential subsidence rates, thereby providing a scale for comparing relative timing of sequences on each guyot. Unfortunately, the absolute magnitude of eustatic sea-level variations cannot be determined from the carbonate banks. We have assumed that a longer duration of exposure would result in a greater degree of diagenesis, thereby producing a relatively thicker and more cemented zone of emergence. Therefore, we have assigned 'major' sequence boundaries to the few exposure surfaces underlain by intervals exhibiting an anomalously pervasive ce mentation (e.g. high resistivity) extending up to 5 m in depth, and relatively 'minor' sequence boundaries to the more frequent emergences underlain by thin ner or less pronounced cementation zones. Long term major second-order trends in sea-level can be vaguely approximated from the relative average thicknesses of parasequences. Carbonate platforms traditionally have been a weak spot of classical stratigraphy because time resolution within these shallow-water facies is gen erally less precise than with open-ocean pelagic biota. Assignment of stage boundaries and associ ated subsidence history for each guyot changed dramatically between the tentative shipboard pale ontological compilations (e.g. Shipboard Scientific Party, 1 993a-e), the revised assignments in the Initial Reports volumes (Premoli-Silva et a/., 1 993; Sager et a/., 1 993), and the post-cruise studies reported in the Scientific Results volumes (Haggerty et a/., 1 995; Winterer et al., 1 995). In the process, a general consensus was reached for the Aptian Albian portions of the guyots. The initiation of the carbonate platform at each guyot is constrained by the radiometric age of the underlying volcanic peak and faunal assemblages within the initial transgressive deposits at each site. For additional inter-guyot correlations (Fig. 6), es pecially to the Aptian succession of Takuyo-Daisan Guyot, the benthic foraminiferal assemblages pro vide approximate age estimates, but it was not possible to assign zonal boundaries or utilize datum levels of index taxa (Arnaud-Vanneau & Sliter, 1 99 5). As another surface of inter-guyot correlation, we made an assumption that the episode terminat ing the carbonate banks at Allison, Resolution and MIT guyots in the upper Albian was a simultaneous
drowning within the Rotalipora appenninica plank tonic foraminifer zone. This assumption is consis tent with the available biostratigraphy, especially the co-occurrence of Rotalipora appenninica, Biti cinella breggiensis and Rotalipora ticinensis in the oldest pelagic sediments overlying the platform of Allison Guyot (Gr6tsch & Fliigel, 1 992), and the observed clustering of drowning of other Tethyan carbonate platforms in the same biozone (Gr6tsch et a/., 1 993). The stratigraphical array of biostratigraphical and isotopic features provides a general framework that constrains correlations among guyots and with Aptian-Albian succession in Europe. We assume that the major sequence boundaries marked by significant exposure episodes are responses to a major eustatic sea-level fall that affected all plat forms. The subsidence rates between major se quence boundaries are assumed to be quasi-linear, therefore minor sequence boundaries should be in approximately the same proportional position among the suite of guyots. There is an implicit assumption that the shallowing or emergent epi sodes at each site are a response to changes in relative sea-level, rather than an independent inter nal shifting of facies patterns. There are three major Aptian sequence bound aries, Apt2, Apt7 and Apt I 0, within the carbonate platforms, punctuated by approximately nine mi nor Aptian sequences (Fig. 6). The Albian contains two extraordinary sequence boundaries followed by major transgressive deepening, Alb2, and the up permost Alb 1 2 sequence boundary, which preceded the drowning of the guyot. Four other major Albian sequence boundaries, Alb5, Alb7, Alb9 and Alb 1 1 , occur between these two levels. The major lower Aptian sequence boundary, Apt2, occurs on all four guyots, but the major transgressive deepening that followed Apt2 caused the terminal drowning of the carbonate platform at Takuyo-Daisan Guyot (Rohl & Ogg, 1 996). The events are not developed to the same extent because of facies variability: in more or less unce mented carbonate sands a clear cyclicity is not always visible. However, as emergence and accom panying cementation happened relatively easily in this case, we found very prominent sequence boundaries. Bioturbated wackestones and mud stones evolving into algal mats that include 'clay', organic matter and slightly cemented cycle tops generally are constructed of pronounced cycles or
I
Allison
I
Hole 865A Ma1n (mbsf) deepemng episodes Depth
I
I
Hole 878A Ma1n (mbsf) deepemng episodes
Depth
Topof carb onate
r-----r---v180 f140
__..,-I- �� II
'Aib12'
240 294
70
I Takuyo-Daisan I
I
MIT
I
Hole 879A Ma1n (mbsf) deepemng episodes
Depth
Topof carbonate platform
�'th < Q _
r / I 264f-'Alb6?'-
492
/I
'Alb7'
392
'Alb4'
672
'Alb3'
·
-"?'
1:! . . ·c: 0 � • :0 · ·
I 624 I--"Aot5"---t'"
794 I--"Alb1"--l
I
I
"'
"'
c::,. 1;
? ?
;:3 � o ;:3 :;;! (;j
no· 6 �3314F'Al,b6' , 153 �--Apt5
/
1
-
579
\ '\.
�
iS'
T�
� ;::
I_;_::
1 - -
">
$::l
'Apt10'
I
641 662 \ -, 691 'Apt7' 714 --·'Apt6'··-·
........,_
�
ro t"'lanKromc Foraminifer
538 -·"Apt11'·-·
I I 641",/,/,/,/, "'""'"''. ... . ...
I .. . .... I
Albian
: � �:�� �
I �: t��;.
'1 759 -- ··"Apt5"··-· 799 837 853 90 1 925
Fig. 6.
�
87186
317 -··'Alb6' 'AibS' 341 363 393 -·'Alb3?'·-· 'Aib2' 414 · -
��� ��
728 I--"Aib2" --r
871 I
'Aib7'
--:�-::6:""o� o! .....
:
___
J 278
Topofcarbonate platform
'Aib2'
:: f·:� : jj �!�: 626
'Aib9'
--------------
'AibS'
�4�
'Alb8'
L I 226 186
'Aib8'
� =v 111 [
o1t :g_ {Calibrated to Aptian ,! foraminifer zones] ""' "' .
Planktonic Foraminifer zones
139
'Aib9'
'Alb4?'
448
Stage
'Aib11'
'Aib10'
'Aib7'_-=
365 r--'Aib9'---r
Hole 866A Matn deepemng episodes
I Composite Biostratigraphy and lsotoiiG-StratigraphyI
: I-
' Aib11' -
1 97
I
24
'Afb12'
�"lUI I � /ll�" [ 'Aib10'
Resolution
"Apt4" -·1 Lr--. . "Aptt"·· •tt. B"
Aptian
600
Ticenella bejaouaensis Hei:Jbergena trochidea -
I
� �
1::;-
;_ �
�e�n� GIJf'rrft!?le!!§iS Leupoldina cabri -
Barremian
1 900
Correlation of sequence boundaries of the guyots for an area of 40• longitude and 20" latitude in the central and north-west Pacific, composite bio and isotope stratigraphy (modified from Rohl & Ogg, 1 996).
0 'D
1 10
U. Rohl & J. G. Ogg
parasequences. They do not provide sharp sequence boundaries, however, as they are covered by the general 'background (parasequence) cyclicity' (com pare Rohl & Ogg, 1 996). AGE-DEPTH MODELS
One advantage of Pacific guyots is their general isolation from tectonic disruptions of thermal sub sidence. The rate of thermal subsidence can be modelled as a function of the relative age difference between the volcanic edifice and the underlying older oceanic crustal basement. After an initial thermal resetting of the relative basement age, the seamount subsidence follows an exponential curve (e.g. Detrick & Crough, 1 978; Crough, 1 984; Epp, 1 984): depth (t) = ('initial depth') + k ..J r.: [ '7 ....,.,_-a_ i-:a1-age (t-:-= e-:g_ )]. in....,.i-t:-
( 1)
Empirical estimates of the subsidence constant k are typically between 300 and 3 60 m Myc1• In this formulation, the 'initial depth' does not correspond to the depth at the time of eruption, but to a projected value corresponding to the older 'initial age'. Even though such models are simple (e.g. McNutt et al., 1 990; Winterer et al. , 1 993), the depth histories of nearly all these guyots are consistent with such a curve (Van Waasbergen & Winterer, 1 993; Winterer & Sager, 1 995; R. Larson, personal communication). Therefore, once a common age calibration horizon is established among the guyots, the relative accumulation rates can be adjusted for
the computed differential subsidence rates, thereby providing a scale for comparing relative timing of sequences on each guyot. Of course, it is conceivable that the subsidence of one or more of these guyots did not follow such a simplified age-depth curve. Superimposed long term uplifts and/or accelerations of subsidence could be caused by late-stage volcanism, passage over geoid anomalies, construction of new nearby seamounts, or other factor (Marty & Cazenave, 1 989). For example, ODP Site 878 at MIT Guyot records a major late"stage volcanic episode that disrupted the carbonate platform, and the accumu lation rate of Takuyo-Daisan Guyot's platform is anomalously slow relative to the other three guyots. Another important factor in subsidence calcula tions is loading by the accumulated carbonates. Such isostatic loadin& may not significantly influ ence the general age-oepf.h relationship, because if the shallow-water carbonate accumulation main tained a constant area and kept pace with thermal subsidence, then its age-loading curve will also have a .../(elapsed age) trend, thereby amplifying the main thermal subsidence curve. In any case, within the resolution of the array of biostratigraphical and isotopic stratigraphy constraints and relative thick nesses among major surfaces of emergence, we are unable to document any significant departures from such simple age-depth relationships. We computed thermal subsidence curves for each of the Aptian-Albian sites by an empirical fit to age-depth constraints of the carbonate platform stratigraphy and of the underlying volcanic edifice (Fig. 7; Table 1). The applicable equation for cur rent depths below sea-level of stratigraphical levels
500 1000
Thickness of
Drowning of carbonates
1- � Submergence, and Initiation of carbonates ' Formation age of seamount, and ·
emergent height
Initial conditions for thermal model e.;... it_ th...;. l d_ e l_ res t ....: e;... e_ th ia_ L-...: :::,_ : ..---- ....; ('r m _ a _ _ _ _ _ ag :... _ '. " .i n p_ "l
_
Flattening of subsidence curv.e after 80 m.y.
1500 2000 2500 3000 3500
____________...L 4000
c "'
� :[ Fig. 7. Example of thermal subsidence model for a guyot and its carbonate cap. For further explanation see text (modified from Rohl & Ogg, 1 996).
Table 1. Constraints and model parameters for reconstruction of subsidence and age-depth models; summary of age assignments and correlations of sequence
boundaries for north-west Pacific guyots.
Boundary
mbsf
Drowning
23.5
Takuyo-Daisan Guyot
Allison Guyot
Resolution Guyot Age (Ma)
Boundary
mbsf
Age (Ma)
1 00.0
Drowning
1 39 . 7
1 00.0
Boundary
mbsf
Age (Ma)
Mean age of sequence boundaries
MIT Guyot Boundary
mbsf
Drowning
3.0
Age (Ma)
Boundary
Age (Ma)
1 00.0
Drowning
100.0
Duration
Comments Set as base of R.
appenninica
foraminifer zone Alb l 2
60.0
101 . 1
Alb12
1 80.4
100.9
Alb l 2?
1 9.0
1 00.5
Alb12
1 0 1 .0
1 .0
Albll
84.4
101.8
Alb! I
239.7
102.2
Albll
69.9
1 02.2
Albll
1 02. 1
1.1
Alb i O
1 38 . 8
1 03.4
AlblO
293.5
103.4
Alb i O
1 09 . 3
1 03 . 5
Alb t O
1 03 . 4
1 .3
Alb9
1 86.4
1 04.8
Alb9
365.0
1 04 . 8
Alb9
1 38 . 8
1 04.4
Alb9
1 04 . 7
1 .2
Alb8?
226.0
1 05 . 8
Alb8
448.4
1 06 . 4
Alb8
1 96 . 7
1 06.3
Alb8?
1 06. 1
1 .5
Alb7
277.6
1 07 . 2
Alb?
491.6
1 07 . 1
Alb7
243.5
1 07 . 7
Alb?
1 07 . 3
1 .2
Alb6
3 1 7.0
108.3
Alb6
55 5.0
1 08.2
Alb6?
263.5
1 08.3
Alb6
1 08.2
0.9
AlbS
34 1 . 2
1 08.9
Alb5?
590.0
1 08 . 8
AlbS
283.7
1 08.9
AlbS
1 08.8
0.6
Alb4?
362.8
109.4
Alb4
626.0
1 09 . 3
Alb4?
306.0
1 09 . 5
Alb4?
1 09.4
0.6
Alb3?
393.0
1 1 0.2
Alb3
672.2
1 1 0.0
Alb3?
1 1 0. 1
0.7
Major event
� S· ;:s :L
(facies) Drowning
0
1 1 0.5
Alb3 i s observed at Resolution and Allison only
Alb2
406 . 1
1 1 0.5
Alb2
727.5
1 1 0.8
Alb2
6
1 1 0.8
Alb!
442.3
1 1 1 .4
Albl
793.5
1 1 1 .6
Alb!
22
I l l .7
Apt l 2
500.0
1 1 2.7
Apt l 2
35
1 1 2. 3
Apt!
I
Alb2 Volcanic
342.4 405.4
1 1 0.6 1 1 2.4
Alb2
1 1 0.6
0.5
Alb !
1 1 1 .5
0.9
Apt l 2
1 1 2.7
1 .2
1 1 3. 6
Apt I I Apt i O
88
579.2
1 1 4.5
640.8
1 1 5.8
Apt8
662.0
1 1 6.2
Apt?
690.5
1 1 6.8
Apt6
7 1 3. 5
1 1 7.3
Apt5
759.0
1 18.1
AptS
1 52 . 7
1 1 8.1
Apt5
Apt4
799.2
1 1 8.9
Base o f
1 64.4
1 1 8. 6
Apt4
Apt?
1 25 . 6
0.8
AptlO
1 1 4.5
0.9
1 1 5.8
1.3
Apt8
1 1 6.2
0.4
Apt7
1 1 6.8
0.6
1 1 7. 5
Apt6
1 1 7. 3
0.5
623.5
1 1 8.0
Apt5
1 1 8. 1
0.9
663.0
1 1 8.9
Apt4
1 1 8.9
0.7
1 1 6.8 Volcanic
1 1 3. 6
Apt9
1 1 5.0
AptlO Apt9
Apt i -Apt l 2 from Resolution only
submergence 538.0
Major event
602.5
(3:: i:;• ;:s
'I> :::: c.,
B"
�·
c.,
�'
� ""' 'I>
c::;-
base
carbonate Apt3
837.2
1 1 9. 5
Apt3
1 1 9.5
0.7
Apt2
853.0
1 1 9. 8
Apt2
699.7
1 1 9.8
Apt2
1 1 9.8
0.3
Apt!
895.0
1 20 . 5
Base o f
727.5
1 20.4
Apt !
1 20.5
0.7
LtB
1 2 1 .0
0.5
900.0
1 24.0
carbonate LtB
924.8
1 2 1 .0 Emergent island
-
1 12
U. Rohl & J. G. Ogg
(by convention, a depth is given as a positive number) is: depth of horizon with age 'Ma' = ('initial depth') - k,,['initial age' - age (Ma)].
(2)
This form of the subsidence equation is appropriate for the current depth of the age-horizon below sea-level. Three variables must be determined for each site: 'initial depth', 'initial age' and k; therefore it is necessary to incorporate either three age-depth pairs or two age-depth pairs and an assumed value for k. The computed subsidence curve is considered to be valid for the first 80 Myr following volcanism, and the models implied that further subsidence of old (> 80 Myr after volcanism) seamounts has been only about 1 00 m (Fig. 7). Following the Cretaceous time-scale of Gradstein et al. ( 1 994), an age of 1 2 1 .0 Ma was used for the base of the Aptian (assigned to the transgressive deposits above sequence boundary 'Lt. Barrem.', and 1 00.0 Ma for the drowning of Resolution, Allison and MIT platforms (boundary of R. appen ninica and R. ticinensis planktonic Foraminifera zones, which occurs at the top of the Mortoniceras (M.) inflatum ammonite zone, or c. 1 Myr before the end of the Albian stage (Table 2). An Aptian Albian boundary age of 1 1 2 . 2 Ma (Gradstein et al. , 1 994) was assigned to the base of the Allison guyot succession (below sequence boundary Alb 1 ) . How ever, for Resolution, MIT and Takuyo-Daisan guyots, this Aptian-Albian boundary age was merely compared with its predicted age (below sequence boundary Alb 1 ), and these computed assignments are within 0.4 Myr of the Gradstein et al. ( 1 994) age. This procedure can be applied to the depth of each sequence boundary at each guyot. Details on the calculations for each guyot have been given by R6hl & Ogg ( 1 996). This process allows direct comparison of 'subsidence corrected' se quence boundaries among guyots, and derivation of a common sequence scaling (Table 1 ). COMPARISON WITH EUROPEAN ALBIAN-APTIAN SEQUENCES
The Aptian-Albian portion of the sequence chrono stratigraphy compilation by Haq et al. ( 1 987, 1 988) was primarily calibrated with numerous sections in France. Aptian-Albian successions were also exam ined in the US Western Interior in Colorado and
central Texas. The composite 'Exxon' sequence series was rescaled to an equal-duration ammonite biozone scale, and assigned corresponding esti mates of absolute ages. Sequence stratigraphical interpretations for sev eral of these French sections were later published by other groups, and the results are generally consis tent with the compilation by the Exxon group for the major sequence boundaries, but Amedro ( 1 992) and Fries & Rubino ( 1 990) added several minor sea-level excursions. Sequence stratigraphy inter pretations of Lower Cretaceous successions in Western Europe and North Africa have been made by Amedro ( 1 992), Arnaud-Vanneau & Arnaud ( 1 990), Bralower et al. ( 1 993), Breheret ( 1 988), Fries & Rubino ( 1 990), Grafe ( 1 994), Hunt & Tucker ( 1 99 3), Jacquin et al. ( 1 99 1 ) and Robaszyn ski et al. ( 1 993). For a comparison 'standard' sequence scale, we have used a composite sequence stratigraphy de rived from those French and English sections that have detailed biostratigraphy (Breheret, 1 988; Fries & Rubino, 1 990; Amedro, 1 992; Bralower et al. , 1 993) coupled with strontium-isotope stratigraphy (Jones et al. , 1 994) or carbon-isotope stratigraphy (Weissert & Breheret, 1 99 1 ; Jenkyns, 1 995) which can be partially correlated with the combined bio stratigraphy and isotope stratigraphy of the Leg 1 43- 1 44 sites (Fig. 8). The relative thicknesses of major sequences in the Cantabrian Basin (Grafe, 1 994) also correlate fairly well with the relative thicknesses observed on the guyot composite. We retained the Haq et al. ( 1 987, 1 988) assignment of 'age-labels' for their major sequence boundaries (e.g. ' 1 07.5'), which were generally based upon this same region. However, the ages used by Haq et al. ( 1 98 7, 1 988) require minor adjustment to more recent Cretaceous time-scales (e.g. Obradovitch, 1 993; Gradstein et al., 1 994), so we consider these labels to be a numbering system. Correlations with ammonite-zoned stratotypes and depositional se quences interpreted on European margins rely on the basal ages of the carbonate succession, the R. appenninica biozone and approximate ages from benthic foraminifers. Aptian
The Aptian of the guyot composite has two major sequence boundaries, Apt2 and Apt7, which sug gests a preliminary correspondence to the two major Aptian sequence boundaries of the Haq et al.
Aptian-Albian eustatic sea-levels
1 13
Table 2. Time-scale calibration; for further explanation see text.
Sequence boundary
Ammonite and foraminifer zones calibration
Age (Ma)
Base of Cenomanian
Calibration age top Inflatum Zone (base R. appenninica foram) = (estimated I Myr after Alb 1 2) '(SB I I of Amedro; combined in "98" of Haq?)' base Injlatum Zone = Alb 1 2 '(SB 1 0 o f Amedro; combined i n "98" o f Haq?)' R. ticinensis (foram) base = upper Pricei subzone R. subticensis (foram) base = lower Pricei subzone base Pricei subzone = just after Alb I I '(SB 9 of Amedro; "99" of Haq)' 'base Cristatum subzone (base Inflatum Zone; base L. Alb.) = slightly after Alb lO' (SB 8 of Amedro) T praeticinensis (foram) base = before Inflatum Zone base base Niobe subzone = Alb9 '(SB 7 of Amedro; " I 00.5" of Haq)' (may be SB 6 of Amedro, but relative age seems too old) base Steinmanni subzone = after Alb? '(SB 5 of Amedro; " 1 03" of Haq)' base Bulliensis subzone = before Alb? (probably this is a parasequence set, not a well-developed sequence) T primula (foram) base = lower Mammillatum Zone '(SB 4 of Amedro; " I 06" of Haq)' base Kitchin subzone (base Mammillatum Zone) = before Alb5 (SB 3 of Amedro) base Regularis subzone = slightly before Alb4 (SB 2 of Amedro) base Milletoides subzone = slightly before Alb3 '(SB I of Amedro; " 1 07.5" of Haq)' base Famhamensis subzone = before Alb2 (gap in Europe sequence sections) est. slightly before Alb I Hedbergella panispira (foram) base = uppermost Jacobi Zone (SB d8 of Vocontian)
98.9 1 00.0 1 00.0 1 0 1 .0 1 0 1 .0 1 0 1 .2 101.8 1 02.0 1 02 . 1 1 03.2 1 03.4 1 03 . 5 1 04.7 1 04.7 1 06 . 1 1 07 . 1 1 07.3 1 07.6 1 08.2 1 08 . 5 1 08.8 1 09.2 1 09.4 1 09 . 5 1 1 0. 1 1 1 0.2 1 1 0.6 1 1 1 .0 1 1 1 .5 1 1 1.9 1 1 2.5 1 1 2.7 1 1 3.6 1 1 4.5 1 1 5.3 1 1 5.7 1 1 5.8 1 1 5. 8 1 1 6.2 1 1 6.7 1 1 6.8 1 1 7.0 1 1 7.3 1 1 7.7 1 1 8. 1 1 1 8.3 1 1 8.6 1 1 8.9 1 1 8.9 1 1 9.5 1 1 9.7 1 1 9. 8 1 20.5 1 2 1 .0 1 2 1 .0
Widespread drowning Alb 1 2 (very major)
Alb I I (major) Alb l O Alb9 (major) Alb8? Alb? (major) Alb6 Alb5 (major) Alb4 Alb3? Alb2 (very major) Alb ! Base of Albian Apt 1 2 Apt I I Apt l O (major)
Apt9 Apt8 Apt? (major) Apt6
(SB d7 of Vocontian) 'Jacobi' black shale = transgression (estimated 0.5 Myr after Apt9)' T bejaouensis (foram) base = about Apt9 base Jacobi Zone = Apt9 (SB d5 of Vocontian) (probably this is a parasequence set, not a well-developed sequence) Hedbergella trochidea (foram) base = just after Apt? '(SB d4 of Vocontian; " 1 09.5" of Haq)' base Subnodosocostatum (base Nutfieldensis Zone) = just before Apt? (probably this is a parasequence set, not a well-developed sequence) Glob. algeriana = approximately midway between Apt5 and Apt6
Apt5
Apt4 Apt3
Glob. ferreolensts (foram) base = just before Apt5 'Selli' black shale = transgression after Apt4 ( = Apt4-0.3) Leupoldina cabri (foram) base = Apt4 (SB d2 of Vocontian) (probably this is a parasequence set, not a well-developed sequence) top Prodeshayesites (Fissicostatus) Zone slightly after Apt2 '(SB d I of Vocontian; " 1 1 2" of Haq)' =
Apt2 (major) Apt! Base of Aptian LtB (major)
'Calibration age; assigned just after "LtB" sequence boundary'
Amedro refers to Amedro ( 1 9 92), and Haq refers to Haq et al. ( 1 987). See text for further explanation.
� II
Composite guyot stratigraph y
(Amedro, 1992)
Event
.
...... .,.
(Haq et a/. 1987, 1988)
Aube (F) - Folkstone (GB)
PlanktOniC Foraminifer Zones
Stage
European margin sea-level curve
Sequence (Supercycle l a nd cyc e)
Cen.
;;-= ��
g
ns
�§� £.8 E �
Sea level curve {short-tenn cycles and long-term trend)
Chronostrati-
graphy (Serie�
and Stage boundaries with age in Me)
low
CEN.
96
r:::
.!2
--tl 1--
.c
·�·
C: «<
:c < f--
F.
r- - r----
·'Apt1 2' . , ,."., -·' - Io. Apt1 1 '· mn. oorr-'•tJon• """' "'"""" J.atopfS:
..
�:s.�- �:�: :: �rEC!!�! Ap
r::: ctl :;::: c.
_
G/. •1-IMM
GJ;�-;,�·= IiUpolillni
-
.ubd. Glob. blow/
_
f: •
Apt7"
•
.
t6"
AptS"
"
Apt4"
•Apt3?"
-"Apt2"
-
·.. � :
I
Fig. 8.
Vocontian Basin
1.2
(fries &Rubino, 1990)
--
- -- -
--
--
- ---
--
--
-- - -
CIUtUIIIll!:jllltJIIJ Event foraminifer Zont1 "
--
- --
1--da---l
bejaouensis
trocoidea
I
jacobi
nutfieldensis
- - subrlOdoa/geriana socostatum n s si e eo_ ..: ' :: ".:.: ' r:..: l___-j. L �� =::::w:: ""1 cabri
�
I
LZB 4
4.1
3.5 LZB-3 � 113.5 I 3.4
�L
I
Correlation with sequences of European continental margins (France, UK) and with the 'Haq ( 1 987) curve' (modified from Rohl & Ogg, 1 996). : ;��
:·
�
o. ;::......
Re :.....
� O
4.2 I
�
s;J
1.1
--
block d6 ,_
' M
Aptian-Albian eustatic sea-levels
( 1 987) compilation. The carbon-isotope curve of Resolution Guyot provides a calibration to stan dard planktonic foraminifer zones (Jenkyns, 1 995), thereby providing a correlation with depositional sequences mapped in the Vocontian Basin (Fries & Rubino, 1 990). The detailed correlation of guyot sequence boundaries with those of Haq et a!. ( 1 987) have been documented by Rohl & Ogg ( 1 996). The correlation of the sequence boundaries observed in the guyot successions with those in English and French sections is shown in Fig. 8. A major emergence and sequence boundary 'Apt2' occurs in the middle of the Early Aptian with a projected age of approximately the middle of the Globigerinelloides blowi planktonic foraminiferal zone (Jenkyns, 1 995). We correlate this event with the major sea-level fall ' 1 1 2 ' of Haq et a!. ( 1 987) and middle Lower Aptian sequence boundary 'd 1 ' of Fries & Rubino ( 1 990) in the Vocontian Basin. This major fall of sea-level exposed the Urgonian platforms of south-eastern France, producing 50-m deep incised valleys into that shelf margin (Arnaud Vanneau & Arnaud, 1 990). The minor sequence boundary 'Apt3' at Resolution and MIT guyots probably represents a minor oscillation during the major eustatic lowstand following the ' 1 1 2' sea level fall. The following rapid major transgression drowned these Urgonian platforms and has been invoked as a factor in producing the condensed organic-rich horizons of 'Livello Selli' of Italy, 'Niveau Goguel' of southern France (Bn!heret, 1 988), and 'Oceanic Anoxic Event Ia' of the global oceans during the middle of the G. blowi foraminiferal zone (Bralower et a!., I993). Jenkyns ( 1 995) correlated the organic rich clay with plant material at the base of an oolite succession near this position at Resolution Guyot with the 'Livello Selli' episode, but this is only a single horizon within a larger transgressive trend. The trend of the carbon isotope curve at Resolution Guyot towards higher values (Fig. 6) indicates that significant global removal of light organic carbon occurred both before and after the sequence bound ary 'Apt4'. The minor sequence boundary 'Apt3' at Resolution and MIT guyots probably represents a minor oscillation during the major eustatic low stand following the ' 1 1 2' sea-level fall. The correla tion of the sequence boundaries observed in the guyot successions with those in English and French sections is shown in Fig. 8. In Hole 866A at Reso lution Guyot, the Aptian-Albian boundary is as signed to be c. 1 50 m higher than the sequence
115
boundary Apt9 (Arnaud-Vanneau & Sliter, 1 995; Jenkyns, 1 995; Jenkyns et a!., 1 995), which is equivalent to nearly one-third of the Aptian sedi mentary thickness. The relative spacing of the major Aptian sequence boundaries recorded by a shallow-water carbonate bank on a smoothly subsiding guyot should corre spond to the relative spacing of the same sequence boundaries on a thermally subsiding passive margin where sedimentation kept pace with accommoda tion space. The Basco-Cantabrian basin of northern Spain probably represented such a passive margin during the Aptian-Albian (Grafe, 1 994). A detailed measured section in marginal facies in the Basco Cantabrian basin contains c. 5 km of uppermost Barremian to upper Albian strata, terminated by a drowning in the R. appenninica planktonic fora minifer zone (Grafe, 1 994). The Aptian consists of 2000 m of interbedded deltaic sands and clays with shallow-water 'Urgonian' limestones, and is bounded and subdivided by five major sequence boundaries ('AP 1 '-'ALl ') . The approximate rela tive thicknesses of the depositional sequences 'AP 1 '-'AP4' in units of l 00 m are 2 : 5 : 6 : 7 (Grafe, 1 994). In comparison, the relative thicknesses of the major depositional units bound by sequence bound aries lt.B-Apt2-Apt6-Apt 1 O-Alb2 on Resolution Guyot in units of c. 3 5 m are 2 : 5 : 5 : 6. Therefore, the relative spacings of major Aptian sequence boundaries in the Pacific guyots and the Basco Cantabrian margin are similar. Albian
Biostratigraphy and isotope stratigraphy within the Albian successions at the guyots provide only broad correlations with the Albian of Europe. Amedro ( 1 992) identified 1 2 sequence boundaries in his de tailed biostratigraphical and sequence stratigraphy interpretation of Albian strata in the Department of Aube and in the Channel cliffs of Folkestone. Although several intervals within the outcrop suite are condensed or discontinuous, his compilation is currently the best documented study (J. Hardenbol, personal communication). The correlation of the sequence boundaries observed in the guyot succes sions with those in English and French sections is shown in Fig. 8. The basal Albian is generally a gap in sedimenta tion on the northern France-southern England shelf, and was considered to represent a major sequence boundary ' 1 07.5' by Haq et a!. ( 1 987).
1 16
U. Rohl & J G. Ogg
The following major transgressive rise is considered to be the cause of the organic-rich 'Paquier Level' of France (Breheret, 1 98 8) and a coeval widespread anoxic event 'OAE 1b' in the oceans (Bralower et a!., 1 993). The earliest major Albian sequence bound ary and deepening event observed on the guyots is 'Alb2' at Allison and Resolution guyots. Sequences Alb2-Alb4 are developed in a relatively deeper facies at the guyots, and this general deepening trend probably correlates with the major Early Albian transgression that followed ' 1 07 .5'. This major transgression also drowned the carbonate platform of Takuyo-Daisan. As in the Aptian, the relative spacing of the major Albian sequence boundaries in these guyot sections should agree with the relative spacing of the same sequence boundaries in the marginal facies of the Basco-Cantabrian passive margin. The Albian, be fore a major deepening in the R. appenninica plank tonic foraminifer zone, consists of 2 700 m of deltaic facies that is bounded and subdivided by nine major sequence boundaries ('ALl '-'AL9') (Grafe, 1 994). The approximate relative thicknesses ofthe Lower to lower Upper Albian depositional sequences 'ALl ' 'AL7' of the Basco-Cantabrian margin in units of 1 00 m are 3 : 3 : 4 : 4 : 2 : 5 : 5. In comparison, the relative thicknesses of the depositional units bound by sequence boundaries Alb2-Alb5-Alb7-Alb8Alb9-Alb 1 0-Alb 1 1 -Alb 1 2 on Resolution Guyot in units of c. 20 m are 3 : 3 : 2 : 2 : 2 : 5 : 4. The general trends of subequal depositional sequences within the lower and middle Albian and two rela tively thickened sequences in the upper Albian are similar, but the details of the ratios are not entirely consistent. These minor discrepancies may reflect difficulties in assigning precise sequence boundaries in the Basco-Cantabrian section (some covered intervals span 1 50 m in stratigraphic thickness) and in some intervals on the guyot platforms. APTIAN -ALBIAN TIME-SCALE AND SCHEMATIC EUSTATIC CURVE Schematic sea-level curve
The relative magnitude and thickness of successive cemented exposure horizons serve as an approxi mate guide to the apparent significance of the se quence boundary. It is not possible to derive esti mates of the absolute sea-level fluctuations for any of these sequences, but a schematic relative magnitude
chart can be plotted with the ages derived from the thermal subsidence model (left column of eustatic sea-level fluctuations in Fig. 9). Minor and moderate sequences in the Aptian appear to be nearly twice as frequent as sequences identified in the Albian, al though several of the Early Aptian sequences have durations of about 400 kyr and are probably associ-· ated with parasequence sets induced by the 400-kyr Milankovitch cycle of eccentricity. Unfortunately, the absolute magnitude of eustatic sea-level variations cannot be determined from the carbonate banks, but a relative duration of sea-level! lowstand can be assigned from the degree of diage-· netic cementation of the emergent sediments. Long term major second-order trends in sea-level can be: vaguely approximated from the relative average thicknesses of parasequences and general facies characteristics. At each guyot, some successions of sequence boundaries are associated with well developed parasequences in crypto-algal-rich facies, whereas other sets of sequences are dominated by relatively deep facies. A relative abundance of thin, shallow-facies parasequences may indicate that car bonate accumulation was limited by a relatively slow rate of accommodation compared with thicker and less obvious parasequences in a facies domi nated by deeper lagoonal sediments. This philoso phy has been used in other carbonate successions as a proxy for long-term rises and falls of sea-level (e.g. Fischer, 1 964; Read & Goldhammer, 1 988; Sadler et al , 1 993). We incorporated these patterns of facies and parasequences that span groups of depo sitional sequences among the guyots into an esti mate of long-term trends in relative sea-level (middle column of eustatic sea-level fluctuations in Fig. 9). It is not possible to make estimates of the magnitude of these longer-term trends. The schematic curves of relative magnitudes of individual sequences and of the multi-million year trends in sea-level were combined and plotted upon the general trend of transgressive sea-level upon continental margins during the Barremian to Cen omanian (e.g. Haq et a!., 1 987, 1 988) (right column of eustatic sea-level in Fig. 9). This composite curve is only a schematic representation of sea-level trends, and does not imply a linear scale. .
Guyot drowning
The drowning of guyots m the earliest Albian (transgression after Alb2 sequence boundary at Takuyo-Daisan) and in the latest Albian (transgres-
Major and minor sequence boundaries
Ma
-
(sges derived from subsidence model)
Guyot 11 Hl!lh
� . . . 1-
100 - Drowning
105 -
Alb 1 2 Alb1 1
110 -
Alb10
Alb9
Alb7 Alb6
Alb5 Alb4
120 -
<;
� � ( � c
�
<
Alb3?
�
Alb1
�
Alb2
� - - -�
115 -
· · · · ·
Is
Alb8?
.
Apt12
Apt11
Apt1 0 Apt9 Apt8
Apt7 Apt6
Apt5 Apt4 Apt3
Apt2 Apt1
Lt. Bar
Low
�
(
-
(
� <
<.
-
c
r"'"
Fig. 9 . Composite
Schematic eustatic curve (short-term cycles;
superimposed on long-term trend from Hsq et sl. 1987)
1 03.4 1 04.7
)
1 06.1 1 07.3 1 08.2 1 08.8
Q) co ...1
.� - - - - -
1 1 5.8 1 1 6.2
(
1 1 6.8 1 1 7.3 1 1 8.1 1 1 8.9
1 1 9.5 1 1 9.8
120.5
1 2 1 .0 ·
-�
1 1 1 .5
1 1 3.6
-
k
1 1 0.6
1 1 4.5
·-
�.
1 1 0.1
1 1 2.7
c: m .c
��
1 09.4
r
··
----�-� f)_ · · · ·fP ·
·
• • Stoi;C'zk'alS disnar
-
\}?""" �-
c: m
·-
...
c.
-
-
:::2:
-
>'i: co w
1 12.2 ± 1 . 1 Ma• •
-
• 121.0 ± 1.4 Ma• •
�
58 -+ 4 109. 1 58 3 -+ 109.5 58 2 -+ 1 10.2 58 1 -+ - 1 1 1.0
js8
Hypacanthoplites jacobi
-+
T.(T.) bowerbsnkl D. deshsyes/ D. forbes/
Prodeshayesite� fissicostatus - -
M.(M.) price/
--
D. crlstatum
{ equlv. to D. blpflcstus ]
- - - - - - D. niobe A. Intermedius - -H-:(HJ cientst";;s- _ /f:.(/f:.) !!."!!!!'!!•!!.u! P.(l.) steinmsnnl 0.
-
B. bidentatum
5B d5-+
1 15.8
58 d4-> - 1 17.0 biiiCk shaM "d3" [S./11 tov•IJ 5B d2-+
0. arxltlfonnla; 0. IMchMI; P. puxoaMnua; C. florldum S. kltchlnl
- - - - - - 103.5
Ticinella primula
�
�S· �
:l - - - - - - - 108.5
M. millet/odes
6,-P In Englloh
anrwl a.ctlona}
· - - - - -
•
- - -
H. jacobi
N. nolsnl
<::;:: �S · �
�
Hedbergella planispira
F. farnhsmensls L. schrammenl
§
r;·
- - - - - - - 1 12.5
�;;;�
Ticinella bejaouensis
c;;-
Hedberge/ls 1 15· trochidell Giobigerinelloides1 16· 7 slgerisns 1 1 7_ 7 3 s.z. In msrtlnoldes & Gl. ferreo/ensls forbes/; 2 s.z. In 1 18·3 Leupoldins csbrl bowerbsnkl & deshsyes/ 1 18.9
P. cunn/ngtonl
7
T.(E.) subsrcticum
58 d1-+ 1 19.7 121.0 ± 1.4 Ma
Ticinella praeticinensis
bulllens/s
H. rubrlcosus
biM:k shaM "d6" {JtteobiMwl}
C.(E.) msrtlnoldes
-
d8
100.0 Rotaliapora ti:i'!..e!!s.!_s _ 101_2 _ � s.!!b!!.c!!!e!!.sf! 101_8
M.(M.) lnflstum
5B d7-+
Parahoplites nutfie/densis
- - -
· · · · - - · - - · Rota/iapora • • appenninica M.(D.) perlnflstum M.(M.) fsflsx
58 9 -+ 102.0
103.2 Euhopliles lsutus 58 8-+ - - - - - Euhoplites 58 7 -+ 104.7 loricatus - fiopilies (H.) - 8 61'--+ dentatus - Douvilleiceras 58 5 -+ 107. 1 107.6 mammillatum
- - -
Subzone
58 11)--.+ 101.0
�-------�------·
_.
w
Guyot Drowning 8 1 1-+ 100.0
Leymeriella tardefurcata
�
� co
� 98.9 ± 0.6 Ma•
Mortoniceras (M.) inflatum
,
·
Calibration
=��fliles
• ' 98.9 ± 0.6 Ma • •
101.0 102.1
Zone
Neog�
Cenomanian
1 00.0
Planktonic Foraminifer Zones
(correlations to Amedro, 1992, and Fries & Rubino, 1990; N.W. Europe ammonite scale modified from Hsrdenbol et sf., In press)
(boundary ages from Gradste n l et a l., 1994)
Low
Rlslna Falllna Hlah
Ammon ite Zones and Subzones
Geologic Stages
. .� - - - - -
....
� � (. �
AGE
Secona; order sea-level trends
P. obsoletus
P. bode/ - - - - - - - - - ·
Barremian
Globiqer• inello1des ' • • blowi
--
-.]
sequence-biostratigraphy-isotope scale for Aptian to Late Albian (modified from R6hl & Ogg, 1 996).
1 18
U. Rohl & J G. Ogg
sion following Alb 1 2 sequence boundary) corre sponds to intervals when a major sea-level fall and successive transgressive rise are interpreted to be superimposed upon a longer-term trend of rising sea-levels. As a result, the sea-level during the transgression after the major sequence boundary crests at a much higher level than during the previous highstand. When coupled with the accu mulated thermal subsidence during the major low stand (Alb2 or Alb 1 2), the rapid rate and enhanced magnitude of the succeeding transgressive rise can apparently overwhelm the accumulation rate of the carbonate system, and the guyot platform is deep ened beyond a critical depth where carbonate pro duction of bioclastic debris can build upwards faster than the rate of subsidence. The susceptibility to drowning may be enhanced when the carbonate platform is in near-equatorial latitudes where higher nutrient levels, El Nino-type warming events, Intertropical Convergence Zone precipita tion, strong equatorial currents, or other factors may inhibit the rate of carbonate accumulation (e.g. Menard, 1 9 82; Schlager, 1 989; Ogg et al., 1 995b). Aptian-Albian time-scale
The correlation of major sequences observed on the suite of guyots with major sequences interpreted in north-west Europe (Fries & Rubino, 1 990; Amedro, 1 992) allows the associated absolute age scale for the guyots to be applied to coeval biostratigraphical events in Europe (Table 2). The relative position of ammonite subzonal boundaries and planktonic foraminifer boundaries to the recorded European sequences allows assignment of ages to approxi mately 1 5 ammonite events and six foraminifer events, and an estimate of the ages of four other foraminifer events can be derived from their asso ciation with ammonite zones (right half of Fig. 9). In these summary tables and chart, the correspond ing named sequence boundaries as denoted by Amedro ( 1 992) and Fries & Rubino ( 1 990) are also shown. Our age assignments to the Aptian-Albian am monite zonation display a few significant depar tures from previous time-scale estimates that incorporated a working assumption of 'equal sub zone' scaling (e.g. Gradstein et al., 1 994; Hardenbol et a!., 1 997). In the Late Aptian, the duration of the Hypacanthoplites jacobi ammonite zone is equal to the four ammonite zones that make up the Early Aptian. The apparent series of overlapping trans-
gressive pulses in the late Early Aptian of the guyots (Apt2-Apt6) may have produced a more rapid overturn of nearshore ammonite fauna in the Euro pean margin sections, and therefore an increase in the number of identified ammonite subzones. In contrast, the overall regressive nature of the latest Aptian to earliest Albian on the guyots (Apt 1 1 Alb2) corresponds t o a n abundance of hiatuses or tidal sands in European shelf sections (e.g. Ruffell, 1 99 1 ; Amedro, 1 992), and therefore may have contributed to the lack of ammonite subzonal re finement. The regressive trend spanning the Barremian-Aptian boundary interval on the guyots (Lt. Bar.-Apt2 on Resolution) also corresponds to a common hiatus or condensation in European sec tions (e.g. Ruffell, 1 99 1 ). The correlation of the guyot succession in the Albian implies brief ammonite subzones in a por tion of the middle Early Albian, where six subzones (L. regularis to 0. auritiformis) span only 2 Myr. The early Late Albian is implied to have three ammonite subzones spanning over 3 Myr (D. crista tum to M. injl.atum subzones; 1 Myr each). In contrast, the Middle and Early Albian have 1 6 ammonite subzones spanning 8 Myr (0.5 Myr each). As in the Aptian, this interval of expanded ammonite subzones was also interpreted as a gen eral shallowing trend (Alb 1 0-Alb 1 2). However, an other interpreted shallowing trend in the late Early Albian of these guyots (Alb3-Alb7) did not display an expansion of duration of ammonite subzones; and this interval has been interpreted as a transgres sive trend in north-west Europe (J. Hardenbol & M. Farley, personal communication). This composite sequence-ammonite-foraminifer chronostratigraphy can be improved in the future by utilizing detailed counts of Milankovitch cycles incorporated within individual ammonite subzones or sequences. SUMMARY AND CONCLUSIONS
Four carbonate banks of Aptian-Albian age kept pace with the steady subsidence of the underlying volcanic edifices and therefore maintained a depo sitional environment close to sea-level over several million years. These carbonate banks record rapid falls of relative sea-level as emergent surfaces and rapid rises of sea-level as deepening or drowning events. Simultaneous episodes of emergence or deepening over this portion of the Pacific Ocean
1 19
Aptian-Albian eustatic sea-levels
should represent major eustatic sea-level events that are important on a global scale. Our sequence stratigraphical interpretations are based on a single 'outcrop' of 1 0 em diameter drilled through the central platform at each guyot. Downhole logging coupled with recovered carbon ate facies allowed delimitation of depositional facies and surfaces of cementation, and recognition of a hierarchy of parasequences, parasequence sets and sequences. We assigned geological ages from biostratigraphical assemblages and chemostrati graphical (carbon and strontium isotope) curves and scaled the detailed sequence stratigraphy by compensating for thermal subsidence rates. The number of upward-shallowing cycles or parase quences was used to compare relative durations of sequences among sites. The results show that the Aptian-Albian carbonate banks in the Pacific con tain 1 2 Aptian and 1 2 Albian significant shallowing events, of which a third were associated with major episodes of emergence. Major transgressions which follow prolonged emergences were a major factor for the widespread drowning of the Albian platforms (e.g. sequence boundary Alb2 in the earliest Albian, sequence boundary Alb 1 2 in the late Albian). The major Aptian and Albian sequence bound aries recognized in the suite of Pacific guyots can be correlated within the constraints of carbon-isotope stratigraphy and biostratigraphy with the major events recognized in the French stratotype regions by Haq etal. ( 1 987), Breheret ( 1 9 88), Fries & Ru bino ( 1 990) and Amedro ( 1 992). Sequence strati graphy interpretations for several of the French sections are generally consistent with the compila tion of Haq et al. ( 1 987) for the major sequence boundaries, but added several minor sea-level ex cursions. The number and relative spacing of major events within the Aptian are in agreement with coeval events recognized on the thermally subsiding Basco-Cantabrian margin by Grafe ( 1 994). These correlations, coupled with the assumptions of a thermal subsidence model for the guyots during the Aptian-Albian, allow us to propose a revised sequence chronostratigraphy and scaling of associ ated biostratigraphical and isotopic curves for the Aptian to mid-Late Albian. This chronostratigra phy model will have minor distortions caused by relative changes in the rate of accommodation space caused by second-order eustatic sea-level trends, but the magnitudes of these eustatic trends are relatively insignificant over the 900-m thickness of the main Aptian-Albian platform succession.
Our study indicates that guyot sequences in the Pacific are consistent both in relative timing and relative magnitude with those independently inter preted in Europe. This implies a common global sea-level variation, and hence a eustatic signature. Therefore, sequence stratigraphy recognition based on such sea-level changes provides a chronostrati graphical scale with ability to obtain fairly precise global correlation horizons. The mid-Cretaceous had eustatic, high-amplitude sea-level changes, im plying that even a greenhouse 'hothouse' world with elevated greenhouse conditions can experience sig nificant oscillations of sea-level. Methods are re quired for temporary removal of large quantities of the global ocean water. We suggest that evidence for lack of continental ice during this period in the interiors of high-latitude continents should be re examined. ACKNOWLEDGEMENTS
This research was sponsored by the Ocean Drilling Program, USSAC and the Deutsche Forschungsge meinschaft (DFG). U.R. acknowledges support by the Bundesanstalt fUr Geowissenschaften und Rohstoffe (BGR), Hannover. Many of our interpre tations of the depositional history are based upon the observations of our shipboard colleagues. Jan Hardenbol, Hubert Arnaud, Hugh Jenkyns, Annie Arnaud-Vanneau, Patricia Cooper, Jiirgen Grotsch, Kai-Uwe Grafe and Bill Sliter provided preprints of their papers. Valuable reviews and critiques were provided by Annie Arnaud-Vanneau, Gilbert Cam oin, Angela Coe, Felix Gradstein, Jan Hardenbol, Hugh Jenkyns and an anonymous reviewer. Por tions of the text and figures from 'Aptian-Albian sea-level history from guyots in the Western Pacific' by Rohl & Ogg ( 1 996) are used with permission of the American Geophysical Union. APPENDIX Resolution Guyot
Aptian Latest Barremian sequences (964. 0 mbsf to Lt.B. (924. 8 mbsf) to Aptl (895. 0 mbsf)). This 69-m inter
val is characterized by high-amplitude simultaneous fluctuations in resistivity and natural gamma-ray in-
1 20
U. Rohl & J. G. Ogg
tensity (Fig. 5). Recovered core lithologies (upper Lithologic Subunit VIC) are very heterogeneous and dominated by a bioturbated peloid wackestone mudstone facies with variable amounts of reworked ooids, blackened elements, rare micritic-algal onc olites, rudist shells, laminations, bird's eyes, partial dolomitization, algal-microbial mat structures and some desiccation cracks (Shipboard Scientific Party, 1 993b; Arnaud eta!. , 1 995). The combined logging signature and recovered facies indicate a dominance of peritidal parasequences with common shallowing into algal-mat flats or brief emergences. The resistivity-gamma peaks are interpreted as en hanced cementation and uranium scavenging at this algal-microbial mat to emergent surfaces. Arnaud considered this interval to represent the onset of an Aptian second-order transgression to more open ma rine conditions following the restricted marine con ditions of the underlying dolomite-rich successions. However, the array of facies suggests that accommo dation space was still limited, so we place such a main transgressive episode where significant deep ening of facies occurs at 8 9 5 mbsf. Arnaud eta!. ( 1 995) tentatively assigned possible sequence boundaries of 'Sb 1 3' at 964 mbsf, where a relatively high gamma-ray and resistivity peak and possible associated desiccation cracks are followed by temporary disappearance of tidal flat environ ments, and of 'Sb 1 4' at c. 922 mbsf, where another pronounced gamma-resisitivity peak is followed by a restricted facies of algal-microbial mats, flat pebble layers and bird's eye features. We denoted their 'Sb l 4' sequence boundary as 'Lt.B.' in Fig. 5 . The next 2 1 m consist of a series of parasequences with brief emergence indicated by desiccation cracks (Arnaud eta!., 1 995). These parasequences climax at 901 -895 mbsf in a 6-m bed of sustained high resistivity and natural gamma-ray intensity, followed by a pronounced drop in natural gamma ray intensity accompanying the onset of a thick series of wave-reworked grainstone beds. We inter pret this 6-m layer as a peritidal facies with slowed rates of accommodation, thereby indicating a se quence boundary. In contrast, Arnaud eta!. ( 1 99 5) considered this bed and the underlying 20 m as a transgressive systems tract, although this interpre tation is difficult to reconcile with their observa tions of desiccation cracks within the same interval.
prinid rudist fragments (upper Lithologic Unit VIB). This shell hash of open-marine conditions contains numerous peaks in resistivity, which we interpret as shallowing-upward cycles. The nine main parase quences are grouped into four sets. The lowest parasequence set suggests a retrograding trend, the next two thickened sets are assigned to an early highstand systems tract, and the highest compact set culminates in well-cemented parasequences as sociated with peaks in natural gamma-ray intensity. The corresponding Core 9 1 R from this uppermost portion contains algal-mat facies, bird's eyes, on coids and some clay layers (Shipboard Scientific Party, 1 993b). Therefore, we interpret this algal rich zone as a peritidal to emergent facies with either a progradation of distal clastics from a weathered volcanic edifice or episodes of altered volcanic ash. This rapid shallowing to emergence and associated major sequence boundary 'Apt2' is identical to sequence boundary 'Sb 1 5' of Arnaud et a!. ( 1 995).
Aptian sequence 1 (Aptl to Apt2, 895.0-853.0 mbsf).
This 38-m sequence has a parasequence signature in restricted shallow-subtidal to peritidal and algal mat facies similar to that of the underlying se-
Aptian sequence 2 (Apt2 to Apt3, 853.0-83 7.2 mbsf).
This 1 6-m sequence is dominated by peloidal pack stone-wackestone alternating with thin algal mats, representing a type of lagoonal-peritidal parase quence (Strasser et a!., 1 99 5). Six main para sequences are clearly displayed in the downhole logs and FMS images that culminate in typical resistivity-gamma peaks associated with microbial algal mats. These parasequences are grouped into two sets, culminating in the maximum resistivity gamma peak observed in the Aptian succession. At this level there is also a prominent clay horizon with a relatively high organic carbon content (Baudin et a!., 1 99 5). Strasser et a!. ( 1 99 5) considered such organic-rich clays as transgressive deposits, but we presume that the source of such a clay must be a weathered hinterland, therefore the deposit repre sents a progradation to the site by distal clastic influx. As a result, we assign a sequence boundary 'Apt3' to this well-cemented horizon with overlying clay. However, sequence boundary 'Apt3' is of minor im portance compared with underlying event 'Apt2', indeed, it may represent a continuation of a larger scale 'regressive' to 'lowstand' phase of a second order sequence. Aptian sequence 3 (Apt3 to Apt4, 83 7.2- 799. 2 mbsf).
The majority of this 42-m sequence consists of well washed grainstone-rudstone with abundant ca-
Aptian-Albian eustatic sea-levels
quence Apt2. Seven major parasequences are grouped into three sets. Volcanic debris and glass are recorded in the lower portion (Cores 89R to 8 7R). The upper sequence boundary 'Apt4' is assigned to the top of a pronounced resistivity peak consid ered to be the product of increased cementation caused by freshwater diagenesis (see Rohl & Strasser, 1 99 5; Strasser et al. , 1 99 5). This level coincides with the shipboard lithological boundary between Lithologic Unit VI (restricted lagoonal peritidal parasequences) and Unit V (white oolitic grainstone) (Shipboard Scientific Party, 1 9 9 3b). Aptian sequence 4 (Apt4 to Apt5, 799. 2- 759. 0 mbsf). This 40-m sequence begins with a deepening to c.
775 mbsf, where the corresponding Core 83R con tains a layer with various corals. The overlying succession displays aggrading oolitic parasequences followed by a shallowing to peloidal wackestone packstone with oncoids and associated increase in natural gamma-ray intensity. The cycles within the sequence can be grouped into four sets of parase quences. We assigned sequence boundary 'Apt5' to a resistivity peak, interpreted as a cementation horizon, underlying a low resistivity-gamma facies. Arnaud et al. ( 1 995) assigned a sequence boundary 'Sb 1 6' about 9 m higher, where natural gamma-ray intensities again increase, but we interpreted this as the onset of the next higher oolitic highstand facies. Aptian sequence 5 (Apt5 to Apt6, 759. 0- 713. 5 mbsf).
This 45. 5-m sequence is characterized by oolitic upward-shallowing cycles (Jenkyns & Strasser, 1 995; Strasser et al. , 1 99 5). Seven major parasequences observed in the FMS logs can be grouped into three parasequence sets. The upper parasequences are ter minated by hardgrounds exhibited as sharp resistiv ity peaks (e.g. at 7 1 8 mbsf) or as horizons bored by bivalves (Arnaud et al., 1 995). The upper sequence boundary ' Apt6' was assigned to a prominent peak in resistivity, corresponding to an oolitic hardground surface and algal-microbial mat facies overlain by peloidal packstones in Core 76R. Aptian sequence 6 (Apt6 to Apt7, 713.5-690.5 mbsf).
The lower portion ofthis 23-m sequence has peloidal and oolitic packstone-wackestone, followed by an upward increase in abundance of keystone vugs and indications of beach environments. Seven upward shallowing parasequences were grouped into two sets. Sequence boundary 'Apt7' was assigned to the
121
top of these deposits. This major change was also recognized as sequence boundary 'Sb 1 7' by Arnaud et al. ( 1 99 5) and marks the approximate top ofLitho logic Unit V. Aptian sequence 7 (Apt7 to Apt8, 690. 5-662. 0 mbsf).
This 28.5-m sequence is dominated by a prograding series of shallow subtidal to intertidal parasequences of oolitic packstone to mudstone-wackestone and algal mats (Shipboard Scientific Party, 1 99 3b). Com pared with the underlying oolite-dominated open marine sequences, this facies change represents a shift to a more restricted and shallow lagoonal facies. We grouped the nine main parasequences into three sets, and assigned a minor sequence boundary 'Apt8' at the sharp top of a resistivity peak corresponding to algal-laminated mudstone overlain by wackestone with miliolid foraminifers in Core 7 1 R. Aptian sequence 8 (Apt8 to Apt9, 662. 0-640. 8 mbsf).
This 2 1 -m sequence also consists of wackestones to packstones deposited in shallow subtidal and re stricted marine environments. We identified five parasequences in two parasequence sets. Sequence boundary 'Apt9' is assigned to the sharp top of a pronounced resistivity peak, interpreted as en hanced cementation. Aptian sequence 9 (Apt9 toAptlO, 640.8-579.2 mbsf).
This 62-m sequence contains 1 6 parasequences grouped in four sets. Sequence boundary 'Apt l O' is assigned to the top of a relatively high-resistivity peak, below another interval of relatively low resistivity parasequences. A 1 5-cm vertical coalified band of possible bark in Core 62R (Fig. 1 4 in Ship board Scientific Party, 1 993b) may be associated with an emergence at this sequence boundary. The general lower resistivity and reduced levels of natu ral gamma-ray intensity suggest a deeper facies than Sequence Apt7 or Apt8. Arnaud et al. ( 1 995) grouped these three sequences into a single trans gressive systems tract of a major 1 40-m-thick se quence, with the maximum flooding placed at the base of our overlying Sequence Apt l O. Aptian sequence 10 (AptlO to Aptl 1, 5 79.2-538. 0 mbsf). This 4 1-m sequence contains 1 0 major
parasequences defined by resistivity and natural gamma-ray fluctuations, which we have grouped into three sets. The top of the sequence is assigned where a pronounced peak in natural gamma-ray in tensity coincides with a 4-m-thick zone of relatively
1 22
U. Rohl & J G. Ogg
higher resistivity. This upper interval corresponds to algal-coated grains, blackened bioclasts, organic flecks and clay influx in Core 58R (Shipboard Scien tific Party, 1 993b). Arnaud et a!. ( 1 995) placed their sequence boundary 'Sb 1 8' at the same major peak in gamma-ray intensity. Aptian sequence 1 1 (Aptl l to Apt12, 538. 0-500. 0 mbsf). This 38-m sequence contains numerous high
amplitude fluctuations in resistivity and natural gamma-ray intensity. Eight of the major cycles were grouped into four sets. The cycles of packstone wackestone also contain algal laminations, organic rich clays and blackened intraclasts. We interpret the majority of this sequence to represent peritidal conditions. The upper sequence boundary 'Apt 1 2 ' i s assigned t o a simultaneous peak i n resistivity and natural gamma-ray intensity, which appears to cor respond to the upper limit of clay enrichment. At this level, a sharp decrease in resistivity and natural gamma-ray intensity corresponds to a sudden change from well-cemented organic-rich packstone to less-cemented, open-marine grainstone-pack stone (Core 54R). Alternatively, Arnaud et a!. ( 1 995) selected a position 1 2 m lower, placing their sequence boundary 'Sb 1 9' at the onset of this clay enrichment with an implicit assumption that these clays are associated with a minor transgression. Albian Aptiaf!-Albian sequence 12 (Apt12 to Alb1, 500.0442. 3 mbsf). This 58-m sequence has a similar char
acter to Sequence Apt 1 1 , with numerous fluctua tions in resistivity and natural gamma-ray intensity. The recovered lithologies include burrowed wacke stone, grainstone-rudstone with blackened 'pebble' lithoclasts, and horizons of organic-rich packstone. Only about 5% of the cored intervals were recovered. Arnaud et a!. ( 1 995) suggested that some ofthe zones of organic enrichment, blackened pebbles and other clasts may be reworking in subtidal to intertidal channels in a lagoon adjacent to marshes. Alterna tively, the array of lithological changes may repre sent cyclic depositional environments with periodic reworking. The downhole logs indicate 1 5 main cy cles or parasequences in four sets. At 442 mbsf is a bed of unusually high resistivity sharply overlain by a low-resistivity facies. This sharp-topped well-cemented bed appears to be the earliest emergence event preceding the later epi sodes of extended emergence between 430 and
4 1 5 mbsf. We assigned a minor sequence boundary 'Alb 1 ' to this level, and a major boundary 'Alb2' to the sharp upper surface of the next suite. Arnaud et a!. ( 1 99 5) also recognized a major sequence bound ary 'Sb20' in this interval, but assigned it to the base of the upper series of well-cemented beds at c. 430 mbsf. Albian sequence 1 (Alb1 to Alb2, 442.3-406. 1 mbsf).
The upper portion of this 36-m sequence appears to incorporate several episodes of emergence. A lower facies of packstone-grainstone with peloids (upper most Lithologic Unit IV) has a rapid transition to a middle to upper facies of dense mudstone wackestone with calcrete horizons and incipient calichification (lower Lithologic Unit IIIC) (Ship board Scientific Party, 1993b). The sequence con tains at least seven major parasequences, divided into two sets by a densely cemented bed (neutron porosity near zero), which probably represents an emergent horizon. Some of the parasequences in the upper set are also capped by major cementation horizons. The extremely high resistivity (5001 000 Q m) associated with these densely cemented mudstone-wackestone beds is offscale in Fig. 5. The sequence terminates in another episode of major emergence, one of the most important expo sure events within the Albian. The various lithification horizons may represent different episodes of emersion of the platform, or alternatively, some may be subsurface cementation zones produced by the variable position of a fresh water lens associated with the two or three main exposure periods. The second possibility is sug gested by the pre-lithification facies of bioturbated limestone, a typical highstand facies, in Cores 46R and 47R. Albian sequence 2 (Alb2 to Alb3, 406. 1-393. 0 mbsf).
This 1 3-m interval was interpreted to be a distinct sequence, but may also represent a 'parasequence set' superimposed on a large-scale deepening trend. Four parasequences are grouped into a lower 'retro grade' set, in which the capping bed of each succes sive cycle has lower resistivity, and a upper 'prograding' set that terminates in the sharp upper surface of a very high-resistivity peak. The upper most well-cemented horizon is interpreted to be a minor emergent event. A fragment from the upper portion of this interval has incipient calichification of the wackestone-mudstone (Shipboard Scientific Party, 1 993b).
Aptian-Albian eustatic sea-levels Albian sequence 3 (Alb3 t o Alb4, 393.0-362.8 mbsf).
Within this 30-m sequence are 10 parasequences grouped into two equal sets. As in the underlying two sequences, the parasequences are defined by variable cementation of the mudstone-wackestone facies in FMS imagery. A minor sequence boundary was placed at the sharp upper surface of a major well-cemented bed (resistivity peak). Albian sequence 4 (Alb4 to Alb5, 362. 8-341 . 2 mbsf).
This 22-m sequence terminates in major emergent episodes with formation of calcrete crusts (e.g. Fig. 1 1 in Shipboard Scientific Party, 1 993b) and associated peaks of extremely high resistivity. Four cycles within this generally well-indurated interval were grouped into two sets. The uppermost cycle may represent a lesser exposure horizon above the major event. Albian sequence 5 (Alb5 to Alb6, 341.2-31 7. 0 mbsf).
This 24-m sequence is similar to the underlying 'Sequence Alb4'. In FMS imagery, eight cycles of variable cementation of the mudstone-wackestone can be grouped into two sets. The sequence bound ary was assigned to the sharp upper surface of a major cementation horizon, which we interpret as an episode of emergence. This sequence boundary is less major than 'Alb5' or 'Alb7'. Albian sequence 6 (Alb6 to Alb7, 31 7.0-277.6 mbsf).
This 39-m sequence terminates in a major emergent episode. The uppermost level is also the top of Lithologic Unit IIIC of white, dense mudstone wackestone with repeated calcrete horizons (Ship board Scientific Party, 1 993b), and is overlain by chalky facies. Ten cycles were distinguished in this interval; these display grouping into four sets. Albian sequence 7 (Alb7 to Alb8, 277.6-226. 0 mbsf).
This 52-m sequence is dominated by wackestone with a chalky texture and high mouldic porosity (lower Lithologic Unit IIIB). Peaks of high resistiv ity (well-cemented layers) define 1 1 parasequences grouped into four sets. The lower parasequence set has a 'retrograde' aspect, in which the cap of each successive parasequence has a lesser resistivity than the previous one. The upper sequence boundary 'Alb8' is characterized by a pronounced resistivity swing to enhanced cementation related to fresh water diagenesis. An emergent episode with meteoric-phreatic diagenesis is also indicated in corresponding Core 25R, which includes a promi-
123
nent level of grainstone-rudstone with planar beach lamination, keystone vugs and thin calcrete crusts (Shipboard Scientific Party, 1 993b). Albian sequence 8 (Alb8 to Alb9, 226. 0-186. 4 mbsf).
This 40-m sequence contains peaks in resistivity (enhanced cementation) defining eight parase quences in two sets. Sequence boundary 'Alb9' was assigned to the sharp top of the highest resistivity peak. This level appears to have the most significant cementation between 260 and 90 mbsf, and there fore is considered to represent a relatively major sequence boundary. However, core recovery from the associated 1 0-m interval was only a single tiny fragment of yellow-stained wackestone. Albian sequence 9 (Alb9 to Alb 10, 186.4-138.8 mbsf).
Only a few fragments were recovered from this 48-m sequence, and these range from mudstone to peloidal grainstone. The Lithologic Unit IIIB to IliA bound ary was arbitrarily placed within this interval (Ship board Scientific Party, 1 99 3b), but is not apparent in FMS or other logs. FMS imagery and peaks in natu ral gamma-ray intensity (especially in the upper half) allow subdivision into eight main parasequences, which were grouped into four sets according to rela tively higher resistivity peaks. However, the abun dance ofhigh-amplitude peaks in natural gamma-ray intensity suggests more parasequences may be present. The high-intensity fluctuations of natural gamma-ray in the upper portion suggest an algal microbial component, but there are no unambiguous algal-mat features apparent in FMS imagery. Se quence boundary 'Alb l O' is assigned to the sharp termination of this upper facies of high-intensity gamma-ray fluctuations. Yellow staining of the frag ment of packstone-wackestone with intraclasts in Core 1 7R and manganese-oxyhydroxide staining of the piece of wackestone with dasycladacean algae and sponge spicules in Core 1 6R suggest a minor emergence or condensation event. Albian sequences 10 (Alb10 to Albll, 138.8-84.4 mbsf) and 1 1 (Albl l to Alb12, 84.4-60. 0 mbsf). The
lower two-thirds of these two sequences had negli gible recovery, but FMS imagery allowed identi fication of nine main parasequences, which are arbitrarily grouped in four sets. A narrow, but major peak in resistivity occurs at 84.4 mbsf, which was interpreted as a brief emergence and assigned as sequence boundary Alb 1 1 . The base of the drillpipe above 78 mbsf precluded FMS imagery and
1 24
U Rohl & J. G. Ogg
resistivity logging. However, a major sequence boundary 'Alb 1 2' was assigned to the 60-mbsf level based on the combination of core recovery and the sharp upper termination of a high-porosity zone ap parent on the neutron porosity log. Core 8R of Hole 866A and corresponding Core 8M of Hole 866B in dicate levels of incipient calcretization, small desic cation fissures and circumgranular cracking. We interpret this combination as an emergent episode and the formation of a well-cemented diagenetic layer above a zone of enhanced mouldic porosity. Downhole logs in Hole 867B, located a short dis tance northward of Site 866 on the 'outer rim' of Resolution Guyot, penetrated a high-resistivity layer below c. 55 mbsf (Fig. 20 in Shipboard Scientific Party, 1 993c). The top of this hole is at nearly the same depth as the top of the carbonate platform at Hole 866A, therefore the top of this zone may be equivalent to the tentative sequence boundary Alb 1 1 . The associated Cores 8R and 9R in Hole 867B contain intervals of mudstone-wackestone and rudstone-floatstone, with features interpreted as vadose-zone cement and a speleothem at a vadose phreatic boundary (Shipboard Scientific Party, 1 993c). Albian sequence 12 (Alb12 to top, 60. 0-23. 5 mbsf).
The upper 60 m of the carbonate platform at Site 866 consist of burrowed wackestone to mudstone with abundant moulds of gastropods and with dasycladacean algae and bivalves. The limited core recovery did not indicate any further episodes of emergence. The top of the carbonate platform is a mineralized submarine hardground with solution cavities infilled by phosphatic sediment penetrating to a subsurface depth of at least 1 8 m at Site 867 on the low 'outer rim' (Shipboard Scientific Party, 1993c). Microfacies and isotopic ratios within this uppermost 20-m interval (Rohl & Strasser, 1 995) could not determine whether these cavities and later cements are associated with submarine solution and diagenesis or were formed during a postulated uplift and subaerial karst episode (e.g. Winterer et a!., 1 993). In contrast to the 'incipient calcretization', desiccation features, algal-mats and other peritidal to emergent features associated with underlying se quence boundaries and surfaces of emergence, there are no definite exposure-related sedimentary or dia genetic features found within the uppermost 60 m. Also, in contrast to this uppermost zone, the lower episodes of emergence did not appear to produce significant solution cavities to such a depth below the
surface of exposure. Therefore, either the upper sur face and underlying cavities are an exceptional karstic feature or they are related to an extended interval of submarine hardground formation and penetrative partial dissolution. Takuyo-Daisan Guyot
Aptian Aptian sequence 4 (partial) (base ofplatform to Apt5, 164. 4- 152. 7 mbsf). The earliest sediments at
Takuyo-Daisan Guyot consist of clayey sands to claystone with relict volcanic breccia texture over lying the deeply weathered basalt (Shipboard Scien tific Party, 1 993e; Ogg et a!. , 1 995b). The onlapping carbonate sediments contain an early-Late Aptian calcareous-nanofossil assemblage (middle of Par habdolithus angustus Nanofossil Zone, NC7; Cores 1 7R and 1 8R) and a possible Late Aptian plank tonic foraminifer assemblage (Cores 1 6R to 1 8R) (Shipboard Scientific Party, 1 99 3e; Erba, 1 99 5). The transgressive onlap occurred with a low-energy environment, with organic-rich clayey sands of the weathered volcanic edifice overlain by a lagoonal facies of bioturbated, gastropod-rich peloid wackestone-packstone with intervals containing coated grains or intraclasts. No significant storm or wave reworking was observed. The palaeogeogra phy appears to be a coastal marsh adjacent to a lagoon that is semi-protected by an offshore carbon ate shoal. The change to a carbonate facies may represent the gradual encroachment of the lagoon onto the subsiding island, rather than an association with a transgressive systems tract (Ogg et a!. , 1 995b), although the projected timing is equivalent to the deepening that followed sequence boundary 'Apt4' on Resolution Guyot. The lowest 39 m of the carbonate platform con sists of 1 2 major upward-shallowing parasequences. Each of the first four parasequences terminate in a progradation of volcaniclastic clay and sand from the island (Lithologic Subunit I C). The highest level of prominent clay and associated thorium enrich ment is at 1 52.7 mbsf. The next two parasequences have a more subdued character, suggesting deposi tion in a relatively deeper setting. The combination of this deepening in facies and the apparent flooding of the source of the clastic influx indicates the onset of a transgressive pulse; therefore, we assign · a minor sequence boundary 'Apt5' to the top of the highest terrigenous influx.
Aptian-Albian eustatic sea-levels
1 25
Aptian sequences 5 and 6 (Apt5 toApt7, 152. 7-125.6 mbsf). This 20-m interval contains eight main
Aptian sequences 7 and 8, and 9 (Apt 7 to AptlO, 125. 6-88. 0 mbsf). The lower half of this 38-m se
parasequences in three sets (Lithologic Subunit IC). Each parasequence displays an upward trend of in creasing resistivity coupled with increasing uranium enrichment, with a sharp termination to the low resistivity, low-gamma base of the next parase quence. Each parasequence set ends in a more pro nounced resistivity peak with a sharp upper surface. Core recovery and FMS imagery indicates that the uranium enrichment cycles which characterize the lower five parasequences are probably associated with oncolites, because there are no indications of a peritidal algal-mat facies or clay influx. The presence of fenestral features in core recovery indicates epi sodes of peritidal to emergent conditions, and the tops of the upper two parasequence sets have en hanced cementation that is usually associated with exposure. Above the highest surface of emergence at 1 2 5.6 mbsf is the initiation of a sustained deepening trend to a quiet lagoon facies. We interpret each of these three parasequence sets to be an individual depositional sequence, and assign sequence boundaries to the sharp upper surface of each group. A major sequence boundary 'Apt7' is assigned to the highest major surface of emergence at 1 25.6 mbsf. The assignment 'Apt7' is based on a proposed correlation to a similar major sequence boundary at Resolution Guyot. The un derlying sequence boundary at 1 42.9 mbsf is a relatively less major event, and we assign this level to minor sequence boundary 'Apt6' of Resolution Guyot. The overlying major surface of exposure at 1 2 5.6 mbsf is considered equivalent to major se quence boundary 'Apt8' of Resolution Guyot. The interval from 'Apt4' to 'Apt8' on Resolution Guyot exhibits well-developed parasequences and parasequence sets, similar to this interval on Takuyo-Daisan Guyot. Aptian Sequence 'Apt7' is very thin (only 7 m at Takuyo-Daisan Guyot), which suggests several possibilities: (i) there was significant erosion during the emergence to form the sequence boundary 'Apt8' which removed much of the earlier depositional sequence, (ii) the short Sequence 'Apt7' is superimposed on a longer-term major eustatic sea-level fall, so that the rise in sea-level during transgressive and highstand stages after the basal sequence boundary was only suffi cient to drown the carbonate platform for a brief time, or (iii) the upper 'sequence' is a condensed parasequence set developed during the late high stand to early lowstand phase of a main sequence.
quence is a major progressive deepening trend to a poorly cemented foraminifer-gastropod wackestone deposited in a quiet lagoon. The initial stages of deepening are developed in a 'give-up' parasequence set, which displays a sharp upper surface to a rela tively well-cemented bed at 1 20.0 mbsf. We consider this level to be equivalent to minor sequence bound ary 'Apt9' at Resolution Guyot, which displays a similar superposition on a general major deepening trend that was initiated after sequence boundary 'Apt8'. Parasequences within the 'maximum flooding' interval are mainly indicated by cyclic variations in porosity, rather than resistivity. At 1 02 mbsf begins a rapid shallowing to an assemblage of shallow water facies that includes bioclastic rudstone, rede posited skeletal grainstone with coral fragments, and wackestone with fenestral structures. This up per 1 4-m interval also has episodes of relative uranium enrichment indicating possible algal microbial activity (Ogg et al. , 1 99 5b ). Some of the four main parasequences within this upper interval terminate in a sharp upper surface, but there is no evidence for periods of significant emergence. A minor sequence boundary 'Apt 1 0' is assigned to the uppermost resistivity peak (and porosity low), which is overlain by less-cemented foraminiferal packstone of a relatively deeper, quiet lagoonal environment. Aptian sequences 10 'and 1 1 (AptJO to Apt12, 8835 mbsf). The majority of this interval consists of
lagoonal facies, but a widened borehole and lack of core recovery from the lower 30 m inhibit identifi cation of facies trends. Resistivity and porosity logs indicate zones characterized by cyclic variability, which probably represent parasequences. The up permost 4 m of the interval is a low-porosity oncolite-rich facies according to core recovery and natural gamma-ray intensity logs, but this zone and higher strata did not have a resistivity log because of the suspended drill-pipe. A minor sequence bound ary 'Apt l 2' was assigned to the top of this oncolite rich zone. Another minor sequence boundary may be within the middle of this interval. Possible levels include sharp tops of low-porosity (well cemented?) layers at 6 1 or 52 mbsf, or a sharp resistivity decline (flooding surface?) at 69 or 58 mbsf. We prefer a tentative placement at the major porosity contrast
1 26
U. Rohl & J G. Ogg
at 6 1 mbsf, and suggest a correspondence to the minor 'Apt l l ' at Resolution Guyot. Aptian-Albian sequence 12 (Apt12 to Alb1, 3522 mbsf). The upper portion of the carbonate plat
form is dominated by interbedding of wave reworked oyster-algal-coral-mollusc floatstone and of peloid grainstone to skeletal packstone. Major in tervals of contrasting porosity are probably associ ated with these two facies. At 20-22 mbsf there is a sharp-topped layer of very low porosity, which we interpret as a well-cemented horizon and possible surface of exposure. Sequence boundary 'Alb 1 ' is assigned to the top of this layer. Albian Albian sequence 1 (Albl to Alb2, 22-6 mbsf). This
1 6-m interval contains two alternations of high to low porosity, each with several internal cycles. These resemble the parasequence sets of lower shallow water sequences, and the associated core recovery indicates storm debris accumulations of bioclastic floatstone. Site 879 is on a low outer rim to the carbonate platform. Therefore, we interpret this suc cession as episodes of barrier-shoal deposition and construction of this outer rim. An implication is that there was backstepping of the barrier-shoal complex over the underlying lagoonal deposits of the upper Aptian succession. A major sequence boundary 'Alb2' is tentatively assigned to the sharp upper surface of a low porosity zone, which we interpret as enhanced cementation on an emergent bioclastic shoal. The core recovery indicates that this interval is the highest occurrence of coarse bioclastic debris. Albian sequence 2 (partial) (Alb2 to top of carbonate platform, 6-0 mbsf). Coring of the uppermost 6-m
of the platform succession recovered a gastropod bivalve floatstone to coquina with a matrix of peloidal packstone. The uppermost layers do not appear to have undergone any unusual cementation or secondary porosity, and the upper surface is encrusted by a silica- and phosphate-rich hard ground. There is no indication of an emergence before drowning of the guyot, therefore we interpret this facies to be a 'give-up' transgressive facies of wave- and current-winnowed deeper-water carbon ates deposited before the guyot surface was sub merged below the 20-m effective depth limit of
active carbonate production. A similar 'give-up' terminal facies is observed at MIT and Allison guyots. There is no pelagic sediment at this site; the surface remains swept by currents. The final stages of Takuyo-Daisan Guyot con sisted of an emergence during major Albian se quence boundary 'Alb2' followed by a major deepening. On MIT, Resolution and Allison guyots, the 'Alb2' transgressive flooding terminated shallow-water parasequences, but was succeeded by a deep lagoonal facies. However, the same event terminated the slower-subsiding Takuyo-Daisan carbonate bank. A possible explanation is that this carbonate platform was already stressed by other environmental conditions at its approach to the palaeo-equator, such as water temperature fluctua tions, precipitation-evaporation impact on salinity, nutrient availability or other factors, which hin dered production of sufficient carbonate sediment to keep pace with this extended sea-level rise. MIT Guyot
Aptian Aptian sequence 1 (partial) ('Aptl ' to Apt2, 727. 5699. 7 mbsf). In contrast to the lower carbonate
facies at Allison and Resolution guyots, the initial flooding (727.5 mbsf) of the weathered volcanic edifice of MIT Guyot terminated any further influx of terrigenous clay or organic material. The lowest sequence, 28-m thick, contains six upward-shal lowing parasequences in three sets, and encom passes the varied lithologies of Lithologic Subunit VB. The lower two parasequences in skeletal-oolitic grainstone with algal rhodoliths terminate in emer gent surfaces associated with yellow to red staining, enhanced cementation and redox concentration of uranium. The middle set represents a further deep ening into an algal-rich facies with storm beds of bioclastic grainstone-rudstone followed by a shal low-lagoonal facies of gastropod-peloid wacke stone. The upper set has a shallowing trend indicated by pronounced cementation of the upper portions of parasequences and an increased contri bution from algal-microbial colonies. A major se quence boundary 'Apt2' is assigned to the sharp upper surface of a well-cemented layer, which we interpret as enhanced cementation associated with a brief emergence before rapid deepening into a quiet-water lagoonal setting.
Aptian-Albian eustatic sea-levels Aptian sequences 2 and 3 (Apt2 to Apt4, 699. 7-663. 0 mbsf). The lower portion of this 3 7-m sequence is in
a relatively homogeneous protected-lagoon facies of poorly cemented, gastropod-peloid wackestone, which does not display significant development of parasequences. The upper portion displays higher amplitude fluctuations of resistivity and an increase in natural gamma-ray intensity. FMS imagery and core recovery within this upper portion indicates parasequences developed in a facies of algal-coral horizons, winnowed grainstone and fenestral wacke stone. At least six major parasequences are present. Sequence boundary 'Apt4' was assigned to the sharp top of the thickest (c. I m) well-cemented bed, but this is probably a minor event or may be a termina tion of a parasequence set. Aptian sequence 4 (Apt4 to Apt5, 663. 0-623.5 mbsf).
This 39.5-m sequence is dominated by resistivity cycles developed in a restricted, algal-rich shallow lagoon facies. Core recovery indicates that an upward increase in mud-supported lithologies ac companies the increase in amplitude of resistivity cycles. Approximately seven main parasequences were grouped into two sets. The trend culminates at 623.5 mbsf where the higher of two major high resistivity peaks is sharply overlain by a more open-lagoon facies. Yellowish staining in a nerineid-gastropod rudstone recovered just below this level (Core 68R) suggests one or both of these well-cemented beds are associated with emergence. Aptian sequence 5 (partial) (Apt5 to base of volcani clastic breccia, 623.5-602.5 mbsf). The final 2 1 m of
the lower Aptian carbonate platform at Site 878 represent an open-marine lagoon with wave winnowed peloid grainstone, storm beds and some algal encrustations. Three parasequences may be present, but they do not indicate any shallowing trend before eruption of the volcanic rocks. In the middle of the Aptian, MIT Guyot was devastated by two major explosive volcanic erup tions, which transformed a large portion of the carbonate platform into enormous quantities of ejecta of intermixed carbonate and volcanic clasts. The interior platform setting at Site 878 was buried beneath 200 m of steaming volcanic-limestone breccia derived from these volcanic blasts. After these two main episodes, there were at least three secondary ash-rich eruptions. The volcanic breccia blanket transformed this portion of MIT Guyot into
1 27
a large emergent island. The island underwent subaerial weathering before continued subsidence of the edifice renewed marine conditions during the late Aptian. Albian Upper-Aptian-lowermost Albian sequence group (volcaniclastic breccia to Alb2, 405. 4-342. 4 mbsf).
The resubmergence of MIT Guyot was a progres sion from a thin layer of organic-rich clay overlain by few beds of wave-winnowed carbonate sands, followed by a lime-mud-rich lagoonal facies within thin cyclic variations in resistivity. This mud-rich facies is sharply overlain at 392 mbsf by an 80-m interval dominated by homogeneous peloid grain stone. Only a 20-m interval in the middle displays possible cyclicity in the downhole logs; otherwise the lack of core recovery and homogeneous log response do not allow resolution of parasequences or sequences. A minor sequence boundary 'Alb2' is assigned to 342.4 mbsf, where a sharp termination of a rise in resistivity appears to coincide with the presence of Orbitolina benthic foraminifers in the peloid grainstone. This level may represent a brief shoaling followed by a temporary lessening of ma rine restriction. The resistivity logs suggest that three parasequences may precede this episode, which is consistent with a shallowing trend before a sequence boundary. The numbering as 'Alb2' is based on its relative position with respect to the spacing of numbered sequence boundaries on Res olution and Allison guyots. Albian sequences 2-4 (Alb2 to Alb4, 342. 4-306. 0 mbsf). The majority of this sequence is homoge
neous peloid grainstone lacking an unambiguous ex pression of parasequences. In the uppermost portion is an algal-sponge bioherm with algal rhodoliths within the restricted lagoon facies of peloid grain stone (Core 3 3M). This episode, which marked the Lithologic Subunit IliA to IIIB boundary (Shipboard Scientific Party, 1 993d) corresponds to an increase and associated vague banding in resistivity and FMS imagery. The sharp termination of this 6-m thick unit at 306.0 mbsf is followed by a lagoonal facies of peloid-foraminifer grainstone. A minor sequence boundary 'Alb4' is assigned to this level, with the numbering based on correlation of the next higher and major sequence boundary to 'Alb5' of Resolution and Allison guyots. Alternatively, this bioherm
·
1 28
U. Rohl & J. G. Ogg
may be a local development within the lagoon facies and independent of eustatic sea-level changes (Ogg, 1 995). Albian sequence 4 (Alb4 to Alb5, 306. 0-283. 7 mbsf).
This sequence is dominated by lagoonal peloid grainstone lacking expression of parasequences. At 285.7 mbsf begins a 2-m-thick, well-cemented, bio clastic bed ofabundant bivalves, possible rudists and gastropods. The very high resistivity and sharp up per surface of this bed suggest that it may have un dergone enhanced cementation associated with emergence. The significance of this cementation ho rizon and its position relative to other major emer gence levels suggest a correlation to major sequence boundary 'Alb5' recorded on Resolution and Allison guyots. Albian sequence 5 (Alb5 to Alb6, 283. 7-263.5 mbsf).
The lower half of this 20-m depositional sequence is a lagoonal facies of foraminifer-peloid grainstone. A shallowing trend in the upper half is suggested by features in FMS imagery interpreted as an increased abundance of beds of storm-redeposited and win nowed bioclastic debris. A minor sequence bound ary 'Alb6' is assigned to the sharp termination of a set of such beds. The higher resistivity of this set is interpreted as a slightly enhanced cementation, but there is no indication of an emergence. The charac ter of downhole logs (low resistivity, low gamma) and FMS imagery suggests a minor deepening above this surface. Albian sequence 6 (Alb6 to Alb7, 263.5-243.5 mbsf).
After a basal deepening, this 20-m sequence con sists of a major shallowing-upward trend. Algal oncolite facies with rudists are overlain by a 1 2-m thick unit interpreted from FMS imagery and core recovery as a accumulation of storm-reworked de bris. The uppermost metres of this deposit have undergone both enhanced cementation and local leaching during an episode of subaerial exposure. The sharp upper surface of this island episode is a major sequence boundary 'Alb7'. Albian sequence 7 (Alb7 to Alb8, 243. 5-1 96. 7 mbsf).
This 4 7 -m sequence consists of approximately 1 3 major parasequences (Lithologic Subunit liD). The upward-shallowing parasequences are developed in shallow lagoon to intertidal facies, ranging from gastropod wackestone to algal-rich grainstone and storm-debris beds. The cyclicity is apparent in
resistivity and FMS imagery, and to a lesser degree in natural gamma-ray intensity. Relative peaks in resistivity suggest subdivision into three parase quence sets. Sequence boundary 'Alb8' is assigned to the termination of these parasequences. Albian sequence 8 (Alb8 to Alb9, 1 96. 7-138. 6 mbsf).
This 58-m sequence in lagoonal facies may have a minor intermediate shallowing and deepening epi sode in the middle ( 1 66.7 mbsf). The majority of this 58-m-thick sequence is a quiet to restricted lagoonal facies of peloid packstone to gastropod wackestone with variable concentrations of storm beds. A concentration of storm beds overlain by a relatively protected (deeper?) lagoon facies at 1 66. 7 mbsf may indicate a minor shallowing and deepening episode in the middle of the main se quence. The sequence ends in a 3-m-thick zone of enhanced cementation associated with the greatest resistivity horizon within the entire carbonate plat form. The sharp upper surface of this well cemented interval is interpreted as an exposure surface developed on an island of storm-reworked debris. Albian sequence 9 (Alb9 toAlblO, 138. 6-109.3 mbsf).
The lagoonal facies overlying the basal flooding sur face consists of peloid-rich wackestone with storm beds and possible algal-layered intervals. Resistivity fluctuations within this 29-m-thick sequence appear to be caused both by shallowing-upward parase quences and by storm-bed layers. Two upward shallowing algal-rich parasequences are in the upper most 6 m of the sequence, and a minor sequence boundary 'Alb l O' was assigned to the sharp transi tion to a low-resistivity lime-mud-rich facies. Albian sequence 10 (Alb1 0 to Albll, 1 09.3-69. 9 mbsf). The lower half of this 40-m sequence is a
wackestone-packstone lagoonal facies with storm beds. The upper half consists of three major upward shallowing parasequences in a peritidal setting in cluding episodes of algal-flat development, storm re deposition and fenestral structures. The sharp topped, high-resisitivity cemented caps of the upper two parasequences indicate brief episodes of emer gence, and the sequence boundary 'Alb 1 1 ' is as signed to the highest and most major surface of ex posure. Albian sequence 1 1 (Albll toAlb12, 69. 9-1 9. 0 mbsf).
The lower third of this 5 1-m sequence is a progres-
Aptian-Albian eustatic sea-levels
sive deepening from peritidal parasequences into an algal-rich lagoonal facies. The middle third is a la goonal facies with variable abundances of storm beds and lime mud. The upper third is a rapid shal lowing to a lagoonal setting dominated by storm beds with minor interbeds of wackestone-packstone with gastropods. The FMS imagery and resistivity suggest approximately eight major upward shallowing parasequences within the sequence, but with no apparent grouping into sets. A well cemented layer at 30 mbsf and double-layer cemen tation feature at 20 mbsfwere apparently not recov ered during coring, but FMS imagery suggests these levels may represent brief exposure horizons. Se quence boundary 'Alb l 2' was assigned to the top of the uppermost recorded interval of exceptionally high resistivity. Albian 'sequence ' 12 (Alb12 to top of carbonate platform, 1 9. 0-3. 0 mbsf). The uppermost possible
surface of emergence is overlain by mudstone wackestone with gastropods and cyanobacteria algal concretions called Orthonella. The suspended drill-pipe at 1 7 mbsf precluded acquisition of direct downhole logs, but an approximation to cementa tion of carbonates outside the drill-pipe was ac quired by a calibration of the calcium channel of the geochemical logging tool. The relative calcium con centrations indicate that the cemented horizon at 1 9 mbsf is overlain by a series of three lesser peaks (parasequences?) in relative cementation, with the uppermost layer at the top of the carbonate plat form. Drilling penetration rates support this inter pretation, and indicate that the uppermost 2 m of the carbonate platform is a relatively well-cemented bed, with a sharp surface at 3 mbsf to a thin covering of pelagic chalk. Our interpretation of the combined core recovery and log proxies is a la goonal facies with variable storm-redeposited de bris, similar to the underlying 'Sequence Alb l 2'. The uppermost core recovery does not display any unambiguous indication of shallowing or emer gence, therefore this uppermost 1 6-m interval is considered to be a 'catch-up' to 'give-up' facies and submergence with current-storm reworking epi sodes during the initial phases of a large-scale deepening trend. The cementation of the upper most 2 m may be related to a penetrative submarine hardground. The last significant emergent episode was at the 'Alb l 2' base of this sequence. The final succession is similar to what is postulated for Allison Guyot.
1 29
The development of the irregular surface mor phology of MIT Guyot may be related to the periodic exposures associated with sequence bound aries 'Alb 1 1 ' and 'Alb 1 2' causing a succession of superimposed karstic features and/or to prolonged submarine dissolution during the past 90 Myr. Al ternatively, erosion at this site may have removed all lithological and diagenetic evidence of a final stage uplift, extensive karst development and suc ceeding drowning (e.g. Van Waasbergen & Winterer, 1 993; Winterer et a!., 1 993). Allison Guyot
The boundaries and internal structure of the lowest three sequences will be discussed in detail to illus trate the interpretation procedure. Summaries of major features will be given for the overlying sequences. Aptian-Albian sequence 12 (partial) (base ofcarbon ate platform to Alb1, 839.0- 793. 5 mbsf). This
45.5-m sequence is dominated by bioturbated and organic-rich clayey wackestone to packstones, inter preted to represent a lagoonal depositional environ ment adjacent to a marshy shoreline of the weathered volcanic edifice. Within this basal se quence, an overall upward trend of decreasing clay and organic matter influx is observed in the recov ered lithologies and logs. This general slowing of clay influx is interpreted to represent a progressive retreat of the shoreline as the volcanic edifice steadily subsided. Superimposed on this general pattern of a retreating shoreline are periods of temporary encroachment of the clay-organic influx toward the site, which contribute to the character of the parasequences and sets of parasequences. In this respect, the lowest sequence is a dialogue of the land and sea influenced by small fluctuations in eustatic sea-level. In contrast to higher sequences, where cementation and resistivity variations are most important, the fine-scale structure of this basal sequence is primarily indicated by clay and associ ated natural gamma-ray intensity fluctuations. The base of the sequence is placed where marine carbonates overlie clayey oyster-rich limestone with basaltic sills (Lithologic Unit IVD). This flooding surface may be a local response to rapid subsidence, rather than an association with a transgressive stage of eustatic sea-level. The sequence contains 1 3 shallowing-upward parasequences grouped in four parasequence sets.
1 30
U Rohl & J G. Ogg
The average parasequence identified in the logs and FMS images is about 3 . 5 m. These shallowing upward cycles were also observed in core recovery, but there were also some shallowing-upward beds as thin as 0.5 m (e.g. in Core 89R, Section 5). In contrast to the overall trend of decreasing clay within the sequence, each individual parasequence typically has clayey bioturbated limestones at the base, followed by an upward increase in clay and organic matter. This parasequence succession is interpreted as the progressive shallowing of the fringing carbonate complex, thereby allowing in creased progradation of terrigenous clastics from the volcanic island. Within a parasequence set, each progressively higher parasequence contains a greater content of clay. In each of the lowest three parasequence sets, this upward increase of clay reaches a peak in the uppermost metres with con centrations of both thorium and uranium and associated increase in natural gamma-ray intensity. In this regard, the main trend within each parase quence set mirrors the increasing-upward clay in flux within the component parasequences. The lower two parasequence sets are thicker than the upper two sets, which is consistent with the trans gressive to early highstand depositional systems tracts of the main sequence. The fourth and highest parasequence set displays a lesser terrigenous influx (less clay, and lower natural gamma-ray intensity) and an increase in open marine influence (many small oyster fragments, Orbitulinids and planktonic foraminifers in the packstones). The upper boundary ('flooding surface') of the sequence is positioned at an abrupt contact of clay-rich packstones overlain by clay-poor lime stones (observed within Core 85R). This deepening event is associated with a sharp reduction in both resistivity and natural gamma-ray intensity at 793.5 mbsf. It is possible that the onset of increased rate of accommodation space occurred within the preceding parasequence set where the character of parasequences changes (i.e. at about 801 mbsf), and the main flooding event is at 793.5 mbsf. Therefore, the sequence boundary, as defined by the minimum creation rate of accommodation space, may be slightly below this flooding surface. Albian sequence 1 (Albl to Alb2, 793. 5-727.5 mbsf).
This 66.0-m sequence is dominated by packstone and wackestone with bioturbation and minor amounts of ostracods, benthic foraminifers, clay, organic matter and pyrite. It is interpreted as a
fringing lagoon offshore from a central volcanic island and partially protected from wave action by an outer barrier shoal. An overall decrease in clay and organic-matter content compared with the un derlying sequence is indicated by the lighter color ation of the limestone and a lower intensity of thorium-produced natural gamma-rays. The base of the sequence is a 1 2-m-thick interval of very low resistivity and low natural gamma-ray intensity, which we interpret as a poorly cemented packstone associated with a major deepening event. The main body of the sequence contains 1 5 parase quences. The average parasequence cycle is approx imately 4.5 m thick. The top of each parasequence was assigned to a high-resistivity layer, interpreted as an enhanced cementation horizon associated with the shallowest depth of deposition or minor emergent episode. Narrow peaks in natural gamma ray intensity caused by uranium concentrations generally coincide with these high-resistivity layers. Thorium concentrations are of declining impor tance, therefore most of these uranium concentra tions are probably associated with organic matter, redox diagenetic horizons or microbial-algal layers. These 1 5 parasequences were grouped into five sets according to progressions of increasing peaks of both resistivity and natural gamma-ray intensity. These parasequence sets average 1 3 m in thickness. However, the lower two sets, which are character ized by an upward increase in background natural gamma-ray intensity and resistivity, are thicker than the upper three sets, therefore we interpret this lower interval as the transgressive systems tract. The top parasequence in the upper three sets is capped by a very high resistivity layer, which we interpret as enhanced cementation during an emer gent episode. Partial dolomitization and the pres ence of black pebbles observed in some core fragments are probably associated with such emer gent episodes. The upper boundary of this sequence is placed at the termination of these high-amplitude parase quences, which we interpret as the uppermost epi sode of emergence. At this level, the well-cemented cap of the highest major parasequence is sharply overlain by an extended interval of very low resis tivity. Planktonic foraminifers recovered in Core 79R may be associated with this flooding event. This major change in lithological and parasequence character represents one of the major sea-level rises at this site ('Alb2').
Aptian-Albian eustatic sea-levels Albian sequence 2 (Alb2 to Alb3, 72 7. 5-6 72. 2 mbsf).
This 5 5 . 3-m sequence is probably largely equivalent to Lithologic Subunit IVB (Shipboard Scientific Party, 1 993a), which is distinguished from the un derlying Subunit IVC by the absence of terrigenous clay and carbonaceous fragments. This depositional sequence contains patches of dolomitization. The parasequences making up this sequence and higher sequences are more subdued relative to the ex tremely high-amplitude fluctuations of resistivity and natural gamma-ray intensity in Sequence Alb 1 . However, several parasequences terminate in sharp topped layers of relatively high resistivity, and the presence of black pebbles and desiccation cracks (Cores 74R and 75R) indicates brief emergent epi sodes. We tentatively identify 1 1 parasequences, with a possible grouping into four sets defined by those parasequences with relatively high peaks of resistiv ity. The average parasequence thickness is 5 m, the average parasequence set is 1 4 m. In contrast to the underlying and overlying se quences, Sequence Alb2 does not display a distinct trend of pronounced upward-shallowing. Therefore, the upper boundary was assigned as the top of a high-resistivity cap of a parasequence set at 672.7 m, based on lithological evidence of a deepening near Core 72. The overlying Subunit IVA begins with gastropod wackestone with no dolomitization. This upper sequence boundary is a minor event lacking significant emergence. Albian sequence 3 (Alb3 to Alb4, 6 72. 2-626. 0 mbsf).
This 46.2-m sequence coincides with Lithologic Sub unit IVA of wackestone-packstone without partial dolomitization. However, the lithological succession is uncertain, because only scattered fragments were recovered during drilling. The parasequence signa ture is very similar to that of the underlying Se quence Alb2, and 1 0 parasequences were grouped into four sets. The parasequences average 5 m in thickness; the sets are about 1 2- 1 3 m thick. This is the highest sequence containing a minor clay com ponent. A peak in both resistivity and natural gamma-ray intensity marks the top ofthis sequence. The overlying flooding surface Alb4 coincides with the base of Lithologic Unit III. Albian sequence 4 (Alb4 to Alb5, 626. 0-590. 0 mbsf).
This 36 .0-m sequence contains six parasequences of rhythmic fluctuations in resistivity. Core recovery was very limited (less than 3%), but indicates a
131
quiet lagoonal environment dominated by wacke stone with dasycladacean algae, ostracods, sponge spicules, sponges, some benthic foraminifers and gastropods. A sequence boundary was assigned to '590.0 mbsf' based upon recovery of red-stained grainstone in Core 1 43-865A-64R. The individually selected parasequences average 6 m in thickness, whereas Cooper et a!. ( 1 995) obtained a 7.5-m cyclicity using spectral analysis. The parasequences were grouped into two sets, although this division is not as well defined as in lower sequences. Albian sequence 5 (Alb5 to Alb6, 590. 0-555. 0 mbsf).
This 3 5 .0-m sequence culminates in one of the highest resistivity peaks within the hole. We inter pret the uppermost peak of resistivity as a promi nent emersion event with associated intensive leaching, reddish to yellowish staining and cemen tation of limestones in a freshwater environment (Shipboard Scientific Party, 1 993a; see R6hl & Strasser, 1 995). Parasequences are as in Albian sequence 4; they were grouped into two sets. This sequence boundary Alb6 is followed by a major deepening event and deposition of mudstones. This level marks one of the major eustatic events during the Albian. Albian sequence 6 (Alb6 to Alb7, 555. 0-491. 6 mbsf).
This 63.4-m sequence is predominantly a facies of very low resistivity and relatively low natural gamma-ray intensity. The basal parasequence ends in a layer of relatively high resistivity, which may correspond to a shallowing or brief emergence dur ing the initial stages of deepening. The limited core recovery within this sequence is consistent with a thick interval of poorly cemented wackestone mudstone deposited in a relatively deep, quiet water environment. The sequence was subdivided into seven parasequences by low-amplitude fluctuations in resistivity. These parasequences have a vague grouping into three sets. The lack of distinct cyclicity is consistent with a sustained deep-water setting. Sequence boundary Alb6 is placed where sedi mentary or diagenetic cycles become difficult to recognize on the FMS imagery, and we interpret this change to indicate a further deepening episode. A sequence boundary, if present, is probably repre sented at this site by only a minor slowing of the rate of accommodation space formation. Albian sequence 7 (Alb7 to Alb8, 491. 6-448.4 mbsf).
There is no absolute resistivity log for the interval
1 32
U Rohl & J. G. Ogg
between 506 and 4 1 0 mbsf. For display purposes, the FMS grey-scale imagery of resistivity over this span was rescaled such that the amplitude fluctua tions are consistent with the actual values of resis tivity in adjacent intervals. However, as for all intervals, the parasequence and sequence boundary interpretations are based on the high-resolution FMS imagery and natural gamma-ray intensity fluctuations. Sequence Alb7 is 43.2 m thick and contains four major parasequences grouped in two sets. The general character of the facies and parasequence signature are similar to those of the adjacent se quences Alb6 and Alb8. Natural gamma-ray intensities progressively in crease in the upper 1 5 m, then undergo a sharp drop to background levels. If this trend in natural gamma is indicative of uranium concentrations associated with microbial-algal redox conditions (a common signature at other sites), then this interval may correspond to a shallowing-upward succession pre ceding a minor-scale deepening event. Within this interval, FMS resistivity and neutron porosity show a high-frequency cyclicity in the degree of cementa tion of the mudstones to wackestones, which also suggests an algal-mat facies. Yellowish staining and 'chalky' texture on wackestone fragments in Cores 50R and 5 1 R suggest post-depositional meteoric phreatic influence. We tentatively assign a sequence boundary to the top of the major peak in natural gamma-ray intensity. Albian sequence 8 (Alb8 to Alb9, 448.4-365. 0 mbsf).
This 83.4-m sequence has 1 3 main parasequences defined by changes in cementation that are grouped into three sets. Core recovery was very low within Sequence Alb8, but indicates continuation of re stricted lagoon facies of wackestone-mudstone with mouldic porosity, sponge spicules, dasycladacean algae and foraminifers. A sharp-topped resisitivity peak at 365.0 mbsf represents the highest degree of cementation within the sequence. This level appears to coincide with a rapid change observed in core recovery from wackestone-mudstone (Lithologic Unit IIIB) to a more open-marine facies of alternating grainstone to wackestone (Lithologic Unit IliA). A minor sequence boundary Alb8 was assigned to the sharp top of this well-cemented layer, followed by an interpreted deepening. Albian sequence 9 (Alb9 toAlb10, 365. 0-293.5 mbsf).
This 7 1 . 5-m sequence contains seven major cycles
grouped into two sets. This lower part of Lithologic Unit IliA contains beds of grainstone within a facies ofwackestone-packstone with gastropods and rudist fraginents. The grainstone beds are interpreted to represent episodic storm activity winnowing the la goonal bottom and redepositing coarser grains. FMS imagery also displays sharply defined higher resistivity beds of variable thickness. Shallowing upward parasequences were interpreted according to the apparent relative proximity of the storm depos its, assuming that packets of thicker beds represent shallower depths. No storm deposits were recog nized in the 1 0-m-thick, low-resistivity, low-gamma ray intensity zone at 3 1 8-308 mbsf. This 'storm free' interval is overlain by a 1 5-m-thick interval of high resistivity and associated high-gamma-ray in tensity. We interpret this cemented interval as a shallow-water deposit, and assign sequence bound ary Alb9 to the sharp top of the highest-resistivity bed. This level appears to be a major episode of emergence or condensed shallow-water deposition. Albian sequence 10 (Alb1 0 to Alb1 1, 293. 5-239. 7 mbsf). This 5 3. 8-m sequence was included in the
same Lithologic Unit IliA as the underlying se quence (Shipboard Scientific Party, 1 993a), but the FMS and resistivity signature indicates that storm redeposition is of lesser importance. Eight parase quences, averaging 6. 7 m in thickness, are defined by cyclic variations in induration and are grouped into three sets. The top of Sequence Alb I 0 is a 5-m-thick, well lithified, requiniid-rudist limestone with minor yel lowish staining, which exhibits the most prominent resistivity peak within the platform succession. We interpret the intensive cementation event as the result of freshwater diagenesis during a major emer gence episode. It is sharply overlain by the low resistivity facies of typical lagoonal wackestone. Albian sequence 1 1 (Albl l to Alb12, 239. 7- 180.4 mbsf). This 59.3-m sequence is characterized by an
overall upward increase in resistivity. The upper half of the sequence (above 2 1 0 mbsf) is marked by a series ofhigh-intensity peaks in natural gamma rays, most of which are mirrored by peaks in resistivity. The initiation of the suite of high gamma-ray peaks was used to define Lithologic Unit II, which was interpreted by the Shipboard Scientific Party as phosphatization-induced concentrations of ura nium. However, later analysis of the recovered facies failed to indicate any phosphate enrichment (Rohl & Strasser, 1 995), and the FMS imagery indicates an
1 33
Aptian-Albian eustatic sea-levels
upward change from relatively homogeneous wacke stone to a more laminated facies. Therefore, we in terpret this shift as a shallowing from lagoonal wackestone to a suite of subtidal-peritidal cycles with algal-microbial laminations and associated re dox scavenging of uranium and enhanced cementa tion. Three minor peaks in thorium are the only significant thorium concentrations within the upper 400 m ofthe carbonate platform, and are interpreted to be possible volcanic ash layers from late Albian volcanism elsewhere in the mid-Pacific seamount province. Parasequences were difficult to interpret within the lower lagoonal wackestone, but are well defined in the upper subtidal-peritidal succession. The parasequences are grouped into three sets, of which the upper and lower set contain three to four parasequences and the middle set is less distinct. Sequence boundary Alb 1 1 is assigned to the sharp top of a 5-m-thick well-cemented bed, which is overlain by a brief return to low-resistivity la goonal wackestone. An episode of emergence is supported by the yellowish staining on the few fragments recovered from the upper portion of this sequence (Core 22R). Albian sequence 12 (Albl 2 to top, 180.4-139. 7 mbsf).
The middle to upper portion of this 40. 7-m sequence is characterized by parasequences defined by sharp peaks in resistivity and natural gamma-ray intensity, and is interpreted as a series of upward-shallowing subtidal to peritidal cycles. The seven major parase quences are grouped into two sets, and the lower 'lagoonal' portion of the sequence appears to consti tute a third set of less well-defined parasequences. The high degree of cementation of the capping beds ofparasequences within the middle set is interpreted as brief episodes of emergence. This sequence terminates in the drowning of the guyot. Parasequences within the upper set have a lesser degree of cementation than the lower set, which suggests that a slow deepening may have already been initiated during the final stages of platform accumulation. The irregular surface morphology of the carbon ate platform on Allison Guyot may indicate forma tion of karst relief during an extended interval of emergence (Van Waasbergen & Winterer, 1 993; Winterer et a!. , 1 993; Winterer & Van Waasbergen, 1 995). However, the uppermost metres of the car bonate platform do not exhibit either an anomalous cementation or an elevated natural gamma-ray intensity that would indicate a final stage of expo sure diagenesis and soil development. The logging
and lithological data suggests that the carbonate platform at this site (i) was a peritidal environment with cyclic episodes of brief emergence during the formation of the middle parasequence set, (ii) began to deepen during the upper parasequence set, then (iii) underwent a rapid deepening and termination of shallow-water carbonate deposition at the final stage of the uppermost parasequence without an intervening emergence. The uppermost cores con tain a post-drowning partial phosphatization and infilling of cavities by fine-grained pelagic sediment. However, the drowning succession at this site may not be characteristic of the termination process of the entire guyot platform.
REFERENCES F. ( 1 992) L'Albien du bassin Anglo-Parisien: ammonites, zonation phyletique, sequences. Bull. Cent. Rech. Explor., Prod. Eif-Aquitaine, 16, 1 8 7-23 3 . ARNAUD, H., FLOOD, P . & STRASSER, A . ( 1 995) Resolution Guyot (Hole 866A, Mid-Pacific Mountains): facies ev olution and sequence stratigraphy. In: Proceedings ofthe Ocean Drilling Program, Scientific Results, 1 43 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 1 3 3- 1 59. Ocean Drilling Program, College Station, TX. ARNAUD-VANNEAU, A. & ARNAUD, H. ( 1 990) Hauterivian to Lower Aptian carbonate shelf sedimentation and sequence stratigraphy in the Jura and northern Subal pine chains (southeastern France and Swiss Jura). In: Carbonate Platforms: Facies, Sequences and Evolution (Eds Tucker, M.E., Wilson, J.L., Crevello, P.D., Sarg, J.R. & Read, J.F.), Spec. Pubis Int. Assoc. Sediment. No. 9, 203-2 3 3 . Blackwell Scientific Publications, Ox ford. ARNAUD-VANNEAU, A. & SLITER, W.V. ( 1 995) Early Creta ceous shallow-water benthic foraminifers and fecal pel lets from Leg 1 43 and comparison of coeval faunas from the Pacific basin, Central America and the Tethys. In: Proceedings of the Ocean Drilling Program, Scientific Results, 143 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 5 3 7-564. Ocean Drilling Pro gram, College Station, TX. BAUDIN, F., DECON!NCK, J.-F., SACHSENHOFER, R.E., STRASSER, A. & ARNAUD, H. ( 1 995) Organic geochemis try and clay mineralogy of Lower Cretaceous sediments from Allison and Resolution guyots (Sites 865 and 866), Mid-Pacific Mountains. In: Proceedings of the Ocean Drilling Program, Scientific Results, 143 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 1 7 31 96. Ocean Drilling Program, College Station, TX. BERGERSEN, D. ( 1 995) Physiography and architecture of Marshall Island guyots drilled during ODP Leg 1 44: geophysical constraints on platform drowning. In: Pro ceedings of the Ocean Drilling Program, Scientific Re sults, 1 44 (Eds Haggerty, J., Premoli-Silva, I., Rack, F.R. & McNutt, M.K.), pp. 5 6 1 - 5 8 3 , Ocean Drilling Pro gram, College Station, TX. AMEDRO,
1 34
U. Rohl & J. G. Ogg
R.A. ( 1 994) GEOCARB II: A revised model of atmospheric C02 over Phanerozoic time. Am. J. Sci. , 294, 56-9 1 . BRALOWER, T.J., SLITER, W.V. ARTHUR, M.A., LECKIE, R.M., ALLARD, D.J. & SCHLANGER, S.O. ( 1 993) Dysoxic/ anoxic episodes in the Aptian-Albian (Early Creta ceous). In: The Mesozoic Pacific: Geology, Tectonics, and Volcanism (Eds Pringle, M.S., Sager, W.W., Sliter, W.V. & Stein, S.), Geophys. Monogr., Am. geophys. Union, Washington, DC, 77, 5-37. BREHERET, J.G. ( 1 988) Episodes de sedimentation riche en matiere organique dans les marnes bleues d'age aptien et albien de Ia partie pelagique du bassin vocontien. Bull. Soc. geo/. Fr. , 8, 349-356. CLOETINGH, S. ( 1 988) Intraplate stresses: a tectonic cause for third-order cycles in apparent sea level. In: Sea-level Changes: an Integrated Approach (Eds Wilgus, C.K., Hastings, B.S., Kendall, C.G.St.C., Posamentier, H.W., Ross, C.A. & van Wagoner, J.C.), Spec. Pub!. Soc. econ. Paleontol. Mineral., Tulsa, 42, 1 9-30. COOPER, P. ( 1 99 5) Milankovitch cycles from Fourier analysis of logs from Site 865 and 866. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 43 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 3 1 7-327. Ocean Drilling Program, College Station, TX. COOPER, P.A., ARNAUD, H. & FLOOD, P. ( 1 995) Formation MicroScanner log responds to lithology in guyot carbon ate platforms and their implications: Site 865 and 866. In: Proceedings ofthe Ocean Drilling Program, Scientific Results, 1 43 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 329-372. Ocean Drilling Pro gram, College Station, TX. CROUGH, S. T. ( 1 978) Thermal origin of mid-plate hot-spot swells. Geophys. J. R. astron. Soc. , 55, 45 1-469. CROUGH, S.T. ( 1 984) Seamounts as recorders of hot-spot epeirogeny. Geo/. Soc. Am. Bull. , 95, 3-8. DARWIN, C. ( 1 842) The Structure and Evolution of Coral ReefS. Smith, Elder, London. (Reprinted in 1 962 by Univ. California Press, Berkeley, Los Angeles.) DETRICK, R.S. & CROUGH, S.T. ( 1 9 78) Island subsidence, hot spots, and lithospheric thinning. J. geophys. Res., 83, 1 236-1 244. EPP, D. ( 1 984) Possible perturbations to hotspot traces and implications for the origin and structure of the Line Islands. J. geophys. Res., 89, 1 1 273- 1 1 286. ERBA, E. ( 1 995) Quantitative nannofossil biostratigraphy of Quaternary sequences from guyots in the central and western Pacific Ocean. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J., Premoli-Silva, I. Rack, F.R. & McNutt, M.K.), pp. 3-20. Ocean Drilling Program, College Station, TX. FISCHER, A.G. ( 1 964) The Lofer cyclothems of the alpine Triassic. Kans. geo/. Surv. Bull., 169, 1 07- 1 49. FRIES, G. & RUBINO, J.L. ( 1 990) Testing the application of sequence stratigraphy to Aptian deposits in southeast France. In: Cretaceous Resources, Events and Rhythms: Background and Plansfor Research (Eds Ginsburg, R.N. & Beaudoin, B.), NATO Advanced Study Institute Series, Series C, Mathematical and Physical Sciences, 304, pp. 47-62. Kluwer Academic, Dordrecht. GALLOWAY, W.E. ( 1 989) Genetic stratigraphic sequences in basin analysis: I. Architecture and genesis of flooding-
BERNER,
surface bounded depositional units. Bull. Am. Assoc. petrol. Geo/., 73, 1 25- 1 42. GRADSTEIN, F.M., AGTERBERG, F.P., 0GG, J.G., H ARD EN SOL, J. VAN VEEN, P., THIERRY, J. & HUANG, Z. ( 1 994) A Mesozoic time scale. J. geophys. Res., 99(B 1 2), 2405 1 24074. GRAFE, K.U. ( 1 994) Sequence stratigraphy in the Creta ceous and Paleogene (Aptian to Eocene) of the Basco Cantabrian Basin (N Spain). Tiibinger Geowissenschaft liche Arbeiten, A, 18. GROTSCH, J. & FLOGEL, E. ( 1 992) Facies of sunken Early Cretaceous atoll reefs and their capping late Albian drowning succession (northwestern Pacific). Facies, 27 , 1 5 3- 1 74. GROTSCH, J., SCHROEDER, R., No£, S. & FLOGEL, E. ( 1 993) Carbonate platforms as recorders of high-amplitude eustatic sea-level fluctuations: the late Albian appennin ica event. Basin Res. , 5, 1 9 7-2 1 2. HAGGERTY, J., PREMOLI-SILVA, 1., RACK, F. & MCNUTT, M.K. (Eds) ( 1 995) Proceedings ofthe Ocean Drilling Pro gram, Scientific Results, 1 44. Ocean Drilling Program, College Station, TX, 1 05 9 pp., and CD-ROM appendix. HAQ, B.U., HARDENBOL, J. & VAIL, P.R. ( 1 987) Chronology of fluctuating sea levels since the Triassic. Science, 235, 1 1 56- 1 1 67. HAQ, B.U., HARDENBOL, J. & VAIL, P.R. ( 1 988) Mesozoic and Cenozoic chronostratigraphy and cycles of sea-level change. In: Sea-level Changes: an Integrated Approach (Eds Wilgus, C.K., Hastings, B.S., Kendall, C.G.St.C., Posamentier, H.W., Ross, C.A. & van Wagoner, J.C.), Spec. Pub!. Soc. econ. Paleontol. Mineral., Tulsa, 42, 7 1 - 1 08. HARDENBOL, J. et a/. ( 1 997) The Mesozoic-Cenozoic chronostratigraphic framework. In: Mesozoic-Cenozoic Sequence Stratigraphy ofEuropean Basins (Eds de Gra ciansky P.C. et a/.), Spec. Pub!. Soc. econ. Paleontol. Mineral., Tulsa, (in press). HESS, H. H. ( 1 946) Drowned ancient islands of the Pacific Basin. Am. J. Sci. , 244, 772-79 1 . HOMEWOOD, P., GUILLOCHEAU, F., ESCHARD, R. & CROSS, T.A. ( 1 992) Correlations haute resolution et stratigra phie genetique: une demarche integree. Bull. Cent. Rech. Exp/or., Prod. E/fAquitaine, 16, 3 5 7-38 1 . HuNT, D . & TuCKER, M.E. ( 1 993) Sequence stratigraphy of carbonate shelves with an example from the mid Cretaceous (Urgonian) of southeast France, In: Se quence Stratigraphy and Facies Associations (Eds Posamentier, H.W., Summerhayes, C.P., Haq, B.U. & Allen, G.P.), Spec. Pubis, Int. Assoc. Sediment., No. 1 8, 307-34 1 . Blackwell Scientific Publications, Oxford. JACQUIN, T., ARNAUD-VANNEAU, A., ARNAUD, H., RAVENNE, C. & VAIL, P.R. ( 1 9 9 1 ) Systems tracts and depositional sequences in a carbonate setting: a study of continuous outcrops from platform to basin at the scale of seismic lines. Mar. petrol. Geo/. , 8, 1 22- 1 39. JANSA, L.F. & ARNAUD-VANNEAU, A. ( 1 995) Carbonate buildup and sea-level changes at MIT Guyot, Western Pacific, In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J., Premoli-Silva, 1., Rack, F.R. & McNutt, M.K.), pp. 3 1 1 - 3 3 5 . Oce!ln Drilling Program, College Station, TX. JENKYNS, H.C. ( 1 995) Carbon-isotope stratigraphy and paleoceanographic significance of the lower Cretaceous
1 35
Aptian-Albian eustatic sea-levels shallow-water carbonates of Resolution Guyot, Mid Pacific Mountains. In: Proceedings ofthe Ocean Drilling Program, Scientific Results, 1 43 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 99- 1 04. Ocean Drilling Program, College Station, TX. JENKYNS, H.C. & STRASSER, A. ( 1 995) Lower Cretaceous oolites from the Mid-Pacific Mountains (Resolution Guyot, Site 866). In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 43 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 1 1 1 - 1 1 8. Ocean Drilling Program, College Station, TX. JENKYNS, H.C., PAULL, C.K., CUMMINS, DJ. & FULLAGAR, P.D. ( 1 995) Strontium-isotope stratigraphy of lower Cretaceous atoll carbonates in the Mid-Pacific Moun tains. In: Proceedings of the Ocean Drilling Program, Scientific Results, 143 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 89-97. Ocean Drilling Program, College Station, TX. JOHNSON, C.C., BARRON, E.J., KAUFFMAN, E.G., ARTHUR, M.A., FAWCETT, P.J. & YASUDA, M.K. ( 1 996) Middle Cretaceous reef collapse linked to ocean heat transport. Geology, 24, 376-380. JONES, C.E., JENKYNS, H.C., COE, A.L. & HESSELBO, S.P. ( 1 994) Sr-isotopic variations in Jurassic and Cretaceous seawater. Geochim. Cosmochim. Acta, 58, 306 1 -3074. KENDALL, C. G. & SCHLAGER, W. ( 1 9 8 1 ) Carbonates and relative sea level. Mar. Geol. , 44, 1 8 1 -2 1 2. LARSON, R.L. ( 1 9 9 1 ) Geological consequences of super plumes. Geology, 19, 963-966. LARSON, R.L., ERBA, E., NAKANISHI, M., BERGERSEN, D.D., & LINCOLN, J.M. ( 1 994) Stratigraphic, vertical subsid ence, and paleolatitude histories of ODP Leg 1 44 guyots. EOS, Trans. IAGU, 75, 582. MARTY, J.C. & CAZENAVE, A. ( 1 989) Regional variations in subsidence rate of oceanic plates: a global analysis. Earth planet. Sci. Lett. , 94, 3 0 1 -3 1 5. McNUTT, M.K., WINTERER, E.L., SAGER, W.W., NATLAND, J.H. & I To , G. ( 1 990) The Darwin Rise: a Cretaceous superswell? Geophys. Res. Lett. , 17, 1 1 0 1 - 1 1 04. MENARD, H.W. ( 1 982) The influence of rainfall upon the morphology and distribution of atolls. In: The Ocean Floor (Eds Scrutton, R.A. & Talwani, M.), Bruce Heezen Commemorative Volume, pp. 305-3 1 1 . John Wiley, Chichester. MOLINIE, A.J. & OGG, J.G. ( 1 992) Formation MicroScanner imagery of Lower Cretaceous and Jurassic sediments from the western Pacific (Site 80 I ). In: Proceedings ofthe Ocean Drilling Program, Scientific Results, 1 29 (Eds Lar son, R.L., Lancelot, Y. et a!.), pp. 67 1 -692. Ocean Drill ing Program, College Station, TX. NAKANISHI, M., TAMAKI, K. & KOBAYASHI, K. ( 1 992) Magnetic anomaly lineations from Late Jurassic to Early Cretaceous in the west-central Pacific Ocean. Geophys. J. Int., 109, 70 1 -7 1 9. OBRADOVICH, J.D. ( 1 993) A Cretaceous time scale. In: Evolution of the Western Interior Basin (Ed. Caldwell, W.G.E.), Spec. Pub!. geol. Assoc. Can., 39, 379-396. 0GG, J.G. ( 1 995) MIT Guyot: depositional history of the carbonate platform from downhole logs at Site 878. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J., Premoli-Silva, I, Rack, F.R. & McNutt, M.K.), pp. 3 37-359. Ocean Drilling Program, College Station, TX.
J.G., CAMOIN, G., ARNAUD-VANNEAU, A. ( 1 995a) Limalok Guyot: depositional history of the carbonate platform from downhole logs- Site 8 7 1 (Lagoon). In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J., Premoli-Silva, I., Rack, F.R. & McNutt, M.K.), pp. 233-2 5 3 . Ocean Drilling Program, College Station, TX. OGG, J.G., CAMOIN, G. & JANSA, L. ( 1 995b) Takuyo Daisan Guyot: depositional history of the carbonate platform from downhole logs at Site 879 (outer rim). In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 44 (Eds Haggerty, J., Premoli-Silva, I. Rack, F.R. & McNutt, M.K.), pp. 3 6 1 -380, Ocean Drilling Program, College Station, TX. OLSEN, P.E. ( 1 986) A 40-million-year lake record of early Mesozoic orbital climatic forcing. Science, 234, 842848. OLSEN, P.E., KENT, D.V., CORNET, B., WITTE, W.K. & ScHLISCHE, R.W. ( 1 996) High-resolution stratigraphy of the Newark rift basin (early Mesozoic, eastern North America). Geol. Soc. Am., Bull., 108, 40-77. PREMOLI-SILVA, I., HAGGERTY, J., RACK, F. et a/. (Eds) ( 1 993) Proceedings of the Ocean Drilling Program, Ini tial Reports, 1 44, Ocean Drilling Program, College Station, TX, 1 084 pp., and CD-ROM appendix. READ, J.F. & GOLDHAMMER, R.K. ( 1 988) Use of Fisher plots to define third-order sea-level curves in Ordovi cian peritidal cyclic carbonates, Appalachians. Geology, 16, 895-899. REVELLE, R. (Ed.) ( 1 990) Sea Level Change. National Research Council, Studies in Geophysics. National Academy Press, Washington, DC. ROBASZYNSKI, F., HARDENBOL, J., CARON, M. et af. , ( 1 993) Sequence stratigraphy in a distal environment: the Cenomanian of the Kalaat Senan region (central Tuni sia). Bull. Cent. Rech. Explor., Prod. E/fAquitaine, 17, 3 9 5-433. R6HL, U. & 0GG, J.G. ( 1 996) Aptian-Albian sea-level history from Guyots in the Western Pacific. Paleocean ography, 1 1 , 5 95-624. Ro HL, U. & STRASSER, A. ( 1 99 5) Diagenetic alteration and geochemical trends in Early Cretaceous shallow-water limestones of Allison- and Resolution-Guyots, Sites 865 to 868. In: Proceedings of the Ocean Drilling Program, Scientific Results, 1 43 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 1 97-229. Ocean Drilling Program, College Station, TX. RuFFELL, A. H. ( 1 9 9 1 ) Sea-level events during the Early Cretaceous in Western Europe. Cret. Res., 12, 527-55 1 . SADLER, P.M., OSLEGER, D.A. & MONTA Ez, I.P. ( 1 993) On the labeling, length, and objective basis of Fischer plots. J. sediment Petrol. , 63 , 360-368. SAGER, W.W., WINTERER, E.L., FIRTH, J.V. et af. (Eds) ( 1 993) Proceedings of the Ocean Drilling Program, Ini tial Reports, 1 43. Ocean Drilling Program, College Station, TX 724 pp. and CD-ROM appendix. SARG, J.F. ( 1 988) Carbonate Sequence Stratigraphy. Spec. Pub!. Soc. econ. Paleont. Miner., Tulsa, 42, 1 55-1 8 1 . SCHLAGER, W. ( 1 989) Drowning unconformities on carbon ate platforms. In: Sea Level Changes-An Integrated Ap proach (Eds Wilgus, C.K., Hastings, B.S., Kendall, C.S., Posamentier, H.W., Ross, C.A. & van Wagoner, J.C.), Spec. Pub!. Soc. econ. Paleont. Miner., Tulsa, 42, 1 5-25.
OGG,
'
1 36
U Rohl & J. G. Ogg
W. ( 1 9 9 1 ) Depositional bias and environmen tal change-important factors in sequence stratigraphy. In: The record ofsea-level fluctuations (Eds Biddle, K.T. & Schlager, W.). Sediment Geol. , 70, 1 09- 1 30. SCHLAGER, W. ( 1 992) Sedimentology and Sequence Strati graphy of Reefs and Carbonate Platforms. Am. Assoc. petrol. Geol., Cont. Educ. Course Note Ser., 34, 7 1 pp. SCHLANGER, S.O., JENKYNS, H.C. & PREMOLI-SILVA, I. ( 1 9 8 1 ) Volcanism and vertical tectonics in the Pacific basin related to global Cretaceous transgressions. Earth planet. Sci. Lett. , 52, 43 5-449. SHIPBOARD SCIENTIFIC PARTY ( 1 993a) Site 865. In: Proceed ings of the Ocean Drilling Program, Initial Reports, 1 44 (Eds Sager, W.W., Winterer, E.L. & Firth, J.V.), pp. 1 1 1 - 1 80, 3 8 1 -463, and CD-ROM appendix. Ocean Drilling Program, College Station, TX. SHIPBOARD SCIENTIFIC PARTY ( 1 993b) Site 866. In: Proceed ings of the Ocean Drilling Program, Initial Reports, 1 43 (Eds Sager, W.W., Winterer, E.L. & Firth, J.V.), pp. 1 8 1 -27 1 , 465-605, and CD-ROM appendix. Ocean Drilling Program, College Station, TX. SHIPBOARD SCIENTIFIC PARTY ( 1 993c) Site 867/868. In: Proceedings of the Ocean Drilling Program, Initial Re ports, 1 43 (Eds Sager, W.W., Winterer, E.L. & Firth, J.V.), pp. 273-296, 607-62 1 , and CD-ROM appendix. Ocean Drilling Program, College Station, TX. SHIPBOARD SCIENTIFIC PARTY ( 1 993d) Site 878. In: Proceed ings of the Ocean Drilling Program, Initial Reports, 1 44 (Eds Premoli-Silva, I., Haggerty, J. & Rack, F.F.), pp. 3 3 1 -4 1 2, 7 8 1 -972, and CD-ROM appendix. Ocean Drilling Program, College Station, TX. SHIPBOARD SCIENTIFIC PARTY ( 1 993e) Site 879. In: Proceed ings of the Ocean Drilling Program, Initial Reports, 1 44 (Eds Premoli-Silva, I., Haggerty, J. & Rack, F.F.), pp. 4 1 2-44 1 , 973- 1 009, and CD-ROM appendix. Ocean Drilling Program, College Station, TX. SOREGHAN, G.S. & DICKINSON, W.R. ( 1 994) Genetic types of stratigraphic cycles controlled by eustasy. Geology, 22, 759-76 1 . STRASSER, A., ARNAUD, H., BAUDIN, F. & ROHL, U . ( 1 995) Small-scale shallow-water carbonate sequences of Reso lution Guyot (Sites 866, 867, and 868). In: Proceedings
SCHLAGER,
of the Ocean Drilling Program, Scientific Results, 143 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 1 1 9- 1 3 1 . Ocean Drilling Program, College Station, TX. VAIL, P.R. ( 1 987) Seismic stratigraphy interpretation pro cedure. In: Atlas of Seismic Stratigraphy (Ed. by Bally, A.W.), Am. Assoc. petrol. Geol., Stud. Geol., 27, 1 - 1 0. VAIL, P.R., MITCHUM, R.M., JR, TODD, R.G. et a/. ( 1 977) Seismic stratigraphy and global changes of sea level, In: Seismic Stratigraphy-Applications to Hydrocarbon Ex ploration (Ed. Payton, C.E.), Mem. Am. Assoc. petrol. Geol., 26, 49-2 1 2 . VAN WAASBERGEN, R.J. & WINTERER, E.L. ( 1 993) Summit Geomorphology of Western Pacific Guyots. Geophys. Monogr., Am. geophys. Union Washington, DC, 77, 3 35-366. VAN WAGONER, J.C., MITCHUM, R.M., CAMPION, K.M. & RAHMANIAN, V.D. ( 1 990) Siliciclastic Sequence Strati graphy in Well Logs, Cores, and Outcrops. Am. Assoc. petrol. Geol., Tulsa, Methods in Explor. Ser., 7. WEISSERT, H. & BRtHI':RET, J.G. ( 1 9 9 1 ) A carbon-isotope record from Aptian-Albian sediments of the Vocontian trough (SE France). Bull. Soc. geol. Fr. , 162, 1 1 33-1 1 40. WINTERER, E.L & SAGER, W.W. ( 1 995) Synthesis of drilling results from the Mid-Pacific Mountains: regional con text and implications. In: Proceedings of the Ocean Drilling Program, Scientific Results, 143 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 4975 3 5 . Ocean Drilling Program, College Station, TX. WINTERER, E.L., NATLAND, J.H., VAN WAASBERGEN, R.J. et a!. ( 1 993) Cretaceous Guyots in the Northwest Pacific: an overview of their Geology and Geophysics. Geophys. Monogr., Am. geophys. Union, Washington, DC, 77, 307-334. WINTERER, E.L., VAN WAASBERGEN, R J ., MAMMERICKX, J. & STUART, S. ( 1 995) Karst morphology and diagenesis of top of Albian limestone platforms, Mid-Pacific Mountains. In: Proceedings of the Ocean Drilling Pro gram, Scientific Results, 143 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 433-470. Ocean Drilling Program, College Station, TX. .
Spec. Pubis int. Ass. Sediment. (1998) 25, 137-144
Origin of white sucrosic dolomite within shallow-water limestones, ODP Hole 866A, Resolution Guyot, Mid-Pacific Mountains: strontium isotopic evidence for the role of sea water in dolomitization P. G. F L O O D University o fNew England, Division o fEarth Sciences, Armidale, NSW 2351, Australia
ABSTRACT Strontium isotopic (87 Sr/86 Sr) measurements obtained from a 30-m-thick unit of massive white sucrosic dolomite at a depth of 1250-1280 mbsf (metres below sea-floor) within the 1620-m-thick shallow-water limestone sequence recovered by the Ocean Drilling Program in Hole 866A at Resolution Guyot indicate that the dolomite was precipitated from slightly modified sea water at 24 Ma, some I 00 Myr after deposition of the enclosing limestones. The temperature of formation of the dolomite is consistent with the temperature of the pore-waters recorded by the downhole temperature probe. Oxygen and carbon stable isotopes also indicate a marine source for the Mg-rich dolomitizing solutions. Solution of the precursor limestone below 1000 mbsl (metres below sea-level) is responsible for the development of solution cavities and mouldic porosity which is infilled by the dolomite precipitating fluids. In addition to the formation of new crystals of dolomite within the cavities, occasionally the original grains are mimically replaced and then the original texture is preserved. Conditions of dolomitization include the following: (i) a permeable substrate; (ii) contiguous deep 2 ocean water supplying an abundant source of Mg + ions and capable of dissolving aragonite and calcite and replacing them with dolomite; (iii) a sufficiently thick limestone pile to impose a thermal gradient allowing a geothermal convective process to operate. Such a process has been called geothermal endo-upwelling. Although some researchers have indicated that sea-water can be responsible for the transformation of a carbonate precursor by secondary replacement dolomite, the strontium isotopic data obtained from ODP Hole 866A, Resolution Guyot, provide confirmatory evidence of the marine water dolomitization model.
INTRODUCTION
The origin of dolomite continues to be a controver sial topic in sedimentology. As yet, no consensus has been reached to explain the origin and occur rence throughout the geological record (see Braith waite, 1991; Fowles, 1991; Tucker & Wright, 1991; Sun, 1994). However, one model of dolomitization involves convective flow of marine waters and it has gained support from an increasing number of re searchers (Schlanger, 1963; Saller, 1984; Aissaoui et al., 1986; Aharon et al., 1987; Hardie, 1987; Vahr enkamp & Swart, 1990; Wilson et al., 1990; Vahr enkamp et al., 1991; Hein et al., 1992; Flood & Chivas, 1995; Flood et al., 1996).
Results obtained from this study of massive white sucrosic dolomite interbedded within early Creta ceous (Hauterivian to late Albian) shallow-water limestones recovered from Ocean Drilling Program (ODP) Hole 866A on Resolution Guyot (Sager etal., 1993) provide insight into the potential of Kohout convection (Kohout, 1967; Kohout et al., 1977). This process is similar to 'endo-upwelling', which is a geothermally driven convective process operating within the upper part of the volcanic foundation beneath the overlying carbonate pile in atolls and guyots (Rougerie & Wauthy, 1988, 1993; Rougerie et al., 1992; Rougerie & Fagerstrom, 1994).
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
137
138
P. G. Flood
The purpose of this paper is to alert sedimento logical researchers interested in dolomite to the opportunity for additional research on material recovered from ODP Hole 866A.
regional location of the hole. Situated at 21' 19.953'N, 174'18.844'£, it was drilled in April 1992. The hole was spudded in a water depth of 1361 m and it recovered representative core mate rial from the 1620-m-thick shallow-water limestone sequence above a volcanic basement.
ODP LEG 143 Sequence
ODP Leg 143 was part of a larger effort, including Leg 144, designed to explore a number of sunken, flat-topped seamounts-guyots-in the north-west Pacific. A historical account of the research con cerning guyots, originally described by H. H. Hess, has been provided by Winterer & Sager (1995). Before ODP Leg 143 the geomorphology of the summit regions of the guyots posed a major ques tion: are the atoll- and barrier-reef-like forms con structional or erosional? Also, it was of interest to determine the degree to which the limestone succes sion had remained open to circulation of water from the adjacent open ocean and to what extent circulation within the limestone had been influ enced by heat flow from the underlying volcanic basement, and, if it had been, what were the consequences?
Sager eta!. (1993) have provided detailed descrip tions of the lithological units (Fig. 2), encountered in Hole 866A, and Arnaud et a!. (1995) have described the depositional sequence within a se quence stratigraphical framework. One interval of particular interest occurs within Unit VII B (Cores 143-866A-133R to -135R) at a depth of 1251.61280.2 mbsf (metres below sea-floor). The lithology consists of massive white sucrosic dolomite, in part stained light red, which has replaced the original peloidal grainstone. The sedimentary facies inferred by Arnaud eta!. (1995) is an open-marine subtidal environment grading up-sequence to a restricted intertidal envi ronment. Primary depositional features still recog nizable in the core include keystone vugs.
Hole 866A
Subs idence history
ODP Hole 866A is located on the north rim of the summit of Resolution Guyot in the western Mid-Pacific Mountains (MPM). Figure 1 shows the
The subsidence history curve for Hole 866A is shown in Fig. 3. This is based on the age estimates for stage boundaries obtained from strontium iso-
170°
1800
Fig. I. Location ODP Leg 143 Hole 866A, Resolution Guyot, Mid-Pacific Mountains.
Origin of white sucrosic dolomite
Q) Ol
0
19.6-
A
100
200
300
topic data (Jenkyns et a!., 1995) and the radiomet ric data of Pringle & Duncan (1995). The time-scale of Harland eta!. (1990) is used. The curve assumes that the carbonate sedimentation on the platform was nearly always close to sea-level and the form of the curves is exponential. Estimates (Winterer & Sager, 1995) of the subsid ence (=sedimentation rates) decrease from the time of deposition of the base of the carbonate sequence upwards. Values range from 200 m Myc1 at the base to about 25 m Myr-1 at the top of the succes sion and continue to decrease towards the present day. The widely accepted expression is:
:'.:: c ::::J .0 ::::J (/)
<(
174.5 c
ro :c <(
Ill
B 271.0-
d(m)
c 434.5
-
•
IV 600 c
700
ro
·� o._
<( c
.2 s
676.6 t---
v
a. Q)
The borehole temperature log from Hole 866A (Sager eta!., 1993; Paull eta!., 1995) shows that the temperature increases downhole at an average rate of 8 · c km-1. Thermal differences between the geo thermally heated pore-water within the limestone sequence and the guyot-surrounding cooler sea wa ter (4 · C) would result in a lateral and vertical fluid density gradient. Such a thermal gradient would provide a mechanism to stimulate upward fluid flow (Kohout, 1967; Kohout et a!., 1977; Rougerie & Fagerstrom, 1994).
A 859.2
.r::
900
B
r--
917.o-
0
c 1100 c
ro
1200
•
VI
1000
.E � ...... ro
(!) 1300
1203.
VII
A
1251.6B 1299.5-
c
1357.5-
D
1399.7-
1400
1500l -r-.,_; ::::J
A
VIII
- ro I 16 g1 3. 1600) - ?-- 1620. v v v v v v v v IX v v v v v
1700
,____
1743.
K,JT
Borehole temperature
791.8
800
=
where T is the elapsed time in million years since thermal subsidence began and K. is a constant (Winterer & Sager, 1995). The curve does not display any deviations from the idealized curve that could have resulted from eustatic sea-level fluctua tions, which are known to have occurred during the deposition of the shallow-water carbonate sequence (Winterer & Sager, 1995).
400
500
139
v
v
v
v
v
• • • • • •
i
0
Pore-waters
Paull eta!. (1995) have indicated that there is active exchange between interior pore-waters within the 'drowned' carbonate platform sequence and sur rounding sea water. Major element compositions
• •
:
Fig. 2. (Opposite) Stratigraphical sequence ODP Leg 143 Hole 866A Resolution Guyot. Occurrences of brown dolomite indicated by solid rhomb; white sucrosic dolomite indicated by open rhomb. Lithologies and core recovery details are indicated in Fig. 58 of Sager et a/. (1993).
140 130
P. G. Flood 120
110
100
90
ao
70
60
so
40
30
20
10
o Ma B.P.
PULSE I BROWN DOLOMITE
500
I '' ,, ,' ', ,
PULSE II WH ITE DOLOMITE
l ::::
HOLE 866A
, , , ,
--
tOOO
2500
_
_ _ _ _
_
__ _
__ _
__ _
3000
m.b.s.l.
(Sager eta!., 1993), 87Sr/86Sr values and 8180 values (Paull eta!., 1995) of the pore-waters, although not identical, are similar to those of modern sea water. Some flow restriction is indicated by the fact that the pore-water temperature increases downhole. Paull et a!. (1995) concluded that the rate of fluid flow varies from once every 104 yr within the upper porous lime stone to once every 106 yr in the deeper less porous limestone. Flushing rates of more than 10 000 pore volumes of sea water could have occurred since the platform drowned almost 90 Myr ago. Flushing by sea water, chemically modified, dur ing its residence time within the limestone, aided by the geothermal temperature rise with depth could have a profound effect on the extent and nature of carbonate diagenesis. Aragonite and calcite become more soluble as lower temperatures were encoun tered as the limestone platform subsided. Below about 500 mbsf, modern Pacific water becomes undersaturated with respect to first aragonite and then calcite (Scholle et a!., 1983). With increasing depth, dissolution can proceed and increased poros ity or permeability will facilitate fluid flow.
Fig. 3. Inferred subsidence curve ODP Leg 143 Hole 866A, Resolution Guyot, Mid-Pacific Mountains. White sucrosic dolomite formed at about 24 Ma in water depths of about 2500 m.
METHODS
Polished thin sections were prepared from represen tative samples and textural relationships observed using a petrological microscope. Carbonate miner alogy was determined using X-ray diffraction. Only monomineralic samples were analysed for isotopic composition. Samples were examined for lumines cence using cathode luminescence equipment at the Research School of Earth Sciences, Australian Na tional University (RSES, ANU), Canberra. · The strontium isotope values of the sucrosic white dolomite have been determined by the Pre cise Radiogenic Isotope Services of the RSES, ANU, using a Finnigan MAT 261 multicollector mass spectrometer. Also the o 180 oxygen and o 1 3C carbon values were measured at the RSES, ANU, using the Kiel preparation device manufactured by Finnigan MAT of Bremen, Germany, and a Finni gan MAT 251 mass spectrometer. Results of the analyses have already been published by Flood & Chivas (1995).
RESULTS
Dolomite
Dolomite is ubiquitous in the Barremian sediments at depths below 1200 mbsf in cores from Hole 866A. Two distinct dolostones occur; one brown, the other white. The brown dolomite occurs over a 400-m interval (1200-1620 mbsf), whereas the white dolomite is restricted to a 50-m interval (1250-1300 mbsf) enclosed within the brown dolo mite. Material selected from Core 133R is consid ered representative of the white dolomite and is the subject of this paper.
Dolomite textures
Dolostone rock textures observed in Cores 143866A-133R to -135R display coarsely crystalline grains (up to 10) which show well-developed inter crystalline porosity (up to 30%; Kenter & Ivanov, 1995, Plates 1 & 8) together with mouldic porosity. Occasionally, the original depositional textures· are preserved in spite of the pervasive replacive dolo mitization.
Origin of white sucrosic dolomite
141
nearly stoichiometric with regard to its composition. The dolomite is non-luminescent. However induc tively coupled plasma atomic emission spectrometry (ICP-AES) analyses (see Flood & Chivas, 1995) have detected minor traces of Fe (39 p.p.m. maximum), Zn (33 p.p.m. maximum), Cu (3 p.p.m. maximum) and Mn (< 10 p.p.m. maximum). Oxygen and carbon is otopes
Flood & Chivas (1995) and Jenkyns (1995) reported o 180p08 values up to + 3.7%o for the white sucrosic dolomite. The same authors reported a uniform o13Cp08 isotopic value of +3.4%o. These values are consistent with dolomite deposition from slightly modified sea water. The temperature of dolomite crystallization, using either the equation published by Land (1985) or the graph published by Rao (1996, Fig. 12.18), was about 17·c, which is consistent with the modern pore-water temperature (Sager et al., 1993). Strontium is otopes
Fig. 4. Photomicrograph of the white dolomite showing an equigranular mosaic of crystals containing ghosts of darker pre-existing allochems, mainly peloids (sample l43-866A-133R-1, 145-147cm).
The white sucrosic dolomite has an Sr content of 109 p.p.m. and an 87Sr/86Sr ratio of 0.708215± 0.000013 to 0.708217 ±0.000015. Using published 87Sr/86Sr sea water curves (Jones, 1992; McArthur, 1993) or the equation developed by McKenzie eta/. (1993, Table 3) this white sucrosic dolomite formed at about 24 Ma. That is about 100 Myr after the sedimentation age of the enclosing limestone (Jen kyns et a/., 1995).
INTERPRETATION
All crystals are euhedral to anhedral and exhibit planar boundaries (Fig. 4). Textures include unimo dal planar-s types with prominent sucrosic features and bimodal planar-s types with non-mimically replaced allochems. Silbey & Gregg (1987) have suggested that the predominance of planar crystal boundaries having well-formed crystal faces indi cates their development at low temperature (< 50 ·q and low saturation. This might or might not be the situation, and well-formed crystals might simply indicate slow crystal growth. This idea war rants detailed investigation. G eochemistry
Microprobe analyses indicate that the dolomite is
There is no obvious explanation for the preferential selection of stratigraphical interval VIIB to be dolomitized. The original sedimentary facies indi cates deposition of porous, permeable peloidal grainstones deposited in open-marine to subtidal environments, including beach deposits, as indi cated by the keystone vugs (Arnaud et a/., 1995). Conditions operating during the dolomitization process must have included the following: (i) a permeable carbonate substrate; (ii) contiguous deep-ocean water such as Antarctic Intermediate Water, supplying an abundant source of Mg2+ ions (see Land, 1985); (iii) deep ocean water, located below depths of 400 m, capable of dissolving arago nite and calcite; and (iv) a geothermal gradient
142
P. G. Flood
D D D '
@
Dolomite
t
Limestone Volcanic Basement
Impermeable Apron Geothermal Heat Flux
(S �
Endo-Upwelling Flow Paths
(Weak);lf
--
Fig. 5. Schematic cross-section of a carbonate platform or submerged atoll, the thermal convection process associated with geothermal endo-upwelling circulation, and the occurrence of dolomite. Modified after Aissaoui et al. (1986, Fig. I B).
capable of inducing convective flow through the porous or permeable substrate. Such conditions (Fig. 5) would conform to the geothermal endo-upwelling model (Rougerie & Wauthy, 1988, 1993; Rougerie et a!., 1991; Roug erie & Fagerstrom, 1994). When these conditions combine there is the potential for massive dolomi tization. It is known that sea water cannot dissolve calcite or aragonite at depths shallower than 500 m. There fore either the fluids responsible for dolomitization must be different from normal ocean water or the sea water has been sourced from deeper than 500 m, or a combination of the two. Similar massive dolomite has been reported pre viously from the subsurface of Funafuti (Cullis, 1904), Enewetak (Saller, 1984), Mururoa (Aissaoui eta!., 1986; Guille eta!., 1996) and Niue (Aharon et a!., 1987). Perhaps a set of conditions similar to those that operated during the dolomitization at Hole 866A also occurred in the open ocean atolls and carbonate platforms? The proposal of Guille et a!. (1996) that the dolomite below Mururoa and Fangataufu atolls is produced from brackish waters during emergence need not be correct.
866A provide confirmatory evidence of the slightly modified marine water dolomitizing model. The suggestion of thermal convective flow of marine waters throughout the porous limestone edifice is a plausible explanation considering the timing of the dolomitization event, which occurred when the guyot was immersed in c. 2500 m of oceanic waters. The massive dolomite is a by-product of the guyot plumbing system. The complexities of dolomite formation are a topic of continuing debate (Fowles, 1991). In addi tion to the sabkha model (in which sea water supplies the magnesium), the Dorag model (where sea water and ground water mix), the burial model (where compaction of basinal sediments provides the magnesium-containing fluids), the thermal con vection (Kohout or endo-upwelling circulation) al lows deep oceanic water to provide the magnesium responsible for dolomitization. Further studies of the dolomite recovered from ODP Hole 866A are warranted because some of the variables (timing, water depth, temperature) oper ating during the dolomitization process may be constrained.
ACKNOWLEDGEMENTS CONCLUSION
Strontium isotopic (87Sr/86Sr) measurements from massive white sucrosic dolomite from ODP Hole
The opportunity to work with Co-chief scientists Ed Winterer and Will Sager and shipboard scientific staff of ODP Leg 143 is gratefully acknowledged.
Origin of white sucrosic dolomite
Also, critical advice provided by H.M. Arnaud, A. R. Chivas, J.A. Fagerstrom, J.A. McKenzie, F. Rougerie and V.C. Vahrenkamp has enhanced this paper. Financial assistance in 1993 was provided by the Australian Research Council.
REFERENCES AHARON, P., SocKJ, R.A. & CHAN, L. (1987) Dolomitiza tion of atolls by sea water convective flow: test of a hypothesis at Niue, South Pacific. J Geol. , 95, 187203. AISSAOUI, D.M., BUIGUES, D. & PURSER, B. H. (1986) Model of reef diagenesis: Mururoa Atoll, French Polynesia. In: Reef Diagenesis (Eds Schroder, J.H. & Purser, B.H.), pp. 27-52. Springer-Verlag, Heidelberg. ARNAUD, H.M., FLOOD, P.G. & STRASSER, A. (1995) Reso lution Guyot (Hole 866A) Mid-Pacific Mountains: facies evolution and sequence stratigraphy. In: Proceed ings of the Ocean Drilling Program, Scientific Results,143 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 133-157. Ocean Drilling Program, College Station, TX. BRAITHWAITE, C.J.R. (1991) Dolomites, a review of ori gins, geometry and textures. Trans. R. Soc. Edinburgh: Earth Sci., 82, 99-112. CuLLIS, C.G. (1904) The mineralogical changes observed in cores from the Funafuti borings. In: The Atoll of Funafuti (Ed. Bonney, T.G.), pp. 392-420. Royal Soci ety, London. FLOOD, P.G. & CHIVAS, A.R. (1995) Origin of massive do lomite, Leg 143, Hole 866A, Resolution Guyot, Mid Pacific Mountains. In: Proceedings of the Ocean Drilling Program, Scientific Results, 143 (Eds Winterer, E.L., Sager, W.W., Firth J.V. & Sinton, J.M.), pp. 161-169. Ocean Drilling Program, College Station, TX. FLOOD, P.G., FAGERSTROM, J.A. & ROUGERIE, F. (1996) Interpretation of the origin of massive replacive dolo mite within atolls and submerged carbonate platforms: strontium isotopic signature ODP Hole 866A, Resolu tion Guyot, Mid-Pacific Mountains. Sediment. Geol., 101, 9-13. FowLES, J. (1991) Dolomite: the mineral that shouldn't exist. New Sci., 132, 38-42. GUILLE, G., GOUTIERE, G., SORNEIN, J.F., BUIGUES, D., GACHON, A. & GuY, C. (1996) The Atolls of Mururoa and Fangataufa (French Polynesia). Oceanographic Mu seum, Monaco. HARDIE, L.A. (1987) Dolomitization: a critical view of some current views. J. sediment. Petrol., 57, 166-183. HARLAND, W.B., ARMSTRONG, R.L., Cox, A.V., CRAIG, L.E., SMITH, A.G. & SMITH, D.G. (1990)A Geologic Time Scale 1989. Cambridge University Press, Cambridge. HEIN, J.R., GRAY, S.C., RICHMOND, B.M. & WHILE, L.D. (1992) Dolomitization of Quaternary reef limestones, Aitutaki, Cook Islands. Sedimentology, 39, 645- 661. JENKYNS, H.C. (1995) Carbon-isotopic stratigraphy and pa leoceanographic significance of the Lower Cretaceous shallow-water carbonates of Resolution Guyot, Mid Pacific Mountains. In: Proceedings of the Ocean Drilling,
143
Scientific Results, 143 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 99-103. Ocean Drilling Program, College Station, TX. JENKYNS, H.C., PAULL, C.K., CUMMINS, DJ. & FULLAGAR, P.D. (1995) Strontium-isotope stratigraphy of Lower Cretaceous atoll carbonates in the Mid-Pacific Moun tains. In: Proceedings of the Ocean Drilling Program, Scientific Results, 143 (Eds Winterer, E.L., Sager, W. W., Firth, J.V. & Sinton, J.M.), pp. 89-97. Ocean Drilling Program, College Station, TX. JONES, C.E. (1992) The strontium isotopic composition of Jurassic and Early Cretaceous seawater. PhD thesis, University of Oxford. KENTER, J.A.M. & IVANOV, M. (1995) Parameters control ling acoustic properties of carbonate and volcaniclastic sediments at sites 866 and 869. In: Proceedings of the Ocean Drilling Program, Scientific Results, 143 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 287-303. Ocean Drilling Program, College Station, TX. KoHOUT, F.A. (1967) Groundwater flow and the geother mal regime of the Floridian Plateau. Trans. Gulf Coast Assoc. geol. Soc. 17, 339-354. KOHOUT, F.A., HENRY, H.R. & BANKS, J.E. (1977) Hydro geology related to geothermal conditions of the Florid ian Plateau. In: The Geothermal Nature of the Floridian Plateau (Eds Smith, D.L. & Griffin, G.M.), Spec. Publ. Fla. Bur. Geol., 21, 1-41. LAND, L.S. (1985) The origin of massive dolomite. J geol. Educ., 33, 122-125. McARTHUR, J.M. (1993) Strontium isotope stratigraphy in and out of ODP. UK ODP Newslett., 16, 8-9. McKENZIE, J.A., ISERN, A., ELDERFIELD, H., WILLIAMS, A. & SwART, P. (1993) Strontium isotope dating of palaeoeanographic and dolomitization events on the northeastern Australian margin, Leg 133. In: Proceed ings of the Ocean Drilling Program, Scientific-Results, 133 (Eds McKenzie, J.A., Davies, P.J. & Palmer-Julson, A.), pp. 489-498. Ocean Drilling Program, College Station, TX. PAULL, C.K., FULLAGAR, P.D., BRALOWER, T.J. & ROHL, U. (1995) Seawater ventilation of Mid-Pacific· Guyots drilled during Leg 143. In: Proceedings of the Ocean Drilling Program, Scientific Results, 143 (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 231-241. Ocean Drilling Program, College Station, TX. PRINGLE, M.S. & DUNCAN, R.A. (1995) Radiometric ages of basaltic lavas recovered at Sites 865, 866 and 869. In:
Proceedings of the Ocean Drilling Program, Scientific Results, 143. (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 277-283. Ocean Drilling Program, College Station, TX. RAo, C.P. (1996) Modern Carbonates, Tropical, Temper
ate, Polar: Introduction to Sedimentology and Geochem istry. Carbonates, Howrah, Tas. ROUGERIE, F. & FAGERSTROM, J.A. (1994) Cretaceous history of Pacific Basin guyot reefs: a reappraisal based on geothermal endo-upwelling. Palaeogeogr. Palaeoc/i matol. Palaeoecol. , 112, 239-260. ROUGERIE, F. & WAUTHY, B. (1988) The endo-upwelling concept: a new paradigm for solving an old paradox.
Proceedings of the 6th International Coral Reef Sympo sium, Townsville, Qld, 3, 21-26.
144
P. G. Flood
ROUGERIE, F. & WAUTHY, B. (1993) The endo-upwelling concept: from geothermal convection to reef construc tion. Coral Reefs, 12, 19-30. ROUGERIE, F., FAGERSTROM, J.A., & ANDRIE, C. (1992) Geothermal endo-upwelling: a solution to the reef nu trient paradox. Continent Shelf Res., 12, 785-798. SAGER, W.W., WINTERER, E.L., FIRTH, J.V. et al. (1993)
Proceedings of the Ocean Drilling Program, Initial Re ports. 143, Ocean Drilling Program, College Station, TX. SALLER, A.H. (1984) Petrologic and geochemical con straints on the origin of subsurface dolomite, Enewetak Atoll: an example of dolomitization by normal sea water. Geology, 12, 217-220. ScHLANGER, S.O. (1963) Subsurface geology of Enewetak Atoll. US geol. Surv. Prof Pap., 260-BB, 991-1066. SCHOLLE, P.A., ARTHUR, M.A. & EKDALE, A.A. (1983) Pelagic environments. In: Carbonate Depositional Envi ronments (Eds Scholle, P.A., Bebout, D.G. & Moore, C.H.), Mem. Am. Assoc. petrol. Geol., Tulsa, 33, 620691. SILBEY, D.F. & GREGG, J.M. (1987) Classification of dolo mite rock textures. J. sediment. Petrol., 57, 967-975.
SuN, S.Q. (1994) A reappraisal of dolomite abundances and occurrences in the Phanerozoic. J. sediment. Res., A64, 396-404. TuCKER, M.E. & WRIGHT, V.P. (1991) Carbonate Sedimen tology. Blackwell Scientific Publications, Oxford. VAHRENKAMP, V.C. & SWART, P.K. (1990) New distribu tion coefficient for the incorporation of strontium into dolomite and its implications for the formation of ancient dolomites. Geology, 18, 387-391. VAHRENKAMP, V.C., SWART, P.K. & RUIZ, J. (1991) Epi sodic dolomitization of late Cenozoic carbonates in the Bahamas: evidence from strontium isotopes. J. sedi ment. Petrol., 61, 1002-1014. WILSON, E.N., HARDIE, L.A. & PHILLIPS, O.M. (1990) Dolomitization front geometry, fluid flow patterns, and the origin of massive dolomite: the Triassic Latemar Building, northern Italy. Am. J. Sci., 290, 741-796. WINTERER, E.L. & SAGER, W.W. (1995) Synthesis of drill ing results from the Mid-Pacific Mountains: regional context and implications. In: Proceedings of the Ocean Drilling Program, Scientific Results, 143, (Eds Winterer, E.L., Sager, W.W., Firth, J.V. & Sinton, J.M.), pp. 497-535. Ocean Drilling Program, College Station, TX.
Spec. Pubis int. Ass. Sediment. (1998) 25, 145-161
Computer simulation of a Cainozoic carbonate platform, Marion Plateau, north-east Australia K. L I U*1, C . J . P I G RA Mt, L. P A TE RSON* and C. G. St C. KENDA Lq *Australian Petroleum Cooperative Research Centre, CSIRO Division ofPetroleum Resources, PO Box 3000, Glen Waverley, Vic. 3150, Australia; tMarine, Petroleum and Sedimentary Resources Program, Australian Geological Survey Organisation, Canberra, ACT, 2601, Australia; and tDepartment ofGeological Sciences, The University ofSouth Carolina, Columbia, SC 29208, USA
ABSTRACT The characteristics of the Marion Plateau carbonate platform in north-east Australia were simulated from the Early Miocene to the present using a depositional modelling computer program, SEDPAK. The simulation mimics the platform architecture and geometry seen on seismic lines and it supports an existing depositional model constructed from the study of the seismic stratigraphy and the sedimen tological and palaeontological data obtained from ODP cores. The simulation sequentially unravels the evolutionary history of the initiation, development and demise of the platform, in particular the two Miocene platform events (MP2 and MP3). The Early Miocene MP2 platform was initiated around 20 Ma during a sea-level rise. It evolved through four major platform-building phases in response to sea-level rises and highstands of third-order cycles. The growth of the MP2 platform was dominated by platform progradation. The Late Miocene platform (MP3) was initiated on the basinal facies of the MP2 platform during the Late Miocene lowstand around 1 0.2 Ma. The MP 3 platform developed through four platform-building stages during the Late Miocene second-order sea-level rise. It was dominated by platform aggradation. The MP3 platform was drowned in the early Pliocene (c. 4 Ma) by an abrupt tectonic pulse coupled with a sea-level rise. The MP2 platform was exposed during most of the Middle and Late Miocene ( 14-6 Ma) and was reflooded in the latest Miocene. It was drowned in the early Pliocene as a result of the tectonic pulse. Since the drowning in the early Pliocene, the Marion Plateau has largely remained in a bathyal environment and it is now covered by Pliocene-Holocene hemipelagic sediments. The simulation has shown that the architecture and geometry of the Marion Plateau carbonate platform can be modelled using the third-order cycles of the EXXON global sea-level curve. The platform-building phases are shown to have occurred during periods of sea-level rises and highstands. During periods of sea-level falls and lowstands, platforms were marked by hiatuses. The sea-level positions in the Late Miocene (from 10 to 6 Ma) in the Marion Plateau region are found to be over I 00 m lower on average than that suggested by the EXXON sea-level curve. This study has also demonstrated that sedimentary simulation is useful in testing seismic interpretations and quantitatively estimating the influences of various factors on carbonate platform evolution, namely carbonate production rate, sea-level, tectonic pulse, initial basin shape and depositional processes.
INTRODUCTION
(GBR) is probably the best known platform. Car bonate platforms are present throughout the region and have a history that extends back to at least the Eocene (Davies et a!. , 1989). The Marion Plateau is located between 1 8 S and c. 23 S off north-east Australia (Fig. 1). It is the most southerly of the marginal carbonate plateaux in north-eastern
Geological setting
The north-eastern region of Australia is a vast carbonate province where the Great Barrier Reef
o
1Present address: School of Earth Sciences, James Cook University, Townsville, Qld 4811, Australia.
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
145
o
146
K. Liu et a!. 15o•
152• E
E
154• E
CARBONATE PLATFORMS
TOWNSVILLE TROUGH
816 826
•
Dredge site
o
ODP drill site
CATO TROUGH
2o•s
22° s
0
100 km
Australia. The Marion Plateau is c. 77 000 km2 in area and the present plateau suiface forms a deeper extension of the Queensland continental shelf with water depths ranging from 100 m along its south western margin to 500 m along the northern and eastern margins. The plateau is bounded along its northern margin by the Townsville Trough, by the Cato Trough along the eastern margin, and by the central GBR to the south-west (Fig. 1). Tectonically, the Marion Plateau is situated on a passive continental margin. The Marion Plateau and the surrounding region is part of a complex upper plate margin (Symonds et a!., 1988). The Marion Plateau is structured along its margins and has the features characteristic of a marginal plateau formed on an upper plate by midcrust detachment, and lower-crust and mantle ductile extension (Pi gram, 1993). The Marion Plateau basement proba bly formed during Cretaceous extension that led to the opening of the northern Tasman Sea and Coral Sea basins (Mutter & Karner, 1980). It has under-
Fig. I. Locality map showing the bathymetric and geometric features of the Marion Plateau, the position of the Great Barrier Reef, the simulated seismic profile position AA' and ODP drilling sites (after Pigram et at., 1992).
gone c. 1000-m subsidence since it was first flooded about 25 Myr ago, and approximately half of the subsidence has taken place since the early Pliocene (Pigramet a!., 1992). The Marion Plateau was first systematically sur veyed in the 1980s by the Bureau of Mineral Re sources (BMR; now Australian Geological Survey Organisation) R.V. Rig Seismic. In 1990, the Ocean Drilling Program (ODP) Leg 133 occupied 16 sites in the region, three of which were on the north-western margin of the Marion Plateau (Fig. 1). The ODP Leg 133 objective was to examine the evolution of major carbonate platforms in a rifted passive margin set ting. The results of the project, together with the earlier seismic data, provide a vast data set on the nature and history of carbonate platforms in the re gion (Davies et a!., 1991), part of which forms the data base for this study. As carbonate platforms develop during the rising and highstand phases of a sea-level cycle, the com bined effects of eustatic sea-level fluctuations and
Computer simulation of a carbonate platform subsidence have a marked effect on the architecture of the platform, and determine whether the plat form aggrades, progrades, backsteps, or dies. As warm-water carbonate platforms commonly build to sea-level, they may therefore preserve a record of the location of relative sea-level at the time of platform development (Kendall & Schlager, 1981; Davies & Montaggioni, 1985). It is thus possible to construct a regional sea-level curve from carbonate platform study if it is possible to separate eustatic from tectonic influences. Forward computer simu lation provides an effective way to estimate such influences quantitatively by modelling the platform architecture and geometry, and reconstructing the platform-building events (Kendall et al., 1991; Schroeder & Greenlee, 1993). In this paper we employed a computer program, SEDPAK, to simu late the evolutionary history and the architecture of the Marion Plateau carbonate platform. The SEDPAK computer program
The SEDPAK program has been developed by the Stratigraphic Modelling Group of the University of South Carolina (Strobel et al., 1989; Cannon et al., 1994). SEDPAK is a two-dimensional simulation program that forward models the sedimentary fill of basins through time, using linear differential equa tions to represent geological assumptions (Cannon et al., 1994). The program simulates the develop ment of sedimentary basins in two dimensions by considering principally four major geological pro cesses: eustatic sea-level, tectonic movement, sedi ment accumulation, and the initial and evolving basin geometry. SEDPAK is capable of simulating the infilling of a sedimentary basin from two sides with either clastic or carbonate deposition, or a combination of both. The simulation starts from an initial basin surface geometry and models the dep ositional geometry in a series of time steps forward to the present-day configuration. The SEDPAK program has been applied in sim ulations of a number of depositional settings in clastic, carbonate and mixed depositional environ ments (e.g. Helland-Hansen et al. , 1988; Kendall & Lowrie, 1990; Aodulrahman & Kendall, 1991; Eberli et al., 1994; Liu et al., 1994), and its algorithm has been tested in varied situations. The carbonate algorithm within SEDPAK is par ticularly sophisticated. It mimics the growth of reefs and platforms as it happens in nature, and has the capacity to simulate various scenarios of carbonate
147
deposition with the fluctuations of relative sea-level and carbonate production rates. It is also able to mimic carbonate talus deposition, hardground de velopment, pelagic deposition, and the influence of wave, lagoon and clastic damping. Carbonate plat form modelling includes considerations of progra dation, downslope apron, keep up, catch up, backstep and drowned settings. SEDPAK invokes 15 combined input variables that influence the geometry and evolution of sediment accumulation for the specified time interval set by users (Cannon et al. , 1994; Eberli et al., 1994). In the simulation of the Marion Plateau carbonate platform, we used all variables except clastic deposition, winnowing and wave damping. Carbonate platform modelling and the EXXON global sea-level curve
Since its publication, the EXXON global eustatic sea-level chart (Haq et al., 1987) has drawn much attention regarding its validity and universal appli cability. Many people have taken the chart for granted, and applied it unconditionally in regional and global sequence stratigraphical correlation. Others, however, have expressed reservation and doubt about the reliability and the universal appli cability of the curve (e.g. Carteret al., 1991; Miall, 1991, 1992). It has recently been suggested that the EXXON global sea-level chart should be treated only as a working model rather than an accurate record in sequence stratigraphical correlation (Posamentier & Weimer, 1992), until more substan tial data on sea-level variations are documented in various continents. The warm-water Marion Plateau carbonate plat form provides an ideal case for testing the applica bility of the EXXON sea-level chart in the Australian region in the Miocene, because the carbonate plat form was formed in a tectonically relatively stable passive margin setting. Indeed, using seismic strati graphical analysis and backstripping techniques, Pigram et al. (1992) were able to estimate the ampli tude of the sea-level fall at the Mid-Late Miocene lowstand around 10 Ma on the Marion Plateau. Their result indicated that the sea-level fall at c. 10 Ma was at least 180 m, much greater than that suggested by the EXXON sea-level curve. With the help of high-resolution computer simu lations, we will quantitatively examine the applica- . bility of the EXXON sea-level curve in the Marion Plateau region in the Late Miocene by forward
148
K. Liu et a!.
modelling the evolution of the carbonate platform. The aims of this paper are twofold: to document the evolution of the Marion Plateau since the early Miocene, and to evaluate the applicability of the EXXON global sea-level curve in the Australian region in the Miocene. In particular, we attempt to: (i) reconstruct the evolution history of the Marion Plateau platform by simulating the platform archi tecture and the geometry of individual platform events; (ii) test the previous seismic interpretations and the depositional model; (iii) understand the influence of various geological factors on the car bonate platform development; (iv) construct a sea level curve that can best model the Marion Plateau platform evolution; and (v) demonstrate an alterna tive approach to seismic or sequence stratigraphical analysis.
sequences was subdivided into a number of seismic sequences. Pigram (1993) also recognized four ma jor platform-building events (MP1 to MP4) within the cover sequence, corresponding to the megase quences (Fig. 2). Because of the scope of this study, the discussion on the stratigraphy and sedimentol ogy of the Marion Plateau will be largely confined to the Miocene-Recent sequences, with particular ref erence to the two Early-Late Miocene platform events (MP2 and MP3). Detailed discussion on the seismic and sequence stratigraphy of the Marion Plateau region can be found elsewhere (Davies et a!., 1989, 1991; Pigram et a!., 1992; Pigram, 1993). Seismic sequences and platform events
The four seismic megasequences and platform events in the Marion Plateau cover sequence recog nized by Pigram (1993) correspond to the major depositional episodes on the Marion Plateau. They constitute the major part of a ?Late Oligocene Recent carbonate platform. The Early to Late Mi ocene platform events (MP2 and MP3) are particu larly distinct on seismic lines (Pigram et a!., 1992). These two platform events have been sampled by sea-floor dredging and ODP drilling (Davies et a!., 1991).
SEQUENCE STRATIGRAPHY AND SEDIMENTOLOGY OF THE MARION PLATEAU CARBONATE PLATFORM
Based on seismic stratigraphical analyses and ODP core study, Pigram (1993) recognized four seismic megasequences within the Marion Plateau platform cover sequence (Fig. 2; Table 1). Each of the mega-
.,
.,
.!!
Stages
a;
rn
�llAZZIAN
PLEIST. w
z
8
�
::J ...J
�
1l: "
CA
P-11 !AN
•�c
AlAN "'
ZANCLEAN MESSINIAN
z
w
u 0
�
"' " ::!;
5.2 6.3
TOATONIAN
15.2 16.2
20.0 AQUITANIAN
MSD
� :::w:a::: N17
-
� � NS
� � MP3b
MP3a
MSB
� � �
�
Pi:
MP2a
r-? MP1e
r-N4
P22
GHAniAN
?
f--
:..ilili..: ..1!.'1....
25.2 r--
� "
MSC
1-----P21
Upper MSA
Marion Plateau sea-level curve
E!!
0 c =., .. > ii:"
N20 N19
N12
BURDIGALIAN
1
0
.,
N16
SERRAVALLIAN
LANGHIAN
::!;
.,
"
"'" ., ., :=;" r::r .,
N22
3.6
10.2 w
..
" c 0 N LL 0..
MP1d
MP�c
f-MP1b
f-MP1a
Depth (m)
200
100
0
'
'-� 200m
(__
c:_
c_
Fig. 2. Stratigraphical column
(_ c.. c._ <..._ c._
showing the ages of the mega seismic sequences and platform events and a sea-level curve derived from the Marion Plateau platform. PF Zones, planktonic foram biochronozones. Modified after Pigram (1993).
Computer simulation of a carbonate platform
149
Table 1. Mega-seismic sequences and corresponding platform events recognized in the cover sequence of the Marion Plateau (modified after Pigram, 1 993) Megasequence
Sequence
Platform event
Maximum thickness
A
A I to AS
MP l
B
B l to B6
c D
Age
Palaeo-environment
>2000 m
?Late K to early M i ocene
Floodplain to clastic shelf and slope, cool and temperate carbonate shelf
MP2
I OOOm
Early to Middle M iocene
Warm-water carbonate platform
C l to C5
MP3
600 m
Late Miocene
Warm-water carbonate platform with outer shelf mounds
D l to D6
MP4
300 m
Pliocene to Recent
Variable; modern reefs to starved and winnowed plateau
MSA and the Early Miocene and older platform event (MP1) The MP1 platform in the upper part of MSA is in part a Lower Miocene cool to temperate water car bonate platform. The age of the MP1 platform is not known because it has not been sampled. On the basis of its stratigraphical position and the regional tec tonic history, Pigram (1993) suggested that the MSA probably has an age range between the latest Creta ceous and the earliest Miocene. The MP1 platform is mainly confined to the eastern and northern slopes of the Marion Plateau. On seismic lines it onlaps the plateau basement. This MP1 platform is not well documented, and was not particularly considered in the SEDPAK simulations. MSB and the Middle Miocene platform event (MP2) The MP2 platform (MSB) is seen on seismic lines across the Marion Plateau and has a thickness up to 1000 m (Pigram, 1993). It covers an area of c. 26000 km2. The bio-assemblages from this sequence comprise shallow, warm-water, carbonate sediments which are dominated by Halimeda, coral, coralline algae and larger foraminifer, indicating deposition under warm surface water conditions. Samples from the MP2 platform range in age from late Early Mi ocene to Middle Miocene (Zones N. 7 to N. l 0-12; Fig. 2). On seismic lines, the MP2 platform is char acterized primarily by platform margin prograda tion. The top of the MP2 platform was exposed for c. 8 Myr from the Middle Miocene to the Early Pliocene (end of Zone N.l 0 until the beginning of
Zone N. l 9-20; Fig. 3). The MP2 platform evolved through four major platform-building phases (MP2a-2d in Fig. 2; Pigram, 1993). It was finally drowned in the early Pliocene and is now overlain by Pliocene-Holocene hemipelagic sediments. MSC and the Late Miocene platform event (MP3) The MP3 platform (MSC) is primarily confined to the eastern and northern parts of the plateau. It varies considerably in thickness across the plateau, and has a maximum thickness of about 600 m. Samples from ODP drilling and dredging on the MP3 platform suggest that it is a Late Miocene (Zones N. l 5 to N. l 7) warm-water carbonate plat form (Shipboard Scientific Party, 199 l a). The MP3 platform lies directly on the hemipe lagic basinal facies of the MP2 platform. It is overlain by a thin layer of Pliocene-Holocene hemi pelagic sediments. On seismic lines this platform is marked by aggradation. Four stages of platform growth were recognized within the MP3 platform (Fig. 2). The platform build-up lasted for c. 6 My, and the platform was drowned in the early Pliocene. MSD and the Pliocene-Recent platform event (MP4) The MP4 platform (MSD) is a Pliocene-Holocene sediment layer that covers most of the plateau. It overlies directly both the MP2 and MP3 platforms. The MSD comprises primarily hemipelagic sedi ments and some Holocene reefs. The maximum thickness of MSD is about 300 m.
150
K Liu et al. Slope
East
MS MSD
t
Warm
Current
I
w
t
Ql
1ii ...J
RSL SWT Events
Exposed
� &J l2:d Shallow water platform S Slope 1---+---1 § Hemipelagic/pelagic c:i .!2> [ZJ Siliciclastic 5
Warm
I + t
MP2 platform Platform flooding
• Exposed
MSA
I
MP3 platform
Cool
+
Cool water platform MP1
I
Fig. 3. Chronostratigraphical diagram for the Marion Plateau with a summary of the major factors controlling development of these carbonate platforms during the last 25 Ma. Modified after Pigram (1993).
Depositional model for the Marion Plateau since the Miocene
The evolutionary history of the Marion Plateau since the Miocene is schematically depicted by the model shown in Fig. 4. This depositional model was constructed by Pigram (1993), based on seismic stratigraphical analysis and sedimentological and palaeontological information from dredged samples and ODP cores. Before the early Miocene, a cool to temperate water carbonate platform (MP1) existed at the then shelf margin (Figs 3 & 4a). This platform was confined to the eastern and northern slopes of the plateau and covered an area of 2000 km2. During the Early Miocene, warm surface waters entered the Marion Plateau region because of the combined effect of the northward movement of Australia (Davies et a!., 1987) and the development of boundary currents in the Pacific Ocean as a consequence of the closure of the Indonesian sea way (Kennett, 1980). The influx of warm waters coincided with the first flooding of the entire pla teau as a result of a relative rise in sea-level (Fig. 2). The change in the surface water temperature caused a rapid change in the carbonate-producing biota, and the rise in sea-level led to the initiation of the warm-water platform, MP2 (Fig. 4b). The MP2
platform evolved through four major phases, each apparently related to a third-order relative rise and highstand of sea-level followed by a relative fall (Fig. 2; Haq et a!. , 1987). During the first three phases of MP2 platform growth, the platform mar gin prograded an average of 10 km (maximum 16 km) into the Townsville Trough and an average of 20 km (maximum 30 km) across the eastern plateau towards the Cato Trough. The fourth phase is marked by aggradation and slight backstepping. The Late Miocene MP3 carbonate platform (Fig. 4c) was initiated during the Late Miocene lowstand maximum on the hemipelagic basinal facies of the MP2 platform. It evolved through four platform building stages during a major second order rise in sea-level (Fig. 2; Haq et a!., 1987). The MP3 platform was primarily characterized by aggradation (Fig. 4d). During the development of the MP3 platform, much of the Early Miocene MP2 platform was exposed and ceased growing (Fig. 3). The MP3 platform was abruptly drowned in the early Pliocene by a rapid rise in sea-level (Fig. 4e), possibly because of the combination of a tectonic subsidence pulse coupled with a rise in sea-level (Pigram et a!., 1992). The entire Marion Plateau was drowned as a result of this event, but the MP2 platform was unable to re-establish itself during the reflooding of the plateau.
151
Computer simulation of a carbonate platform During the late Pliocene and Quaternary, most of the Marion Plateau was under water. Apart from isolated Holocene reefs, the whole plateau is now covered by a thin layer of hemipelagic and pelagic sediments. Since the drowning in the early Pliocene, the Marion Plateau has remained largely in an upper bathyal environment. The plateau is now between 100 and 500 m below sea-level.
(a) PRIOR TO EARLY MIOCENE
Plateau exposed
(b) EARLY-MIDDLE MIOCENE
Plateau floodedMP2 platform deposited
(c) EARLY LATE MIOCENE
Sea level faiiMP2 exposed and MP3 initiated on basinal facies of MP2
(d) LATE MIOCENE-PLIOCENE
MP3 platform deposited
SEDPAK SIMULATION OF THE MARION PLATEAU
Determination of the input parameters
The variables required for the SEDPAK simula tions were primarily derived from analyses of seis mic and ODP data. Using the backstripping technique, Pigram et al. (1992) restored the deposi tional thickness of the MP2 and MP3 platforms at two sites on BMR seismic line 75/64. The same seismic cross-section was used for the SEDPAK simulation in this study (Fig. 5). The sedimentary facies and age information was primarily derived from the palaeontological data obtained from ODP drilling and sea-floor dredging (Pigram, 1993). The internal seismic velocity used for the time-depth conversion was derived from (i) examination of stacking velocities, (ii) seismic refraction studies of the pre-Holocene GBR basement, and (iii) ODP downhole logging (Pigram, 1993). The backstrip ping method and procedure were described in detail by Pigram et al. ( 1992). The following paragraphs briefly describe the variables used in the simulation.
(e) PLIOCENE-QUATERNARY
SL
Plateau drowned
Fig. 4. Schematic depositional history for MP2 and MP3 phases of the Miocene to Pliocene carbonate platforms on the Marion Plateau illustrating the lowstand nature of the initial phase of MP3 and its relationship with the MP2 platform. Modified from Pigram (199 3).
Initial basin surface The initial basin surface of the Marion Plateau along the seismic profile AA' at 25 Ma is shown in Fig. 6a. It was largely estimated from the structural contour map published by Pigram et al. (1992), after the removal of the platform cover sequence and water column by backstripping. The shape of the initial basin surface profile also reflects the overall passive margin tectonic setting in the Mar ion Plateau region. To minimize the marginal ef fects, the actual simulated profile was 200 km wide, although only 160 km of the profile was used to mimic the seismic cross-section shown in Fig. 5. Tectonic subsidence The Marion Plateau is known to have been sub-
sided 1000 m since 25 Myr ago (Shipboard Scien tific Party, 1991b). Approximately half of this subsidence has occurred since the early Pliocene (Pigram, 1993). If the subsidence rates remained the same throughout the two periods, the average subsidence rates for the Miocene and Pliocene Holocene would be 0.025 m kyc1 and 0.1 m kyc1 respectively. ODP drilling at Site 826 revealed that an early Pliocene tectonic subsidence pulse oc curred in the Marion Plateau. The structural con tour map of Pigram et al. ( 1992) suggests that tectonic subsidence in the plateau occurred along a hinge line parallel to the present Great Barrier Reef, with increasing subsidence towards the east and north-east.
·
152
K. Liu et a!. NW
0.0
SE
A
<;;
s.
� 05 >--
1.0
0
<;;E �
20
km
05
0.5
1.0
1.0
Fig. 5. Seismic profile AA' and its geological interpretation. Note the unique structural and stratigraphical relationship between MP2 and MP3 platforms. See Fig. I for the position of the seismic profile. From Pigram et a/. (1992).
The subsidence curves shown in Fig. 6b reflect the tectonic setting and subsidence history of the Marion Plateau. As the simulated seismic profile AA' is south-east of the tectonic hinge line, in creasing subsidence rates were assigned for loca tions away from the hinge line. To create the accommodation space required by the MP3 deposition, a relatively high subsidence rate (0.04 m kyc 1) was given for the Late Miocene
(a)
period compared with 0.014 m kyc' for the Early and Middle Miocene. A tectonic pulse was intro duced at the early Pliocene to honour the observa tion from ODP data (Fig. 6b). Eustatic sea-level Three sea-level curves (Fig. 7) were employed to simulate the Marion Plateau platform: (i) a locally
(b)
Basin Surface at 25 Ma
Subsidence Rates
0.00
I c 0
�
> ·w
A'
0
0.01
C?
1oo
..;,::
:s
200
225 km 0.02
Pulse
) 10km -
2 km
·-
-
170km
Cll
300
;;; a:
[iJ 400
0.03
500 0
50
100
150
200
Distance {km)
250
0.04
25
20
15
10
5
0
Age (Ma before present) Fig. 6. (a) The initial basin surface used in the Marion Plateau simulation. The actual simulated part is from 40 km to 200 km. AA' marks the approximate position of the simulated seismic profile. (b) Tectonic subsidence curves for various parts of the plateau simulated, for example, I I 0 km curve refers to the subsidence for the location at I I 0 km in (a). Arrow indicates the Early Pliocene tectonic pulse.
153
Computer simulation of a carbonate platform
0
Marion Plateau sea-level curve
Stochastic sea-level curve
EXXON sea-level curve
-5 (?
� Fig. 7. Three sea-level curves used in the final simulation of the Marion Plateau. The Marion Plateau sea-level curve i s essentially a modified EXXON sea-level curve. It has the same third-order cycles as the EXXON sea-level curve but its amplitudes between 10.2 and 5 Ma differ significantly from the EXXON curve.
Q) Ol <(
-10
-15
-20
100
0
-100
Depth (m)
derived Marion Plateau sea-level curve; (ii) the EXXON global eustatic sea-level curve (Haq et a!., 1987); and (iii) a stochastic sea-level curve. The Marion Plateau curve was derived from seismic stratigraphical analysis (Fig. 2). It is similar to the EXXON sea-level curve in terms of the relative positions and numbers of third-order cycles but differs significantly in the magnitude, in particular for the Late Miocene to Pliocene part (Fig. 7). The Marion Plateau sea-level curve was primarily based on the relative sea-level curve of Pigram ( 1993) (Fig. 2) but the amplitudes of the curve were ob tained by empirically adjusting during many pre liminary simulation runs aiming to match the platform architecture and geometry observed on the seismic profile AA'. The stochastic sea-level curve was generated with the same maximum and mini mum magnitudes as the EXXON sea-level curve, but with different numbers of cycles. It was de signed to test the validity and sensitivity of third order cycles in the EXXON sea-level curve in the simulation of the Marion Plateau platform.
-200
100
0
-100
-200
100
Depth (m)
0
-100 -200
Depth (m)
Fig. 8a honour decreasing carbonate production rates with increasing water depth (Schlager, 1981), as most of the carbonate sediment is produced by organisms that are dependent upon light. In the simulations, two carbonate production curves were used (Fig. 8a): carbonate productivity for the initial period before 20 Ma was assigned a much lower rate than that for the period since 20 Ma, because the sea-water temperature was probably lower dur ing the earliest Miocene period (see Fig. 3; Feary et a!., 1991). The carbonate production rates used in the simulation for the MP2 and MP3 platforms are comparable with those used by Eberli et a!. (1994) in the Bahama Bank simulations. Apart from carbonate-secreting organisms and chemical precipitation, a second source of sediment is the water column. SEDPAK uses a special func tion, pelagic deposition, to simulate this form of carbonate deposition. The pelagic rain varies as a function of time and comprises primarily mud-size carbonate. The pelagic deposition curve shown in Fig. 8b was derived empirically during preliminary simulation runs.
Carbonate accumulation rates The carbonate production rates for the Marion Plateau platform used in the simulation are largely speculative. There is not much information avail able regarding the growth rates through the Mi ocene to the present in the Marion Plateau region, although modern measurements from the Great Barrier Reef and Bahama Bank provide some con straints. The carbonate production curves shown in
Hardground or depositional lag time It is known from Holocene reef and carbonate platform studies that rapid sea-level rises suppress carbonate production (Schlager, 1981, 1991). This decrease in carbonate sedimentation results in a lag time before the platform resumes full production. To simulate such a natural scenario, the SEDPAK program has a special function, hardground, to
154
K. Liu et al.
(a)
(b)
0.6
E �
Carbonate Production
�
0.4
:§.. Q) £
g Q) 1il a:
0.2
0.0 -200
-100
-50
-10
0
Pelagic Deposition 0.006 0.004 0.002 0.000
+--�-�----�--.J
-25
-20
-15
Depth (m)
Q) .!!l Cii Q) a: Q) Cl ctl c Q) e Q)
-5
0
(d)
(c) "0
-10
Age (Ma)
20
Hardground
100 80
Lagoon Damping
15
60
.g>
·a. E
40
ctl 0
20
10
5
a..
10.0
20.0
30.0
40.0
50.0
Sealevel Change (m)
0.0
20.0
40.0
60.0
80.0
Distance (km)
Fig. 8. Input variables used in the simulation: (a) rates of carbonate production; (b) pelagic accumulation rates through time; (c) hardground plot; (d) lagoon damping plot.
simulate the development of hardground. This function (Fig. 8c) exerts additional suppression on carbonate growth rates when there is a rapid rise in sea-level. The hardground function allows a carbon ate margin first to fail when sea-level is rising too fast, and then catch up when the rate of sea-level rise is slowing down.
lagoon (Stockman et a!., 1967; Neumann & Land, 197 5). SEDPAK uses a special function, lagoon damping, to mimic this phenomenon. The damping effects decline away from reefs, which allows the creation of buildups at the platform margin. The lagoon damping curve shown in Fig. 8d was empir ically derived during preliminary simulation runs to mimic the platform architecture and geometry.
Lagoon damping Carbonate production is highest in areas of open circulation at the shelf margin, and decreases to wards the more restricted environments of the Table 2. Carbonate variables used for the Marion Plateau simulation Carbonate variables Talus or turbidite depositional angle Percentage to sea Percentage to talus Talus penetration distance (km) Turbidite penetration distance (km) Clastic suppression of carbonate (km)
20.0" 50% 50% 10 50 0.0
Carbonate variables The carbonate variables in SEDPAK exert con straints on the carbonate deposition; they pre determine the slope limit (angle of repose) of carbonate deposition on the platform slope, the distribution of excessive carbonate production, the penetration distance for talus and turbidite, and the impact of siliciclastic deposition (Table 2). SEDPAK simulation
Numerous preliminary simulation runs were car ried out before the final simulations to fine tune the input variables. These input variables were adjusted
Computer simulation of a carbonate platform interactively during simulation runs within the ranges given above, to obtain the best match of thick ness and facies types for each platform event, as well as the platform architecture and geometry observed on seismic lines. After the input variables were final ized, the evolution of the Marion Plateau was simu lated alorig seismic line 75/64 (Fig. 5) in three sepa rate final runs. During each simulation, all the input variables except for sea-level were kept the same. The three sea-level curves shown in Fig. 7 were em ployed in the simulations. The beginning of each simulation was set at the Early Miocene (25 Ma), and the simulation was continued to the end of the Ho locene. We simulated the carbonate platform evolu tion every 50 kyr through 500 time steps. The hori zontal resolution was kept at 0.5 km (equivalent to one horizontal X-step). Table 3 summarizes the in put variables used in the final simulations. Simulation with the Marion Plateau sea-level curve Plate 1 (facing p. 160) displays the final simulation outputs of the seismic cross-section Line 75/64 in the forms of sequence plot (Plate 1 a) and facies plot (Plate 1b) from SEDPAK using the Marion Plateau sea-level curve. The simulation results match the simulated seismic sections in terms of the platform, carbonate facies variations, the gross architecture of the platform and the geometry of individual plat form events (see Fig. 5 & Plate 1). Table 3. Summary of SEDPAK set-up and input variables used in modelling the Marion Plateau carbonate platform SEDPAK set-up and input variables
Run I, 2, 3
Time interval Time step Duration of time step Simulated horizontal distance Horizontal column Width of horizontal column Initial basin surface Eustatic sea-levels Tectonic subsidence rates Winnowing Clastic deposition Carbonate accumulation rates Pelagic deposition Hardground Lagoonal damping Wave damping Carbonate parameters Overburden
-25 Ma to 0.0 500 50 kyr 200 km 400 0.5 km See Fig. 6a See Fig. 7 See Fig. 6b Turned off Turned off See Fig. 8a See Fig. 8b See Fig. 8c See Fig. 8d Turned off See Fig. 3 Did not apply
155
As shown in Plate 2 (facing p. 160), the evolu tionary history of the Marion Plateau since the Early Miocene (c. 20 Ma) is sequentially modelled by the SEDPAK simulation. Each of the sequence plots in Plate 2 represents a particular platform evolution stage specified by its age. For comparison with the depositional model of Pigram (1993) shown in Fig. 4, only seven snapshots of the plat form evolution stages from the simulation are shown here (Plate 2a-g). The SEDPAK simulation, however, plots the evolution of the platform every 50 kyr. The colours on the sequence plots mark the depositional sequences that we have defined on the sea-level curve shown in Plate Ia. The MP2 platform is shown to have initiated at c. 20 Ma on an existing slope, possible the MP l plat form during a gradual sea-level rise (Plate 2a). The initial stage of the platform development from 20 Ma to 17.5 Ma was dominated by rapid platform progradation (Plate 2b). The basinward platform progradation during the first platform-building phase (20-17.5 Ma) was c. 20 km, comparable with the measurements from the seismic lines (Pigram, 1993). Between 17.5 and 15 Ma, however, the bas inward platform progradation slowed down signifi cantly with a total progradation less than 4 km (Plate 2c; Table 4). During the period between 14 to 12.5 Ma, most of the deposition occurred on the slope, and the MP2 platform was exposed around 14 Ma (Plate 2d; Table 4). The MP3 site remained below sea-level throughout the Early-Middle Miocene (20-10.5 Ma). Apart from a thin hemi pelagic sediment layer (Plate 1b), there was no sig nificant shallow-water platform carbonate accumu lation at this site. This is because the water was too deep to allow the initiation of any reef and platform carbonate. An abrupt drop in sea-level around the Middle Late Miocene lowstand (c. 10 Ma) resulted in the initiation of the MP3 platform (Plate 2e). The MP3 platform was developed directly on an existing topographic high which was covered by the hemi pelagic basinal facies of the MP2 platform. The growth of this platform was dominated by rapid aggradation during most of the platform-building phases between 10 and 5.5 Ma (Plate 2f & g). A lagoonal depositional environment developed be tween the MP2 and MP3 platforms during this period (Plate 1b). By 5.5 Ma the whole Marion Plateau was flooded by sea water, owing to the rapid rise of sea-level (Plate 2g). Deposition on the MP2 platform
K Liu et al.
156
Table 4. Statistics on MP2 and MP3 platforms obtained from SEDPAK simulation outputs Platform stages in SEDPAK simulations
MP2
Stage I Stage 2 Stage 3 Stage 4 ?Stage 5
MP3 Stage I Stage 2 Stage 3 Stage 4
Starting time (Myr BP)
Ending time (Myr BP)
Hiatus on platform crest (Myr BP)
20.8 20.8 17.4 16.1 15.2 14.0
10.8 17.7 16.7 15.6 14.0 10.2
17.7-17.4 16.7-16.1 15.6-15.2 14.0-10.2 10.2-6.0
10.2 10.2 7.6 6.2 5.0
4.8 8.8 6.7 5.8 4.5
26.0 21.0 2.0 0.5 1.5 1.0
Vertical aggradation (m)
30.0 35.0
115.0 150.0 10.0 50.0 45.0
8.8-7.6 6.7-6.2
probably resumed briefly in the latest Miocene and earliest Pliocene but the platform was unable to re-establish subsequently because of a rapid rise in relative sea-level caused by a tectonic pulse in the early Pliocene around 4 Ma (Fig. 6b). The Marion Plateau was drowned in the early Pliocene, and has since remained in a bathyal environment (Plate 1a & b). It is now covered by a thin layer of hemipelagic sediments. The simulation shows that both MP2 and MP3 evolved through four major platform-building phases (Plate 3a & b, facing p. 160). These phases were apparently related to third-order cycles in the EXXON sea-level curve (Haq et al. , 1987). Each of the platform-building phases occurred during a sea-level rise and highstand period. Except for the
A
Horizontal progradation (km)
Platform events recognized on seismic lines
MP2a MP2b MP2c MP2d MP2e
MP3a MP3b MP3c MP3d
contact between the MP3c and MP3d phases, the platform-building phases were interrupted by short periods of hiatuses apparently caused by platform exposure during periods of sea-level falls and low stands (Fig. 9). The simulation results indicate that the MP2 platform ceased growing around 14 Ma when sea level dropped slightly off the platform crest, and the platform remained exposed for about 8 Myr till the latest Miocene (c. 6 Ma), when it was reflooded (Fig. 9). The drowning of the MP3 platform was apparently due to a rapid sea-level rise primarily caused by the tectonic pulse introduced in the early Pliocene around 4 Ma (Fig. 6b). The MP2 was a platform initiated and developed during a second order highstand sea level cycle, whereas the MP3
A'
co
2 0. QJ
'iii
QJ
E --------� F
Distance (km)
Fig. 9. Chronostratigraphical diagram for the simulated part of the Marion Plateau platform showing the temporal and spatial relationship between the MP2 and MP3 platforms and periods of platform exposures (hiatus). The platform-building periods for each individual platform event are also indicated (MP2 a-d and MP3 a-d).
Computer simulation of a carbonate platform platform was initiated and developed during a second-order lowstand sea-level cycle (Plates 1a, 3a & b). Although the MP3 platform is structurally below the MP2 platform, stratigraphically, it over lies unconformably on the basinal facies of the MP2 platform (Plate 1a).
157
platform progradation and the filling of the shelf with shallow-water facies, and a large amount of excessive carbonate deposition on the slope and in the basin, whereas a low rate would cause carbonate platform backstepping and starvation because of the incapacity to fill the available accommodation space.
Simulations with alternative sea-level curves The simulation using the EXXON sea-level curve (Haq et a!., 1987) has reproduced the architecture and geometry of the Early Miocene MP2 platform but not that of the Late Miocene MP3 platform (Plates 4a & 5, facing p. 160). The simulation shows that the sea-level drop at the Middle-Late Miocene lowstand (c. 10 Ma) was not low enough to initiate the MP3 platform (Plate 5). The simulation with a stochastic sea-level curve failed completely to reproduce either the MP2 or the MP3 platform even in terms of gross platform thick ness and architecture, not to mention the details of various platform-building stages (Plate 4b). This sug gests that the carbonate platform growth is very sen sitive to the sea-level variations in both cycles and amplitudes. A comparison of the simulation outputs shown in Plates 1 & 4 indicates that the observed platform architecture and geometry in the Marion Plateau platform can only be best modelled using the Marion Plateau sea-level curve which has the same third-order cycles as the EXXON sea-level curve but with different amplitudes.
Sea-level Sea-level is shown to be one of the most important factors controlling the carbonate platform growth. As seen from the evolution of both MP2 and MP3 platforms (Plate 3), the platform-building phases occurred only during periods of sea-level rises and highstands, whereas a hiatus and/or erosion oc curred during sea-level falls and lowstands (Fig. 9 & Plate 3). Simulations with different sea-level curves (Plates 1 & 4) produced completely different plat form architecture and geometry. Accommodation space
Compared with the seismic interpretation (Fig. 5), the SEDPAK simulation using the Marion Plateau sea-level curve (Plate 1) has accurately reproduced the architecture and geometry of the Marion Pla teau carbonate platform. The results generally sup port the depositional model constructed by Pigram (1993) (Fig. 4). The simulation results have also provided some quantitative information regarding how various geological factors affect the initiation, development and demise of the Marion Plateau carbonate platform.
The simulation suggests that the carbonate platform is very sensitive to the accommodation space. A small change in accommodation space can result in an abrupt switch in the platform growth styles (i.e. progradation, aggradation or backstepping). During the initial stage of the MP2 platform development (stages a and b in Plate 3a) the combination of a gradual sea-level rise and a slow subsidence rate (c. 0.014 m kyr-1; see Fig. 6b) could not accommodate the carbonate production and thus resulted in rapid progradation of the platform. During the late stage of the MP2 platform development (stages c and d), relatively rapid sea-level rises resulted in an initial platform aggradation in both stages although pro gradation prevailed again during the subsequent static highstands (Plate 3a). During the first two stages of the Late Miocene MP3 platform growth (a and b in Plate 3b), a relatively high subsidence rate (c. 0.04 m kyr-1; see Fig. 6b) coupled with a gradual sea-level rise had created sufficient accommodation space to allow a rapid aggradation of the MP3 platform (Table 4).
Carbonate production
Tectonic subsidence
Carbonate production rates are found to have dictated the platform growth rate, carbonate facies variations and the geometry of the platform. A high carbonate production rate would result in rapid
Tectonic subsidence and the sea-level fluctuations create accommodation space. A rapid relatively · large-magnitude tectonic subsidence can cause the cessation of carbonate platform growth and thus the
Summary and discussions
1 58
K Liu et al.
drowning of carbonate platforms. As seen in the simulation, the introduction of the tectonic pulse in the early Pliocene resulted in the complete demise of the MP2 and MP3 platforms and the drowning of the Marion Plateau.
mimic the architecture and geometry of the MP3 platform, the Late Miocene part of the EXXON sea level curve would need to drop an average of over 100 m between 10.2 and 5.0 Ma (Table 5). The max imum drop would be 1 60 m around 8.0 Ma.
Initial basin surface
Justification on the SEDPAK simulations
The shape of the initial basin surface can affect the initiation of the carbonate platform. Both MP2 and MP3 platforms were initiated on local topographic highs (Plate 2). A slight alteration of the basin surface can cause a change in the initiation position and result in a different platform geometry in the simulation.
As discussed above, the simulation results generally support the seismic interpretation and the deposi tional model constructed by Pigram ( 1993). How ever, there is a slight difference in the timing of the reflooding of the MP2 platform. In the depositional model (Pigram, 1993), the re flooding of the MP2 platform occurred at least in the early Pliocene (c. 3 . 5 Ma) based on the presence of CN l l nanofossils in the sediments at ODP Site 826. The SEDPAK simulations, however, indicate that the reflooding may have occurred in the latest Mi ocene around 5 . 5 Ma. Sh�llow platform carbonate deposition may have occurred briefly on part of the MP2 platform between 5 . 5 and 4.0 Ma before the whole platform was drowned. This disagreement on the timing ofthe reflooding ofthe MP2 platform may be partly due to the difference in the palaeo elevations between ODP Site 826 and the simulated part of the platform (see Fig. 1 ): the simulated part of the platform was probably much lower than the ODP Site 826 and would therefore be flooded first. As the top of the MP2 platform has not been sampled (Ship board Scientific Party, 199 1 a), the age of the young est sediments in the MP2 platform in the simulated part cannot be verified. The SEDPAK simulation also shows that there was no apparent hiatus between the final two stages of the MP3 platform growth, with the MP3d plat form phase conformably overlying the MP3c plat form phase (Plate 3b & Fig. 9). This scenario is not consistent with the inferred unconformable contact between the two platform-building phases in the depositional model. To model this unconformable relationship, the sea-level fall at 5.4 Ma would need to be lower than the values used in this simulation. Clastic deposition was not included throughout the simulation. Its influence on the carbonate pro duction rates was regarded as minimal. However, Pigram ( 1993) suggested that clastic sediments may have been present in the western part of the Marion Plateau between 1 7. 5 and 14 Ma. If the clastic sediments did enter into the simulated part of Marion Plateau, some of the simulation parameters may need to be readjusted.
Depositional processes During the simulation, it has been found that both the lagoon damping and hardground functions within the SEDPAK program could significantly affect the platform architecture and geometry. The gross architecture and the facies of the Marion Plateau platform could not be reproduced when either of the functions was turned off. EXXON sea-level curve The prograding and aggrading geometries of the Marion Plateau seen on the seismic lines have been successfully reproduced using the third-order cycles similar to the EXXON sea-level curve (see Fig. 7). The numbers of platform-building phases observed from the Marion Plateau platform are the same as the numbers of the third-order cycles in the EXXON global sea-level chart (Figs 2, 7 & Plate 3). Although the timing of those third-order cycles cannot be verified by this study because of the inadequate res olution of the palaeontological data, the EXXON sea-level curve is shown to be a much better working model for the Marion Plateau region compared with a stochastic sea-level curve (Plate 4b). However, the simulation using the EXXON sea level curve itself did not reproduce the Late Miocene MP3 platform (Plate 4a). The simulation shows that to initiate the MP3 platform the sea-level fall at 1 0. 2 Ma lowstand would need to be c. 205 m (sea level changes between 10.8 and 1 0. 2 Ma; see Table 5). This sea-level fall is comparable with the estimation of Pigram et al. ( 1992). A comparison between the Marion Plateau and the EXXON sea level curves (Table 5) suggests that to accurately
159
Computer simulation of a carbonate platform
Table 5. Comparison o f sea-level amplitudes between the Marion Plateau and the EXXON sea-level curves. Numbers under sea-level columns are in metres below present sea-level. The last column (EXXON-MP) gives the amplitude differences between the EXXON and the Marion Plateau sea-level curves
Time (Myr) 1 1 .0 1 0.8 1 0.6 1 0.4 1 0.2 1 0.0 9.8 9.6 9.4 9.2 9.0 8.8 8.6 8.4 8.2 8.0 7.8 7.6 7.4 7.2 7.0 6.8 6.6 6.4 6.2 6.0 5.8 5.6 5.4 5.2 5.0
Time step
Marion Plateau sea-level
EXXON sea-level
EXXON-MP
280 284 288 292 296 300 304 308 312 316 320 324 328 332 336 340 344 348 352 356 360 364 368 372 376 380 384 388 392 396 400
55 40 - 1 25 - 1 55 - 1 65 - 1 55 - 1 45 - 1 35 - 1 25 -1 15 - 1 00 -1 10 - 1 20 - 1 50 - 1 70 - 1 85 - 1 70 - 1 60 - 1 35 - 1 05 - 1 05 -95 - 1 05 - 1 20 - 1 25 - 1 20 -30 -90 - 1 00 -95 5
55 40 -25 -70 -80 -70 -50 -25 -10 -5 0 0 -5 -15 -20 -25 -25 -20 -10 5 10 15 10 -25 -30 -25 40 -35 -45 -30 85
0 0 1 00 85 85 85 95 1 10 1 15 1 10 1 00 1 10 1 15 1 35 1 50 1 60 1 45 1 40 1 25 1 00 95 80 95 95 95 95 70 55 55 65 80
The subaerial erosion of the MP2 platform dur ing the Late Miocene sea-level lowstand was not simulated by SEDPAK, because no substantial evi dence of subaerial erosion by rivers, such as the presence of incised valleys and terrigenous sedi ments, was documented in the simulated part of the Marion Plateau during this lowstand (Pigram et a!. , 1992; Pigram, 1993). Subaerial erosion of the plat form surface by winds would have little influence on the palaeo-elevation and thickness of the MP2 platform, which are critical to the simulation. Limitations of the SEDPAK simulation SEDPAK simulation by itself can only provide necessary but not sufficient solutions to carbonate platform development. It can only mimic what
probably happened but by no means suggests that this is the only way it happened. SEDPAK is unable to simulate depositional processes in a dynamic way because the program uses linear differential equa tions to represent geological assumptions rather than the non-linear differential equations that model the dynamic depositional system (Cannon et a!. , 1994). Regarding the demise of the MP3 platform, for example, SEDPAK considers only one possibility, that is, the rapid rise in sea-level caused by the Early Pliocene tectonic pulse. There may be other causes for the demise of the MP3 platform at this particu lar time, such as a sudden intrusion of cool ocean current or a sudden change in the water chemistry (Pigram, 1993). However, because the early Pliocene tectonic pulse is known to have occurred
1 60
K. Liu et al.
and the carbonate facies types of the Late Pliocene Holocene cover sequence are known to be hemipe lagic to pelagic, to account for all the information, a rapid sea-level rise must have occurred in the early Pliocene. Therefore, there is a higher probability that the demise of the MP3 platform was caused by this event.
CONCLUSIONS
The SEDPAK simulation mimics the architecture and geometry of the Marion Plateau platform seen on seismic lines since the Miocene. It particularly accurately models the initiation, evolution and demise of the two Miocene platform events (MP2 and MP3). The simulation results support the dep ositional model constructed by Pigram (1993) based on the study of the seismic stratigraphy and the sedimentological and palaeontological data from ODP cores and dredged samples. The Early Miocene platform (MP2) was initiated around 20 Ma during a sea-level rise. It evolved through four major platform-building phases corre sponding to the periods of sea-level rises and high stands of third-order cycles. This platform event was dominated by a rapid initial progradation and a minor late-stage aggradation. It ceased growing around 14 Ma, when sea-level fell below the plat form crest. The Late Miocene platform (MP3) was initiated on the hemipelagic basinal facies of the MP2 platform during the Middle-Late Miocene lowstand maximum (c. 10.2 Ma). This platform developed through four platform-building stages during the Late Miocene second-order lowstand. It was dominated by platform aggradation. This plat form was eventually drowned in the early Pliocene (c. 4 Ma) by an abrupt tectonic pulse coupled with a sea-level rise. The MP2 platform was exposed during most of the Middle and Late Miocene from 1 4 to 6 Ma and was probably partly reflooded in the latest Miocene. It was subsequently drowned in the early Pliocene as a result of the tectonic pulse. Since the drowning in the early Pliocene, the Marion Plateau has largely remained in a bathyal environ ment and it is now covered by a thin layer of Pliocene-Holocene hemipelagic sediments at water depths of 1 00-500 m. The simulation suggests that the architecture and geometry and the numbers of platform-building phases observed in the Marion Plateau platform can be best modelled using the third-order cycles in
the EXXON global sea-level chart (Haq et a!. , 1987). However, the simulation indicates that the sea-level positions in the Late Miocene (from 10 to 6 Ma) in the Marion Plateau region were over 1 00 m lower on average than that suggested by the EXXON sea-level curve, with a maximum differ ence of 1 60 m. Carbonate production rates appear to be the most important factor that dictates the growth rates and the facies types of the Marion Plateau platform. Sea-level played an important role in the develop ment of the Marion Plateau platform. All the carbonate-building phases occurred during periods of sea-level rises and highstands. During periods of sea-level falls and lowstands the platform crest was marked by hiatuses. The carbonate build-up styles (i.e. progradation, aggradation or backstepping) are controlled by the availability of the accommodation space. Apart from the influences by the carbonate production rates, sea-level variations and tectonic subsidence, the simulation has also shown that carbonate platform deposition can be affected by the initial basin physiography and the carbonate depositional processes and parameters defined within SEDPAK, such as carbonate depositional slope limit, hardground and lagoon damping. In addition, this study has demonstrated that sedimentary simulations such as SEDPAK can be effectively used to: (i) model carbonate platform evolution; (ii) determine some quantitative infor mation regarding how various factors control the initiation, development and demise of carbonate platforms; and (iii) aid seismic and sequence strati graphical analyses.
ACKNOWLEDGEMENTS
We thank Bob Cannon and Phil Moore of the Stratigraphic Modeling Group, University of South Carolina, for their assistance and advice on the SEDPAK program. Chris Pigram thanks the Direc tor of AGSO for permission to publish this paper.
REFERENCES ABDULRAHMAN, A. & KENDALL, C.G ST.C. (1991) Creta ceous chronostratigraphy, unconformities and eustatic sea-level changes in the sediments of Abu Dhabi, United Arab Emirates. Cret. Res., 12, 379-401. CANNON, R., KENDALL, C.G.ST.C., MOORE, P., WHITTLE, G. & HELLMANN, D. (1994) SEDPAK 4. 0 Manual. The
Computer simulation of a carbonate platform University of South Carolina Stratigraphic Modeling Group, Columbia (unpublished). CARTER, R.M., ABBOTT, S.T., FULTHORPE, C.S., HAYWICK, D.W. & HENDERSON, R.A. ( 1 99 1 ) Application of global sea-level and sequence-stratigraphic models in Southern Hemisphere Neogene strata from New Zealand. In:
Sedimentation, Tectonics and Eustacy: Sea-level Changes at Active Margins (Ed. Macdonald, D.l.M.), Spec. Pubs. int. Assoc. Sediment., No. 12, 41-65. Blackwell Scientific Publications, Oxford. DAVIES, P.J. & MONTAGGIONI, L. (1985) Reef growth and sea-level change: the environmental signature. Proceed ings of the Fifth Coral Reef Congress, Tahiti, 3, 477515. DAVIES, P.J., SYMONDS, P.A., FEARY, D.A. & PIGRAM, C.J. (198 7) Horizontal plate motion: a key allocyclic factor in the evolution of the Great Barrier Reef. Science, 238, 169 7-1700. DAVIES, P.J., SYMONDS, P.A., FEARY, D.A. & PIGRAM, C.J. (1989) The evolution of the carbonate platform of northeast Australia. In: Controls on Carbonate Platform to Basin Development (eds Crevello, P., Sarg, J.F., Reed, J.F. & Wilson, J.L.), Spec. Pub!. Soc. econ. Paleonl. Miner., 44, 233-258. DAVIES, P.J., McKENZIE, J.A., PALMER-JULSON, A. et a!. (1991) Proceedings of the Ocean Drilling Program, Ini tial Reports, 1 3 3. Ocean Drilling Program, College Station, TX. EBERL!, G.P., KENDALL, C.G.ST.C., MOORE, P., WHITTLE, G.L. & CANNON, R. (1994) Testing seismic interpreta tion of Great Bahama Bank with a computer simula tion. Bull. Am. Assoc. petrol. Geol. , 78 , 98 1 -1004. fEARY, D.A., DAVIES, P.J., PIGRAM, C.J. & SYMONDS, P.A. (1991 ). Climatic evolution and control on carbonate deposition in Northeast Australia. Palaeogeogr. Palaeo climatol. Palaeoecol. , 89, 341-361. HAQ, B.U., HARDENBOL, J. & VAIL, P.R. (1987) Chronology of fluctuating sea-levels since the Triassic. Science, 235, 1156-1167. HELLAND-HANSEN, W., KENDALL, C.G.ST.C., LERCHE, l. & NAKAYAMA, K. (1988) A simulation of continental basin margin sedimentation in response to crustal move ments, eustatic sea level change and sediment accumu lation rates. J. math. Geol. , 20, 7 77-802. KENDALL, C.G.ST.C. & LowRIE, A. (1990) Simulation modelling of stratigraphic sequences along the Louisi ana offshore. Trans. Gulf Coast Assoc. geol. Soc. , 40, 355-362. KENDALL, C.G.ST.C. & SCHLAGER, W. (1981) Carbonates and relative changes in sea-level. Mar. Geol. , 44, 181212. KENDALL, C.G.ST.C., STROBEL, J., CANNON, R., BEZDEK, J. & BISWAS, G. ( 1 991) Simulation of the sedimentary fill of basins. J. geophys. Res., 96, 6911-6929. KENNETT, J.P. (1980) Paleoceanography and biogeo graphic evolution of the Southern Ocean during the Cenozoic, and Cenozoic microfossil datums. Palaeo geogr. Palaeoclimatol. Palaeoecol. , 3 1 , 1 23-152. L1u, K., PATERSON, L. & JIAN, F.X. (1994) Depositional
161
modelling of the Gippsland Basin and the Barrow Exmouth Sub-basins. APEA J. , 34, 350-365. MIALL, A.D. ( 1 991) Stratigraphic sequences and their chronostratigraphic correlation. J. sediment. Petrol. , 6 1 , 49 7-505. MIALL, A.D. (1992) EXXON global cycle chart: an event for every occasion? Geology, 20, 787-790. MuTTER, J.C. & KARNER, G.D. (1980) The continental margin off northeast Australia. In: The Geology and Geophysics ofNortheast Australia (Eds Henderson, R.A., Stephenson, P.L.), pp. 47-69. Geological Society of Australia (Queensland Division), Brisbane, Qld. NEUMANN, A.C. & LAND, L.S. (1975) Lime mud deposition and calcareous algae in the Bight of Abaco, Bahamas: a budget. J. sediment. Petrol. , 45, 763-786. PI GRAM, C.J. (1993) Carbonate platform growth, demise
and sea level record: Marion Plateau, Northeast Austra lia. PhD thesis, The Australian National University, Camberra, ACT, 322 pp. PIGRAM, C.J., DAVIES, P.J., FEARY, D.A. & SYMONDS, P.A. (1992) Absolute magnitude of the second-order Middle to late Miocene sea level fall, Marion Plateau, Northeast Australia. Geology, 20, 858-862. POSAMENTIER H.W. & WEIMER, P. ( 1 992) Siliciclastic se quence stratigraphy and petroleum geology-where to from here? Bull. Am. Assoc. petrol. Geol. , 77, 731-742. ScHLAGER, W. (1981) The paradox of drowned reefs and carbonate platforms. Geol. Soc. Am. Bull. , 92, 19 7-211. SCHLAGER, W. (1991) Depositional bias and environmen tal change-important factors in sequence stratigraphy. Sediment. Geol. , 70 , 109-130. ScHROEDER, F.W. & GREENLEE, S.M. (1 993) Testing eustatic curves based on Baltimore Canyon Neogene strati graphy: an example application of basin-fill simulation. Bull. Am. Assoc. petrol. Geol. , 77, 638-656. Shipboard Scientific Party (1991a) Sites 815, 816. In:
Proceedings of the Ocean Drilling Program, Initial Re ports, 1 3 3 (Eds Davies, P.J., McKenzie, J.A., Palmer Julson, A.A et a!.), pp. 243-344. Ocean Drilling Program, College Station, TX. Shipboard Scientific Party (19 9 1 b) Site 826. In: Proceed ings of the Ocean Drilling Program, Initial Reports, 1 3 3 (Eds Davies, P.J., McKenzie, J.A., Palmer-Julson, A.A et a!. ), pp. 805-810. Ocean Drilling Program, College Station, TX. STOCKMAN, K.W., GINSBURG, R.N. & SHINN, E.A. (1967) The production of lime mud by algae in south Florida. J. sediment. Petrol. , 37, 633-648. STROBEL, J., CANNON, R., KENDALL, C.G.ST.C., BISWAS, G. & BEXDEK, J. (1989) Interactive (SEDPAK) simulation of clastic and carbonate sediments in the shelf to basin settings. Computers and Geosciences, 15, 1279-1290. SYMONDS, P.A., DAVIES, P.J., BERNADEL, G., FEARY, D.A. & PIGRAM, C.J. (1988) Deep structure beneath the Towns ville Trough, Northeast Australian Continental Margin. In: Seismic Probing ofContinents and their Margins (Ed. Dooley, J.C.). Bureau of Mineral Resources, Geology and Geophysics Record 1 988/21.
Spec. Pubis int. Ass. Sediment. ( 1 998) 25, 1 63- 1 95
Quaternary and Tertiary subtropical carbonate platform development on the continental margin of southern Queensland, Australia J. F. MARSHALL* 1 , Y. T S U Jit2 , H. M AT S U DAt3, P. J . D A V I E St, Y. IRYU§, N. HONDAt4 andY. SATOHt *Australian Geological Survey Organisation, GPO Box 378, Canberra City, ACT, Australia 2601, tTechnology Research Center, Japan National Oil Corporation, 2-2 Hamada 1-chome, Mihama-ku, Chiba-shi 261, Japan, tDepartment of Geology and Geophysics, The University of Sydney, Sydney, NSW, Australia 2006; and §Institute of Geology and Palaeontology, Faculty of Science, Tohoku University, Aobayama, Sendai 980, Japan
ABSTRACT
The continental margin of eastern Australia extends from the tropics to temperate latitudes ( 1 o· -44 • S), and is one of the largest areas of modern carbonate sedimentation in the world. A joint survey carried out by the Japan National Oil Corporation and the Australian Geological Survey Organisation on the continental shelf off Fraser Island, southern Queensland, has delineated a distinct subtropical biotic assemblage, composed principally of coralline algae, forming a narrow zone between the tropical and temperate assemblages on the shelf. This assemblage is present both as surface sediments and build-ups. The coralline algae form thick crusts and rhodoliths, and are associated with low species diversity hermatypic corals and large benthic foraminifers, as well as bryozoans, Halimeda and molluscs. In addition, seismic reflection profiling and dredging of the outer shelf and upper continental slope has revealed that up to three carbonate platforms are present. Two platforms are considered to be Early to late Middle Miocene in age, whereas the overlying platform, beneath and seawards of the present Gardner Banks, is considered to be Quaternary. Dredged samples consist of a variety of carbonate rocks, in particular, coral-algal wackestone-packstone-boundstones and rhodolith float stone to rudstones. Dolomitic limestones and well-lithified limestones occur on the shelf edge and continental slope. These have diagenetic features suggesting both marine and meteoric environments. Recent studies have postulated that, during the Tertiary and Quaternary, carbonate platform devel opment off north-east Australia has been u nderpinned by the northward movement of the Australian plate, whereby tropical carbonate facies have developed through time as the continental margin has progressively moved to lower latitudes. An essential consequence is that subtropical and temperate carbonate facies are contemporaneous with these tropical facies, but displaced further south. Our studies show this drift hypothesis to be correct. The Fraser Platform is a large subtropical carbonate platform that developed at the same time as the Early to Middle Miocene tropical platform beneath the Marion Plateau to the north, whereas the Quaternary platform beneath and seawards of Gardner Banks is the subtropical equivalent of the Great Barrier Reef. INTRODUCTION
almost exclusively concentrated on tropical envi ronments such as the Bahamas, Florida, Belize, the Persian Gulf, Pacific atolls and the Great Barrier
Detailed models of carbonate depositional environ ments developed since the 1950s and 1960s have
1 Present address: Research School of Earth Sciences, Australian National University, Canberra City, ACT, Australia 260 1. 2 Present address: Geological Survey Dept., Japan Na tional Oil Corporation, 2-2-2 Uchisaiwai-cho, Chiyoda ku, Tokyo 100, Japan.
3 Present address: Department of Earth Sciences, Kuma moto University, 2-39-1 Kurokami, Kumamoto 860, Japan. 4Present address: Cosmo Oil Co., Ltd, 1 - 1 , Shibausa 1 -chome, Minato-ku, Tokyo 105, Japan.
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
163
164
J. F Marshall et a!.
Reef. As late as 1983, definitive texts such as the American Association of Petroleum Geologists Memoir 33 on carbonate depositional environ ments (Scholle et al., 1983) used only tropical examples as modern analogues (even though some of the ancient analogues are distinctly non-tropical). Conversely, studies of carbonate sediments outside the tropics have received scant attention from carbonate sedimentologists. This is despite the fact that sites of extensive carbonate production were reported from non-tropical areas over 25 yr ago (e.g. Conolly & von der Borch, 1967; Wass eta/. , 1970). Recently, an increasing amount of attention has been paid to cool water carbonates because of their distinctive composition; as compared with tropical carbonates (Lees & Buller, 1972; Marshall & Davies, 1978); their detrital (in general) rather than biohermal depositional regime (James et a/., 1992); and their recognition in ancient limestones (Nelson, 1988; James, 1990). Much of the study of cool water carbonates has occurred in the southern hemisphere, particularly from shelf sediments around New Zealand (Nelson, 1988; Nelson eta/. , 1988) and southern Australia (Collins, 1988; James & Bone, 1991; James & von der Borch, 1991; James et al., 1992; Bone & James, 1993; Boreen & James, 1993; Boreen et al., 1993). Bryozoans are com monly the most abundant bioconstituent of these cool water carbonates, often forming 20-50% of the total sediment (Conolly & von der Borch, 1967) and sometimes as much as 70% (Marshall & Davies, 1978). Although there is a greater acceptance these days of the importance of cool water carbonates, and an increasing amount of effort is being devoted to these deposits in relation to tropical carbonates, there is one area of carbonate sedimentation that has been almost entirely neglected. This is the subtropical to warm-temperate region, essentially the interface between tropical and cool water car bonates, which at present occurs roughly between the latitudes of 24 and 30°. However, although the Bahamas and the Florida Keys fall into this zone, the effect of the Gulf Stream maintains a tropical carbonate assemblage at these latitudes in the west Atlantic. One of the few areas where subtropical to warm-temperate carbonate assemblages have been studied in any detail is the Ryukyus Islands. Here, the outer shelf and upper slope are dominated by coralline algae, mainly in the form of rhodoliths (lryu, 1985; Matsuda, 1987; Tsuji et al., 1989; Tsuji, 1993), and similar assemblages are present in the emergent Pleistocene limestones of the islands o
(Konishi et al., 1970; Minoura & Nakamori, 1982; Noda, 1984a,b; Yuki et al., 1988). Marine carbonate sediments are extremely sensi tive indicators of the environment in which they were formed; either by biomineralization (skeletal carbonates) or inorganic(?) precipitation (ooids, cements). Changes in sea-level and climate are often recorded in these accumulations. Some of the stron gest signals of environmental change should be encapsulated in carbonate sediments deposited in marginal climatic environments, i.e. in the subtrop ical and warm-temperate regions, where profound environmental change may see a switching between tropical and temperate carbonate deposition. The continental shelf of eastern Australia, mantled by both tropical and temperate carbonate sediments (Marshall & Davies, 1978), is an ideal location in which to study the transition between these two dominant carbonate environments. The passive continental margin of eastern Austra lia extends for nearly 4000 km from the tropics ( l OoS) to temperate latitudes (44 OS). It forms one of the largest areas of shallow-water carbonate deposi tion in the world, and includes the Great Barrier Reef (Fig. 1) The architecture of the margin has been primarily defined by Late Cretaceous rifting and its sedimentary evolution owes much to the northward motion of the Australian plate during the Cainozoic (Symonds et al., 1983; Davies et al., 1987, 1989). The Great Barrier Reef represents the major feature on the margin, extending for 2000 km and consisting of some 2500 individual reefs. It is widest in the central-south area around latitude 2 1 s where coral reefs and Halimeda-rich sedi ments dominate the outer shelf. To the south of 24 S, surface reefs are absent, and the sediments on the outer shelf are largely composed of coralline algae and bryozoans, with ubiquitous foraminifers and molluscs. These sediments are bioclastic and are dominated by calcite rather than aragonite. The facies and mineralogical attributes clearly reflect the change from tropical warm water to temperate cool water (Marshall & Davies, 1978). This paper is the result of a joint research project carried out by the Technology Research Center of the Japan National Oil Corporation and the Austra lian Geological Survey Organisation (formerly the Bureau of Mineral Resources) on the continental shelf of eastern Australia (Davies et a/., 1992; Tsuji et al., 1994). One of the major objectives of the project was to compare and contrast carbonate depositional facies on the shelf, ranging from tropo
o
,
165
Subtropical carbonate platform development
� •
60-80%
>80%
AUS
Fig. 1. Distribution of carbonate sediments on the continental shelf of eastern Australia (from Marshall & Davies,
1 978).
166
J. F.
Marshall et a!.
ical to temperate environments, and to determine the effects of changes in climate and sea-level on their distribution. The focus of this paper is the transitional zone between tropical and temperate carbonates that occurs on the continental shelf offsouthern Queensland between Fraser Island and Noosa Heads (Fig. 2), directly to the south of the Great Barrier Reef. Regional geology
The continental margin of eastern Australia was formed by the opening of the Tasman Basin in the south between 80 and 60 Myr ago (Hayes & Ringis, 1973) and the Coral Sea Basin in the north between
25'S
26'S
o
o,
Hervey Bay
I
Noosa Head
27'S
65 and 55 Myr ago (Weisse! & Watts, 1979). In the Tasman Basin, sea-floor spreading occurred along a NNW-trending ridge axis, offset by right-lateral transform faults, which separated the Lord Howe Rise from the Australian continent. The south-east Australian margin is unusually steep. The continen tal slope has gradients of 8 -12 and locally may be as steep as 20°. Jongsma & Mutter (1978) attributed this steep slope to the absence of rift stage tectonic elements typical of most Atlantic type margins. They suggested that the entire pre breakup rift basin remained attached to the Lord Howe Rise during the formation of the Tasman Basin. Etheridge et a!. (1990) applied the detach ment model of Wernicke (1981, 1985) to recon-
t
..._
l
c:: <2J
L... L... .::J
0 co co
....
(/) ::J <( (/) (1j w 50km
Fig. 2. Location map of study area off southern Queensland.
Subtropical carbonate platform development struct the passive margin of south-east Australia, which they interpreted as an upper plate margin. The main geological features of the region have been deduced from geophysical surveys both off shore and onshore, geological mapping onshore and well information. The study area is underlain by the Maryborough Basin (Ellis, 1966), an intracratonic basin that covers an area of about 10 000 km2 onshore together with at least 15 000 km2 offshore. It consists of a thick sequence of sediments and some volcanic rocks that are latest Triassic to Early Cretaceous in age. The offshore part of the basin has not been tested by drilling, but a stratigraphical hole �as drilled at the northern end of Fraser Island (GSQ-Sandy Cape 1-3R; Grimes, 1982). The drill hole penetrated 420 m of late Tertiary and Quater nary sandy marine shelf deposits, I 72 m of mid Tertiary interbedded basaltic volcanic rocks, and marine, deltaic and fluvial sediments. The bios tratigraphy shows that Late Oligocene and Early Miocene sediments are unconformably overlain by latest Miocene and younger carbonates and sands (Palmieri, 1984). These sequences can be correlated with the Oligocene and younger rocks in Wreck Island I, and the Capricorn lA and Aquarius 1 wells in the Capricorn Basin further north. At the base of the Sandy Cape I-3R hole are sandstones and shales, which are considered to be correlatives of the Late Jurassic Tiaro Coal Measures. This suggests that the Cretaceous sequence is absent beneath the shelf in this area. Marine geological studies of this part of the shelf have concentrated on the surficial sediments and the shallow (< I s two-way travel time (TWT)) stratigraphy and structure (Marshall, 1977, 1978, 1979, 1980). These data show a seaward-thickening wedge of sediment above a prominent bedrock reflection (S2 of Marshall, 1978, 1979), except off Fraser Island where a strong reflector crops out on the outer shelf-upper slope. Dredged material re covered from a depth of 293 m off Indian Head consists of conglomeratic material. The rounded ferruginous pebbles and cobbles are considered to be calcareous algae of Tertiary age, whereas the carbonate matrix has been dated as Pliocene to Recent (Marshall, 197 1). The sediment distribution on the shelf shows low-carbonate sands inshore becoming increasingly carbonate rich offshore. The major carbonate components are coralline algae, forams, molluscs and bryozoans. Marshall & Dav ies (1978) have shown that the highest concentra tion of coralline algae on the continental shelf of eastern Australia occurs on the outer shelf between
167
Fraser Island and Brisbane. Off Fraser Island, the coralline algae are encrusting, and have bound other skeletal fragments to form banks or hardgrounds (Marshall, 1980). Morphology
Fraser Island, 240 m high, lies adjacent to the study area. It is composed almost entirely of siliceous sand and is the world's largest sandy island. It extends from the mainland in a north-easterly direction across the shelf to Breaksea Spit (Fig. 2), effectively bisecting the shelf and forming the low energy leeward shelf of Hervey Bay to the west and the high-energy open shelf to the east. The width of the shelf is fairly uniform, but because of the orientation of Fraser Island, the high-energy shelf varies from 22 km in the north to about 70 km to the south. Offshore from Fraser Island, the very flat inner shelf extends to a depth of about 45 m, sloping at about 0.2 whereas to the south off Noosa Heads it has an even lower gradient. The mid-shelf extends from about 45 m to around 100 m. Numerous banks and hardgrounds are present on the mid shelf. A series of banks, the Gardner Banks, lies some 16 km east of Indian Head, extending for some 30 km parallel to the I00 m isobath (Fig. 3). Individual banks are about 6 km wide and rise from depths of 60 m to around 24 m. The Gardner Banks, like others in this area, have a gently sloping windward side and a relatively steep leeward side. A series of low-relief (5-15 m) banks are present to the east of the larger banks. These have a generally rugged topography, and have been inter preted as drowned reefs that, during the post-glacial transgression, colonized what were originally low stand aeolian ridges (Marshall, 1977). A shallow linear depression on the mid-shelf, which extends for some 7 5 km along the shelf, forms a fairly subdued feature several kilometres wide, but not more than 4-5 m deep (Jones, 1973). These shallow depressions, and others further south, represent a drowned system of coastal swamps and lagoons similar to those behind present-day beaches. The boundary between the mid-shelf and outer shelf is usually well defined. It occurs at the edge of a prominent terrace or nick point at a depth of I 05 m. Off Fraser Island the nick point is often backed by a small cliff, about 20 m high. The outer shelf forms a gently sloping plain with gradients of 1-3 Between Sandy Cape and Indian Head (Fig. 3) the outer shelf levels off somewhat, forming o,
o .
168
J. F.
Marshall et a!.
. ""Ga'peJ?"'\)
10 km
24'40'
0 Grab site H Dredge site
24'50'
25'00'
!/ / ( '
/
25'10'
0 /
I I
I I I
/
1 o
o
153'20'
25'20'
23/G56/46
Fig. 3. Location of sa mpling stations off Fraser Island.
a slightly concave platform of 6-10 km width, referred to here as the 'outer shelf terrace.' This terrace is higher along its seaward edge, forming a prominent mound, before descending steeply down the slope.
The shelf break off southern Queensland is much deeper than most in the world (Shepard, 1963), varying between 210 and 450 m. The morphology and origin of the shelf break off eastern Australia has been discussed by Jones et a!. (1975), who
Subtropical carbonate platform development concluded that deep shelf breaks are unrelated to eustatic processes, but are controlled by the config uration of the underlying basement surface. Beyond the shelf break the continental slope is relatively steep and in places has been deeply incised by submarine canyons. Two canyons are present on the continental slope off Fraser Island (Marshall, 1972), one at about 25' 30'S, the other off the south ern end of the island at 25'50'S. A complex of can yons, the Noosa Canyon System (Marshall, 1978), occurs on the slope between 26' l O'S and 26'20'S. These head at a depth of about 220 m, although it is possible that part of the canyon system has cut back into the outer shelf. The Moreton Canyon cuts the slope south of the Noosa Canyon System, at about 26'36'S. Evidence of terraces and nick points is wide spread on the shelf in this region (Jones, 1973; Marshall, 1977). However, the best preserved and most continuous terrace is that at 105 m, marking the boundary between the mid- and outer shelf. The widespread preservation of the 105 m terrace in this region suggests that it represents a significant low stand of sea-level on the east Australian shelf. However, it is higher than a terrace at 165 m off the Capricorn-Bunker Groups to the north, dated at 13 600-17 000 yr BP (Veeh & Veevers, 1970) and another at 127 m, dated at 17 900±600 yr BP, off central New South Wales to the south (Phipps, 1970). Climate and oceanography
The east coast of Australia is affected by three main pressure systems: the tropical low-pressure zone, the subtropical high-pressure zone and the subpolar low-pressure zone. The subtropical high-pressure zone dominates along the southern Queensland coast. During winter its axis is at about 25'S, and in summer at about 35'S. As the high-pressure zone moves north during the winter it is replaced by the subpolar low pressure zone. This forms a series of depressions which move across the southern part of the continent in an easterly direction. The regional climate is subtropical, with distinct winter and summer seasons and wind patterns. Winds are predominantly from the south-east dur ing the winter months and are consistently strong. The period September-December (austral spring) is consistently calmer. The summer tendency is for infrequent very strong winds. The effects of tropical cyclones are commonly felt in the region during
169
summer, with higher than average rainfall and large swells. The oceanic circulation system off eastern Aus tralia is dominated by the East Australian Current (Wyrtki, 1960, 1966; Church, 1987), a strong nar row southerly flow that forms the western boundary current of the subtropical gyre in the South Pacific. The East Australian Current (EAC) sweeps over the continental shelf between 20'S and 25'S, and becomes narrower and deeper, before turning southwards. The current is present at all times of the year, but is usually strongest between December and April (Hamon, 1965). From January to March it is supplied by equatorial water masses from the north and north-east, and from April to September by subtropical water masses from the east (Wyrtki, 1966). Circulation is somewhat affected between the southern end of the Great Barrier Reef and Fraser Island because of the large embayments of the Capricorn Channel and Hervey Bay. Here the current forms a series of gyres, but the major part of the current flows southwards. The EAC appears to narrow off Fraser Island (Fig. 2), with surface flows in this region reaching up to 4 knots along the edge of the shelf. It is possible that the current is controlled by the width of the shelf and the morphology of the outer shelf-upper slope in this region. By 27'S the cur rent flows along the edge of the shelf and continues in this manner to about 32'S (Boland & Hamon, 1970), reaching its maximum velocity and strength near Cape Byron (28'25'S). South of 33'S the current tends to separate from the coast, and often returns towards the north or north-east, whereas a series of warm core, anticyclonic eddies of c. 250-km diameter form and move generally south ward. The tides along the Queensland coast are pre dominantly semidiurnal and there is a progressive increase in tidal range towards Broad Sound in central Queensland. The effects of tidal currents are virtually unknown in the study area, but the range at the coast is about 2 m. As a result of the dominant south-easterly waves and swell, the overall direction of longshore drift along the coast is northwards. Longshore currents operate from Noosa Heads along the entire length of the coastline and the eastern side of Fraser Island (Fig. 2), possibly continuing along the shallower parts of Breaksea Spit.
170
J. F.
Marshall et al.
RESULTS
Seismic data
Method Some 450 km of multichannel seismic reflection data were acquired using 80-in3 Seismic Systems Inc. S-80 Water Guns in an array of five guns, each operated at 2000 psi (62 I min-1 at 14 MPa) from up to six A-300 Price air compressors, and a Fjord Instruments transformerless seismic receiving ar ray, configured as 6.25-m group lengths with 96 channels and 600-m active streamer length. The sample rate was 1 ms with a record length of 3 s. The data were acquired on AGSO's own designed and built digital acquisition system and recorded on 6250 bpi tapes in SEG-Y format. Vessel speed was 5 knots and shot interval 5 s. Cable depth was 5 m,
controlled by seven Syntron RCL-3 cable levellers. and gun depth 3 m. Near offset was 52 m, whereas maximum offset was 646 m. After geometry definition, the field tapes were re formatted to an internal Disco format and the data resampled to 2 ms. A common mid-point sort was applied to the data, and velocity analysis conducted for every 25 cdps using an interactive system. After normal moveout correction, a median stack was applied (24-fold). After filtering and automatic gain control (AGC) the data were migrated. In addition to the multichannel data, some 400 km of shallow, high-resolution seismic (boomer) data were acquired along and between the E-W mul tichannel lines, and a 50-km grid of good-quality side-scan sonar was run over the central to northern part of the Gardner Banks and the boundary of the mid- and outer shelf to the east. The high-resolution seismic data were acquired using an ORE boomer
Line 105/21 w-
105/15
-E
Fig. 4. Seismic line showing the various units and carbonate platforms on the outer shelf and upper slope otf Fraser Island.
Subtropical carbonate platform development plate, mounted beneath a small PVC catamaran, and powered by a 400-joule Geopulse power supply. The receiving array was a Benthos multi-element single channel streamer. The data were recorded in ana logue form. The sidescan data were acquired using an EG&G Model 990 sonar system.
171
The positions of seismic lines off Fraser Island are shown in Fig. 4. Three seismic units (Fig. 4) have been identified in the multichannel seismic data.
and to the west, Unit changes from a distinctly bedded series of sequences to a massive, poorly reflecting unit (Figs 4 & 5). The lower sequence boundary becomes indistinct or is lost in the multi ple, but both boomer and multichannel data indi cate that a carbonate build-up (Platform 3 in Fig. 4), extending from the I 05-m cliff to the Gard ner Banks, is present in the shallow subsurface. The I 05-m cliff appears to form the eastern edge of the build-up, and explains the rugged relief of this feature; such a cliff is not observed elsewhere on the shelf at this depth.
Unit I. This unit occurs beneath the mid- and outer shelf. Seaward of the ! 05-m cliff, it consists of at least three seismic sequences (Fig. 5), identified through downlap and onlap terminations. The se quences consist of a series of downlapping packages between the sea-floor and the lower boundary. The boundary between each sequence is an erosional unconformity. All three sequences downlap the lower sequence boundary (Fig. 5), which eventually merges as the sea-floor to the east. The three sequences can be interpreted as reflecting successive periods of outbuilding, presumably during high sea-level events. Their geometry suggests that out building has been restricted by the EAC, whereby the current has continually winnowed sediment from the outer shelf. Although winnowing would be more pronounced during periods of relative low sea-level, current meter measurements recorded from the outer shelf indicate that it is also occurring at the present time (Harris et a!., 1996). Beneath the junction of the mid- and outer shelf,
Unit 2. In this unit, often only the upper sequence boundary is visible so that its thickness is largely unknown. The upper sequence boundary is flat lying beneath the outer shelf terrace and forms a steeply inclined (up to 300) upper slope (Fig. 4). Although it is difficult to trace the upper unit boundary to the west beneath Unit I, it appears to coincide with the erosional unconformity at the base of that unit. The seismic characteristics allow Unit 2 to be divided into two non-reflecting sec tions, either at depth or at either end of the outer shelf, and a section that shows strong reflection characteristics, usually between the non-reflecting sections (Figs 4 & 6). The non-reflecting sections appear to represent the development of carbonate platforms. One platform (Platform I) occurs be neath the edge of the shelf and crops out on the upper slope (Fig. 4). The other platform (Platform 2) occurs leewards of Platform I. In some sections (Fig. 4) the two appear to be connected, whereas in others (Fig. 6) they are distinct. In the latter
Seismic stratigraphy
West
Fig. 5. Part of a seismic line showing the main attributes of Unit l, including the three downlapping sequences in front of Platform 3.
East
1km
J. F. Marshall et al.
172
Line
105/27
Fig. 6. Part of line 1 05/27 showing the separation of Platforms I and
2, with a prograding and downlapping periplatform sequence between them .
situation, there is a sequence of bedded sediments between the two platforms that possibly represents contemporaneous shallow water or periplatform deposition. Some evidence of foreset bedding is apparent in front of Platform 2, whereas there appears to be downlap of sediments onto the lee ward side of Platform 1 (Fig. 6). Unit 3. This upper slope sequence is poorly devel oped in many places, where only a thin (< I 00-m) unit is present. Its thickness appears to be con trolled locally by the underlying 'basement' topo graphy. In places where the basement is steep, Unit 3 is thin and shows chaotic terminations, erosion and slumping. Where the basement is stepped, there is a much larger build-up of Unit 3, but the outer reflections again tend to be chaotic and indicate slumping. Sampling results
In addition to the seismic data, a total of 42 grab and 14 dredge samples were collected from the study area. The locations of these samples are shown in Fig. 3. The sediment data were augmented by 1 8 underwater camera stations, to examine the distinct biological and hydrological features of the shelf. The results from the sample analysis have been divided to reflect their relevance to carbonate platform development through time, based on the location of samples with respect to the seismic interpretation and the age of the samples inferred from palaeontological results: (i) the surficial sedi ments of the present-day carbonate platform; (ii) the Quaternary carbonate platform (Platform 3);
and (iii) the Tertiary carbonate platform (Platforms I and 2). Methods The grab samples were split before analysis. The carbonate content was determined by means of a 'carbonate bomb'. One gram of pulverized sample was weighed and reacted with about I 0 ml of concentrated HCl in a clear acrylic vessel 'bomb'. The volume of C02 generated by the reaction was measured by a digital manometer and then con verted to percentage carbonate, using tables based on standard mixtures. A subsample of about I 00-g weight was washed and sieved using sieves with 2-mm and 0.063-mm mesh sizes. The sieved samples were weighed and the ratio of gravel, sand and mud was determined. Two grams of the sand-size fraction were analysed by settling tube to obtain the mean grain size (¢) of this fraction. After dissolving a similar sized sample in acid, washing and drying, the grain size of the non-carbonate sand fraction was also determined. The compositions of the gravel- and sand-size grains of the grab samples were described both visually and under a binocular microscope on board ship. The samples were described on the basis of their biogenic composition, especially those compo nents reflecting sedimentary environments. Grains of 0.125-4 mm in size were mounted after sieving, and thin-sectioned. Grain composition was deter mined by point counting. Four hundred points were counted for each thin-section at 1-mm intervals over a 2-cm square grid. Dredge samples were cut and thin-sectioned.
Subtropical carbonate platform development Thin sections of samples suspected of containing dolomite were stained with Alizarin-Red S. For rock component analysis, some sections were point counted to 400 points at 1-mm intervals over a 2-cm square grid. Carbonate mineral composition was determined on powdered samples using a Kigaku RINT I 000 X-ray diffractometer. Each sample was scanned from 24o to 35 (29) at 1 min-1, using Cu-Ka: radiation of 40 kV and 50 rnA. The weight ratio of the minerals was determined from the ratio of the peak area of each mineral as proposed by Milliman ( 1974). However, the peak of aragonite used in the calculation was sometimes disturbed by a quartz peak. In these cases, the area of half the visible aragonite peak was measured and doubled. There fore, the mineral content presented here is only semiquantitative. For carbon and oxygen stable isotope analysis of calcite and dolomite, selected samples were pulver ized and dissolved in 100% phosphoric acid at 25oC. The C02 gas evolved was purified through a vacuum line and analysed with a Finnigan MAT 251 mass spectrometer. All dolomite samples are a fine mixture of calcite and dolomite, and it was impossible to separate these mechanically. On the basis of the different reaction rates of dolomite and calcite with phosphoric acid at 25°C, it was as sumed that the C02 gas reieased during the first 30 min of reaction of the calcite-dolomite mixture came from calcite and the C02 gas collected after that was from dolomite. o
o
THE MODERN PLATFORM
The sedimentological characteristics of the surface sediments have been described on the basis of the results of their carbonate content, grain size and composition, in conjunction with the underwater photographs. Carbonate content The surficial sediments in the study area vary between wholly carbonate and wholly siliciclastic, with often steep compositional gradients between these two end-members. The carbonate content varies from 0 to 98% (Fig. 7). The carbonate con tent is low on the inner shelf, less than 30% overall and usually less than I Oo/o closer to the shoreline of Fraser Island. The area of highest carbonate content
173
(> 90%) occurs around the banks, where it is as much as 15 km wide in places. On the outer shelf the carbonate content decreases to less than 50%. Grain size The sediments in the area are predominantly sands. However, many samples have gravel contents in excess of 20%. Although the gravel content is high on the Gardner Banks (> 30%), there are other parts of the outer shelf where the gravel content is significantly higher, with one sample having up to 90%. The mud content on the shelf is generally low (< 5%), but can be as high as 20% on the upper slope. Distribution of grain types The sediments of the study area consist mainly of skeletal carbonate sand and gravel, and siliciclastic sand. The carbonate skeletons consist of the re mains of benthic and planktonic Foraminifera, hermatypic corals, molluscs, coralline algae, bryo zoans, Halimeda, echinoids, brachiopods and bar nacles. These occur as whole skeletons or fragments, and are mostly related to specific environments on the shelf and upper slope. Bryozoans. Although cool water carbonates tend to be dominated by bryozoans, here they make up less than 6% of the sediment. Off Fraser Island they occur in areas of moderately high to high carbonate, but are not necessarily associated with the mid- to outer shelf banks. The bryozoans tend to be more abundant in the southern part of the area, and it is noteworthy that outer shelf sediments some 150 km to the south have more than three times the propor tion of bryozoans as here. Molluscs. The shells of bivalves, gastropods, Dental ium sp. and planktonic pteropods occur as whole skeletons and fragments, ranging from gravel to fine sand, although most tend to be coarse sand-size. Whole skeletons are sometimes encrusted by coral line algae, bryozoans and/or Foraminifera, and are bored. Pteropods are common in sediments from the outer shelf and upper slope. The molluscan content is generally between 5 and I Oo/o on the middle to outer shelf, with highest values in the lee of the banks.
1 74
J. F. Marshall et al.
Sandy Cape
10 km
-
24'40'
Carbonate (%)
24'50'
25'00'
25'10'
25'20'
I
153'20'
23/G56115
Fig. 7. Carbonate content distribution map.
Echinoderms. Fragments of plates or spines of echinoids occur in the granule- to fine sand-size range in both shallow and deep water, forming 1 -17% of samples, with most samples containing at
least 5%. Although high values (> 10%) occur along the line of banks, there are also some high values· at the base of the cliff forming the mid-shelf-outer shelf junction.
175
Subtropical carbonate platform development Coralline algae. Coralline algae, both articulated and �rustose, are present in sediments ranging from gravel to fine sand. However, crustose forms are more abundant in this area. Living crustose algae
are typically red, whereas dead forms are white. Crustose coralline algae are distributed widely throughout the study area (Fig. 8), but are most abundant on the banks and along the cliff bounding
Sandy Cape
10 km
24"40"
Coralline Algae(%)
D C2l D D • .
.
<5 5· 10 10·30 30·50
24"50"
>50
25"00"
25"10"
25"20"
23/G56/23
Fig. 8. Distribution of coralline algae off Fraser Island.
176
J. F.
Marshall et a!.
the mid-shelf. Their abundance varies from 1 to 54%. The highest concentration occurs on the out ermost part of the mid-shelf (Fig. 9a), whereas, both on the inner shelf and on the outer shelf at depths greater than 220 m, values decline to less than 5%. Their abundance on the mid- to outer shelf is higher
than elsewhere along the continental shelf of eastern Australia (Marshall & Davies, 1978). Four types of crustose coralline algae can be recognized: 1 rhodoliths: pebble- or cobble-size and have con centric internal structures (Fig. l Oa & b);
(a)
Fig. 9. Underwater photographs
(b)
showing (a) the high concentration of r hodoliths on the outer part of the shelf off Fraser Island, and (b) the disposition of large plate corals on the Gard ner Banks. The width of the black stripes on the trigger we ight is l em.
Subtropical carbonate platform development
177
(a)
(b)
(c)
(d)
(e)
(f )
Fig. 10. Examples of coralline algal growth off Fraser Island. (a ) Outer view of rounded r hodolith (43 m; x 2). (b) Internal view of r hodolith in (a). Note concentric layering and nucleus (dark). (c) Maerl r hodoliths consisting of Lithothamnion australe (52 m; x2). (d) Calcareous cobble covered by crustose coralline algae and squamariacean algae (28 m; xO.S). (e) & (f) Outer and internal view of an algal-coated pebble ( 40 m; x I).
178
J. F.
Marshall et al.
2 maerl rhodoliths: open-branched crustose coral line algae (Fig. l Oc); 3 cobble- to boulder-size skeletal fragments (mainly corals) encrusted by coralline algae (Fig. 1Od); 4 algal-coated pebbles and granules (Fig. I Oe & f ). The rhodoliths have a fairly limited distribution, but where present they make up the majority of the sediment. They are most commonly found along areas of topographical change and on the banks and their surrounds. Rounded and maerl rhodoliths were recovered from the Gardner Banks and envi rons, and also from an area along the outer mid shelf at depths of 42-117 m. The nuclei of the rhodoliths are variable, consisting of coral, molluscs and other bioclasts. Some rhodoliths are only thin crusts of several layers a few millimetres thick around a large nucleus, but generally all show typical concentric structure. Algal-encrusted cobbles and boulders were recovered from North Gardner Bank and on and around Gardner Bank. Measurement of near-bed flow, using an array of current meters deployed across the mid-shelf at a height of 100 em above the sea-floor, shows a predominantly southerly flow, with the highest measured current speed being 135 em s-1 in a water depth of 72 m (Harris et a!., 1996); this is an area of abundant pebble-size rhodoliths, many of which have outer surfaces of living coralline algae. Veloc ity measurements, in both summer and winter, indicate that currents attain relatively high speeds for periods of 4-5 days, followed by longer periods of slower flow. Rates are generally highest during what are interpreted as eddy intrusions of the EAC onto the mid-shelf. Flume experiments confirm that
rhodoliths of all sizes reach threshold velocities at values similar to or less than those suggested by current meter results (Harris et a!., 1996). Detailed examination of rhodoliths off Fraser Island shows that they are mainly composed of crustose coralline and squamariacean algae. Small amounts of other encrusting organisms, including bivalves, bryozo ans, sponges and Foraminifera, are also present. It is notable that the encrusting foraminifer Acer vulina inhaerens is very limited in this area, in contrast to tropical and subtropical reef regions such as the Ryukyus (Tsuji, 1993), Mascarene Archipelago (Montaggioni, 1979), Caribbean (Reid & Macintyre, 1988; Littler et a!., 1991) and Gulf of Mexico (Minnery, 1990). Some 20 species of coralline algae have been identified in the rhodoliths. Of these, the following are major genera: Lithoporella, Lithophyllum, Mes ophyllum erubescens, Lithothamnion, including L. funafutiense and L. australe, and Sporolithon. Two forms of squamariacean algae were also found in the rhodoliths off Fraser Island. The rhodoliths off Fraser Island can be divided into two types on the basis of their floral composi tion and water depth. Rhodoliths at depths shal lower than 70 m are characterized by abundant Lithothamnion, commonly associated with Litho phyllum and squamarlacean algae. Sporolithon is rarely found. In contrast, rhodoliths at depths greater than 90 m are dominated by squamariacean algae, with Lithothamnion and Sporolithon. Age determinations were carried out on five rhodoliths from across the shelf off Fraser Island (Table 1). All samples still had a red outer coating,
Table 1. Rad iocarbon ages of r hodoliths off Fraser Island
Site
Latitude
Longitude
1 05/DR/007
24'54.5'S
1 53' 3 2 . 5'E
1 05/DR/002
25'04.8'S
1 5 3' 34.9'E
1 05/DR/0 1 8
25'09.5'S
1 5 3' 3 8 . 8'E
Gardner Bankt
25 '04.4'S
1 53' 37.0'E
105/DR/00 1
25' 1 4.0'S
1 5 3'40.9'E
Laboratory reference R20 1 4/ l R20 1 4/2 R20 1 4/3 R20 1 4/4 R20 1 4/5 R20 1 4/6 R20 1 4/7 R20 1 4/8 R20 1 4/9 R20 1 4/IO
Water depth (m) 43 52
D* (mm) 5 I 4 I
60 80 1 00
* Distance from r hodolith surface to the point where the sample was taken. t Environmental correction factor of -450 ± 3 5 yr for marine samples. t Sample collected on 1 3 December 1 992.
12 I 3 I 12 I
Conventional 14C age (yr)
Corrected 1 4C aget (yr)
1 427 18 1 230 513 7020 6424 2230 958 1784 514
977 ± 5 5 Modern 780 ±67 6 3 ± 55 6 570 ± 94 5974 ± 85 1780 ± 85 500 ±66 1 334 ± 80 64 ± 59
±6 5 ± 62 ±76 ± 65 ± 1 00 ± 92 ± 92 ± 75 ± 88 ± 69
Subtropical carbonate platform development indicating that they were living at the time of collection. Two samples were taken from each rhodolith for age determination; the outer sample was collected 1 mm below the surface, and the inner was collected close to the nucleus. Radiocarbon analyses were carried out at the Institute of Geolog ical and Nuclear Sciences Laboratory in New Zealand, and the results are reported using the standard practice outlined by Stuiver & Polach (1977). An environmental correction of 450 ± 35 yr has been subtracted from the conventional age (Gillespie & Polach, 1979). The rhodoliths show a range of corrected ages (Table 1); three show outer cortical ages of< 100 yr, supporting the interpretation that the outer part was living, whereas two give ages that are older. The interpretation for the c. 6000 year-old rhodolith is open to question. This particular rhodolith may have been buried for this length of time and only recently 'resurrected' or may be part of a palimpsest deposit. The current meter data of Harris et a!. (1996) suggest that all rhodoliths are capable of being active at this depth, so at this stage we offer no interpretation for this older date. Taking the three modern rhodoliths only, and calculating their accre tion rates from their ages and the distance between the inner and outer samples, we arrive at rates of 2.8-8. 7 Jlm yr-1 . This is much less than the annual growth rate for coralline algae presented by Mat suda (1989), and implies that the rhodoliths do not form continuously, supporting the results from the current meter experiments (Harris et a!., 1996). Benthic Foraminifera. Whole or fragmented tests of benthic foraminifers range from gravel to fine sand size. Large benthic forams are an important and distinctive component of the sediments, particu larly in the vicinity of the banks. The following larger Foraminifera were identified in this study (H. Nakagawa, personal communication); Amphiste gina, Baculogypsina, Calcarina, Cycloclypeus, Gypsina, Operculina, Heterostegina, Amphisorus, Sorites, Alveolinella and Bore/is. Many of these forams, which are up to 2 em in diameter, are possibly epibenthic because underwater photo graphs indicate that sea grass is fairly common on the shallower parts of the banks. The ratio of larger Foraminifera to total benthic Foraminifera is as high as 48% around the Gardner Banks. Elsewhere, most sediment samples, other than those domi nated by siliciclastic sediments on the inner shelf, contain abundant tests of benthic Foraminifera.
179
They tend to be more abundant (> 15%) on the mid- to outer shelf, whereas they are fairly low (< 5%) on the inner shelf. In general, there is a change in foraminiferal assemblage with depth. Some species, such as Amphistegina sp., Calcarina sp. and some species of Cibicides, are dominant in shallow water, whereas species such as Caccidulina sp. are dominant in deeper water. Planktonic Foraminifera. Planktonic foraminifers occur as whole or fragmented skeletons of medium to fine sand-size grains. They are dominant in the sediments of the outer shelf to upper slope, and show a trend of increasing in abundance with depth. They are usually less than 1o/o on the inner and mid-shelf, but are abundant on the outer shelf (> 20% at depths greater than 200 m). Hermatypic corals. Corals occur as autochthonous colonies and allochthonous gravel-size fragments. The living corals are either large platey and/or encrusting varieties (Montipora sp.; Fig. 9b) or small massive colonies (e.g. Goniopora sp., Porites sp.). Almost no robust branching varieties of coral (e.g. Acropora sp.; Pocillopora sp.) were observed, either from the underwater photographs or the dredge and grab samples. Some delicate branching types (e.g. Seriatopora sp.) were recovered, but they are rare. The species diversity is extremely re stricted, with possibly no more than ten species of hermatypic corals represented. The dead corals are mainly similar to living species, but with possibly a slightly greater diversity. Most dead corals are pebble- to boulder-size fragments, encrusted by living coralline algae, bryozoans and serpulid tubes. Some have been extensively bored by sponges. The living corals have their widest distribution on the Gardner Banks, where they occur at depths of 28-50 m. The predominance of platey and en crusting forms is considered to be a function of the photoadaptibility of the corals at these depths, and the restriction in coral species is considered to reflect both depth and water temperature. Analysis of the point count data suggests that corals are typically under-represented in the sand fraction of the sediment. Point counting shows that coral grains are rare in sediments on the inner shelf (< 5%; commonly< 1%), but around the banks they are more abundant, with one sample in front of the bank as high as 18%. Seawards of the banks the coral content usually decreases with depth, from about 10% at 25 m to less than 1o/o at 250 m.
180
J. F.
Marshall et al.
Halimeda. The codiacean green alga, Halimeda, has a limited distribution in the area, usually occurring as pebble- to sand-size segments or particles. Living Halimeda was recovered at a depth of 28 m on North Gardner Bank. Halimeda is commonly less than 1o/o of the sediment on the inner shelf, and between 1 and 5% on the mid-shelf. The area of relatively abundant Halimeda (> 5%) occurs on the Gardner Banks and on the mid-shelf at a depth of 43 m on the northernmost transect, where it reaches a maximum of 19%.
Facies distribution Facies variations in the sediments have been de fined on the basis of carbonate content, bioconstit uents, grain size and sea-floor photographs. Five sedimentary facies are identified in the study area: (i) inner shelf quartz sand facies; (ii) mid-shelf sand facies; (iii) coral biolithite facies; (iv) rhodolith facies; and (v) outer shelf-upper slope carbonate sand facies (Fig. 12). Inner shelf quartz sand facies. This facies occupies the inner shelf, between the shoreline and approxi mately the 50-m isobath. It is dominated by fine to medium sand-size quartz grains (90%) that are generally subrounded and well sorted. Associated non-carbonate grains include feldspars and lithic fragments. The few bioclastic grains present are fine- to medium sand-size mollusc fragments and benthic Foraminifera. This facies is largely con trolled by the amount of quartz sand supplied by longshore drift along the eastern edge of Fraser Island. The presence of ripples in underwater pho tographs reinforces conclusions from current meter results, which indicate strong currents in this area.
Siliciclastics. Siliciclastic grains are mostly quartz, transported from the south by longshore drift and dispersed over the shelf at various times. They range from 0 to 93% of the sediments in the study area, and are highest (> 60%) on the inner shelf, relatively low (< 5%) on the mid-shelf, but increas ing again on the outer shelf, where, with one exception, they exceed 20%. Mineral composition The major carbonate minerals present in the sedi ments are aragonite, Mg-calcite and low-Mg-calcite (Fig. 11). Aragonite content is high (20%) in shallow water, presumably reflecting the contribution by corals and Halimeda on the shallower parts of the banks, decreasing gradually with depth. Mg-calcite is highest (60-80%) at depths between 50-100 m, centred on the depth of maximum rhodolith devel opment. It then gradually decreases, but it still makes up about 50% of the carbonate fraction as deep as 250 m. Low-Mg-calcite is fairly constant with depth, varying between 20 and 40o/o.Some of the low Mg-calcite is considered to be derived from the underlying platforms.
0 50 1 00 1 50 200 250 300
0
20
Aragonite (%) 60 40
80
1 00
,t})o 0
50
� �
Coral biolithite facies. This facies is characterized by living hermatypic corals in growth position, and
Mg·Catcite (%) 60 80 40
20 0
0
I 0
0 0·
150 200 0
•
Area 1 Area 2
250
•
��
IIBS... § ill � .-
1 00
;, . ... "'
0
0
Mid-shel f sand facies. This consists of coarse to very coarse carbonate sands with subordinate coarse quartz present on the inner side of the banks. The carbonate sand is made up of benthic Foraminifera, intraclasts, mollusc fragments and other unidenti fied bioclastic grains. The sediments also include biogenic gravels composed of larger benthic Fora minifera, bivalves, gastropods, bryozoans, barna cles and echinoid spines.
•• •• •
I
300
Fig. 1 1 . Plot of carbonate mineral composition versus water depth (m).
1 00 0 50 1 00 150 200 250 300
0
20
Low Mg-Calcite (%) 60 80 40
• • oo ..a;, "' -.: i ... . • . •. . .
I
1 00
o
I 23/GSB/30
18 1
Subtropical carbonate platform development
10 km
24'40'
Facies Distribution
D CJJ D D
•
Inner shelf fine sand facies Inner mid-shelf sand facies Coral bio-lithite facies Rhodolith dominated gravel facies Outer shelf I upper slope carbonate sand facies
24'50'
25'00'
25'10'
25'20'
1 53'20'
231G56129
Fig. 12. Sedimentary facies distribution map.
has a very limited distribution, confined mainly to the shallower parts of the banks (Fig. 12). It has been identified by underwater photographs and from dredge samples from the bank tops. The corals
are generally platey or tabular, with associated coralline algae as crusts and/or rhodoliths and sea grasses on some parts of the Gardner Banks.
182
J. F.
Marshall et a!.
Rhodolith facies. The most distinctive component of this facies is coralline algae, either as crusts or, more commonly, rhodoliths. The crusts are usually thin (< 1 em}, but are ubiquitous, coating most fragments such as corals and molluscs. Most crusts have living red algae on their surfaces, and are associated with other encrusting organisms such as bryozoans and serpulids. The rhodolith facies occurs in water depths of 24-1 1 0 m, and is the main facies developed on the deeper parts of the banks, particularly on their outer eastern side. The width of this facies on the shelf varies from 10 km at the centre of the survey area to 3 km in the south and north. The rhodoliths and encrusted corals form biogenic gravels and coarse to very coarse sand. In addition to coralline algae and coral, characteristic constituents of the facies in clude bivalves, larger benthic Foraminifera, Hal imeda, bryozoans, gastropods, echinoid spines, barnacles and articulated coralline algae. Outer shelf-upper slope carbonate sand facies. Sea wards of the bioherms a belt of sediments is com posed of 30-90% carbonate. The facies occurs in water depths of 110 to > 300 m, and consists of fine to medium sands composed of biogenic carbonate and subordinate quartz sand, with some mud (less than 5o/o). The carbonate grains are mainly benthic Foraminifera, molluscs, bryozoans and unidentified bioclasts. Abundant planktonic Foraminifera and pteropod fragments are characteristic constituents of the deeper shelf terrace. The sediments also include granule-size brown carbonate lithic grains, which suggest either reworking of sediments from the bank and/or limestones outcropping on the outer shelf. THE Q U ATERNARY CARBONATE PLATFORM
This platform (Platform 3 in Fig. 4) lies directly beneath the present-day platform and is stratigraph ically continuous with it. Nine dredge sites were occupied on this mid-shelf carbonate platform, in water depths of 28-107 m. The dredge material consists of a variety of carbonate rocks, in particu lar, coral-algal boundstones and rhodolith float stones to rudstones. These limestones are characterized by the dominance of crustose coral line algae and hermatypic corals as constituents. At sites on the steep cliff east of the Gardner Banks, a plentiful supply of in situ coral-algal boundstone samples was obtained.
Petrographic description On the basis of rock texture, grain components, di agenetic features and carbonate mineral composi tion, dredge samples from the mid-shelf were di vided into the following rock types: (i) rhodolith floatstone-rudstone (RH-type limestone); (ii) coral algal boundstone (CA-type limestone); and (iii) bio clastic grainstone-packstone (BC-type limestone). Rhodolith floatstone-rudstone (RH-type limestone). These limestones are composed mainly of rhodoliths in a grainstone to packstone matrix made up of var ious bioclastic grain types (Fig. 13a & b). The lime stones are generally porous and consolidated. Most samples in this type consist predominantly of Mg calcite (average 74o/o), and subordinate low-Mg calcite and aragonite (average 16 and 1Oo/o, respectively). The RH-type limeston�s were recovered from the Gardner Banks in water depths of 28-60 m. The rhodoliths are 1-4 em in diameter, and have a typical concentric structure of crustose coralline algae (Fig. 13a & b). Most have been bored to some extent. The nuclei consist of corals, shell fragments and carbonate rock fragments. Most have smooth surfaces, although some are pinnacled (maerl rhod oliths). In places, the rhodoliths are closely com pacted, and little grainstone to packstone matrix is present. Where it occurs, the matrix consists of medium sand-size to granule-size bioclasts, plus fine calcareous material (Fig. 13c). The bioclasts include both encrusting and articulated coralline algae, benthic Foraminifera, echinoids, Halimeda, mol luscs and coral with minor amounts of angular to sub-angular quartz grains. The fine calcareous ma terial in the grainstone-packstone is mainly micrite and often contains peloids (Fig. 13c). There is little evidence of cementation in the RH-type limestones. Most cement appears to be Mg-calcite micrite in the matrix, plus some filling of pores by peloids. Some endolithic borings in the rhodoliths are filled by peloids and quartz silt. Geopetal fabrics are also present. The distribution of the RH-type limestones coin cides with that of present-day rhodoliths. The lime stones are coarse, poorly sorted, with a relatively low mud content, and resemble the present rhodo lith facies. ·
Coral-algal boundstone (CA-type limestone). These limestones are characterized by the dominance of
Subtropical carbonate platform development
183
(a)
Fig. 13. Rhodolith floatstone-rudstones (RH-type limestones). (a) Rhodolith rudstone showing rounded rhodoliths, of 2-3-cm d iameter, in an open matrix (52 m). (b) Rhodolith floatstone showing scattered rhodoliths of l -2-cm diameter in a packstone matrix (52 m). (c) Photom icrograph of the packstone matrix in the RH-type limestones. I ntergranular pore spaces are partly filled with finer calcareous materials containing peloidal grains. Plane-polarized light.
crustose coralline algae and hermatypic corals. They are widely distributed across the mid-shelf, but are most abundant on the steep cliff at the seaward edge of the platform. The CA-type limestones are brownish yellow to brown boundstones (e.g. Fig. 1 4a), composed mainly of abundant crustose coralline algae, herma typic corals, and encrusting Foraminifera, with minor amounts of benthic forams, gastropods and Halimeda. The coralline algae form thinly layered, thick crusts, and bind other bioclastic grains and finer calcareous material, along with minor quartz grains (Fig. 14b). A matrix, consisting chiefly of the tests of planktonic and benthic Foraminifera, mol luscan shell fragments and quartz grains with finer calcareous materials and terrigenous clay, is present between the algal layers and is bounded by them. The matrix often contains fine peloidal grains. The angular to subangular quartz grains are fine sand size. The carbonate content of this type of limestone is 77-93%. Like the RH-type limestones, the presence of unaltered coralline algae results in a dominantly
-
0.5 m m Mg-calcite mineralogy (average 78. 5%), whereas aragonite and low-Mg-calcite are relatively minor. The abundant boundstones from the cliff to the east of the Gardner Banks in water depths of 90110 m are generally pale yellow to yellowish brown, and are cobble- to boulder-size (Fig. 14a). Most are encrusted by living coralline algae. The layering of the coralline algae is distinct; each layer is less than I mm in thickness. In some of them, two genera tions of growth can be observed (e.g. Fig. 14c), with a very pale brown to yellow older layer overlain by later whitish coralline algae. The upper surface of the recent layer is encrusted by living coralline algae. In the CA-type limestones, cementation is not readily apparent, and most consolidation appears to reflect the binding effect of the coralline algae. Poorly developed fibrous aragonite cements are present in some intraskeletal pore spaces within coral fragments. Porosity is mainly framework po rosity, vuggy porosity and borings. Secondary po rosity, such as vuggy porosity and borings, is often filled with internal sediment consisting of fine cal-
184
JF
Marshall et al. (b)
(a)
-
1 .0 m m (c)
Fig. 14. Coral-algal wackestone-packstone-boundstone type limestones. (a) Algal boundstone showing overlapping sheets of coralline algae with high porosity between the layers (93 m). (b) Photomicrogra ph of an algal boundstone . The encrusting corall ine algae show multiple layering and bind fine calcareous materials, planktonic forams and quartz grains. (c) Algal boundstone. White newer generation of corall ine algae grows on a yellowish brown older ge neration. Note the distinct boundary between the two .
careous material, which sometimes has a peloidal texture, and angular quartz grains. Geopetal struc tures are observed in primary and secondary pores. A few poorly preserved calcareous nanofossils were found in the matrix of one specimen of coral algal boundstone. Most specimens are strongly over grown and difficult to identify. The presence of Ge phyrocapsa caribbeanica indicates that part of the matrix is no older than 1.66 Ma (Biodatum I I of Sato & Takayama, 1990). The absence of Pseudoe miliania lacunosa cannot be confirmed because of the rarity and poor preservation of the nanofossils. Bioclastic grainstone-packstone (BC-type limestone). BC-type rocks are composed mainly of a variety of bioclastic grains, with a minor amount of quartz grains, and show no binding by encrusting organisms (Fig. 15a & b). This type of limestone is not as com mon as the others, and was mainly recovered from the Gardner Banks. The carbonate content varies from 50 to 93%, but average values are around 83%. The major constituents are coralline algae, benthic
Foraminifera, corals, molluscs, Halimeda, bryozo ans and echinoderms. Grain size is generally me dium to very coarse sand- and granule-size, and sort ing is poor. The matrix consists of a lime mud that often contains peloids. This type of rock is poorly cemented and friable. The carbonate mineral com position is dominated by Mg-calcite (64-89%) with subordinate aragonite (2-28%) and low-Mg-calcite (8-22%). THE TERTIARY CARBONATE PLATFORM
Five dredge sites located on the outer shelf to mid-slope, in water depths of 270-600 m, recov ered well-lithified limestones of presumed shallow water origin from Platform 3. Petrographical description On the basis of grain components, diagenetic fea tures, and carbonate mineral composition, the dredge samples from Platform 3 were divided into
Subtropical carbonate platform development
185 (b)
-
1 . 0 mm Fig. 15. Bioclastic grainstone-packstone type limestones. (a) Bioclastic grainstone (60 m). (b) Photomicrograph of the
bioclastic packstone . I ntergranular pore-spaces are partly filled with micr itic matrix containing peloids. Plane-polarized light.
the following rock types: (i) well-cemented lime stone (WR-type limestone); and (ii) dolomitic lime stone (DL-type limestone). The well-lithified limestone and dolomitic lime stone samples are dense and have diagenetic features suggesting both marine and meteoric environments. Calcareous nanofossils suggest Oligocene to early Middle Miocene ages, and benthic forams indicate Early to Middle Miocene ages. The well-lithified limestones contain rhodoliths and large forams (e.g. Lepidocyclina). Well-cemented limestone (WR-type limestone). These limestones are white to pale brown, and com positionally and texturally are mainly rhodolith large foram-molluscan packstones, grainstones and floatstones. The rhodolith-large foram packstones are composed mainly of rhodoliths, encrusting and articulated coralline algal fragments, and large benthic forams with minor amounts of planktonic forams and sometimes quartz grains. The rhodoliths are white, 5-40 mm in diameter, and show typical concentric structure. Large forams (e.g. Lepidocy clina sp.) can be 5-40 mm in diameter. Some lime stones also contain larger coral fragments. One dis tinctive limestone is a yellowish brown to yellow floatstone that is fairly well cemented. This consists of mollusc shells, coralline algae, solitary corals, bry ozoans and forams in a wackestone matrix. Total carbonate content of the WR-type lime stones is very high, from 86 to 96%, and is pre dominantly low-Mg-calcite, with some Mg-calcite but no aragonite.
Diagenetic fabrics include isopachous and equant sparry calcite cements. Isopachous rim cements grow radially from the surfaces of bioclastic grains, and are dominant in grainstones from the upper slope (Fig. 16a). They consist of fibrous or bladed crystals, about ! 50 J.l.m in length. The outer zone of the rim cement is sometimes brown and dusty in appearance. In some examples, two generations of isopachous cement could be observed (Fig. 16b & c). Sparry calcite cement is common as a late-stage pore fill in these limestones. The sparry calcite cement overlies the isopachous rim cement and fills intergranular and intraskeletal pores (Fig. 16a). It occurs as a drusy mosaic of equant, subhedral to anhedral crystals, 15-400 J.l.m in diameter. Dolomitic limestone (DL-type limestone). Dredging of the middle and upper slope off Fraser Island recovered some dolomitic limestone samples. All samples containing dolomite were classified as DL type regardless of other rock components and rock textures. They consist of rhodolith-large foraminif eral floatstone and bioclastic packstone, grainstone and mudstone, and their dolomite content ranges from 15 to 37%. White dolomitic rhodolith-large-foram float stone is well consolidated, with a packstone to wackestone matrix (Fig. 17a). The main bioconstit uents are crustose coralline algae and large benthic Foraminifera (Lepidocyclina sp.). The dolomite oc curs mainly as a cement, lining pores created by dissolution or boring and growing on micritic inter nal sediments which occupy the lower parts of
186
J. F. Marshall et al. (b)
-
0.2 m m
0 .5 m m
Fig. 16. Well-consolidated limestone type (438 m). (a) Photomicrograph of a bioclastic grainstone showing articulated coralline algal fragments and other skeletons lined by isopachous cement. Remaining pore-space has been filled with sparry calcite cement. Plane-polarized light. (b) Photomicrograph of two generations of isopachous rim cements (F 1 and F2 ). The two cement layers are intercalated with micrite (I). Plane-polarized light. (c) Photomicrograph showing the second-stage isopachous rim cement growing directly on the first-stage isopachous rim cement, and followed by internal sediment. Plane-polarized light.
(b)
-
0.2 m m
0.2 mm
Fig. 17. Dolomitic limestone type ( 5 9 2 m). (a) Photomicrograph o f dolomitic limestone. The dolomite occurs as
equant, subhedral to euhedral pore-lining cement. Micritic internal sediment below the dolomite cement is often dolomitized. The dolomite cements are covered by the next stage of internal sediment, equant mosaic sparry calcite . cement and isopachous bladed calcite rim cements. Plane-polarized light. (b) Photomicrograph of euhedral dolomite crystals within a vug or boring. The dolomite crystals are covered by equant mosaic calcite cements. Plane-polarized light.
187
Subtropical carbonate platform development dissolution pores (Fig. 17b). The dolomite cement consists of equant subhedral to euhedral crystals, about 50 Jlm in length. The micritic sediments beneath are also often dolomitized. The dolomite cement is commonly also covered by internal sedi ments and/or fibrous to bladed isopachous calcite cement, which in turn are succeeded by equant mosaic calcite cement (Fig. 17a). The dolomites have an isotopic compositional range of 0 1 80p08 from + 1.58 to +4.27o/oo and 0 1 3CPDB from + 1.69 to + 3.45o/oo (Table 2). The c) 1 3C and c) 1 80 values of the dolomites provide a regular pattern of variation. Dolomites with more 1 80-enriched oxygen tend to have more 1 3C enriched carbon. Dolomitic limestones with a higher dolomite content are also more 1 80- and 1 3C enriched.
DISCUSSION
The environmental significance of subtropical shelf carbonates
The separation of shelf carbonates into chlorozoan and foramol assemblages by Lees & Buller (1972) and their attribution to temperature-salinity changes (Lees, 1975) presents a definite division between tropical and temperate carbonates. How ever, the boundary between the two has not been
investigated in any detail, and the nature of this boundary remains largely unknown. Along the east coast of Australia, Marshall & Davies (1978) deter mined the boundary to be at 24 • S, based on the southernmost occurrence of surface reefs. It is apparent that such a boundary cannot be so clear cut, and that some overlap of assemblages should occur. The skeletal composition of low-relief bio herms on the mid-shelf off Fraser Island supports this contention. The Gardner Banks, although dom inated by coralline algae, contain hermatypic corals and Halimeda, an association that is definitely tropical. However, warm temperate shelf assem blages, some 150 km to the south, contain no trace of Halimeda, and, apart from the presence of coralline algae and corals, have a greater abundance of bryozoans and barnacles, an association that is distinctly temperate (Tsuji et a!., 1994). Conversely, bioherms the same distance to the north form surface reefs, which, although areally restricted, are definitely tropical. This suggests that the boundary between tropical and temperate assemblages con sists of a zone, some 300 km wide, that contains within it a distinctly subtropical assemblage. The narrowness of this boundary (less than 8% of the total length of the east Australian shelf) reflects the sensitivity of these assemblages to subtle changes, presumably in water temperature. How ever, there is no perceptible change in surface water temperature over this distance, other than slight
Table 2. Carbon and oxygen isotopic composition (in PDB) of the Tertiary platform samples
Sample number 1 05/DR/005 - 1 -2 -4 -7 -9 1 05/DR/006 -4 1 05/DR/009 -3 -5 -7 -8 1 05/DR/0 1 0 - 1 -2 -3 -4 -6
Water depth (m) 592
420 27 1
438
Calcite
Dolomite
Rock type
o 1 3C
WR WR DL
1 . 399 1 .693 2.690 2.695 2 . 5 76 1 . 993 2.729
1 .064 1 .059 0.784 0.728 2 . 36 5 1 . 559 2.422
3 .086 3.080 1 .944 1 .285 1 .946 1 .900 2.027 0.822 2. 1 88 2.238
2.835 2.808 0.820 1 . 1 42 -0.06 1 1 .330 1 .420 -0.745 1 .308 1 .992
DL DL DL DL DL DL DL DL WR WR WR WR WR
a t so
o 1 3C
a t so
3.463
3.644
2.76 8 2.444 2.840 1 .689 3.449 3.430
3.259 2.686 3.292 1 .586 4.272 4.257
1 .9 1 2
2.247
1 88
J. F
Marshall et al.
differences in summer and winter ranges. It may be that other, as yet unknown, factors are involved. Possibly seasonal changes in oceanic circulation, involving the EAC and the gyres immediately to the north, are important. Whereas the subtropical and warm temperate assemblages are similar, the ab sence of Halimeda in the south being the most apparent difference, both zones are distinctly dif ferent from coral-Halimeda-dominated tropical assemblages to the north and bryozoan-mollusc foram-dominated cool temperate assemblages further south. The most obvious difference is the dominance of crustose coralline algae, along with subordinate and distinctly species-limited herma typic corals, in both the subtropical and warm tem perate assemblages. This change from coral-Halimeda to coralline algae-coral-Foraminifera to bryozoan-Foramini fera-mollusc is very similar to that identified on the Brazilian shelf off the east coast of South America (Carannante et a!., 1 988). There, four major types of carbonate lithofacies were distinguished on the shelf: 1 Chlorozoan-characterized by hermatypic corals and Halimeda, associated with molluscs, benthic foraminifers, echinoids, bryozoans, sponges and coralline algae. This lithofacies is typical of tropical areas containing well-developed coral reefs. 2 Chloralgal-contains large amounts of calcareous green algae but no hermatypic corals. This litho facies is present in tropical-subtropical zones where coral reefs have not developed. No equivalent lithofacies has been identified in subtropical assem blages from eastern Australia, but it is particularly abundant in the Great Barrier Reef as Halimeda bioherms and biostromes. 3 Rhoda/gal-dominated by abundant encrusting coralline algae, bryozoans and variable amounts of large benthic foraminifers and barnacles. This litho facies is well developed in transitional warm temperate-subtropical zones. Locally, bryozoans can be abundant, especially towards cooler areas. 4 Molechfor-characterized by abundant mollusc fragments, benthic foraminifers (many of them arenaceous) and echinoids. Barnacles can be the main constituents; serpulids and bryozoans may be present. Encrusting coralline algae are generally absent. This lithofacies characterizes cold-temper ate carbonate shelves. Carannante et a!. ( 1 988) also pointed out that the distribution of these lithofacies is obviously subject to complex environmental factors related primarily to latitude and depth, both of which control water
temperature. However, factors affecting salinity and temperature, nutrient levels, light penetration, etc., may also play a fundamental role. This apparent sensitivity to temperature changes within the transitional zone should be reflected in the sedimentary record, particularly during glacial interglacial cycles. We have no evidence at this stage to verify this; presumably the frequency of Pleis tocene glacial cycles is too high to preserve a legible record of varying carbonate assemblages. On the other hand, more gradual changes do appear to be preserved. The Early to Middle Miocene carbonate platform off Fraser Island contains similar skeletal elements to the modern shelf. From the results of our present study, we would interpret this assem blage as being deposited in a subtropical environ ment. This is more appropriate than a cool temperate or tropical interpretation because of the lack of both abundant bryozoans on one hand and corals-Halimeda on the other. The rhodolith-large foram assemblage is more in keeping with the transitional environment of the present study area, where coralline algae, both as crusts and, more significantly, rhodoliths, predominate over all other bioconstituents. Subtropical carbonate platform development
Recent studies (e.g. Davies et a!., 1 987, 1 989; Pigram et a!., 1 993) have detailed substantial work aimed at understanding the evolution of the Great Barrier Reef and its relation to other large carbonate platforms off tropical north-east Australia, namely the Marion and Queensland plateaux. These studies have postulated that the development of carbonate platforms off eastern Australia is underpinned by the northward movement of the Australian plate throughout the Cainozoic. This northward drift of the Australian continent has progressively moved north-east Australia from a temperate to a tropical environment. From an analysis of the rate of drift, in conjunction with palaeoclimate reconstructions for the Tertiary, Davies et al. ( 1 987) postulated that the present tropical carbonate platforms of north-east Australia would be underlain by temperate carbon ate accumulations. An essential consequence of the drift hypothesis is that subtropical and temperate facies should not only underlie the tropical facies of the marginal plateaux, but also that subtropical car bonate platforms would be contemporaneous with their tropical counterparts, but further to the south. Therefore, off eastern Australia subtropical facies
Subtropical carbonate platform development could have occurred south of the Marion Plateau in the Early to Middle Miocene, and have occurred south of the Great Barrier Reef from the Pleistocene to the present day. Miocene subtropical carbonate build-ups Results from ODP Leg 133 off north-east Australia have confirmed that large tropical reef platforms developed beneath the Marion and Queensland plateaux in the Early and Middle Miocene (Davies et al., 1991). It also appears that a very large subtropical carbonate build-up occupied the outer continental shelf to the east of Fraser Island at the same time. We have named this feature the Fraser Platform. On seismic profiles (Figs 4 & 6) the Fraser Platform can be seen as a 10-km-wide feature with poor reflection characteristics and truncated sea wards by the steep upper continental slope. Changes in the reflection character allow the differentiation of an outer (Platform 1) and inner zone (Platform 2) of poor reflections, with dipping reflectors in front of Platform 2, and the progradation of a strongly reflecting bedded upper section over Platform 1. The interpretation of the seismic sequences sug gests: 1 Build-up of the Fraser Platform on the shelf edge with a steep seaward facing slope (Platform I) part of which may be erosional. West of the shelf edge platform, lagoonal or leeside sediments were depos ited, and further west, Platform 2 was sporadically developed. If the platform (non-bedded) and la goonal (bedded) sediments are stratigraphically re lated then they define deposition during a slowly rising sea-level, demonstrated by the upward and outward progradation of the bedded and lagoonal facies over the shelf edge platform. Sediment thick ness is of the order of 450 m, similar to that beneath the Marion Plateau (Pigram et al. , 1993). 2 Exposure of the Fraser Platform with subaerial erosion and the probable development of karst. The steep slope of the margin is probably partly related to erosion at this time. 3 Flooding of the Fraser Platform, with backstep ping and growth of a new platform (Platform 3) on the mid- to outer shelf. Quaternary subtropical carbonate build-ups Seismic, bathymetric, side-scan sonar and sediment data indicate the development of a large Quaternary carbonate platform beneath and seaward of the
189
Gardner Banks. These results indicate that subtrop ical carbonates, as compared with temperate car bonates in general, are capable of forming bioherms of some magnitude. The present-day bioherms of the Gardner Banks stand as much as 20 m above the surrounding sea-floor. They also cover a wide area of the mid-shelf. Seismic reflection data indi cate that they are at least 60 m thick. There appears to have been backstepping of Platform 3 during the Quaternary from the I 05 m cliff at the edge of the mid-shelf to the present location of the Gardner Banks. Nanofossil and planktonic foraminiferal bio stratigraphy shows that limestone samples dredged from this build-up were deposited during the Qua ternary, and nanofossil assemblages indicate an age possibly younger than 0.39 Ma (Biodatum 3). From a consideration of their age and sea-level history, the timing and environment of the formation of these limestones can be considered as follows: 1 Precursor sediments, similar to present-day sedi ments of the area, were deposited during the Qua ternary in an environment basically similar to that which exists today. 2 The limestones have not suffered marked mete oric diagenesis, although the dredge sites are situ ated between supposed highstand and lowstand levels of eustatic oscillations during the Pleistocene. It is possible that the limestones have been lithified since the last postglacial transgression. The lime stones have only been weakly cemented, possibly under relatively quiet water conditions. Diagenetic features of the subtropical platforms
The diagenetic features of the Tertiary limestones of the Fraser Platform suggest that diagenesis occurred in both marine and non-marine environments. Isopachous fibrous to bladed rim cements indicate a submarine environment, whereas bladed to equant sparry calcite cement is considered to have formed in fresh to brackish water. Dissolution features within coralline algal fragments also indicate mete oric diagenesis. The dolomitic limestones have undergone a vari ety of diagenetic processes, such as dolomite cemen tation, dolomitization, internal sediment deposition and cementation, and cementation by isopachous bladed calcite and equant mosaic calcite. The first stage of diagenesis was the formation of vuggy po rosity and/or boring porosity, because no diagenetic fabric is cut by vuggy porosity. Porosity created by
190
J. F.
Marshall et al.
the first stage of dissolution and/or boring was locally filled with micrite. Subsequently, euhedral pore lining dolomite cement deposition and dolomitiza tion of pore-filling sediments occurred. In the re maining pores, deposition of micritic (peloidal) sediments, pore-filling and isopachous bladed calcite cements followed dolomitization, and these were succeeded in turn by equant blocky calcite cement. The oxygen isotopic composition of the dolomite ranges from + 1.58 to +4.27%o. Calcites equili brated with these dolomites are predicted to have 8 1 80 values of +0. 13 to +2.82%o. The isotopic composition of the Fraser Platform dolomites sug gests that they were precipitated from sea-water slightly more 1 80-enriched than normal sea-water at 25 • c. Enrichment in 1 80 is generally considered to correspond to a decrease in water temperature and/or evaporation. There is no evidence that an evaporitic environment existed, whereas the oxygen isotopic value of calcites equilibrated with the dolomites is equal to that of limestones formed in water depths of 0- 150 m. This suggests that the dolomite formed from slightly cool normal sea water. Unfortunately, there are few clues to the origin of the dolomite, so it is very difficult to clarify the mechanism and timing of dolomitization of this subtropical carbonate platform. However, some inferences can be made by comparing the Fraser Platform dolomites with results from ODP Leg 133 off north-eastern Australia (Davies et a!., 1991). D,olomitic intervals exist throughout the Tertiary to Quaternary section of the Marion Plateau (Pigram et al., 1993). From the petrographical and micro palaeontological evidence, the diagenetic sequence observed in the Marion Plateau cores can be di vided into three stages: 1 late Early to early Middle Miocene: the plateau formed a shallow marine environment, with mic rite, isopachous fibrous and botryoidal aragonite cements forming on grains; 2 middle Middle to middle Late Miocene: the pla teau was emergent and underwent meteoric diagen esis, resulting in the formation of syntaidal overgrowths, bladed to granular pore lining ce ments, and mouldic and vuggy porosity; 3 late Late Miocene: the plateau was again flooded and became a deep marine environment, as evi denced by borings, apatite crusts, coarse equant sparry calcite and pore filling with hemipelagic clay. Dolomitization is considered to have occurrecl on at least three occasions: the middle Middle Mi-
ocene, the late Late Miocene and the post-Late Pliocene. Some of the diagenetic textures observed in the carbonates of the Marion Plateau are absent in the dredged limestones off Fraser Island, but there is no significant difference between the order of diage netic events for the two areas. The first diagenetic event was dissolution and/or boring. Both micrite and isopachous fibrous to bladed calcite cements can be recognized forming before and after the dolomite cement and dolomitization. These tex tures are generally considered to have formed under shallow marine conditions. Equant to blocky calcite cements follow the isopachous calcite cement, and are similar to coarse equant spar in the dolomites of the Marion Plateau, which are considered to have formed in a deep marine environment as a late stage cement (Pigram et a!., 1993). The presence of hemipelagic planktonic foraminiferal clay that post dates the equant calcite cement supports this view. Precipitation of dolomite cement and dolomiti zation are considered to have occurred at a stage when the environment was changing from shallow marine to somewhat deeper water. This is sup ported to some extent by the oxygen isotopic values, which suggest that the dolomite formed from cool normal marine waters. This dolomitization appears to correspond to the second (Late Miocene) stage of dolomitization within the Marion Plateau carbon ates, which also occurred in a marine environment changing rapidly from shallow to deep marine. Comparison between the Ryukyus and the southern Queensland shelf
As noted, one of the few areas where subtropical carbonates have been studied in any detail is the Ryukyu Islands (Tsuji, 1993). The Ryukyu Island Arc (Fig. 18), which extends c. 1200 km from Ky ushu south-westward to Taiwan, rims the north western Pacific Ocean, and is bounded by the Okinawa Trough to the north-west and the Ryukyu Trench to the south-east. The Kuroshio boundary current flows to the north-east through the Okinawa Trough and extends the influence of tropical water into the Ryukyus. To the west of Miyako Island, Tsuji (1993) recognized five sedimentary facies on the shelf: 1 reef facies (0-60 m), made mainly of autochtho-· nous hermatypic corals and encrusting algae; 2 near reef sand facies (0-90 m), characterized by tests of shallow benthic Foraminifera such as Cal..
19 1
Subtropical carbonate platform development
.�/1 �1( cc;!HIu§YJ/1 �/E/1
fJIJ rCU�UrC OrC!EI1 fNJ
il
8 124· 0
,.•.
136.
500 km
Fig. 18. Location map of the Ryukyu Islands showing the area studied by Tsuji ( 1 993).
carina and Marginopora, plus Halimeda segments; 3 muddy sand-sandy mud facies (20-60 m), con sisting of a very fine sand-size carbonate fraction and lime mud; 4 rhodolith and large foraminiferal gravelly sand facies (60-200 m), containing pebble- to cobble-size rhodoliths and/or the large benthic Foraminifera Cycloclypeus; 5 bryozoan sand facies (80-200 m), rich in bryo zoan fragments. Tsuji et al. ( 1989) and Tsuji ( 1993) showed that tidal currents dominate the shelf area, and that
current speeds are high in the open sea where the reef, rhodolith and large Foraminifera, and bryo zoan facies are distributed, and low in the restricted area where the muddy sand facies is distributed. The most obvious differences between offshore Miyako Island and offshore Fraser Island lie in their relative tectonic setting and oceanography. The differences in tectonic settings create differences in their respective large- and small-scale topography, and in the amount of terrigenous influx. The large topographical setting created by island-arc develop ment has strongly influenced the deposition of
192
J. F.
Marshall et a!.
carbonates in the Ryukyus. The Okinawa Trough acts as a barrier to siliciclastic input from the continent, and consequently, the Ryukyu arc is a high carbonate area. On the other hand, the conti nental shelf of southern Queensland exists in a passive margin setting where terrigenous grains are constantly supplied from the continent, although Fraser Island acts as a barrier to siliciclastic input. In both areas the influence of strong currents is apparent, particularly with respect to rhodolith formation, but the measurement of near-bottom currents on the shelf and upper slope has demon strated a significant difference in the hydrological regime between the areas. Off Fraser Island, current measurements confirm that the southward-flowing East Australian Current dominates the hydrological regime, whereas near-sea-bed current measure ments off Miyako Island indicate a predominantly tidal influence amplified by the topography and the opposing flow of the East China Sea at high tide and the Pacific Ocean at low tide. However, despite different tectonic and oceano graphic settings, the carbonate sediments in the southern Ryukyus and southern Queensland are very similar. Coralline algae are one of the most important carbonate constituents in both areas. They have created a rhodolith-large Foraminifera facies in the southern Ryukyus and a rhodolith dominated facies in southern Queensland. The rhodoliths in both areas can be classified as deep water rhodoliths (Bosence, 1983) from their occur rence. Rhodoliths are distributed on the outer shelf in the Ryukyus at depths of between 60-150 m, and on the mid-shelf off southern Queensland at depths of between 30 and 140 m. Corals, bryozoans and Halimeda are associated with coralline algae in both areas. It appears that similar settings, in terms of their subtropical latitudinal position and range of water temperatures, have determined their com mon biofacies. The formation of rhodoliths rather than crusts can be attributed to the high-energy oceanographic regime in both areas. The in-situ direct measure ment of near-bottom currents on the shelf and slope has proven the existence of high-velocity currents, up to 130 em s-1 at depth, controlling the distribu tion of coarse sediments both in the Ryukyus and southern Queensland. In the study area off Fraser Island, the sediments on the shelf and upper slope are mud free. The EAC impinges on the continental shelf and the distribution of shelf sediments is strongly influenced by the current (Harris et a!.,
1996). Tsuji ( 1993) described coarse, mud-free sediments off Miyako Island, and indicated that tidal currents in the Ryukyus provide the high energy deeper shelf environment essential to win now the mud fraction and move rhodoliths; a process important to their concentric growth, par ticularly when it is superimposed by processes such as swells or typhoons. The high-energy environ ments in both areas provide the necessary boundary conditions for coarse sediment, specifically rhode lith, formation and accumulation.
CONCLUSIONS
For the past 3 0 yr, substantial emphasis has been placed on depositional models of carbonate rocks, particularly from an exploration viewpoint. The knowledge of sedimentary facies helps to predict the distribution and the heterogeneity of reservoirs, and carbonate sedimentary facies models have aided the exploration for hydrocarbons in carbonate reser voirs throughout the world (e.g. Scholle et al., 1983; Roehl & Choquette, 1985). However, the great majority of these models have been developed on the basis of studies of tropical shelf carbonates. Models based on temperate and subtropical carbon ates are virtually non-existent. This is despite the fact that the sedimentary facies identified in the present study area are very similar to carbonate facies in many parts of the world, some of which are oil-producing. In this paper we make the case for the recognition of extensive subtropical build-ups composed largely of coralline algae, corals, bryozoans and Foramin ifera. The combination of relatively shallow-water hermatypic corals and relatively deep-water rhodo liths would seem to be a contradiction, but it is exactly this association that typifies the subtropical shelf environment. Although coralgal associations are more often tropical, it is the relative abundance of coralline algae, whether as crusts or rhodoliths, together with the restricted distribution and diver sity of the corals, that makes this environment unique. Another attribute that can be associated with subtropical shelf carbonates is the variation between Halimeda towards the tropical boundary and bryozoans towards the temperate boundary.. They can also be separated from temperate shelf carbonates in that they can develop bioherms that produce significant vertical relief. In this sense, subtropical carbonates are a true watershed be-
Subtropical carbonate platform development tween majqr shelf carbonate depositional environ ments. Such build-ups, growing at the same time as their better understood tropical counterparts, occurred in the Miocene and Quat�rnary of north-eastern Aus tralia. This indicates that these assemblages will shift latitudinally in response to climate change, and this is supported by the existence of warm temperate assemblages beneath tropical carbonate platforms further north. We believe that recognition of these environmentally sensitive assemblages in ancient carbonates will prove to be invaluable in palaeogeographical and palaeoclimate reconstruc tions.
ACKNOWLED GEMENTS
We wish to thank the captain and crew of R.V. Rig Seismic for the successful completion of Cruise 105 off southern Queensland. We also thank the AGSO seismic and geological technicians for their usual high standard of support. David Feary and Gary Bickford (both AGSO) contributed their time and expertise to the cruise. We thank Peter Harris (previously Ocean Sciences Institute, Sydney Uni versity, but at present at the Antarctic CRC, Hobart) for his expertise and dedication with the current meter deployments and interpretation. We acknowledge our appreciation to Hiroshi Nakagawa (Japan Oil Engineering Co. Ltd) for his effort in identification of the foraminifers, and Keiko Ha tano and Manami Aikawa (JNOC) for X-ray dif fraction and stable isotope analyses. We thank Colin Braithwaite and Ian Macintyre for their reviews of the manuscript, and Gilbert Camoin for his editorial assistance. This paper is published with the permission of the Executive Directors of the Australian Geological Survey Organisation and the Technology Research Center of the Japan National Oil Corporation.
REFERENCES
BOLAND, F.M. & HAMON, B.V. ( 1 970) The East Australian Current, 1 96 5- 1 96 8 . Deep-Sea Res., 17, 777-794. BONE, Y. & JAMES, N.P. ( 1 993) Bryozoans as carbonate sediment producers on the cool-water Lacepede Shelf, southern Australia. Sediment. Geol., 86, 247-27 1 . BOREEN, T.D. & JAMES, N.P. ( 1 993) Holocene sediment dynamics on a cool-water carbonate shelf: Otway, south eastern Australia. J. sediment. Petrol., 63, 574-588.
1 93
BOREEN, T., JAMES, N., WILSON, C. & HEGGIE, D. ( 1 993) Surficial cool-water carbonate sediments on the Otway continental margin, southeastern Australia. Mar. Geol., 1 1 2, 3 5-56. BOSENCE, D.W.J. ( 1 983) Description and classification of rhodoliths (rhodoids, rhodolites). In: Coated Grains (Ed. Peryt, T.M.), pp. 225-242. Springer-Verlag, Berlin. CARANNANTE, G., ESTEBAN, M., MILLIMAN, J.D. & SIMONE ( 1 988) Carbonate lithofacies as paleolatitude indicators: problems and limitations. Sediment. Geol., 60, 333346. CHURCH, J.A. ( 1 987) East Australian Current adjacent to the Great Barrier Reef. Aust. J. mar. freshw. Res., 38, 67 1 -6 8 3 . COLLINS, L.B. ( 1 988) Sediments and history o f the Rott nest Shelf, southwest Australia: a swell-dominated, non tropical carbonate margin. Sediment. Geol., 60, 1 5-50. CONOLLY, J. R. & VON DER BORCH. C.C. ( 1 967) Sedimenta tion and physiography of the sea floor south of Austra lia. Sediment. Geol., 1, 1 8 1 -220. DAVIES, P .J., SYMONDS, P.A., FEARY, D.A. & P IGRAM, C.J. ( 1 987) Horizontal plate motion: a key allocyclic factor in the evolution of the Great Barrier Reef. Science, 238, 1697-1700. DAVIES, P.J., SYMONDS, P.A., FEARY, D.A. & P IGRAM, C.J. ( 1 989) The evolution of the carbonate platforms of northeast Australia. In: Controls on Carbonate Platform to Basin Development (Eds Crevello, P., Sarg, J.F., Read, J.F. & Wilson, J.L.), Spec. Pub!. Soc. econ. Paleont. Miner., Tulsa, 44, 233-258. DAVIES, P.J., McKENZIE, J.A., JuLSON, A.P. et a!. ( 1 99 1 ) Preliminary Results of the Ocean Drilling Program, Leg 133-Northeast Australia. US Government Printing Of fice, Washington, DC. DAVIES, P.J. & TsuJI, Y. et a!. ( 1 992) Tropical and temper ate carbonate environments: the effects of sea level, climate and tectonics on facies development-post cruise report. Bur. Min. Res. Aust. Rec. 1 9921 17, 85 pp. ELLIS, P.L. ( 1 966) The Maryborough Basin. APEAJ. 3036. ETHERIDGE, M.A., SYMONDS, P.A. & LISTER. G.S. ( 1 990) Application of the detachment model to reconstruction of conjugate passive margins. In: Extensional Tectonics and Stratigraphy of the North Atlantic Margins (Eds Tankard, A.J. & Balkwill, H.), Mem. Am. Assoc. petrol. Geol, Tulsa, 46 , 23-40. GILLESPIE, R. & POLACH, H.A. ( 1 979) The suitability of marine shells for radiocarbon dating of Australian prehistory. In: Radiocarbon Dating (Eds Berger, R. & Suess, H.), pp. 404-42 1 , University of California Press, Berkeley. GRIMES, K.G. ( 1 982) Stratigraphic drilling report-GSQ Sandy Cape 1 -3R. Queensland Gov. Mining J., 83, 224-233. HAMON, B.V. ( 1 965) The East Australian Current 1 9601 964. Deep-Sea Res. , 12, 899-92 1 . HARRIS, P.T., TSUJI, Y., MARSHALL, J.F., DAVIES P.J. HoNDA, N. & MATSUDA, H. ( 1 996) Sand and rhodolith gravel entrainment on the mid- to outer-shelf under a western boundary current: Fraser Island continental shelf, eastern Australia. Mar. Geol. , 1 29, 3 1 3-330. HAYES, D.E. & RINGIS, J. ( 1 973) Seafloor spreading in the Tasman Sea. Nature, 243, 454-459.
194
J. F
Marshall et al.
IRYU, Y. ( 1 985) Study on the recent rhodoliths around the Ryukyu Islands. In: Prompt Reports of the Comprehen sive and Scientific Survey in the Ryuku Archipelago, 2, 1 23-1 3 3 , Kagoshima University (in Japanese with Eng lish abstract). JAMES, N.P. ( 1 990) Cool water carbonates: viable ana logues for Paleozoic limestones. International Sedimen tological Congress, Nottingham, p. 275 (abstract). JAMES, N.P. & BONE, Y. ( 1 99 1 ) Origin of a cool water, Oligo-Miocene deep shelf limestone, Eucla Platform, southern Australia. Sedimentology, 38, 323-34 1 . JAMES, N.P. & VON DER BORCH, C.C. ( 1 99 1 ) Carbonate shelf edge off southern Australian: a prograding open platform margin. Geology, 19, 1 005- 1 008. JAMES, N.P., BONE, Y., VON DER BORCH, C.C. & GOSTIN, V.A. ( 1 992) Modern carbonate and terrigenous clastic sediments on a cool-water, high-energy, mid-latitude shelf; Lacepede Shelf, southern Australia. Sedimentol ogy, 39, 877-904. JONES, H.A. ( 1 973) Submerged shorelines and channels on the east Australian continental shelfbetween Sandy Cape and Cape Moreton. Bur. Min. Res. Aust. Rec., 1 97 3/46, 7 pp. JoNES, H.A., DAVIES, P.J. & MARSHALL, J.F. ( 1 975) Origin of the shelf break off southeast Australia. J. geol. Soc. Aust., 22 , 7 1 -78. JONGSMA, D. & MUTTER, J.C. ( 1 978) Non-axial breaching of a rift valley: evidence from the Lord Howe Rise and the southeastern Australian margin. Earth planet. Sci. Lett. , 39, 226-234. KONISHI, K., SCHLANGER, S.O. & OMURA, A. ( 1 970) Neo tectonic rates in the Central Ryukyu Islands derived from 230Th coral ages. Mar. Geol., 9, 225-240. LEES, A. ( 1 975) Possible influence of salinity and temper ature on modern shelf carbonate sedimentation. Mar . Geol. , 19, 1 59-1 98. LEES, A. & BULLER, A.T. ( 1 972) Modern temperate-water and warm-water shelf carbonate sediments contrasted. Mar. Geol. , 13, M67-M73. LITTLER, M.M., LITTLER, D.S. & HANISAK, M.D. ( 1 99 1 ) Deep-water rhodolith distribution, productivity, and growth history at sites of formation and subsequent degradation. J. exp. Mar. Bioi. Ecol. , 150, 1 63- 1 80. MARSHALL, J.F. ( 1 97 1 ) Phosphatic sediments on the eastern Australian continental slope. Bur. Min. Res. Aust. Rec., 1 97 1 /59, 9 pp. MARSHALL, J.F. ( 1 972) Morphology of the east Australian continental margin between 2l"S and 33"S. Bur. Min. Res. Aust. Rec., 1 972170, 1 8 pp. MARSHALL, J. F. ( 1 977) Marine geology of the Capricorn Channel area. Bur. Min. Res. Aust. Bull. , 163, 8 1 pp. MARSHALL, J.F. ( 1 978) Morphology and shallow structure of the continental shelf of southern Queensland and northern New South Wales. Bur. Min. Res. Aust. Rec., 1 978/ 1 00, 25 pp. MARSHALL, J.F. ( 1 979) The development of the continen tal shelf of northern New South Wales. Bur. Min. Res. J. Aust. Geol. Geophys., 4, 28 1 -288. MARSHALL, J. F. ( 1 9 80) Continental shelf sediments: southern Queensland and northern New South Wales. Bur. Min. Res. Aust. Bull., 207, 33 pp. MARSHALL, J. F. & DAVIES, P.J. ( 1 978) Skeletal carbonate variation on the continental shelf of eastern Australia.
Bur. Min. Res. J. Aust. Geol. Geophys. , 3, 85-92. MATSUDA, S. ( 1 987) Non-articulated coralline algae in the Ryukyus. Earth Monthly (Chikyu), 9, 1 62- 1 68 (in Jap anese). MATSUDA, S. ( 1 989) Succession and growth rates of en crusting crustose coralline algae (Rhodophyta, Cryp tonemiales) in upper fore-reef environment off Ishigaki Island, Ryukyu islands. Coral ReefS, 7, 1 85- 1 95. MILLIMAN, J.D. ( 1 974) Marine Carbonates. Springer Verlag, Berlin. MINNERY, G.A. ( 1 990) Crustose coralline algae from the Flower Garden Banks, Northwestern Gulf of Mexico: controls on distribution and growth morphology. J. sediment. Petrol. , 60, 992- 1 007. MINOURA, K. & NAKAMORI, T. ( 1 982) Depositional envi ronment of algal balls in the Ryukyu Group, Ryukyu Islands, southwestern Japan. J. Geol., 90, 602-609. MONTAGGIONI, L.F. ( 1 979) Environmental significance of rhodolites from the Mascarene reef province, western Indian Ocean. Bull. Cent. Rech. Explor. Prod. Elf Aquitaine, 3, 7 1 2-722. NELSON, C.S. ( 1 988) Non-tropical shelf carbonates modern and ancient. Sediment. Geol., 60, 5 1 -70. NELSON, C.S., KEANE, S.L. & HEAD, P.S. ( 1 988) Non tropical carbonate deposits on the modern New Zealand shelf. Sediment. Geol., 60, 7 1 -94. NODA, M. ( 1 9 84a) Ryukyu Limestone ofOkinoerabu-jima (Part I )-Stratigraphy. J. geol. Soc. Japan, 90, 26 1 -270 (in Japanese with English abstract). NODA, M. ( 1 984b) Ryukyu Limestone of Okinoerabujima (Part 2)-Sedimentary facies. J. geol. Soc. Japan, 90, 3 1 9-328 (in Japanese with English abstract). PALMIERI, V. ( 1 984) Neogene Foraminiferida from GSQ Sandy Cape l -3R bore, Queensland: a biostratigraphic appraisal. Palaeogeogr. Palaeoclimatol. Palaeoecol., 46, 1 65 - 1 8 3 . P HIPPS, C.V.G. ( 1 970) Dating o f eustatic events from core taken in the Gulf of Carpentaria, and samples from the New South Wales continental shelf. Aust. J. Sci. , 32, 329-330. PIGRAM, C.J., DAVIES, P.J. & CHAPRONIERE, G.C.H. ( 1 993) Cement stratigraphy and the demise of the Early Middle Miocene carbonate platform on the Marion Plateau. In: McKenzie, J.A., Davies, P.J., Palmer-Jut son, A. et al., 499-5 1 2. Ocean Drilling Program, College Station, TX. REID, R.P. & MACINTYRE, l.G. ( 1 988) Foraminiferal-algal nodules from the eastern Caribbean: growth history and implications on the value of nodules as paleoenviron mental indicators. Palaios, 3, 424-43 5 . ROEHL, P.O. & CHOQUETTE, P.W. ( 1 98 5) Perspectives on world-class carbonate petroleum reservoirs. Am. Assoc. petrol. Geol. Bull., 69, 1 48 (abstract). SATO, T. & TAKAYAMA, T. ( 1 990) Age determination based on calcareous nannofossils. J. lap. Assoc. petrol. Tech no!., 55, 1 2 1 - 1 28. SCHOLLE, P.A., BEBOUT, D.G. & MOORE, C.H. (Eds) ( 1 983) Carbonate Depositional Environments. Mem. Am. As soc. petrol. Geol., 33, 708 pp. SHEPARD, F.P. ( 1 963) Submarine Geology, 2nd edn. Harper and Row, New York. STUIVER, M. & POLACH, H.A. ( 1 977) Discussion: reporting on C- 1 4 data. Radiocarbon, 19, 3 5 5-363.
Subtropical carbonate plat form development SYMONDS, P.A., DAVIES, P.J. & PARISI, A. ( 1 983) Structure and stratigraphy of the Central Great Barrier Reef. Bur. Min. Res. J. Aust. Geol. Geophys., 8, 277-29 1 . TsuJI, Y. ( 1 993) Tide influenced high energy environ ments and rhodo1ith-associated carbonate deposition on the outer shelf and slope off the Miyako Islands, southern Ryukyu Arc, Japan. Mar. Geol. , 1 13, 25 5-27 1 . TSUJI, Y., SUNOUCHI, H., YAMAMURA, T., KODATO, T., YUKI, T. & IWAMOTO, H. ( 1 989) Present-day carbonate sediments and their depositional environments, western offshore Miyako Island, Ryukyus. Earth Monthly (Chikyu), 11, 6 1 2-6 1 7 (in Japanese). TsuJI, Y., HONDA, N., MATSUDA, H. et a!. ( 1 994) Sedimen tology of subtropical to temperate carbonates and the East Australian Current off eastern Australia. Earth Monthly (Chikyu), 16, 407-4 1 1 (in Japanese). VEEH, H.H. & VEEVERS, J.J. ( 1 970) Sea-level at - 1 75 m off the Great Barrier Reef 1 3600 to 1 7000 years ago. Nature, 226, 5 36-537. WASS, R.E., CONOLLY, J.R. & MACINTYRE, R.J. ( 1 970) Bryozoan carbonate sand continuous along southern
195
Australia. Mar. Geol., 9, 63-73. WEISSEL, J.K. & WATTS, A.B. ( 1 979) Tectonic evolution of the Coral Sea Basin. J. geophys. Res., 84, 4572-4582. WERNICKE, B. ( 1 98 1 ) Low-angle normal faults in the Basin and Range province: nappe tectonics in an extending orogen. Nature, 291, 645-648. WERNICKE, B. ( 1 985) Uniform-sense normal simple shear of the continental lithosphere. Can. J. Earth Sci., 22, 1 05 - 1 2 5 . WYRTKI, K . ( 1 960) The surface circulation in the Coral and Tasman Seas. CSIRO Div. Fish. Ocean. Tech. Pap., 8, 1 -44. WYRTK, K. ( 1 966) East Australia Current. In: Encyclopae dia of Oceanogtophy (Ed. Fairbridge, R.). Reinhold, New York. YuKI, T., IWAMOTO, H., TsuJI, Y. et a!. ( 1 988) Cyclic sedimentary sequences and their lateral variation in the Ryukyu Limestone (Pleistocene), lrabu Island, Oki nawa, Japan. Soc. econ. Paleont. Miner. Annual Midyear Meeting. Abstracts, p. 59.
Spec. Pubis int. Ass. Sediment. (1998) 25, 197-213
Pleistocene reef complex deposits in the Central Ryukyus, south-western Japan Y. I R Y U , T. NAKAMO RI and T. YAMADA Institute of Geology and Paleontology, Graduate Sch ool ofScience, Tohoku University, Aobayama, Sendai 980-77, Japan
ABSTRACT The Ryukyu Group, composed of Pleistocene reef complex deposits that locally pass laterally into terrestrial sediments, is extensively distributed over the Ryukyu Islands. The carbonate rocks are divided into four facies: coral, rhodolith, Cycloclypeus-Opercu/ina, and poorly sorted detrital limestones. Their depositional environments are specified based on the distribution and depth range of the present-day reef biota and associated sediments around the Ryukyu Islands. The stratigraphical succession of the Ryukyu Group is investigated at Toku-no-shima, Okierabu-jima and Yoron-jima, Central Ryukyus. Here reef complex deposits are associated with terrestrial sediments formed at relatively high and low sea-level stands. The highstand deposits are thick, occur extensively, and consist of terrestrial and marine conglomerates, and coral, rhodolith and poorly sorted detrital limestones that are arranged from inland proximal to coastal distal parts. Lowstand deposits are thin, composed mainly of coral limestone, and distributed in very limited areas at elevations less than the highstand deposits. Abundant rhodoliths occur in the deep fore-reef to insular shelf areas in the Pleistocene to present-day Ryukyus, which perhaps indicates that nutrient-rich marine environments observed in Halimeda banks have never prevailed over the shelves on the Ryukyus.
INTRODUCTION The Ryukyu Islands (Ryukyus) are located to the south-west of mainland Japan and consist of several tens of islands and islets, extending from Tanega shima (30° 44'N, 131oO'E) in the north-east to Yonaguni-jima (24o27'N, 123oO'E) in the south west (Fig. 1). These islands are an active island arc (Ryukyu Arc) generated by subduction of the Phil ippine Sea Plate beneath the Eurasian Plate and bounded by the East China Sea on the north-west and by the Pacific Ocean on the south-east. Pleistocene reef complex (coral reef and associ ated fore-reef and moat) deposits accompanied by terrestrial sediments are distributed over most of the islands of the Central and Southern Ryukyus and reach up to c. 200 m in elevation. These deposits have been called the Riukiu Limestone (Yabe & Hanzawa, 1930) or the Ryukyu Group (MacNeil, 1960). The name 'Ryukyu Group' is adopted in this study, because the unit contains both carbonate and siliciclastic rocks.
Fig. 1. Map showing localities of the study area.
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
197
198
Y Iryu, T. Nakamori & T. Yamada
Stratigraphical studies on the Ryukyu Group were carried out from the morphostratigraphical viewpoint in the 1960s (Nakagawa, 1967, 1969). In this paradigm, the Group was regarded as terrace forming deposits: coral reefs were developed at high sea-level stands and the islands have risen to form a flight of terraces from inland to the coast. Uranium series dating was introduced to the field of studying the Group by Konishi and his co-workers (Konishi, 1967; Konishi et a!., 1970, 1974). They revised the stratigraphy of the Group based on the resulting ages, conducted a correlation of the Group with reef deposits in other regions such as Barbados and New Guinea, and estimated rates of vertical displace ment of the Ryukyu Islands. The Okinawa Quater nary Research Group (1976) and Takayasu (1976, 1978) expressed a quite different opinion from the morphostratigraphers. They divided the Ryukyu Group into two basic stratigraphical units: the older, main constituent unit with extensive distri bution; and the younger terrace-forming deposits with limited distribution. The older unit was thought not to be coral reef deposits but to be basinal sediments. In the stratigraphical works of these decades, sedimentological and palaeontologi cal features were not fully analysed, so that neither an anatomy of reef complex deposits nor a palaeo bathymetric interpretation was provided. Stratigraphical and sedimentological studies of the Ryukyu Group since the middle of the 1970s have shown that the Group is composed principally of plural reef complex deposits which are compara ble with those of the present-day Ryukyus (Noda, 1976; Nakamori, 1986; Iryu et a!., 1992). The extensive investigations of the distribution of mod ern biota and sediments around the Ryukyus (Na kamori, 1986; Iryu & Matsuda, 1988; Matsuda et a!., 1992; Tsuji, 1993, Iryu et a!., 1995) have allowed precise palaeoenvironmental determina tions, especially of palaeobathymetry, of the bio and lithofacies of the Group(Nakamori, 1986; Iryu, 1992; Iryu et a!., 1992; Nakamori et a!., 1995a). This paper aims to provide a basic framework for establishing a reef stratigraphy for the Ryukyu Group. We here present a classification of carbonate and siliciclastic rocks of the Group and specify dep ositional environments of bio- and lithofacies based on the known distribution of marine biota and sed iments around the Ryukyus. Using such a frame work, the Group is shown to be divisible into plural reef complex deposits by referring to stratigraphical examples of the Ryukyu Group in the islands of
Toku-no-shima, Okierabu-jima and Yoron-jima, Central Ryukyus (Fig. 1). The Pleistocene reef for mations in the three islands are discussed.
LITHOLOGY OF THE RYUKYU GROUP AND DEPOSITIONAL ENVIRONMENTS The Ryukyu Group consists of both carbonate and siliciclastic rocks (Table 1). Six major facies have been identified in the carbonate rocks: coral, rhod olith, Cycloclypeus-Operc ulina, Halimeda, poorly sorted detrital, and well-sorted detrital limestones (Nakamori et a!., 1995a). Of these, two facies (Halimeda and well-sorted detrital facies) do not occur in the study area. Carbonate rocks Cora/limestone
·
Coral limestone is defined as a limestone in which autochthonous hermatypic corals are contained. The volume of hermatypic corals can be as high as 40%. This facies is up to 50 m thick and occurs at the proximal part of a single reef complex deposit, overlying conglomerate and sandstone or the base ment rocks. It grades laterally into distal rhodolith limestone. This facies can be subdivided into two subfacies. One is characterized by the occurrence of autochthonous corals embedded with bioclasts of corals, coralline algae, foraminifers and molluscs, forming a floatstone-like structure (Fig. 2A). The other comprises massive to encrusting forms of hermatypic corals and nongeniculate coralline algae that accumulate to form a framestone structure (Fig. 2B). Dendritic (i.e. branching) corals act as baffles in places. The former sub facies is much more abundant than the latter. Hermatypic corals are distributed from reef flat to fore-reef slope down to depths of 100 m in the present-day Ryukyu Islands. Iryu et a!.(1995) noted that the corals are common to abundant to depths of 50 m and that they are extremely reduced in number and diversity at greater depths. It follows that the coral limestone largely accumulated at depths up to 50 m. Depositional depth of the coral limestone can be determined more precisely by examining fossil coral communities and non geniculate coralline-algal assemblages. Nakamori (1986) and Iryu (1992), respectively, discriminated
Table 1.
Classification of carbonate and siliciclastic rocks of the Ryukyu Group
Carbonate rocks Coral limestone
Rhodolith limestone
Cycloc/ypeus-Operculina limestone
Poorly sorted detrital limestone Siliciclastic rocks Conglomerate and sandstone
Expected depositional environment
Definition
Sedimentological features
Palaeontological features
Limestone (framestone) with autochthonous hermatypic corals
Massive or bounded by corals and coralline algae to show biological framework structures
Associated with coralline algae, echinoids, molluscs and larger foraminifers
Reef flat and fore-reef slope (0-50 m deep)
Limestone (rudstone) consisting of more than 20% by volume rhodoliths
Massive but exhibiting parallel- to cross-bedding in places
Associated with C. carpenteri and solitary corals
Insular shelf (50-150 m deep)
Limestone (rudstone) with abundant C. carpenteri and/or 0. ammonoides
Massive with micrite
Associated with rhodoliths
Depths from 50 to !50 m where C. carpenteri dominates and I 0-200 m where 0. ammonoides dominates
Poorly sorted limestone (packstone) composed of organic skeletons
Parallel- or cross-bedding conspicuous
Siliciclastic conglomerate with intercalated beds of sandstone
Massive with muddy matrix Parallel- or cross-bedding conspicuous
� <;;· 0 <":> 'I> � 'I>
Rich in bryozoans associated with molluscs and brachiopods
Insular shelf (>50 m deep)
No fossils
Fluvial or coastal plain
Associated with 0. ammonoides, molluscs, bryozoans and nanofossils
From inside of reef flat to insular shelf
� � <":> a
� � �
'1:5 a ""'
�-
�
""= :;;::
� :;;:: ""'
'oO 'oO
200
Y lryu, T. Nakamori & T. Yamada
Fig. 2. Carbonate rocks of the Ryukyu Group. (A) Coral limestone with massive hermatypic corals, Porites sp., on an outcrop at Kunigami-misaki (Cape Kunigami), Okierabu-jima. Note that the corals are 'floating' in surrounding bioclasts. (B) Coral limestone with encrusting form of corals and non-geniculate coralline algae, showing a biological framework structure, exposed at Kametsu, Toku-no-shima. (C) Rhodolith limestone that crops out north of Masana, Okierabu-jima. Note concentrations of rhodoliths. The rhodoliths comprise multiple species of nongeniculate coralline algae and/or contain several individuals of an encrusting foraminifer Acervulina inhaerens. (D) Cycloc/ypeus Operculina limestone dominated by C. carpenteri (arrowed) from Uchijiro, Okierabu-jima. (E) Cycloclypeus-Operculina limestone dominated by 0. ammonoides (small lenticular forams) from Okidomari, Okierabu-jima. (F) Well-bedded, poorly sorted detrital limestone exposed to the west of Tamina, Okierabu-jima.
Pleistocene reef complex deposits, Ryukyus five coral commumties and four coralline-algal assemblages, each of which represents a particular depth range (Table 2). Rh odolith limestone The rhodolith limestone, characterized by rocks with more than 20% of the total volume made up of rhod oliths, is widespread, encloses areas of coral lime stone, and grades laterally into poorly sorted detrital limestone(Fig. 2C). It is mostly less than 20 m thick. This limestone is generally massive, although parallel- to cross-bedding is occasionally observed. The rhodoliths are well rounded, range in mean diameter from 0.5 to 8 em, and consist mainly of multiple species of thin, encrusting nongeniculate coralline algae and an encrusting foraminifer Acer vulina inhaerens, which together form an irregularly concentric internal structure. These rhodoliths occur associated with poorly sorted bioclasts of coralline algae, larger foraminifers ( Cycloc lypeus carpenteri and Operculina ammonoides), molluscs and bryozo ans. Solitary corals and brachiopods are also found. It is well documented that the modern rhodoliths analogous to fossil forms in the Ryukyu Group are extensively distributed on the insular shelves rang ing in depth from 50 to 150 m around Miyako-jima (Tsuji, 1993) and Okinawa-jima (Iryu et a!., 1995). Thus, the rhodolith limestone is thought to have been deposited in such deep fore-reef to insular shelf areas. Cycloclypeus-Operculina limestone This limestone is distinguished from the other facies by the abundance of larger foraminifers, C. carpenteri and/or 0. ammonoides. In most cases, this limestone is found associated with the rhoda lith limestone and exposed sporadically, forming a relatively thin massive bed (less than 5 m). It consists of up to pebble-sized bioclasts of coralline algae, bryozoans, foraminifers and molluscs as well as the tests of C. carpenteri and/or 0. ammonoides, and displays a grain-supported texture. Both fora minifers can occur together but in general one or the other predominates in this facies (Fig. 2D & E). It was confirmed that C. carpenteri inhabits the deep fore-reef to island shelf zones corresponding to the range of rhodoliths off Miyako-jima (Tsuji, 1993) and Okinawa-jima (Iryu et a!., 1995). Kodato & Nakagawa (1993) reported that 0. ammonoides occurs at depths from 30 to 60 m off Miyako-jima,
201
whereas Reiss & Hottinger (1984) found this spe cies in a much broader depth range ( 10-140 m) in the Gulf of Aqaba. It can be concluded that the limestones dominated by C. carpenteri and by 0. ammonoides were deposited at depths of 50-150 m and 10-200 m, respectively. Poorly sorted detrital limestone This limestone consists mainly of poorly sorted, coarse-grained sand- to pebble-sized bioclasts of foraminifers, bryozoans, corals and coralline algae, with less abundant molluscs and brachiopods (Fig. 2F). Autochthonous hermatypic corals are en tirely lacking. Rhodoliths, C. carpenteri and 0. ammonoides are found in places. Intergranular space is filled with micrite. It occurs at the most distal part of a single reef complex deposit of the Ryukyu Group and reaches more than 50 m in thickness. Rhythmic stratification (5-20 em thick) is common, with occasional large-scale (up to 1 m thick and 5-20 m across) cross-bedding. Common to abundant occurrences of bryozoan skeletons in this facies indicate that it was deposited on the insular shelf at depths greater than 50 m, because Iryu et a!. (1995) showed that the skeletons are comparably more abundant on the shelf at depths from 40 to 160 m off Miyako-jima. Siliciclastic sediments The siliciclastic sediments of the Ryukyu Group con sist of terrestrial and marine conglomerate accom panied by sandstone, and their thickness reaches up to 90 m. The terrestrial conglomerate consists of poorly sorted, angular to subangular, up to boulder-sized gravel with massive, muddy to sandy matrix (Fig. 3A). It occurs, overlying the basement rock, at the most proximal part of a single reef complex deposit and grades laterally into the marine con glomerate. The clasts are of metamorphic slate, sandstone and basalt derived from the basement rock. Generally, a clast-supported texture is evident in this facies. The marine conglomerate is characterized by poorly sorted, subangular to subrounded granules and pebbles, and by a matrix of fine- to coarse grained sand(Fig. 3B & C). It overlies the terrestrial conglomerate or covers directly the basement rocks. It is more or less bedded, commonly showing well defined cross-bedding. Fossil marine organisms
N 0 N
Fossil coral communities and nongeniculate coralline-algal assemblages of the Ryukyu Group and their depositional environments; the comparable modem coral communities and algal assemblages were defined by Nakamori (1986) and lryu & Matsuda (1988), respectively
Table 2.
Community or assemblage
Characteristic species or taxonomic groups
Other features
Hermatypic corals Acropora spp. (Branching) Community A Porites spp.
Expected depositional environment
Comparable modem community or assemblage
Moat to reef crest of fringing reef or protected shallow water of patch reefs
Porites cylindrica Com. Porites nigrescens Com. Heliopora coerulea Com.
Community B
Acropora spp. (Tabular)
Reef edge
Acropora hyacinthus Com.
Community C
Acropora spp. (Tabular) Porites spp. Favia spp. Platygyra spp.
Fore-reef slope (0-15 m deep)
Favia stelligera Com.
Community D
Pectiniidae Favia spp. Platygyra spp.
Fore-reef slope (10-30 m deep)
Oxypora lacera Com.
Community E
Leptoseris spp. Pachyseris spp.
Fore-reef slope to insular shelf (>30 m deep)
Leptoseris scabra Com.
� ......
�
F �
�
iS �
Occasionally associated with Cycloclypeus carpenteri, Operculina ammonoides and rhodoliths
0
:::!. � �
;;:: �
Nongeniculate coralline algae Hydrolithon onkodes Assemblage A Lithophyllum insipidum
0-20 m deep
Assemblage I
Assemblage B
Hydrolithon murakoshii Neogoniolithon fosliei Neogoniolithon sp. A sensu lryu & Matsuda (1994) Lithophyllum insipidum
20-35 m deep
Assemblage II
Assemblage C
Mesophyllum purpurascens
35-50 m deep
Assemblage III
Assemblage D
Lithothamnion australe Lithothamnion sp.
Associated with C. carpenteri and " rhodoliths
>50 m deep
�
�
Pleistocene reef complex deposits, Ryukyus
203
Fig. 3. Siliciclastic rocks of the Ryukyu Group. (A) Terrestrial conglomerate with subangular to subrounded clasts of the Mesozoic volcanic or volcaniclastic rocks on an outcrop near Yakomo, Okierabu-jima. (B) Thick, well-bedded marine conglomerate of unit I overlain by the coral limestone (arrowed) of unit 2, exposed at the mouth of Okuna-gawa, Toku-no-shima. Note an arrowhead indicating their boundary. (C) Marine conglomerate with thin layers rich in Operculina ammonoides, at Wanjouhama, Okierabu-jima.
such as corals, molluscs, larger foraminifers (0. am monoides), bryozoans and echinoids are abundant, which makes this conglomerate calcareous(Fig. 3C). Gravel with thin coralline-algal crusts (algal-coated gravel) occasionally occurs, forming concentrations. It is impossible to determine the depositional envi ronment of the marine conglomerate precisely. A broad depth range from 0 to 200 m appears to be the best estimate, and is based on the occurrence of var ious marine fossils including 0. ammono ides (10200 m deep).
STRATIGRAPHY Toku-no-shima In Toku-no-shima, basement rocks of the Ryukyu Group are Mesozoic metamorphic slate, sandstone and basalt, and Tertiary granitic rocks, forming hilly areas(up to 600 m in elevation) at the centre of the island. The Group consists of conglomerate and coral, rhodolith and poorly sorted detrital lime stones, and crops out on the hillside to the coastal plain at less than 200 m elevation (Figs 4 and 5). The following four stratigraphical units (units 1-4) are discriminated.
204
Y.
Iryu, T. Nakamori & T. Yamada
B'
Toku-no-shima
2km
'======---
Legend
AD B-1� C-16 C-2· C-3§ C-40 D-1� . . F --- --- E-2 Do D-2 D E-1 � ...,..'-'"• ,'.•;' . ,..•.. . . . E-3 !§§! E-4·Qlill] � G >-;.,..,_ • "•
·
Fig. 4. Geological map of the southern part of Toku-no-shima. (A) Recent beach and alluvial deposits; (B-E), the
Ryukyu Group: (B) coral limestone of unit 4; (C) unit 3 (C-1, coral limestone; C-2, rhodolith limestone; C-3, detrital limestone; C-4, conglomerate and sandstone); (D) unit 2 (D-1, coral limestone; D-2, conglomerate and sandstone); (E) unit I (E-1, coral limestone; E-2, rhodolith limestone; E-3, detrital limestone; E-4, conglomerate and sandstone). (F) basement rocks; (G) fault.
Unit 1, the lowest unit of the Group, is widely exposed in the southern part of Toku-no-shima and is distributed from 0 to 120 m elevation. The thickness is up to 90 m. It consists of conglomerate and coral, rhodolith and poorly sorted detrital limestones. The terrestrial conglomerate accumu lates on the basement rocks and grades laterally into marine conglomerate with intercalated beds of sandstone. The coral limestone overlies the con glomerate or unconformably covers the basement rocks. It occurs continuously from 0 to 120 m elevation. The rhodolith and detrital limestones are laterally equivalent to the coral limestone. Unit 2 (less than 30 m in thickness) consists of thin coral limestone and conglomerate and crops out at elevations of 20-70 m with very limited distribution. It lies unconformably on the rhodolith limestone and conglomerate of unit 1 (Fig. 3B). Unit 3 is another reef complex deposit, consisting
of conglomerate and coral, rhodolith and poorly sorted limestones, which overlies units 1 and 2. These four facies are distributed from inland prox imal to coastal distal parts in this order, and are arranged more or less parallel to the coast. Unit 3 overlies unit 1 unconformably and covers unit 2 conformably. It is distributed up to 200 m elevation and its thickness reaches 70 m. Unit 3 forms ter races at two levels: a higher terrace at about 160200 m elevation (the Itokina Terrace of Nakagawa (1967)), consisting of coral limestone and corre sponding to the former reef flat; and a lower terrace at elevations from 30 to 70 m(the Kametsu Terrace of Nakagawa (1967 }}, composed of the rhodolith and detrital limestones which accumulated on the insular shelf. Unit 4 crops out in the southern part of Toku-rio shima with limited, sporadic distribution. It con sists mainly of coral limestone associated with
205
Pleistocene reef complex deposits, Ryukyus m 200
100
A'
Fig. 5. Geological cross-sections of
the southern part of Toku-no shima across the lines A-A' and B-B' indicated in Fig. 4. Legend as in Fig. 4.
B' 2km ==--===---
rhodolith limestone and conglomerate containing pebble- to boulder-sized clasts of the limestones assignable to the lower units. It is less than 5 m in thickness. Coral limestone of this unit unconform ably overlies unit 3 at several localities less than 50 m in elevation. Okierabu-jima In Okierabu-jima, basement rocks of the Ryukyu Group are Mesozoic sedimentary rocks (turbidite and volcaniclastic-volcanic rocks), and Tertiary granodiorite and porphyrite. They form a topo graphic high and are unconformably overlain by the Ryukyu Group, which covers most of the island except for O-yama (246 m in elevation) and crops out at 0-200 m elevation. Two sedimentary units are discriminated in the Group (Figs 6 & 7). The lower unit consists of conglomerate with intercalated beds of sandstone, and reef complex limestones. This unit occurs extensively, rising to 130 m elevation, and is up to 90 m thick. Thick terrestrial conglomerate covers the basement rocks in the most proximal part of the unit. Marine conglomerate lies on the basement rocks on the northern coast or on the terrestrial conglomerate to the south-west of O-yama(Fig. 7). The reef complex limestones overlie the terrestrial and marine con glomerates or directly cover the basement rocks.
They consist mainly of coral limestone associated with poorly sorted detrital limestone. The coral limestone occurs continuously from 0 to 130 m elevation. The detrital limestone is laterally equiv alent to the coral limestone. The upper unit ( < 50 m thick) unconformably overlies the lower unit (Fig. 8). It is also composed of underlying conglomerate-sandstone and reef complex limestones. The lowest strata of this unit are proximal marine conglomerate and distal sand stone, which are distributed from 30 to 150 m elevation. The marine conglomerate is rather thin ( < 1.5 m thick) and occurs along a road encircling O-yama at 110-130 m. It passes laterally into the distal sandstone (thickening to 30 m) which is exposed to the north of O-yama and along the northern coast of the island. The sandstone is calcareous in places with abundant 0. ammonoides. Coral limestone occurs at elevations of 120-200 m, encircling O-yama. Rhodolith limestone and poorly sorted detrital limestone with abundant rhodoliths, Cycloclypeus and Operculina extend, surrounding the coral limestone. These strata are laterally equiv alent to the coral limestone. The upper unit forms terraces at two levels: the higher one, from 150 to 200 m elevation (the Shimoshiro Terrace of Naka gawa (1967)), is considered to have been a reef flat of the Pleistocene fringing reef; the lower one, from 30 to 100 m elevation (the Shinjo and Serikaku
206
Y lryu, T Nakamori & T Yamada
Okierabu-jima
n
n
lc
N Legend
-t-
AD B-1 m B-2 1!1 8 B-4 � B-5 G_] C-1 iji(;§ C-2 § B-3
B-
C-3
f}}t��
Em
0
2km
Fig. 6. Geological map of the western half of Okierabu-jima. A, Recent beach and alluvial deposits; B & C, the Ryukyu Group: B, upper unit (B-1, coral limestone; B-2, rhodolith limestone; B-3, Cycloclypeus-Operculina limestone; B-4, detrita1limestone; B-5, conglomerate and sandstone); C, lower unit (C-1, coral limestone; C-2, detrital limestone; C-3, conglomerate and sandstone); E, basement rocks.
Terraces of Nakagawa (1967)), is composed of insular shelf deposits. By stratigraphical position, reef geometry and topographical features (distributional elevations of the Ryukyu Group and terraces), the lower and upper units of Okierabu-jima are correlated with unit 1 and unit 3 in Toku-no-shima, respectively. Yoron-jima The basement of the Ryukyu Group is composed mainly of Mesozoic turbidites and volcaniclastic rocks with limestone olistholiths. This limestone is probably Permian in age. The Group unconform ably overlies the basement rocks and covers most of the island. It is less than 50 m thick and composed of siliciclastic rocks (conglomerate and sandstone)
and reef complex limestones (Fig. 9). There is a difference in stratigraphical succession of the Group between eastern and western parts of the island. In the eastern part, conglomerate and sandstone inter·· calating coral limestone overlie the basement rocks. These strata crop out in very limited areas at the surface, so their nature is uncertain, although litho.. logical data were given by some previous workers (e.g. Noda, 1976). They are overlain by coral, rhodolith and poorly sorted detrital limestones. The coral limestone occurs continuously from 0 to 97 m elevation and laterally grades into the rhodolith and poorly sorted detrital limestones exposed in the northern and western peripheries of the island. In the western part, sandstone and conglomerate ·are lacking and coral limestone overlies the basement rocks.
207
Pleistocene reef complex deposits, Ryukyus 0
0 A
0
250m 200 150 100
Fig. 7. Geological cross-sections of
the western half of Okierabu-jima across the lines 0-A, 0-B and 0-C indicated in Fig. 6. Legend as in Fig. 6.
The Ryukyu Group in Yoron-jima correlates with unit 1 in Toku-no-shima by stratigraphical position and topographical features. Age assignment Geological ages of the Ryukyu Group in three islands are determined by radiometric methods (uranium series and electron spin resonance (ESR) methods) and calcareous nanofossil biostratigraphy. Omura (1982) dated corals from the lower part of unit 1 exposed at the southern tip of Toku-no-shima by the U-series method. Although his resulting coral ages were close to or beyond the limitation of the 230Th/ 234U method(> 300 ka), he estimated the ages to be 387-709 ka from the mean 234UF38U activity ratio (Table 3). Nakamori et al. (199 Sa) measured ESR age at the point where Omura (1982) collected his samples and reported an age of 481 ± 17 ka. This agrees well with the U-series ages. Calcareous nano fossils detected from a lower horizon of unit 1 and lower and middle horizons of unit 3 by Nakamori et al. (1995a) indicate a CN14a zone(Okada & Bukry, 1980), the age of which ranges from 890 to 390 ka (Takayama & Sato, 1987). After considering all the data available, it is concluded that the ages of the Group in Toku-no-shima range widely from c. 400 to 900 ka.
50
f===--""'!1 km
Ikeda et al. (1991) reported ESR ages ranging from 730 to 840 ka for the lowest horizons of the lower unit in Okierabu-jima (Table 3). Calcareous nanofossils were recovered from the lowest horizon of the upper unit, indicating a CN14a zone (Okada & Bukry, 1980). Consequently, the age of the Ryukyu Group in Okierabu-jima is considered to be in a range from c. 400 to 900 ka. The same is true of the Group in Yoron-jima, where calcareous nano fossils detected from the uppermost horizon of the sandstone and conglomerate in the western part of the island indicate a CN14a zone and no radiomet ric age has been reported. As a consequence, the geological ages of the Ryukyu Group in the three islands are not well constrained, and range widely from c. 400 to 900 ka.
DISCUSSION Pleistocene reef formation in the Central Ryukyus Our stratigraphical and sedimentological study re veals that there are two different types of reef complex deposits on the three islands of the Central Ryukyus. One is represented by units 1 and 3 on Toku-no-shima and their correlative deposits on Okierabu-jima and Yoron-jima. The single unit is
208
Y Iryu, T Nakamori & T Yamada
A
Fig. 8. Field photographs
showing stratigraphical relationship between the lower and upper units of the Ryukyu Group in Okierabu-jima. (A) Coral limestone (c) overlain by Cycloclypeus-Operculina limestone dominated by 0. ammonoides (o) with thin, granule- to pebble-sized gravel layer at its base, exposed to the north of Serikaku. (B) Cycloclypeus Operculina limestone dominated by C. capenteri (cc) overlying coral limestone (c) on an outcrop near Ashikyora.
thick (up to 90 m) and extensively distributed on the islan·ds, and shows a wide range of elevations exceeding 100 m. It consists of a complete set of reef complex sediments: showing, from inland to the coast, terrestrial and marine conglomerates and coral, rhodolith and poorly sorted detrital lime stones. This arrangement is conformable with the distribution of biota and sediments in the present day reef complex around the Ryukyu Islands (lryu et a!., 1995). The other type includes units 2 and 4 on Toku-no-shima. Each unit is thin (less than 30 m), sporadically exposed, and composed mainly of conglomerate and coral limestone associated with occasional thin beds of rhodolith limestone. It does not form a complete set of reef deposits. These two types of reef complex deposits accumulated alternately: the former type of deposits lies at
greater elevations above sea-level (elevations of units 1 and 3 are 150 and 2.00 m, respectively) than the latter (units 2 and 4 are at 70 and 50 m, respectively). Taking account of these stratigraphi cal successions and positions, it is supposed that the: reef deposits constitute a record of glacio-eustatic sea-level changes, superimposed on a record of tectonic movement. The reef deposits at greater elevations(units 1 and 3 and their correlatives) may have accumulated during the rises of sea-level or at: highstands during interglacial episodes. Units 2 andl 4 may have been formed at lower stands of sea-level than units 1 and 3. These lowstands happened! twice, following the highstands that resulted in units 1 and 3. It is uncertain whether these lower stands correspond to other highstands in interglacial-· interstadial episodes or lowstands in glacial epi-·
209
Pleistocene reef complex deposits, Ryukyus
Yoron-jima N
t A'
A
Fig. 9. Geological map and
cross-section of Yoron-jima. A, Recent beach and alluvial deposits; B, the Ryukyu Group (B-1, coral limestone; B-2, rhodolith and detrital limestones; B-3, conglomerate and sandstone intercalating coral limestone); C, basement rocks; D, fault.
�:f
t:�m
� � m
A'
o�==-91km Legend
AD
sodes. At present, precise chronological data for the Ryukyu Group are apparently lacking, and thus it is impossible to specify the timing of reef formation in the Ryukyus and to correlate reef deposits with deep-sea oxygen isotope records. Two different types of reef deposits also occur in the Huon Peninsula, Papua New Guinea. It is known that a well-developed flight of terraces along the north-east coast of the Huon Peninsula consists of reef complex deposits with subordinate deltaic gravel formations (Chappell, 1 97 4, 1980), and thus this area is regarded as one of the best fields for studying Quaternary sea-level changes, similar to Barbados (Broecker et al., 1 968) and the Kabola Peninsula on Alor Island, Indonesia (Hantoro et al., 1 994). These terraces are thought to be an off lapping sequence of coral reefs, each comparable with modern reefs on the coast. In reality, the reef
B-1
a
B-2
�
B-3
0
c�
D�
deposits did not accumulate in such a simple manner as previously thought and their deposi tional history is more complex. Nakamori et at. ( 1 995b) showed that the deposits are divisible into several facies comparable with those of the Ryukyu Group, and undertook palaeoenvironmental inter pretation based on investigations of biota in the modern reefs on the Huon Peninsula, referring to the stratigraphical results of the Ryukyu Group. They found that both interglacial and interstadial reefs occur. The interglacial reefs are volumetrically much larger than the interstadial ones, with shallow facies characterized by abundant occurrences of shallow-water corals such as A. h yacinthus and A. palifera composing the interior bulk from inception to terrace surface (Nakamori et at., 1995b, Fig. 6). Consequently, the interglacial reefs maintained the shallow, frame-building communities throughout
2 10 Table 3.
Y.
Iryu, T. Nakamori & T. Yamada
Recently reported, radiometric and ESR ages of the Ryukyu Group in Okierabu-jim and Toku-no-shima
Locality
Sample number
Elevation (m)
Method of dating
Age (ka)
Reference
Toku-no-shima Isen-zaki
80-2-9-5
<9
23oTh/234u
Omura (1982)
lsen-zaki
75-11-10-1
<9
23oTh/234u
Isen-zaki
75-11-10-2
<9
23oThf234u
lsen-zaki
<9
23oTh/234u
lsen-zaki
80-2-9-5 75-11-10-1 75-11-10-2 817-9
>468 423�n 486�13 4 529��9 >407 301��� 387-709
<9
ESR
481 ± 17
Nakamori eta!. (1995a)
Okierabu-jima Kunigami-misaki Kunigami-misaki Kunigami-misaki Kunigami Kunigami
OK-6 OK-9 OK-10 OK-4 OK-11
I l I
ESR ESR ESR ESR ESR
820 ± 80 790 ± 110 770 ± 160 730 ± 70 840 ± 100
Ikeda eta!. (1991)
30 30
the sea-level rise and appear to be keep-up reefs in the sense of Neumann & Macintyre (1985) and Davies et a!. ( 1985). In marked contrast, the inter stadial reefs are much smaller and thinner, and lie on the distal part of the interglacial reef deposits, showing a shallowing-upward sequence. Pleistocene reef deposits younger than unit 4 crop out on none of the three islands, although those deposits which were formed in the penultimate and last interglacial (oxygen isotope stages 7 and 5, respectively) and subsequent interstadial (oxygen isotope stage 3) episodes occur in Kikai-jima, which is located c. 120 km north-east of Toku-no-shima and is the closest island to the Ryukyu Trench (c. 80 km) in the Central Ryukyus (Fig. 1). This island has been rising rapidly for the last 130 kyr. The rate of vertical displacement is estimated as 2.5-2.6 and 1.7-1.8 mm yc1 for the last 83 and 129 kyr, respec tively (Omura, 1988). If the three islands had been uplifted like Kikai-jima, the older limestone (units 1-4 and their correlatives) would have been sur rounded by reef deposits formed in the successive episodes of highstands, correlated to stages 7 to 3. Non-occurrence of such younger reef deposits may suggest that the three islands appear to have been neotectonically stable after they attained their present elevation. The absence (Okierabu-jima) or very limited occurrences (Toku-no-shima and Yoron-jima) of the raised Holocene reefs, in marked contrast to their good development in Kikai-jima, may reinforce the evidence for this explanation. The
reef deposits formed during the last c. 200 kyr are thin (generally less than 10 m thick) and overlie the thick (up to 50 m) older deposits dated at 400600 ka by the ESR method (Koba et al., 1985; Omura, 1988) which are extensively distributed covering the Pliocene to earliest Pleistocene silici clastic rocks (the Somachi Formation). Recent strati graphical investigations show that the reef deposits formed in the last interglacial episode do occur in Ie-jima located about 10 km west of Okinawa but that they are found in extremely small rock bodies or clots on older limestones (N. Fujishiro & K. Sasaki, unpublished data). These findings suggest that there have been considerable variations in the develop ment of reef complexes from the middle Pleistocene to the present in the Central Ryukyus, and raise an other possibility to explain the non-occurrence of the younger reef deposits in the three islands: after the deposition of unit 4, coral reefs may not have been formed, or were only poorly developed and then worn away by subsequent erosion. Fore-reef rhodoliths of the Ryukyu Group A notable feature of the Ryukyu Group is the extensive occurrence of rhodoliths in the distal part of reef complex deposits, enclosing the proximal lateral equivalent, coral limestone. The modem counterparts of fossil rhodoliths of the Ryuk:yu Group are distributed in deep fore-reef to insular shelf areas throughout the Central and Southern
Pleistocene reefcomplex deposits, Ryukyus Ryukyus (Tsuji, 1993; Iryu et a!., 1995), but their distributions do not extend to the Northern Ryukyus. Although modern and semi-fossil ( 1015 ka) rhodoliths were found off the Northern Ryukyus to southern Kyushu at depths of c. 40440 m, they differ from the rhodoliths around the Central and southern Ryukyus in being constructed mainly of nongeniculate coralline algae and bryozo ans associated with peyssonneliacean algae and in abundant occurrences of fruticose coralline algae. The rhodoliths occurring on tropical reef-associated shelves, banks and seamounts were termed fore-reef rhodoliths by Bosence ( 1983), and have been found in the Mascarene Archipelago (Montaggioni, 1979), Gulf of Mexico (Rezak et a!., 1985; Minnery, 1990), eastern Caribbean (Reid & Macintyre, 1988) and San Salvador (Littler et a!., 199 1 ), as well as the Ryukyus (Tsuji, 1993; Iryu et a!., 1995). These modern fore-reef rhodoliths and fossil forms of the Ryukyu Group have the following characteristics: (i) the rhodoliths are mostly pebble- to cobble-sized and spheroidal to ellipsoidal in shape; (ii) they are composed mainly of several individuals or species of nongeniculate coralline algae and several individ uals of encrusting foraminifers (Acervulina in haerens in the Pacific and Indian Oceans and Gypsina plana in the Atlantic), forming concentric to irregular foliation interiorly; and (iii) the rhodolith-forming algae are mainly encrusting to warty to lumpy rarely associated with fruticose forms (Woelkerling et a!., 1993). The abundant occurrences of the fore-reef rhodoliths having these common features suggest that they can be excellent indicators of depositional environments not only for the Ryukyu Group but also for any of the other Cainozoic reef complex deposits. This is supported by the occurrence of fore-reef rhodoliths with simi lar morhology and constituents in the Pleistocene reef complex deposits on the Huon Peninsula (Mat suda, 1995; Nakamori et a!., 1995b). The abundant occurrences also indicate critical amounts of car bonate production and accumulation on the deep fore-reefs or shelves. In the case of the Ryukyu Group (unit 3 on Toku-no-shima), the rock volume of the rhodolith limestone reaches 15% of the volume of the coral limestone. It should be noted that, in the Ryukyu Islands, Halimeda is a minor constituent of the modern and Pleistocene fore-reef sediments in contrast to abun dant occurrences of the rhodoliths, although it produces biohermal build-ups (Halimeda banks) in some tropical, reef-associated shelves and banks:
211
the northern Great Barrier Reef (Marshall & Dav ies, 1988), the eastern Java Sea (Roberts et a!., 1988) and the northern Nicaraguan Shelf (Hallock et a!., 1988; Hine et a!., 1988). As 1 1 species of Halimeda occur in the modern Ryukyus (Tsuda & Kamura, 199 1), including major constituent species of the Halimeda banks, the floristic characteristics are not considered to be a critical factor controlling the development of shelves dominated by rhodo liths or Halimeda. Iryu et a!. ( 1995) stated that trophic conditions may be of critical significance because the incursion of cold, nutrient-rich water onto the Halimeda-dominant banks appears to be responsible for remarkable algal growth at the ex pense of reef-building corals (Marshall & Davies, 1988; Roberts et a!., 1988) in spite of the possible absence of conspicuous upwelling around the Ryukyus (M. Koga and K. Arakawa, unpublished data).
CONCLUSIONS 1 The Ryukyu Group is represented by siliciclastic rocks and limestones. The limestones are subdivided into four facies: coral, rhodolith, Cycloclypeus Operculina and poorly sorted detrital limestones. Palaeoenvironments and palaeobathymetry of these facies are determined by reference to the areal and depth distribution of the living reef biota and sedi ments in the reef complex around the Ryukyus. 2 Stratigraphical and sedimentological investiga tions reveal that the Ryukyu Group in Toku-no shima, Okierabu-jima and Yoron-jima comprises four reef complex deposits. Units 1 and 3 in Toku no-shima and their correlatives in the other islands were formed at relatively high sea-level stands, whereas units 2 and 4 were formed at relatively low sea-level stands. Each of the highstand reef deposits is thick and occurs extensively, consisting of a com plete set of reef complex deposits: terrestrial and marine conglomerates and coral, rhodolith and poorly sorted detrital limestones, arranged from in land proximal to coastal distal parts of the islands. In contrast, the lowstand deposits are thin, composed mainly of coral limestone, and distributed in very limited areas at elevations less than the deposits rep resenting the period of highstands. 3 In the Pleistocene to present-day Ryukyus, the deep fore-reef to insular shelf areas are dominated by rhodoliths which have similar morphology, internal structure and constituent organisms to those on
212
Y Iryu, T Nakamori & T Yamada
tropical reef�associated shelves, banks and sea mounts elsewheFe. It is probable that the deep fore reefs and insular shelves on the Ryukyus have been in a very different oceanographic (trophic?) condi tion from the nutrient-rich Halimeda-dominant shelves.
ACKNOWLEDGEMENTS We are grateful to Dr K. Mori and Dr K. Ishizaki for their valuable suggestions. We thank Dr A. Eisen hauser and Pr L. Montaggioni for helpful comments on our original manuscript. Thanks are also ex pressed to Mr J. Nemoto for photography and to Mr T. Hatsugai for drawing geological maps.
REFERENCES D.W.J. ( 1 983) The occurrence and ecology of Recent rhodoliths-A review. In: Coated Grains (Ed. Peryt, T.M.), pp. 225-242. Springer-Verlag, Berlin. BROECKER, W.S., THURBER, D.L., GoDDARD, J., Ku, T.-L., MATTHEWS, R.K. & MESOLELLA, K.J. ( 1 968) Milanko vitch hypothesis supported by precise dating of coral reefs and deep-sea sediments. Science, 159, 297-300. CHAPPELL, J. ( 1 974) Geology of coral terraces, Huon Peninsula, New Guinea: a study of Quaternary tectonic movements and sea-level changes. Geol. Soc. Am. Bull., 85, 553-570. CHAPPELL, J. ( 1 980) Coral morphology, diversity and reef growth. Nature, 286, 249-252. DAVIES, P.J., MARSHALL, J.F. & HOPLEY, D. (1985) Rela tionships between reef growth and sea level in the Great Barrier Reef. Proceedings of the 5th International Coral ReefCongress, Tahiti, 3, 95- 1 03. HALLOCK, P., HINE, A.C., VARGO, G.A., ELROD, J.A. & JAAP, W.C. ( 1 988) Platforms of the Nicaraguan Rise: examples of the sensitivity of carbonate sedimentation to excess trophic resources. Geology, 16, 1 1 04- 1 1 07. HANTORO, W.S., PIRAZZOLI, P.A., JOUANNIC, C. eta/. ( 1 994) Quaternary uplifted coral reef terraces on Alar Island, East Indonesia. Coral Reefs, 13, 2 1 5-223. HINE, A.C., HALLOCK, P., HARRIS, M.W., MULLINS, H.T., BELKNAP, D.F. & JAAP, W.C. ( 1 988) Halimeda bioherms along an open sea way: Miskito Cannel, Nicaraguan Rise, SW Caribbean Sea. Coral Reefs, 6, 1 73- 1 78. IKEDA, S., KAsuYA, M. & IKEYA, M. ( 1 99 1 ) ESR ages of Middle Pleistocene corals from the Ryukyu Islands. Quat. Res., 36, 6 1 -71. IRYU, Y. ( 1 992) Fossil nonarticulated coralline algae as depth indicators for the Ryukyu Group. Trans. Proc. palaeontol. Soc. lap., N S. , 167, 1 1 65- 1 179. IRYU, Y. & MATSUDA, S. ( 1 988) Depth distribution, abun dance and species assemblages of nonarticulated coral line algae in the Ryukyu Islands, southwestern Japan. Proceedings of the 6th International Coral Reef Sympo sium, Townsville, Qld, 3, 1 0 1 - 1 06.
BosENCE,
Y., NAKAMORI, T. & YAMADA, T. ( 1 992) A unit of lithostratigraphic classification of the Ryukyu Group, Pleistocene reef complex deposits. J. sedimentSoc. lap., 36, 57-66 (in Japanese, with English abstract). IRYU, Y., NAKAMORI, T., MATSUDA, S. & A BE, 0. ( 1 995) Distribution of marine organisms and its geological significance in the modern reef complex of the Ryukyu Islands. Sediment. Geol., 99, 243-258. KOBA, M., IKEYA, M., MIKI, T. & NAKATA, T. ( 1 985) ESR ages of the Pleistocene coral reef limestones in the Ryukyu Islands, Japan. In: ESR Dating and Dosimetry (Eds Ikeya, M. & Miki, T.), pp. 93- 1 04. Ionics, Tokyo. KODATO, T. & NAKAGAWA, H. ( 1 993) Recent benthic fora miniferal assemblages off Miyako Island, the Ryukyu Islands, southwest Japan. Rep. Techno!. Res. Center, Japanese National Oil Corporation, 24, 93- 1 1 0 (in Japanese, with English abstract). KoNISHI, K. ( 1 967) Rate of vertical displacement and dating of reefy limestones in the marginal facies of the Pacific Ocean-application by natural alpha-radioactive nuclides in biogenic carbonate rocks up to 150,000 years old. Quat. Res., 6, 207-233 (in Japanese, with English abstract). KONISHI, K., SCHLANGER, S.O. & OMURA, A. ( 1 970) Neo tectonic rates in the Central Ryukyu Islands derived from 230Th coral ages. Mar. Geol. , 9, 225-240. KONISHI, K, OMURA, A. & NAKAMICHI, 0. ( 1 974) Radio metric coral ages and sea level records from the late Quaternary reef complex of the Ryukyu Islands. Pro ceedings of the 2nd International Coral Reef Sympo sium, Brisbane, Qld, 2, 595-613. LITTLER, M.M., LITTLER, D.S., & HANISAK, D. ( 1 991) Deep-water rhodolith distribution, productivity and growth history at sites of formation and subsequent degradation. J. exp. Mar. Bioi. Ecol., 1 50, 163-182. MACNEIL, F. S., ( 1 960) The Tertiary and Quaternary Gastropoda of Okinawa. US. geol. Surv., Prof Pap., 339, 1- 1 48. MARSHALL, J.F. & DAVIES, P.J. ( 1 988) Halimeda bioherms of the northern Great Barrier Reef. Coral Reefs, 6, 1 39- 1 48. MATSUDA, S. (1995) Quaternary limestones on Huon Peninsula, Papua New Guinea with special reference to nonarticulated coralline algae (Rhodophyta, Coralli naceae). J. Geogr. , 104, 706-7 1 8 (in Japanese, with English abstract). MATSUDA, S., IRYU, Y. & NOHARA, M. ( 1 992) Rhodoliths on the deep forereef of Okinawa-jima, Ryukyu Islands. J. sedimentSoc. lap., 37, 1 09- 1 1 1 (in Japanese). MINNERY, G.A. ( 1 990) Crustose coralline algae from the Flower Garden Banks, northeastern Gulf of Mexico: controls on distribution and growth morphology. J. sediment Petrol. , 60, 992-1007. MoNTAGGIONI, L.F. ( 1 979) Environmental significance of rhodolites from the Mascarene reef province, western Indian Ocean. Bull. Cent. Rech. Explor., Prod. Elf Aquitaine, 3, 713- 723. NAKAGAWA, H. ( 1 967) Geology of Tokunoshima, Okier abujima, Yoronto, and Kikaijima, Amami Gunto. Part 1. Contrib. Inst. Geol. Paleontol. Tohoku Univ., 63, 1 -39 (in Japanese, with English abstract). NAKAGAWA, H. ( 1 969) Geology of Tokunoshima, Okier abujima, Yoronto, and Kikaijima, Amami Gunto. Part
IRYU,
·
Pleistocene reef complex deposits, Ryukyus 2. Contrib. Inst. Geol. Paleontol. Tohoku Univ., 68, 1 - 1 5 (in Japanese, with English abstract). NAKAMORI, T., ( 1 986) Community structures of Recent and Pleistocene hermatypic corals in the Ryukyu Is lands, Japan. Sci. Rep. Tohoku Univ., 2nd Ser. (Geol.), 56, 7 1 - 1 33. NAKAMORI, T., IRYU, Y & YAMADA, T. ( 1 995a) Develop ment of coral reefs of the Ryukyu Islands (southwest Japan, East China Sea), during Pleistocene sea-level change. Sediment. Geol. , 99, 2 1 5-23 1 . NAKAMORI, T., MATSUDA, S., OMURA, A. & 0TA, Y. ( 1 995b) Depositional environments of the Pleistocene reef lime stones at Huon Peninsula, Papua New Guinea, on the basis of hermatypic coral assemblages. J. Geogr. , 104, 725-742 (in Japanese, with English abstract). NEUMANN, A.C. & MACINTYRE, 1., ( 1 985) Reef response to sea level rise: keep-up, catch-up or give-up. Proceedings of the 5th International Coral Reef Congress, Tahiti, 3, 1 05- 1 1 0. NODA, M. ( 1 976) Ryukyu Limestone of Yoron-to (Yoron Island). J. geol. Soc. lap., 82, 367-381 (in Japanese, with English abstract). OKADA H. & BUKRY, D. ( 1 980) Supplementary modifi cation of code numbers to low-latitude coccolith bio stratigraphic zonation (Bukry, 1 973, 1 975). Mar. Micropaleontol., 5, 321 -325. OKINAWA QUATERNARY RESEARCH GROUP ( 1 976) Quater nary system of Okinawa and Miyako Gunto, Ryukyu Islands-especially on the stratigraphy of "Ryukyu Limestone". Earth Sci., 30, 1 45- 1 62 (in Japanese, with English abstract). OMURA, A. ( 1 982) Uranium series ages of the "Kametsu Formation", Riukiu Limestone on Tokuno-shima, Ryukyu Islands. Trans. Proc. palaeontol. Soc. lap., N S. , 101, 327-333. OMURA, A. ( 1 988) Geologic history of the Kiaki Island, Central Ryukyus, Japan: summary of uranium-series dating of fossil corals from the Riukiu Limestone. Mem. geol. Soc. lap. , 29, 253-268 (in Japanese, with English abstract). REID, R.P. & MACINTYRE, I. G. ( 1 988) Foraminiferal-algal nodules from the eastern Caribbean: growth history and
213
implications on the value of nodules as paleoenviron mental indicators. Palaios, 3, 424-435. REISS, Z. & HOTTINGER, L. ( 1 984) The Gulf of Aqaba. Springer-Verlag, Berlin. REZAK, R., BRIGHT, T.J. & McGRAIL, W. ( 1 985) Reefs and Banks of the Northern Gulf of Mexico. Wiley Interscience, New York. ROBERTS, H.H., AHARON, P. & PHIPPS, C.V. ( 1 988) Mor phology and sedimentol ogy of Halimeda bioherms from the eastern Java Sea (Indonesia). Coral Reefs, 6, 1 6 1 - 1 72. TAKAYAMA, T. & SATO, T. ( 1 987) Coccolith biostratigraphy of the North Atlantic Ocean, Deep Sea Drilling Project Leg 94. In: Initial Reports of the Deep Sea Drilling Project, 94 (Ruddiman, W.F., Kidd, R.B., Thomas, E. et a!.), pp. 65 1 -702. US Government Printing Office, Washington, DC. TAKAYASU, K. ( 1 976) The Quaternary limestones in the northern part of the Motobu Peninsula, Okinawa Is land. J. geol. Soc. lap., 82, 1 53- 1 62 (in Japanese, with English abstract). TAKAYASU, K. ( 1 978) "Ryukyu Limestone" of Okinawa jima, south Japan-a stratigraphical and sedimentolog ical study. Mem. Fac. Sci., Kyoto Univ., Ser. Geol. Mineral., 45, 1 33- 1 75. TSUDA, R.T. & KAMURA, S. ( 1 99 1 ) Floristic and geographic distribution of Halimeda (Chlorophyta) in the Ryukyu Islands. lap. J. Phycol., 39, 57-76. Tsun, Y. ( 1 993) Tide influenced high energy environ ments and rhodolith-associated carbonate deposition on the outer shelf and slope off Miyako Islands, south ern Ryukyu Island Arc, Japan. Mar. Geol. , 1 13, 25527 1 . WOELKERLING W M.J., IRVINE, L.M. & HARVEY, A.S. ( 1 993) Growth-forms in non-geniculate coralline red algae (Corallinales, Rhodophyta). Austr. Syst. Bot. , 6, 277293. YABE, H. & HANZAWA, S. ( 1 930) A stratigraphic study of Tertiary foraminiferous rocks in Taiwan. In: Jubilee Publication in Commemoration of Professor Tamaki Ogawa 's 60th Birthday, pp. 83- 1 26. Kobundo-shobo, Kyoto (in Japanese; original title translated).
Oceanic Reef Case Histories
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
Atolls and Volcanic Islands
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
Spec. Pubis int. Ass. Sediment. ( 1 998) 25, 2 1 9-236
Morphology and sediments of the fore-slopes of Mayotte, Comoro Islands: direct observations from a submersible W. -Ch. DULLO*, G. F. C A M OINt, D. BL OMEIER*, M. C OLONNAt, A. EISEN H A U E R§, G. F A U REII, J. CASAN O V At and B.A. T H OMASSIN�
*GEOMAR, Wischhofstr. 1-3, D-24148 Kiel, Germany; tCEREGE, URA 132 du CNRS, Univer site Aix Mar seille III, B P 80,
F-13545 Aix en Pr ovence, Cedex 4, France; tB RGM , B P 6009, Avenue de Concyr , F-45060 Orle ans Cedex 2, France; §Institut fur Geochemie, Goldschmidtstr. 3, D-37077 Gottingen, Germany; IILaboratoir e d'Hydr obiologie Marine, Univer site M ontpellier II, Case 93, F-34094 M ontpellier Cedex, France; and �Centr e d'Oce anologie de Mar saille, Traverse B atterie des Lions, F-13007 Mar seille, France
ABS T R ACT
Mayotte fore-slopes exhibit a distinct pattern in overall morphology, starting in the deep with an unlithified sedimentary wedge and slope, followed upwards by a cemented slope, and finally by a steep, almost vertical wall. On top of the wall, drowned reefs occur. Dated corals may reveal the history of sea-level changes indicating pristine reef growth during late isotope stage 3 (at 5 5-24 ka) at a present-day water depth greater than 80 m. A maximum sea-level drop of 1 50 m occurred during the last glacial maximum, around 20 ka. This lowering of sea-level is documented by karst features such as small caves and corroded and jagged surfaces. The phase of deglaciation is recorded by two give-up reef levels at 100-90-m water depth and 65-55-m water depth which we may relate to the Be�lling ( 1 4 ka) and post Younger Dryas ( 1 1 . 5 ka) meltwater pulses, known from the deep-sea record.
INTRO DUC TION
Reefs are recorders of environmental changes (Montaggioni & Macintyre, 199 1). Since the classi cal work of Walther (1888), raised reefs or reef terraces are well-known recorders of changing sea levels and have been the subject of numerous investigations (e.g. Broecker eta!., 1968; Mesolella et a!., 1969; Bloom et a!. , 1974; Chappell, 1974; Shackleton & Matthews, 1977; Fairbanks & Mat thews, 1978; Bender et a!., 1979; Kaufman, 1985; Aharon & Chappell, 1986; Dullo, 1990; Pirazzoli et a!., 1991; Vollbrecht & Meischner, 1993; Gvirtz man, 1994; Hantoro eta!., 1994). These studies are concentrated mainly on the uplifted and at present emerged part of reefs and reef terraces, whereas this paper focuses on the morphology of submerged reefs and tries to relate these features of the present day fore-slope to past sea-level changes. Following earlier studies using bathymetric pro files, bottom cameras and dredges (e.g. Macintyre,
1972), the invention of small submersibles has triggered the in situ investigation of fore-slopes. The sites that have been visited by submersibles for geological purposes are still small in number, al though the first deep diving expeditions date back to the early 1970s. The majority of geological investigations have been concentrated around the Caribbean (Moore et a!., 1976; Adey et a!., 1977; Land & Moore, 1977; Lighty et a!., 1978; James & Ginsburg, 1979; Ginsburg et a!., 1991; Grammer, 1991; Macintyre et a!. , 1991; Grammer & Gins burg, 1992), whereas few studies are known from the Pacific reef province (Harris & Davies, 1989; Lambert & Roux, 1991). Equivalent data for the Indian Ocean are limited to the Red Sea (Fricke & Landmann, 1983; Reiss & Hottinger, 1984; Brach ert & Dullo, 1990, 1991, 1994; Dullo et a!. , 1990). Broecker et a!. (1968) were the first to use dated reef terraces to support the orbital theory. The
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
219
220
W-Ch. Dullo et al.
paper of Gvirtzman (1994) linked the position and topographical occurrence of emerged and of sub merged reef terraces in the Sinai to the global SPECMAP data set (Pisias et a!., 1984; Prell et a!., 1986) and absolute ages. The submarine terraces seem to fit this data set well; however, many age determinations and references to other known sites in the Red Sea are missing. The present paper deals with the reconstruction of the late Pleistocene and Holocene evolution of the fore-slopes of the island of Mayotte based on direct observations and sampling from a research submersible combined with U/Th dating of corals. The global sea-level history during the last 130 kyr (e.g. Bard et a. t , 1990) provides the opportunity to interpret the morphology of the fore-slopes and the sedimentary record with respect to Quaternary sea-level fluctuations. The morphological similari ties between the fore-slopes of Mayotte and other studied fore-slopes are discussed herein, to interpret
•
188
•
some of the observed features in relation to global sea-level changes.
G EOLOGICAL S ETTIN G
The Cornaro Islands make up an island chain located at the northern end of the Mozambique Channel between Madagascar and Africa (Fig. 1). The four main islands of this archipelago, Grande Comore, Moheli, Anjouan and Mayotte, are nearly aligned along a NW-SE axis. The isolated main volcanic complexes vary greatly in age and cover a time-span ranging from the Miocene and Pliocene (Esson et a!., 1970) to the most recent eruption in 1977 (Krafft, 1982). The variation in age, increas ing south-eastwards from Mayotte to Grande Co more, is also reflected by the varying coastal morphology, including the development of living coral reefs, the general weathering of the volcanic
l Grande Comoro
187
Moheli �Anjouan Mayotte .-
Iris Bank
� Canal de Mozambique
• 189
�
12° 40'S
190·
� 0
191•
192
45° 10'E 0 s
10
45° 20'E 20
30 km
---====---
Fig. 1. Locality map, showing the Comoro Islands as an inset in the general map of. Mayotte. All dive sites are indicated by an asterisk and a number.
Mayotte for e-slope mor phology and sediments rocks, and the general elevation of the islands above sea-level (Guilcher, 1965, 1971). The easternmost island, Mayotte (Fig. 1), repre sents the oldest part of the volcanic chain above present sea-level, if the volcanic rocks outcropping at the northern tip of Madagascar are neglected (Nougier et a!., 1986; Stieltjes, 1988). The present elevation is 660 m, and long-lasting erosion and weathering processes have removed much of the ancient volcanic core. The oldest volcanic rocks (subaerial flows of basalts and nephelinites) have been dated to 7.7 ± 1.0 Ma (K/Ar, Nougier et a!., 1986), and the youngest volcanic rocks, including phonolitic domes and flows, give an age of 1.49 ± 0.04 Ma. Extended coastal plains and mangrove forests are widespread around the island, whereas volcanic cliffs are missing, except on the eastern side of Pamandzi islet. Fringing reefs flourish all around the island. More spectacular, however, is the almost continuous circular barrier reef that is interrupted only by a few deep passages (60-80 m) in the east and incomplete parts in the west. Coral framework is not developed up to a barrier reef on the north western side; however, the reef rim grades into a relatively wide coral bank, such as the Iris bank (Guilcher, 1965). Furthermore, true lagoonal reefs are locally present, either isolated or forming a double barrier reef system as on the south-western side of the island (Guilcher, 1971). The enclosed lagoon is very large (more than 15-km wide) and is more than 50 m deep (Masse et a!., 1989; Tho massin et a!., 1989). Drowned karst features and rivers record the Pleistocene drop in sea-level, when this area was subaerially exposed. In addition to the mature reef development, thick laterite and even kaolinite profiles indicate the older age of the island geomorphologically. Esson eta!. ( 1970) assumed that Mayotte subsided like all other oceanic islands. The only morphologi cal indication is the lack of any raised coral reefs or reef terraces. However, we have some idea from drill-cores through the barrier reef south of Pam andzi. There, the Holocene-Pleistocene boundary occurs around 20 m deep in the core below the sed iment surface (Camoin et a! ., 1997) which corre sponds to modern sea-level at low tide. Dating of the top of the Pleistocene failed because of intense re crystallization of the coral material. If we assume that this Pleistocene top is coeval with isotope stage 7 then we have to envisage a subsidence rate of 12 em kyc 1• If we assume stage Sa, subsidence is
221
25 em kyr-1• This means that for the last 20 000 yr (last glacial maximum), the island subsided between 2.4 m and around 5 m (Camoin eta!., 1997).
B ASIC CLI M A T E AN D HY D RO G R APHY
The hydrographic regime around the Comoros is controlled by the southern equatorial current and the Mozambique current, which themselves are gov erned by the twice-yearly change of the monsoon winds (Saertre & Da Silva, 1984; Ehny, 1987). How ever, all these ocean-wide changes do not affect the counter-clockwise direction of the local currents around the Comoros throughout the year, because they are mainly driven by the southern equatorial current. Sea surface temperature does not decrease below 24oC and its annual average is around 28oC (Pitton et a!., 1981). Precipitation is monsoon con trolled, with high values between November and May, reaching an average maximum of around 400 mm month-1• Peaks of almost 800 mm oc curred once during a decade (airport data, Pam andzi). The prevailing swell is from the SE and on the southern and windward margins of the island. This long-period swell was still felt in the submersible down to 90-100 m depth!
M A T E RIALS AN D M E THODS
A total of 19 dives were performed around Mayotte (Fig. 1) using the two-man submersible JAGO. JAGO has a diving limit of c. 400 m and was operated from the surface vessel DEEP SAL VAGE I. The bathymetry of diving sites was first checked by simple echo sounding on an analogue record<;r. Contact between the surface vessel and the sub mersible was maintained throughout each dive by a powerful underwater radio system. JAGO has a large front window with a diameter of 60 em, five spotlights and two powerful flashes for underwater photography, and is equipped with a manipulator and a sample box. A simple but heavy duty chisel was mounted to the keel for collecting hard rock samples. By ramming the submersible at full speed, we were able to recover samples that weighed up to 5 kg, even from the well-cemented slopes and the drowned reefs. Therefore we were able to obtain coral material and sedimentological information from the older reef rocks without using
222
W-Ch. Dullo et al.
explosives. The corals we sampled fall into two groups: the first represents in situ specimens, which show no signs of transport and fracturing and still display a framework fabric, whereas the second comprises coral debris, exhibiting fractures. Dated specimens from the latter group cannot provide any hint for ancient sea-level; however, these dates provide an idea of the timing of reef growth and talus sedimentation. The inclination of the slope was measured using a simple clinometer. The depth was continously re corded with a fathometer. Sedimentological and biological information was mapped on these simple profiles. In addition, we recorded simultaneously the water temperature and the light penetration in the shallower parts. Furthermore, we documented each change in slope geometry and facies on video tapes and slides. Collected in situ corals were dated by U/Th-alpha counting (Barnes et a!., 1956) at the Departement de Geochimie du BRGM in Orleans (Table 1) and by thermo-ionization mass spectrography (TIMS, Edwards eta!., 1987) at the Department of Umwelt physik, Heidelberger Akademie der Wissenschaften (Table 2) after petrographic examination and X-ray analyses to check the skeletal aragonite/calcite ratio. The procedures for chemical separation and purifi cation of U and Th, as well as dating procedures for U/Th TIMS are similar to those described by Edwards et a!. ( 1987). Coral skeletons are not diagenetically altered, but intraskeletal porosity is partially filled either by ara gonitic and calcitic cements or by micritic sediment. Thus, to check the validity of the closed-system no tion, it was necessary to compare Uranium and Thorium concentrations and ages on each indepen dent mineralogical phase and whole coral fragments. The separation of mineralogical phases was achieved by using a drill to remove the main fraction of void filling cement, and by exposing broken coral frag ments in a pH-neutral hydrogen peroxide solution. The mineralogical phases were then separated through granulometric filters and recovered under a binocular microscope. Micritic sediments are characterized by lower concentrations in Uranium ([Ulmean 1.9 p.p.m.) than in coral skeletons ([Ulmean 2.8 p.p.m.) with an initial isotopic ratio 234 U F38 U (8234 U(T)) lower than the sea-water value. Aragonitic fibrous ce ments display higher concentrations of Uranium ([Ulmean 3.7 p.p.m.) and higher 8234 U(T) than modern sea-water. These results show that average concentrations of Uranium and Thorium observed =
=
=
in whole coral fragments are generated by an appar ent dilution or enrichment processes. U/Th ages obtained on each mineralogical phase are very close, the differences being lower than the calcu lated age errors; thus, analysis of whole coral frag ments may give reliable ages. Thorium content of coral fragments ranges between 2.45. x 10-2 and 26.9 x 10-2 p.p.m. (average 13.7 x 1 0-2 p.p.m. on 39 samples). 232Th activity ranges between 5.9 x 1 0-3 and 166 x 1 0-3 d. p.m. g-1. These data, coupled with some values of 230ThF32Th ratio not greatly exceeding 20 (the standard cut-off oflvanov ich & Harmon ( 1992)), suggest the entrapment of detrital components. In most analysed samples, initial 234 U F38 U ratios match those recorded in modern sea water (1.15 ± 0.3 or 1.14 ± 0.014 according to Thurber (1962) and Koide & Goldberg ( 1965), respectively); few samples exhibit lower values of 234 U/238 U, probably related to a slight loss in 234 U. This confirms that the Mayotte corals grew in open sea water and did not experience significant alteration after deposition. The average analytical uncertainty ( lo) of U/Th-alpha measurements is 950 yr (1000 yr for samples ranging in age from 35 to 1 0 ka and 200 yr for Holocene samples).
R ESULTS
The fore-slopes of Mayotte are characterized by a general morphology (Fig. 2) that displays striking similarities to other coral reef fore-slopes in the world. We can distinguish, as in the Bahamas, three major morphological and sedimentary units which form a deeper sediment slope, a shallower cemented slope, and a wall or steep cliff (Grammer & Ginsburg, 1992). In addition, we have observed drowned reefs as distinct feature on the fore-slopes in 90 m and 60 m water depth, respectively. All 19 investigated sites of the fore-slopes (Fig. 1) exhibit this overall morphology, which is displayed by five representa tive profiles in Fig. 2. Sediment slope
The deepest part of the sediment slope below 300-m depth is characterized by silty to muddy carbonates which may exhibit ripples on their surface (Fig. 3a) oriented normally to the gently inclined slope (20°35 ) This pattern is related to a contour current around the island which we experienced in the submersible by drifting. When ripples are lacking, intense bioturbation is indicated by open burrows. °
.
Table 1. U-Th geochemistry and ages based on
a
counting Calcite .;;5 %
Depth (m)
Aragonite
Reef of isotope stage 3 J-69 Unidentified J-93 Unidentified
-1 1 0 -105
I I
I I
263.697 265.927
1 7.407 22.678
296.272 295.681
Cemented slope J-65 Unidentified J-60-11 Unidentified Goniopora J-54 J-84 Unidentified
- 1 75 - 1 80 - 1 65 -245
I I I I
I I I
377.473 235.808 392.674 1 139.682
24.609 40.818 2.737 1 03.849
Sample
Taxon
Cement
21su1232Th
230Th/2 32Th
lcr
Age (kyr) lcr
1 9.461 25. 1 29
79. 1 89 1 1 9.227
0.5242 10.077
33.6 ± 1 . 0 5 5 . 6 ± 2. 1
402.903 278.706 430.0 1 2 1 304.682
28.382 4.821 29.884 118.8 1 2
9.070 7.529 1 2 5.244 38 1 .929
0.6444 1 3.791 0.866 34.9 1 5
27.6 34.0 37.3 37.4
1 03.5 86 22.514 1 4.288
1 335.398 375.527 279.506
1 2 1 .372 25.442 1 6. 6 5 7
21.269 85.794 74.656
1 9.302 0.5681 0.4356
1 8.8 ± 0.4 28.0 ± 1 .0 33.5 ± 1.0 1 8.4 ± 0.5
lcr
234U/232Th
lcr
Cora/gal fabric on the cemented slope J-91 Unidentified - 1 80 I J-6 1 -111 Unidentified - 1 60 I J-67 Acropora -135 I
I I
1 136.898 330.275 237.405
In situ shallow-water cora/gal veneer J-96-1 -152 Acropora I
I
242. 1 1 6
1 1 .052
280. 1 92
1 2. 5 5 8
43.707
0.1786
I
1650.559 650.74 1 1 04.516 1 05 1 .483
173.249 52.956 123.573 263.645
1 827.903 755 .046 1 1 92.897 1 20 1 .0 1 8
191.452 61. 1 03 140.794 300.996
278.859 1 1 5. 786 1 95 . 6 5 7 1 93.617
28.737 0.9224 23.243 49.481
Reworked corals of the LGM cemented to the wall or slope J-95-11 -205 Acropora sp. I J-98A Porites -285 I J-95-11 Unidentified -205 I J-58-11 Unidentified -200 I
18.0 1 8. 1 19.4 19.1
± ± ± ±
± ± ± ±
0.9 2.4 1 .1 0.8
0.5 0.6 0.6 1 .1
�
�
�
�
� ..... ')'> c..,
� I'll
� Cl 'ti
;::... Cl
� $::l �
$::)..
c..,
� �
I'll �
Drowned reef J-57 Leptoseris J-57-III Cyphastrea J-83 Porites
-90 -90 - 1 12
I I I
Recent talus J-88 Acropora J-3 Porites J-81 Porites
- 1 00 -200 -230
I I I
I
968.623 2 1 9 1 . 1 26 23.879
155.473 243.491 22.75 1
1 1 18.012 2 5 0 1 .648 274.088
179.362 277.483 26.039
29.976 222.517 32.351
0.5492 24.655 0.3438
2.9 ± 0.3 10.1 ± 0.3 1 3.6 ± 0.7
961.291 631.744 1478.525
47.292 43.065 240.654
1 108.799 722.576 17 1 2.904
8 . 545 48.947 278.546
1 5.472 1 4.676 39.679
0. 1 48 0.1192 0.7022
1 .5 ± 0. 1 2.2 ± 0. 1 2.5 ± 0.2
1:;'
Errors are given in I cr values.
N N w
N N .j>.
Table 2. U-Th geochemistry and ages based on TIMS
activity ratio (%o) decay corrected
23oTh/z3su activity ratio
Ages (ka) TIMS
Ages (ka) TIMS (corr) 2cr
234u;zJsu
232 Th (p.p.b.)
230 Th (p.p.t.)
23Bu (p.p.m.)
234u;z3su activity ratio (%o)
1 4 . 18 (23)
8. 73 (23)
3 . 3 7 {I)
144 ± 5
151 ± 5
0. 1 586 (4 1 )
16.2 (5)
1 6. 1 (5)
Reworked corals of the LGM cemented to the wall or slope 5 .46 (9) J95I 8.69 ( 1 9) Unidentified -205 9 . 70 { I I ) -220 62.89 (25) Porites J36
3 . 3 79 (5) 3 .04 { I )
1 43 ± 4 1 47 ± 6
1 50 ± 4 1 56 ± 7
0. 1 577 (36) 0. 1 95 9 (22)
1 6 . 1 (4) 20.3 (4)
1 6.0 (4) 1 9.7 (4)
Drowned reef J99/1 Porites
- 1 05
1 5.02 { I I)
6.8 ( 1 2)
3.35 {I)
1 39 ± 4
1 45 ± 4
0. 1 350 (3 1)
1 3.8 (4)
1 3.6 (4)
Recent talus J88 Acropora
- 1 00
1 3.30 ( 1 3)
0.93 1 (49)
3 .484 (9)
153 ± 6
153 ± 6
0.0 1 639 (86)
1 .5 6 (8)
1.46 (8)
Sample
Taxon
Depth (m)
In situ shallow-water cora/gal veneer J6 1 I Acropora - 1 60
�
� tl
:;:::
� (I)
�
Errors are given in 2cr values in parentheses referring to the last digits. Our 232Th concentrations are much higher than those found in corals from oceanic islands, which typically have less than 0.5 p.p.b. (Edwards et a/. , 1 987; Chen et a/. , 1 99 1 ). We therefore attribute the high 232Th content to the possible entrapment of detrital or marine materials (clays) from the adjacent areas. To account for this entrapment, our values are corrected for inherited U and Th assuming a U/Th ratio of 3.8. As can be seen in the last column, the corrections to the original ages are minor to negligible.
�
225
Mayotte for e-slope mor phology and sediments ·50-
platform edge
JAGO Dive 187, Mayotte North of Grand Recif du Nord
drowned reel
·100 sand slope 2.'i"
JAGO Dive 190, Mayotte Barrier Reef West of Pte. Mohila
50
100
·150
·200
11lQ
caves, cliffs, ledges,
-150
wall
..
vertk;.r���-� 9uuie� "
·200
1l
�.'i"
cemented slope
JAGO Dive 201, Mayotte -so-
Grand Recif du Nord-Est
platform edge
-250 . 100
1
.
'"
300 (
talus and large cipit boulders
. 150 -350 �u·
cemented slope . 200
begin of
. 250
sediment slope . 250
. so ·150
.!..
vertical chutes and gullies '
cemented slope
JAGO Dive 203, Mayotte Recif Bandele
platform edge
J.'i" caves, cli�s ledges,
drowned reel
. 100 wall
·200 - 150 talus and large cipit boulders
. 250
cemented slope - 200
erosional clifl with internal bedding
.1.�·
U
.'iU
IIMl
I�U
2UJ
L... L...LLI�L...J . __J_ � __L_l_.l
C .w - 250
latus and cipit boulders
sediment stope
Fig. 2. The morphology of the dive sites exhibits an overall pattern which consists of a sediment slope, a cemented slope and a steep wall. Note the slight dominance of cipit boulders on the windward margins. Drowned reefs occur on top of the wall and in some places even shallower, as at dive site 1 87. The five selected profiles are representative of all sites investigated (site 1 87 is representative of sites 1 88 and 1 89 plus a tectonic escarpment as at site 1 93 below 250 m; site 1 90 is similar to site 1 91; site 1 93 represents sites 1 92, 194 and 1 97; site 20I represents sites 1 86, 206, 1 98 and 207; site 203 is representative of sites 202, 204 and 205). They all are drawn to the same scale with no vertical exaggeration.
Upslope, grain size changes abruptly from silt to sand and even gravel around 250-m depth. The sedi ment consists of a mixture of recent unconsolidated material and reworked carbonates (Fig. 3b) charac terized by corroded surfaces and distinctive colours (i.e. white versus grey, respectively). The recent particles are mostly derived from the present-day shallow-water environment and include bioclasts of scleractinians, molluscs and echinoderms. The pre dominant portion of Halimeda plates among these skeletal grains, however, is derived from the living crops on top of the terraces between 70 and 100 m deep. Fossil material that is grey represents eroded
carbonates and reworked corals from the cemented slope and wall above (Fig. 3b). Accumulations of huge blocks occur on top of the sediment slope, up to several cubic metres in size (Fig. 3c). We have seen them in nine dives concen trated along the eastern and south-western margin of the barrier reef (Figs 1 & 2), which is the site most hit by cyclones. These blocks do not display any distinctive arrangement according to depth. They provide a hard substrate for various benthic organisms on a sediment slope where unlithified material prevails. At two sites, 192 and 193 (Fig. 1 ), tectonic es-
·
226
W-Ch. Dullo et a!.
Mayotte for e-slope mor phology and sediments carpments (Fig. 2, profile 193) provide exposures of the underlying cemented limestones. These natural outcrops exhibit well-bedded carbonates gently in clined downslope, which support horizontal ledges (Fig. 3d). The well-bedded carbonates are predom inantly composed of skeletal grainstones with mi nor packstones. The prime skeletal grains consist of two types of Halimeda, which can be differentiated according to the size of their cortical tubes (Fig. 3e). Other skeletal grains include fragments of coralline algae, and shallow- and deeper-water benthic fora minifers dominated by amphisteginids; however, bryozoans, gastropods and echinoids are rare. Cemented slope and wall
Upslope around 200 m water depth, the sediment slope grades into the cemented slope (Fig. 3t) com posed of well-cemented grainstones with a shallow water derived biota; reworked shallow-water corals include Acr opora, Por ites and Goniopora. The incli nation of the cemented slope is predominantly around 60', with a minimum inclination of 40'. The surface of the cemented slope corresponds to a submarine hardground, as indicated by the occur rence of multiple generations of borings. The rock surface becomes increasingly flaky upslope until the occurrence of the typical ledge rocks (Fig. 4a) on the steep wall generally between 190 and 90-m water depth. The ledges have a dense cover of living benthic organisms among which sponges prevail macroscopically. Unlithified sediment may accu mulate on top of the ledges or in small depressions,
3. (Opposite) (a) Contour ripples seen from the submersible; width of field of view is 1 m. Dive 188, 290-m water depth. (b) Sediment slope, exhibiting present skeletal grains represented by bioclasts of scleractinians, molluscs and echinoids as well as ancient clasts exhibiting a corroded surface (differentiated by colour) seen from the submersible; width of field of view is 1 m. Dive 1 90, 250-m water depth. (c) Cipit boulders seen from the submersible; the block is about 1 .5 m across. Dive 197, 3 1 0-m water depth. (d) Ledges on the deeper cliff from the older cemented slope system seen from the submersible; note the linear arrangement indicating bedding geometries. Dive 1 93, 260-m water depth. (e) Ha/imeda-packstone. Sample J76, dive 193, 250-m water depth, thin-section. (f) Cemented slope seen from the submersible. The cemented beds are steeply inclined (around 40') showing less than 5 em of loose sedimentary veneer. Dive 1 97, 240-m water depth; the inclination is predominantly around 60' with a minimum inclination of 40'.
Fig.
227
suggesting that most sediment is usually bypassing these steep slopes. A sharp increase in inclination occurs between 190 and 160 m depth, where the cemented slope steepens and forms an almost vertical wall (75 90') as a prominent part of the cemented slope. This wall is a typical feature of most of the investi gated slope sites of the island (Fig. 2). The surface of the wall is covered by irregularly arranged ledges, which may protrude up to half a metre from the wall. Along the SW side of the island, the cemented slope may continue up to 120-m water depth (Fig. 2, profile 193). Here, the local absence of ledges is probably related to a strong bypass of sediments, which suppresses their development be cause of episodic erosion. There are two karst systems within the bathy metric range of the cemented slope and the wall (Fig. 2). Their occurrence is limited to distinct levels. The first level is located between 150- and 155-m water depth and consists of small solution caves smaller than 3 m (Fig. 4b). These caves may con tinue up to 2 m horizontally into the wall or the cemented slope. Furthermore, the surfac� of the wall exhibits small-scale solution features, such as karren and kamenitza morphologies. However, no petro graphic evidence of subaerial exposure was ob served. A thin coralgal veneer has started to grow over the karst features in a few sites. In situ shallow water scleractinians (Acr opora, probably A. danai, and Por ites) derived from the initial framework of this coralgal facies at 152-m water depth were dated at 18.4 ± 0.5 ka (Table 1), and at -160 m at 16.1 ± 0.5 ka by TIMS (Table 2). The corals were partly encrusted by the Foraminifera Acer vulina in haer ens and by the coralline algae Lithophyllum sp. and Por olithon onkodes. Because of the very limited space for coralgal growth during this time of low stand of sea-level, most of the biota broke off and were transported downslope. Therefore, reworked corals and associated skeletal material derived from this ancient shallow-water reef environment were trapped below 180-m present-day water depth on top of the ledge rocks or on the cemented slope. These corals gave ages ranging from 19.1 ± 1.1 to 18.0 ± 0.5 ka. (Table 1), and from 19.7 ± 0.4 to 16.0 ± 0.4 ka by TIMS (Table 2). The second karst horizon occurs between 120and 125-m water depth, where we found caves more than 3 m deep and wide. Furthermore, karst chan nels and solution pipes may connect different caves, which we could prove by chasing fish from one to ·-
228
W-Ch. Dullo et a!.
Mayotte for e-slope mor phology and sediments another. The surface of the carbonates exhibits karren and kamenitza features as well. Two major lithofacies types were sampled from the wall. One is composed of Halimeda grainstones and packstones or skeletal grainstones and rud stones. They represent a reef talus facies forming the rocks of the wall or of the cemented slope. Therefore, corals occur only as fragments. Ages obtained from these cemented reef talus sediments range between 37.4 ± 0.8 and 27.6 ± 0.9 ka (Table 1). Although we did not see unequivocal internal bedding during all our dives, we could observe a clear indication of an inclined internal bedding in dive 197 (Fig. 3f). The other lithofacies comprises a thin veneer of a coralgal fabric which grows on the surface of the cemented rocks of the steep slope and the wall. This lithofacies type is made up of a scleractinian frame work forming the nucleus of the ledge rocks. Por ites and Acr opora are common; however, additional platy-growing species of Pavona were also recovered (Fig. 4c). This primary biogenic fabric, still in situ, acted as a trap for skeletal sands produced in the nearby shallow-water reef environment. These cor als, together with encrusting organisms represented by red algae (corallinaceans and peyssonneliaceans), foraminifers and especially vermetid gastropods, bathymetrically indicate a reef environment less than 30 m deep and probably within the upper 10 m. Ages obtained on these in situ corals sampled from the coralgal veneer of the cemented slope
Fig. 4. (Opposite) (a) Ledge rock seen from submersible. Note the tiny threads at the margin of the ledge, which may contribute to the stabilization of the sediment. Width of field of view is 0.6 m. Dive 20 1 , 1 80-m water depth. (b) Karst features seen from submersible at 1 5 0-m water depth. Note the solution caves and karren features. Some parts of the corroded surfaces are covered with a thin sedimentary veneer, less than I em. Width of field of view is 0.9 m, dive 1 98. (c) Ledge rock showing Pavona sp. ( 1 ) as a major constituent encrusted by coralline algae (2). Note the open sponge boring in the upper right-hand corner (3). Sample 1 1 0 1 , dive 203, 1 7 5 m water depth, thin-section. (d) Drowned reefs seen from the submersible. Note the typical plate-like growth form of shallow-water Acropora colonies, now completely overgrown by coralline algae. The fish are c. 1 5 em long for scale. Dive 1 97, 90-m water depth. (e) Porites ( 1 ) with encrusting algae (2). Sample J 1 02, dive 203, 95-m water depth, thin-section. (f) Leptoseris fragilis ( 1 ) framework composed of this special scleractinian and several crusts of Acervulina inhaerens (2) coral line algae (3), and peyssonneliaceans (4). Sample J71A, dive 292, 90-m water depth, thin-section.
229
and the wall range from 33.5 ± l .O to 18.8±0.4 ka (Table 1). From two sites (19 1 and 199) we were able to ram off samples from the upper part of the wall at around ll0-m and l05-m present water depth. These sam ples can be characterized as framestones represent ing a real reef facies. The state of preservation of the corals within these rocks is good to moderate. There fore, we obtained only two ages of 55.6 ± 2.1 and 33.6 ± l.l ka (Table 1), which provide only a hint for the time of reef growth. As these reef rocks form the uppermost part of the wall, we assume that they correspond to the talus facies described above, which occurs bathymetrically deeper and forms the lower part of the wall and the cemented slope. Another group of samples, which is not yet ce mented, consists of loose coral debris deriving from the Holocene reef. As most of the reefs produce more carbonate than they can accommodate within their shallow-water environment, carbonate material is transported off the reef downslope and occurs as debris at various depths. Hence, the ages obtained from these samples are young, indicating the still continuing formation of the Holocene talus (Tables l & 2). Drowned reefs
On top of the wall, drowned reefs are found concen trated bathymetrically between l00 m and 90 m. In contrast to the coralgal veneers recovered from the cemented slope and the wall, they form small mounds, elevated by up to 3 m in comparison with the surrounding sediment. They are at present covered by living zooxanthellate scleractinians (Leptoser is fragilis and L. explanata) especially adapted to this twilight zone (Fricke et a/., 1987), Halimeda, sponges and calcareous red algae, as well as gorgonians and antipatharians. Some convincing examples exist where the ancient growth morphol ogy of the constituent shallow-water scleractinians is still seen below the intense encrustation of mainly coralline algae (Fig. 4d). These drowned reef mounds are fragile and it is relatively easy to ram off huge parts with the chisel mounted to the keel of the submersible. Therefore, we obtained samples from the internal parts. The lower part, up to 2 m in height, is composed of shallow-water corals (e.g. Por ites, Pocillopora) with associated encrusting organisms (calcareous red al gae (Fig. 4e), vermetid gastropods and foramini fers). Skeletal sand is trapped within this fabric. It is
230
W-Ch. Dullo et a!.
composed of coated grains and foraminifers among which miliolids are common, indicating a shallow water origin. This coral community is gradationally replaced upward by deeper-water platy growth forms at between 2 and 3 m of mound height. The uppermost part of the mounds, 20 em thick, com prises a lithified L eptoser is framework (Fig. 4f), representing the present-day calcareous commu nity. One in situ shallow-water coral (Por ites) was dated at 13.6 ± 0.4 ka by TlMS (Table 2). Two other corals were dated at 10.1 ± 0.2 (Cyphastr ea) and at 2.9 ± 0.3 ka (Leptoser is) (Table 1). They probably represent the transition from shallower to deeper environments, as the latter is still living at this depth. In a few dives (189, 192, 193 and 204, Fig. 1; 193, Fig. 2), we recognized a second level of drowned reefs between 65 and 55 m, already covered by a veneer of living platy scleractinians. The base of these coralgal associations is composed of shallow water corals such as branching Pocillopora and Acr o pora, although these were recorded as small pieces. Bathymetrically shallower terrace steps have not been included in this survey because of severe swell conditions.
parable ages (Table 1). According to published sea level curves for the last 130 kyr (e.g. Bard et a!., 1990), there is a prominent sea-level lowstand oscillating around 80 m deep during late isotope. stage 3 (Fig. 5). During that time (Table 1 ), shallow water reefs developed together with their associated talus facies on the present-day deeper fore-reef and may have contributed to the overall morphology of this prominent terrace (Fig. 2). This implies that the: cemented slope and parts of the wall record pro-· cesses during sea-level lowstand, a condition which was responsible fo� the equivalent formation on the: fore-slopes of the Bahamas (Grammer & Ginsburg, 1992). The overall morphology of the wall, mainly re-· lated to intense erosion with the formation of reef outrunners (Land & Moore, 1977; James & Gins-· burg, 1979; Grammvr, 199 1), even known in the: fossil record as cipjt p�:mlders (Bosellini, 1989;; Biddle et a!., 1992), and subsequent karstification,, was created during a rapid lowering of sea-level starting around 26 ka (Fig. 5) and approaching the: last glacial maximum (LGM). As sea-level fall stopped around 22 ka, the karstification processes bathymetrically moved down to 150-m present
DISCUSSION
The oldest structure seen during the dives is repre sented by the submarine outcrops of the volcanic basement at site 197, 320 m deep. On top of these volcanic rocks, a series of well-bedded, Halimeda dominated packstones and grainstones (James & Ginsburg, 1979; Ginsburg eta!., 1991) were depos ited and are exposed in two tectonically formed cliffs at sites 193 and 198, between 250 and 300 m deep. Comparable inclined packstones have been reported from Bahamian fore-slopes by Grammer eta! . (1993). As we did not obtain any absolute ages from these rocks, we may only assume a late Pliocene to early Pleistocene age; this is the most likely age, consistent with the subsidence history of the island and the age of the volcanic rocks (Nougier et a!., 1986). The oldest samples we could date to be derived from the uppermost part of the wall have ages of 55.6 ± 2. 1 and 33.6 ± 1.1 ka, respectively. These carbonates represent framestones that belong to a reef facies. Associated with this reef facies is a reef talus forming the deeper part of the wall and the cemented slope. Samples from this part show com-
20 0 -20
Qi > Q)
� Q)
en
-4 0 -60 -80 -100 -120 -140 0
20
40
60
Age (ky BP) 5. Sea-level curve based on normalized 8180 curve determined by Shackleton ( 1987) (continuous line) and determined by Labeyrie et a!. ( 1987) (dotted line) after Bard et a!. ( 1990). The isotope stages are indicated by numbers.
Fig.
231
Mayotte for e-slope mor phology and sediments water depth, creating small caves, karren and ka menitza features. After most parts of the karst had been formed and sea-level may have started to rise slowly at the end of the LGM, scleractinians started to grow on the cemented slope. The overall slope inclination around 150-m water depth is very steep except between sites 192 and 197 (Figs 1 & 2). There is almost no space to accrete a real rock onto these steeply inclined walls. Therefore, the coral veneer growing in this ancient shallow-water envi ronment during the end of the LGM produced more bioclastic debris than was recorded as in situ fabric. Accordingly, we collected reworked corals as frag ments already cemented to the deeper parts of the fore-slopes exhibiting an age corresponding to the LGM (Table 1) (Colonna et al., 1996). This is in good agreement with the dates of 17 ± 1 ka (by U/Th a counting) obtained by Veeh & Veevers ( 1970) on corals (Galaxea clavus) 17 5 m deep in the Middle Great Barrier Reef. Similar dates, 17.595 ± 0.07 ka and 15.58 ± 0.05 ka (by U/Th TlMS) have been published by Bard etal. (1992) from the flanks of Mururoa, and Fairbanks (1989) has presented an age of 18 200 ya (by 14 C, which corresponds to 2 1.93 ± 0.15 ka by U/Th) for Porites aster oides 124 m deep from Barbados.
Sea-level rise recorded during the last deglacia tion, averaging around 8 mm yr-1 (Fairbanks, 1989), interrupted sediment transport off the wall and created new space for accommodation as soon as the wall was flooded. During this interruption, hard grounds and laminar micritic crusts were formed and lithified. The results of Brachert & Dullo (1991) demonstrate that this type of hardground and lami nar micrite was formed under conditions of rapidly rising sea-level. We also assume similar conditions and timing for the formation of laminated micrites occurring on Belize fore-slopes (James & Ginsburg, 1979). During this rapid sea-level rise, the surface of the starved cemented slope would correspond to a diastem, thus recording a typical drowning uncon formity as reported by Grammer & Ginsburg (1992) and Grammer et al. (1993). This is verified by tex tural evidence (i.e. boring) indicating that the surface of the slope is a submarine hardground. The build-ups located on top of the deeper forereef terrace at 90 m-100-m water depth represent a cor algal succession, frequently beginning with shallow water stony corals and ending in an encrustation by a L eptoser is framework. This indicates a classical drowning event. According to our present knowl edge, the process of deglaciation is marked by pulses
0 .------. •
•
U/Th age
Post Younger Dryas Melting Pulse
�
MWP 18
Older Dryas
. -so
=
Younger Dryas Aller¢d
� -60 .
GIVE UP
.
Belling Melting Pulse =
MWP 1A
-100 -120 -140 +---�----.---�--� 25 20 15 10 5 0 AGE IN kyr Fig. 6. Sea-level during the last deglaciation based on dated Acropora palma/a from Barbados. The U-Th age errors
are indicated at the 2cr level. Note the two pulses of rising sea-level, which may have caused the observed drowning events of the reef mounds at I 00-90 m and 65-55 m on Mayotte fore-slopes. Modified after Bard et al. ( 1 9 90).
232
W-Ch. Dullo et a!. -50
3
11.5 - 11 Melt-Water Pulse 1 B
t
. 100
kyr - Present
t
. so
-150
-100
-200
. 150
. 250
FOR MATION OF SEDIMENT SLOPE
. so
6
14 kyr Melt-Water Pulse 1A -100
t
-150
(I) > (I) I
10 (I) Cl)
..... ....
c (I) Cl) (I) a.
�
0 (I) .D ..c
.....
a. (I) 0
9
kyr
Beginning of Holocene Barrier R"f
-200 . 100
-250 ·150 . 50 -200
ONSET OF HOLOCENE BARRIER REEF GROWTH
. 100
. 150
�
LGM
22- 18 kyr
. 200
. 250
50-25 kyr Late Isotope Stage
3
33.8 kyr J 67: 33.5 kyr -> J SSt:
->
. 150
. 200
. 250
7. Summarizing sketch of the recorded sea-level changes on the fore-slopes of Mayotte. ( I ) Reefs grew during isotope stage 3. The dated samples indicate a reef growth between 55 and 2 7 ka for a sea-level around -80 to -90 m. The shallower part might have formed during the early part of stage 3. (2) Maximum drop in sea-level occurred between 22 and 18 ka. The emerged parts were karstified and the steep wall was formed because of erosion, which led to the accretion of coarse material and cipit boulders. (3) Rise of sea-level and formation of ledges on the wall. The new creation of space for accommodation of sediment reduced sedimentation in the 'deep'. This favoured the formation of a hiatus and the coeval cementation of the talus (cemented slope). Reefs started to grow on top of the wall, and were drowned as a result of the B0lling meltwater pulse (meltwater pulse l A). (4) Continuous rise in sea-level. Reefs started to grow on the shallower cliff around 60 m below the present-day sea-level, and were drowned as a result of the Post Younger Dryas meltwater pulse (meltwater pulse I B). (5) Onset of Holocene reef growth on top of the karstified Pleistocene limestones. (6) From 3 ka, sea-level has remained in the present position. As a consequence, new sediment has started to accumulate downslope, forming the sedimentary slope. Selected locations of samples are shown by sample number and age obtained.
Fig.
-100
�<- J 65: 27.15 kyr JSOII.3 .5kyr-> ?"A
#
7 ,
_.r_..-:<- J 84: 37.4 kyr
%?"//
5
. 250
FORMATION OF REEFS AND TALUS
1
Mayotte for e-slope mor phology and sediments of meltwater input. They are recorded both in deep sea sediments (Olausson, 1965; Falls & Williams, 1 980; Berger, 1990) and in drilled coral reefs (Fair banks, 1989; Bard et a/., 1990). The drowned reefs are bathymetrically located at around 90-100-m depth in the Comoro Islands and are present at the same depth in the Red Sea (Dullo et a/., 1990). In Barbados, a prominent fossil reef ridge occurs at 80 m water depth (Macintyre eta/., 1991), whereas it is developed at 60 m in the Bahamas (Grammer, 199 1 ), at 65 m in Jamaica (Goreau & Land, 1974) and around 60-70 m in Tobacco Cay, Belize (James & Ginsburg, 1979). This situation is very similar to that of the Great Barrier Reef, where a distinctive terrace level occurs at a present-day depth of 60 m (Harris & Davies, 1 989). The drowning of reefs in Barbados at 80-m water depth and of those on top of the wall in the Bahamas at 60 m has been attributed to the Post Younger Dryas meltwater pulse (Macin tyre eta/., 1991; Grammer & Ginsburg, 1992), also known as termination 1B in the deep-sea record at around 1 1.5 ka (Sarnthein & Tiedemann, 1992). As there are also drowned reef structures in the Comoro Islands occurring at 55-60-m water depth, we as sume the existence of an older meltwater pulse, which is also seen in the Barbados curve (Fig. 6) of Bard et a/. ( 1 990). In accordance with the deep-sea record, we would like to use the term BeJ!ling pulse, which occurred at around 14 ka (Berger, 1990) and which was responsible for the drowning of the reefs (Neumann & Macintyre, 1985) which are at present at 90- 100-m water depth in the Comoros. The onset of the Holocene reef growth occurred at 9 ka, as deduced from dates obtained on drilled reefs of Pamandzi island (Camoin et a!., in 1997). As soon as the rate of sea-level rise decreased at around 6 ka (Davies & Marshall,1980; Ruddiman & Duplessy, 1985; Eisenhauer et a!., 1993), poten tial space for the accumulation of reef-derived sediments became more and more limited. Modern reefs produce much more carbonate than they can store in their related shallow-water environments. Therefore, carbonate sediments have to be exported into deep-water environments. The interrupted sed imentation during a rapid rise in sea-level started again at that time. Since then, the unlithified sedi mentary wedge (fine-grained skeletal silty to muddy carbonates) has started to accumulate at the base of the cemented slope, as demonstrated by dates ob tained on 1000-3000-year old corals occurring as loose bioclasts on the cemented slope and on the sediment slope (Table 1).
233
CONC LUSIONS
The absolute ages we have obtained from corals may not be sufficient to base the reconstruction of sea-level history entirely upon them; however, to gether with the observed slope morphologies, they may provide valuable hints for a scenario. Figure 7 summarizes the various episodes of changing sea level, which may be responsible for the present fore-slope morphology. 1 Shallow-water reefs grew between 80- and 90-m present water depth during late isotope stage 3 (50-26 ka) forming the uppermost part of the wall and probably the prominent terrace at that depth. Related talus sediments occur deeper within the lower part of the wall or of the cemented slope. 2 Maximum lowering of sea-level because of LGM conditions led to karstification and dissolution fea tures down to 150-m water depth. A thin shallow water coral veneer started to grow close to the end of the LGM, and is recorded as in situ fabrics as well as bioclasts within coeval talus sediments. 3 Rapidly rising sea-level after the onset of deglaci ation soon flooded the top of the wall and created new space for sediment accommodation. Therefore, sediment transport off the wall was interrupted, leading to the formation of hardgrounds on the ledges because of sediment starvation. 4 Coevally reefs started to grow on top of the wall at 100- and 90-m water depth, forming small-scale mounds 3 m high. These are covered now by a deep-water coral community. These reefs were pre sumably drowned as a result of the B0lling meltwa ter pulse (at 14 ka). Shallower reef mounds at present at 65-55 m may have drowned because of the meltwater pulse of the Younger Dryas at l l .S ka.
AC KNO W L E D G E M ENTS
The authors W.C.D., D.B. and B.T. thank the pilots of the submersible JAGO-Ji.irgen Schauer and Robert Ki.irsteiner-for their enthusiastic help, as well as Karen Hismann and Captain Peter Willmot and his crew for their support on board DEEP SAL VAGE I. We greatly appreciate the help we have received from the French authorities in Mayotte, namely from M. le Prefet de Mayotte, the harbour master Charly Tango and Jean-Michel Maggiorani from the Service des Peches. Substantial scientific input was provided by R.N. Ginsburg and by our
W. -Ch. Dullo et a!.
234
friends Thomas Brachert and Michael Grammer. Valuable remarks by two anonymous reviewers have considerably improved the manuscript. Finan cial support was provided within the German pri ority research programme 'Evolution of Reefs' (Du 129/6) and by the French research programme 'Programme National Recif coralliens', and this is gratefully acknowledged.
R E F E RENCES
ADEY, W.H., MACINTYRE, 1., STUCKENRATH, R. & DILL, R.F. ( 1 977) Relict barrier reef system off St. Croix: its impli cations with respect to Late Cenozoic coral reef develop ment in the Western Atlantic. Proceedings of the 3rd In ternational Coral ReefSymposium, Miami, 2, 1 5-2 1 . AHARON, P. & CHAPPELL, J. ( 1 986) Oxygen isotopes, sea level changes and the temperature history of a coral reef environment in New Guinea over the last I 05 years. Palaeogeogr. Palaeoclimatol. Palaeoecol., 56,3 37-379. BARD, E., HAMELIN, B. & FAIRBANKS, R.G. ( 1 990) U/Th ages obtained by mass spectrometry in corals from Barbados: sea level during the past 1 30,000 years. Na ture, 345, 405-408. BARD, E., FAIRBANKS, R.G., ARNOLD, M. & HAMELIN, B. ( 1 992) 230Th/234U and 14C ages obtained by mass spectrometry on corals from Barbados (West Indies), lsabela (Galapagos) and M ururoa (French Polynesia). In: The Last Deglaciation: Absolute and Radiocarbon Chronologies (Eds Bard, E. & Broecker, W.S.), NATO ASI Series I, Global Environmental Change, 2, pp. 1 0 3- 1 1 0. Kluwer Academic, Dordrecht. BARNES, J.W., LANGE, E.J. & PORTRAZ, H.A. ( 1 956) Thorium-uranium content in coral limestone. Science, 1 24, 1 75-1 76. BENDER, M.L., FAIRBANKS, R.G., TAYLOR, F.W., MAT THEWS, R.K., GODDARD, J.G. & BROECKER, W.S. ( 1 979) Uranium series dating of the Pleistocene reef tracts of Barbados, West Indies. Geol. Soc. Am. Bull. , 90, 577594. BERGER, W.H. ( 1 990) The Younger Dryas cold spell-a quest for causes. Palaeogeogr. Palaeoclimatol. Palaeo ecol., 89, 21 9-237. BIDDLE, K.T., SCHLAGER, W., RUDOLPH, K.W. & BUSH, T.L. ( 1 992) Seismic model of a progradational carbonate platform, Pico di Vallandro, the Dolomites, Northern Italy. Bull. Am. Assoc. petrol. Geol., 76, 1 4-30 BLOOM, A.L., BROECKER, W.S., CHAPPELL, J.N.A., MAT THEWS, R.K. & MESOLLELA, K.J. ( 1 974) Quaternary sea-level fluctuations on a tectonic coast: new 230Th/ 234U dates from the Huon Peninsula, New Guinea. Quat. Res. , 4, 1 85-205. BosELLINI, A. ( 1 989) Progradation geometries of carbonate platforms: examples from the Triassic of the Dolomites, northern Italy. Sedimentology, 3 1 , 1 -24. BRACHERT, T.C. & DULLO, W.-CHR. ( 1 990) Correlation of deep sea sediments and forereef carbonates in the Red Sea: an important clue for basin analysis. Mar. Geol. , 92, 255-267.
BRACHERT, T.C. & DULLO, W.-CHR. ( 1 99 1 ) Laminar miC·· rite crusts and associated foreslopes processes, Red Sea. J. sediment. Petrol. , 61, 3 54-3 6 3 . BRACHERT, T.C. & DULLO, W.-CHR. ( 1 994) Micrite crusts on Ladinian foreslopes of the Dolomites seen in the light of a modern scenario from the Red Sea. Jahrb. Geol. BA Wien, 50, 57-68. BROECKER, W.S., THURBER, P.L., GODDARD, J., Ku, T.L., MATTHEWS, R.K. & MESOLLELA, K.J. ( 1 968) MilankO·· vitch hypothesis supported by precise dating of coral reefs and deep-sea sediments. Science, 159, 297-300. CAMOIN, G.F., COLONNA, M., MONTAGGIONI, L.F., CASANOVA, J., FAURE, G. & THOMASSIN, B.A. ( 1 997) Holocene sea level changes and reef development in southwestern Indian Ocean. Coral Reefs, 16, 247-259. CHAPPELL, J. ( 1 974) Geology of coral terraces, Huon Peninsula, New Guinea: a study of Quaternary tectonic movements and sea-level changes. Geol. Soc. Am. Bull. , 85, 5 5 3-570. CHEN, J.H., CURRAN, H.A., WHITE, B. & WASSERBURG, G.J. ( 1 99 1 ) Precise chronology of the last interglacial period; 234U/230Th data from fossil coral reefs in the Bahamas. Geol. Soc. Am. Bull. , 103, 82-97. COLONNA, M., CASANOVA, J., DULLO, W.-CH. & CAMOIN, G.F. ( 1 996) Sea level changes and 8180 record for the past 34,000 yr from Mayotte reef, Indian Ocean. Quat.Res. 46, 335-339. DAVIES, J.P. & MARSHALL, J.F. ( 1 980) A model of epicon-· tinental reef growth. Nature, 287, 37-38. DuLLO, W-CHR. ( 1 990) Facies, fossil record, and age of Pleistocene reefs from the Red Sea (Saudi Arabia). Facies, 22, 1 -46. DuLLO, W.-CHR., MoussAVIAN, E. & BRACHERT, T.C. ( 1 990) The foralgal crust facies of the deeper forereefs in the Red Sea: a deep diving survey by submersible.Geo bios, 23, 26 1 -28 1 . EDWARDS, R.L., CHEN, J.H. & WASSERBURG, G.J. ( 1 987) 238U-234U-230Th systematics and the precise measure ment of time over the past 500,000 years. Earth planet. Sci. Lett. , 81, 1 75-1 92. EHNY, F. ( 1 987) Sedimentologie et diagenese precoce en milieu perirecifal. Les pentes de quelques fles vo/caniques coralliennes Ouest-lndopacifique: I. Mayotte, Bane du Geyser-Zelee et du Leven (N O. canal de Mozambique, Oc. lndien) et I. Chesterfield (Mer de Corail, Oc. Paci fique). These Doct. Universite Aix- Marseille II. EISENHAUER, A., WASSERBURG, G.J., CHEN, J.H. et al. ( 1 993) Holocene sea-level determination relative to the Australian continent: U/Th (TIMS) and 14C (AMS) dating of coral cores from the Abrolhos Islands. Earth planet. Sci. Lett., 1 14, 529-54 7. ESSON, J., FLOWER, M.F.J., STRONG, P.F., UPTON, o.G.J. & WADSWORTH, W.J. ( 1 970) Geology of the Comores Archipelago, Western Indian Ocean. Geol. Mag. , 107, 549-557. FAIRBANKS, R.G. ( 1 989) A 1 7,000-year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep ocean circulation. Nature, 342, 6 37-642. FAIRBANKS, R.G. & MATTHEWS, R.K. ( 1 978) The marine oxygen isotope record in Pleistocene corals, Barbados, West Indies. Quat. Res., 10, 1 8 1 - 1 96. FALLS, W.F. & WILLIAMS, P.F. ( 1 980) Glacial meltwater
Mayotte for e-slope morphology and sediments inflow into the Gulf of Mexico during the Illinoian glacial and Sangamonian interglacial. EOS Trans. Am. geophys. Union, 61, 258. FRICKE, H.W. & LANDMANN, G. ( 1 983) On the origin of Red Sea submarine canyons. Naturwissenschajien, 70, 1 95 - 1 96. FRICKE, H.W., VARESCHI, E. & SCHLICHTER, P. ( 1 987) Photoecology of the coral Leptoseris fragilis in the Red Sea twilight zone: an experimental study by submers ible. Oceanologia, 73, 37 1 -38 1 . GINSBURG, R.N., HARRIS, P.M., EBERL!, G.P. & SWART, P.K. ( 1 99 1 ) The growth potential of a bypass margin Great Bahama Bank. J. sediment. Petrol. , 61, 976-987. GoREAU, T. & LAND, L.S. ( 1 974) Fore-reef morphology and depositional processes, North Jamaica. In: Reefs in Time and Space, (Ed. Laporte, L.F.), Spec. Pub!. Soc. econ. Paleont. Miner., Tulsa, 18, 77-89. GRAMMER, G.M. ( 1 99 1 ) Formation and evolution of Qua ternary carbonate fares/opes, Tongue of the Ocean, Bahamas. PhD dissertation, University of Miami. GRAMMER, G.M. & GINSBURG, R.N. ( 1 992) Highstand vs lowstand deposition on carbonate platform margins: insight from Quaternary foreslopes in the Bahamas. Mar. Geol. , 103, 2 1 -45. GRAMMER, G.M., GINSBURG, R.N. & HARRIS, P.M. ( 1 993) Timing of deposition, diagenesis, and failure of steep carbonate slopes in response to a high-amplitude/high frequency fluctuation in sea level, Tongue of the Ocean, Bahamas. In: Carbonate Sequence Stratigraphy-Recent Events and Applications (Eds Loucks, R.G. & Sarg, J.F.), Mem. Am. Assoc. petrol. Geol., Tulsa, 57, 1 07- 1 32. GuiLCHER, A. ( 1 965) Coral reefs and lagoons of Mayotte Island, Comoro Archipelago, Indian Ocean, and of New Caledonia, Pacific Ocean. In: Submarine Geology and Geophysics (Eds Whittard, W.F. & Bradshaw, R.), pp. 2 1 -45. Butterworth, London. GuiLCHER, A. ( 1 97 1 ) Mayotte barrier reef and lagoon, Comoro Islands, as compared with other barrier reefs, atolls and lagoons in the world. In: Symposium on Regional Variation in Indian Ocean Coral Reefs (Eds Stoddart, D.R. & Jonge, M). Symp. zoo!. Soc., London, 28, 65-86. GviRTZMAN, G. ( 1 994) Fluctuations of sea level during the past 400 000 years: the record of Sinai, Egypt (northern Red Sea). Coral Reefs, 13, 203-2 14. HANTORO, W.S., PIRAZZOLI, P.A., JOUANNIC, C. et a/. ( 1 994) Quaternary uplifted coral reef terraces on Alor Island, East Indonesia.Cora/ Reefs, 13, 2 1 5-223. HARRIS, P.T. & DAVIES, P.J. ( 1 989) Submerged reefs and terraces on the shelf edge of the Great Barrier Reef, Australia. Morphology, occurrence and implications for reef evolution. Coral Reefs, 8, 87-98. IVANOVICH, M. & HARMON, R.S. ( 1 992) Uranium-Series Disequilibrium. Applications to Earth. Marine and Envi ronmental Sciences. Oxford Science Publications, Ox ford. JAMES, N.P. & GINSBURG, R.N. (Eds) ( 1 979) The Seaward Margin of Belize Barrier and Atoll Reefs. Spec. Pubis. int. Assoc. Sediment., No. 3, 1 9 1 pp. Blackwell Scien tific Publications, Oxford. KAuFMAN, A. ( 1 985) The distribution of 230Th/234U ages in corals and the number of last interglacial high-sea stands. Quat. Res., 25, 5 5-62.
235
KOIDE, M. & GOLDBERG, E.D. ( 1 965) 234U/238U ratios in sea water. In: Progress in Oceanography, 3 (Ed. Sears, M.), pp. 1 73 - 1 77. Pergamon, Oxford. KRAFFT, M. ( 1 982) L'eruption volcanique du Karthala en avril 1 977 (Grande Comore). C. R. Acad. Sci. Paris, 294, 7 5 3-758. LABEYRIE, L.D., DUPLESSY, J.C. & BLANC, P.L. ( 1 987) Variations in mode of formation and temperature of oceanic deep waters over the past 1 25,000 years. Na ture, 327, 477-482. LAMBERT, B. & Roux , M. ( 1 99 1 ) L 'environnement car bonate bathyal en Nouvelle Caledonie. Doc. trav. IGAL, Paris, 15, 2 1 3 pp. LAND, L.S. & MooRE, C. H. ( 1 977) Deep forereef and upper island slope, north Jamaica. Am. Assoc. petrol. Geol., Stud. Geol., 4, 5 3-65. LIGHTY, R.G. , MACINTYRE, I.G. & STUCKENRATH, R. ( 1 978) Submerged early Holocene barrier reef, south-east Flor ida shelf. Nature, 276, 59-60. MACINTYRE, I.G. ( 1 972) Submerged reefs of the eastern Caribbean. Bull. Am. Assoc. petrol. Geol., 56, 720-738. MACINTYRE, J.G., RUTZLER, K., NORRIS, J.N. et a/. ( 1 9 9 1 ) An early Holocene reef i n the western Atlantic: sub mersible investigations of a deep relict reef off the west coast of Barbados, W.l. Coral Reefs, 10, 1 67- 1 74. MASSE, J.-P., THOMASSIN, B.A. & ACQUAVIVA, M. ( 1 989) Bioclastic sedimentary environments of coral reefs and lagoon around Mayotte Island (Comoro Archipelago, Mozambique Channel, SW Indian Ocean). J. coast. Res. , 5, 4 1 9-432. MESOLELLA, K.J., MATTHEWS, R.K., BROECKER, W.S. & THURBER, D.L. ( 1 969) The astronomical theory of cli matic changes: Barbados data. J. Geol. , 77, 250-274. MONTAGGIONI, L.F. & MACINTYRE, I.G. (Eds) ( 1 99 1 ) Reefs as recorders of environmental changes. Coral Reefs, 10, 53-54. MOORE, C.H., GRAHAM, E.A. & LAND, L.S. ( 1 976) Sedi ment transport and dispersal across the deep fore-reef and island slope (-5 5 m to -305 m) Discovery Bay, Jamaica. J. sediment. Petrol., 46, 1 74- 1 87. NEUMANN, A.C. & MACINTYRE, I.G. ( 1 985) Reef response to sea level rise; keep-up, catch-up or give-up. In: Proceedings of the 5th International Coral Reef Sympo sium, Tahiti, 3, pp. I 05- 1 1 0. NOUGIER, J., CANTAGREL, J.M. & I
236
W-Ch. Dullo et al.
Graphic correlation of oxygen isotope stratigraphy ap plication to the late Quaternary. Paleoceanography, 1, 1 37- 1 62. REISS, Z. & HOTTINGER, L. ( 1 984) The Gulf of Aqaba: Ecological Micropaleontology. Ecological Studies, 50. Springer-Verlag, Berlin. RUDDIMAN, W.F. & DUPLESSY, J.-C. ( 1 985) Conference on the last deglaciation: timing and mechanism. Quat. Res., 23, 1 - 1 7. SAERTRE, R. & DA SILVA, A.J. ( 1 984) The circulation of the Mozambique Channel. Deep-Sea Res. , 31, 485-508. SARNTHEIN, M. & TIEDEMANN, R. ( 1 992) Younger Dryas style cooling events at glacial terminations I-IV at ODP site 658 associated benthic o 1 3C anomalies constrain meltwater hypothesis. Paleoceanography, 5, I 04 1 1 05 5 . SHACKLETON, N.J. ( 1 987) Oxygen isotopes, ice volume and sea level. Quat. Sci. Rev., 6, 1 83- 1 90. SHACKLETON, N.J. & MATTHEWS, R.K. ( 1 977) Oxygen isotope stratigraphy of late Pleistocene coral terraces in Barbados. Nature, 268, 6 1 8-6 1 9.
SnELTJES, L. ( 1 988) Mayotte (Archipel des Comores). Carte geologique de France. Echelle 1:50 000, notice explicative. BRGM, Orleans. THOMASSIN, A., ARNOUX, A., COUDRAY, J. et a/. ( 1 989) La sedimentation actuelle dans Ia lagon de Mayotte (ile volcanique a recif-barriere, SW ocean lndien) et son evolution recente en liaison avec les apports terrigenes. Bull. Soc. geol. Fr., 8, 1 23 5- 1 2 5 1 . THURBER, P.L. ( 1 962) Anomalous 234U/238U in nature. J. geophys. Res., 67, 45 1 8-4520. VEEH, H.H. & VEEVERS, J.J. ( 1 970) Sea level at - 1 75 m off the Great Barrier Reef 1 3,000 to 1 7,000 years ago. Nature, 266, 5 36-537. VOLLBRECHT, R. & MEISCHNER, D. ( 1 993) Sea level and diagenesis: a case study on Pleistocene beaches,Whale bone Bay, Bermuda. Geol. Rundsch., 82, 248-262. WALTHER, J. ( 1 888) Die Korallenriffe der Sianai Halbinsel. Geologische und biologische Beobachtungen. Abh. math. naturwiss. Kl. Konig. Sachs. Ges. Wissensch. , 14, 439-506.
Spec. Pubis int. As s . Sediment. (1998) 25, 237-248
Tectonic and monsoonal controls on coral atolls in the South China Sea WANG GUOZHONG Labo rato ry o fMa rine G e ology, Tongji Unive rsity, Shang hai 200092, China
ABSTRACT
There are three types of coral atolls in the South China Sea: continental shelf atolls, continental slope atolls and oceanic atolls. This classification is based mainly on tectonic settings. These atolls form a complete system, which has not been described in other seas and oceans. Major tectonic trends seemingly control the distribution and directions of continental slope atolls. The size of reef-crests is closely related to that of antecedent platforms. Based on the width of the reef-crest, the continuity of the reef-wall and the existence of a lagoon, modern continental slope atolls can be classified into three types: large atolls with separate table reefs, medium atolls with a continuous reef wall, and small atolls or isolated table reefs. Slope atolls are larger than shelf atolls in the Great Barrier Reef, and reef-crest width for various classes must range between 5.5 and 2.5 km. The alternating influence ofNE and SW monsoons, waves and currents on the antecedent platforms of continental slope atolls leads to a relatively symmetrical morphology with a more developed NE reef-crest. This bi-directional formation of slope coral atolls contrasts with the formation of shelf and oceanic atolls, which are controlled by unidirectional trade winds.
INTRODUCTION
The South China Sea is a marginal sea of the Pacific Ocean, lying in the centre of the East Asian mon soon belt. It shares some features with other tropical seas and oceans, but also has distinctive character istics, which do not occur elsewhere. One of these special features is that in the South China Sea there is a complete sequence of atolls, including continen tal shelf, continental slope and oceanic atolls. Since the late 1970s, exploration for petroleum and other mineral resources has encouraged re search on the coral reefs and associated carbonates of the South China Sea. The author and a research group from the Department of Marine Geology, Tongji University, have carried out annual field investigations of coral reefs and sedimentary facies on the northern continental shelf and the Xisha Islands of the South China Sea (Wang Guozhong et al., 1979, 1982, 1986, 1988, 1990; Lu Bingquan et al., 1984). We have measured submarine relief by measuring staff, lead line or theodolite, observed reef landscapes and hydrodynamic conditions, doc umented bottom characteristics and determined biotic cover. Coral genera and species coverage and
growth characteristics have been determined along transects. Sediment samples and bottom photo graphs were taken by scuba divers. Sedimentary facies were established and examined. The analysis of the grain-size distribution of loose sediments was performed using standard sieve and settling tech niques. Weight percentages of fractions coarser than 0.5 mm were estimated using stereoscopic micro scope and balance. Based on the results of our investigations and published data, the features of coral atolls of the South China Sea, and particularly continental slope atolls, are summarized. The con trol factors and the bi-directional formation model of the slope atolls are described.
GEOLOGICAL BACKGROUND
In the late Oligocene or early Miocene (about 32 Ma) many micro-continental blocks, including the present Palawan Island, the Nansha and Xisha Islands, etc., were separated from the palaeoconti nent and drifted southward. The ancestral South
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
237
238
Wang Guozhong
China Sea developed in the middle Miocene, 17 Myr ago, after extensive transgressions (Wang Chongyou et a!., 1979; Katili, 1981) (Fig. 1). Unlike other tropical seas, the South China Sea includes three types of submarine geomorphology: a deep-sea basin, a continental slope and a continen-
tal shelf (Fig. 2; Feng Wenke et a!., 1988; Liu Yixuan & Mao Shuzhen, 1989), which are not seen associated in other tropical seas. The central deep-sea basin is deeper than 3400 m in the east and more than 4200 m in the south-west; the maximum depth is 4577 m (Fig. 2). The depth
(a)
A
bE D 2 0 3 a 4 1----1 5 B X
continental slope
-----
[2] EJ EJ C2j
L slope
PT
s·N
6 7 8 9
Pl
A
/'----- M
(b) �
0
100
200km
Fig. 1. Tectonic map (a) and schematic cross-section (b) of crust structure in the South China Sea (based on Katili,. 1981; Taylor & Hayes, 1983). I, coastline; 2, 200-m isobath; 3, 2000-m isobath; 4, deep basin boundary; 5, line of cross-section; 6, location of relict rift; 7, linear magnetic anomaly; 8, fault; 9, consumption zone; M, Mohorovicic discontinuity; C, Conrad discontinuity; X, Xisha Islands; L, Lilu Bank; PI, Palawan Island; PT, Palawan Trough.
Contr ols on co ral atolls in South China Sea
239
!08E
104E
Fig. 2. Distribution of some
modern coral reefs in the South China Sea (based on Feng Wenke et a!., 1988; Liu Yixuan & Mao Shuzhen, 1989). a, Isobath; b, national boundary; c, boundary between shelf and slope; d, boundary between slope and basin; e, fringing reef; f, atoll. 1, Shalao Fringing Reefs; 2, Luhuitou Fringing Reefs; 3, Daichau Barrier Reef; 4, Lingchang Barrier Reef; 5, Weizhou Fringing Reef; 6, Raining Shelf Atoll; 7, Haian Shelf Atoll; 8, Nanbing Shelf Atoll; 9, Huangyan Oceanic Atoll; Slope atolls: 10, Xuande Atoll (including Yongxing Island); 11, Yongle Atoll; 12, Huaguang Atoll; 13, Panshiyu Atoll; 14, Lilu Atoll; 15, Yongshu Atoll.
0--.!.£2.2.001crn
� a Oct E]t [3 e Be I ·�:·If
of the MohoroviCic discontinuity below the basin varies from 8 to 14 km and the basin is probably floored by oceanic crust (Fig. 1, Ludwig eta!., 1979; Katili, 1981; Taylor & Hayes, 1983). Around 15 · N, there is a series of isolated submarine volcanic seamounts, some of which extend from the deep-sea floor to near present sea-level (Xu Zhongfan & Zhong Jinliang, 1980). The continental slopes extend from the shelf edge to the deep-sea basin and are cut by submarine troughs and canyons, which promote the develop ment of upwelling currents, providing the hydrody namic condition favoured by reef-building corals and other organisms. The crust here is estimated to be 24-26 km thick (Fig. 1, Ludwig et a!., 1979; Katili, 1981; Taylor & Hayes, 1983). Drilling re-
suits indicate that the basement is Precambrian gneiss and the overlying reefal limestone has a thickness of about 1 250 m on Yongxing Island in the Xisha Islands (Wang Chongyou et a!., 1979). However, beneath the reef limestone of Lilu Bank (or Lilu Atoll) in the Nansha Islands, the basement is composed of early Tertiary and Cretaceous terres trial sediments (Taylor & Hayes, 1983). These results demonstrate that the Xisha, Zhongsha, Dongsha and Nansha Islands are carbonate plat forms developing on a transitional crust. The continental shelves of the South China Sea are usually taken as less than 250 m below present sea-level. They are wider in the north and south than in the east and west. The thickness of the continental crust beneath the shelves is estimated to
240
Wang Guozhong
be beween 26 and 30 km (Katili, 1981; Taylor & Hayes, 1983; Feng Wenke et a!., 1988; Figs 1 & 2).
TYPES OF ATOLLS AND GEOMORPHOLOGY
Since the Miocene, the crust of the South China Sea has been uplifted. Following the transgressions on the relict positive relief of various geomorphologi cal elements, coral reefs developed and formed carbonate platforms; many of them built up to present sea-level. The thickness of reef carbonates is more than 1200 m (Wang Chongyou et a!., 1979). When the sea-level was around 130 m below its present position in the Quaternary, these carbonate platforms were exposed, partly corroded, and karstified. Then, as sea-level rose over the anteced ent platforms during the Holocene, coral atolls and other reefs developed. Based on the three major submarine geomorphological elements mentioned above, the coral atolls in the South China Sea may be subdivided into three types: continental shelf atolls, continental slope atolls and oceanic atolls. The geomorphological expression of these settings has determined both the type and distribution of the atolls (Fig. 2; Table 1 ). Table 1 lists some of the types of coral atolls and their locations. Continental shelf atolls occur mainly on the southern (Xunta) continental shelf in depths of 100-200 m, for example, Haining Atoll and Nan bing Atoll. The ring-like or horseshoe-like atolls vary from 0. 7 to 2 km in diameter, and the depths of the shallow lagoons vary from 4 to 11 m (Chen Shijian & Zhong Jinliang, 1990). Oceanic coral atolls are located in the central deep basin around 15 ·N. For example, Huangyan Atoll is based on an isolated submarine volcanic seamount which has a base deeper than 4000 m. It
has an isosceles triangular form with a perimeter of about 46 km and its apex towards the south-west. The lagoon is about 20 m deep, surrounded by a continuous reef-wall. Huangyan Atoll is similar to oceanic atolls in the Pacific and Indian Oceans (Xu Zhongfan & Zhong Jinliang, 1980). Continental slope atolls dominate in the South China Sea and correspond to a special type of coral reef (Wang Guozhong et a!., 1990), not reported in other seas and oceans. They are unusual and in clude about 200 shoals, reefs, islets and islands in the Nansha Islands and more than 30 reefs and islands in the Xisha Islands (Wang Guozhong et a!., 1994).
CONTINENTAL SLOPE ATOLLS AND TECTONICS
In the southern, central and northern parts of the South China Sea, extensive continental slopes were fragmented by tectonics and then formed a series of submarine ridges and peaks, isolated by troughs and canyons about 2000 m deep (Figs 1 & 2). The trends of continental slopes and submarine ridges and troughs are controlled by the main tectonic trends, which are predominantly NE -SW. In tht:: middle of the median ridge of expansion the E-W direction has a great significance (Figs 1 & 2, Feng Wenke et a!., 1988; Liu Yixuan & Mao Shuzhen, 1989). Organic reefs developed on the positive relief of the continental slopes and closely mim icked the underlying topography. The morphology, trends and size of the continental slope atolls were seemingly controlled by the NE-SW and E-W regional tectonic trends. The rose diagram of the long dimensions of the atolls clearly shows that this trend is predominant (Fig. 3). The continental slope atolls of the South China
Table 1. Types of coral atolls in the South China Sea Type of atoll
Type of crust
Named examples
Location
Continental shelf
Continental
Haining Atoll Haian Atoll Nanbing Atoll
s·oo'N, s·o1'N, s·21'N,
112.37'E 112.3l'E 112.39'E
Continental slope
Transitional
Haikou Atoll Panshiyu Atoll Yongle Atoll
9•tt'N, 16.03'N, t6•3t'N,
116.27'E 111.48'E 111.40'E
Oceanic
Oceanic
Huangyan Atoll
1s·o9'N,
117"45'E
241
Contr ols on co ral atolls in South China Sea N
0
Large atolls
Medium atolls Small atolls
m---�
-----
20 n=70
40 '-----E
60
Fig. 3. Rose diagram of long dimensions of continental
slope atolls in the South China Sea.
Sea are walled-reef complexes with a rigid organic framework. Marine geological studies show that modern reefs grew on previously exposed and eroded Pleistocene carbonate platforms and formed a thin veneer of Holocene reef sediments with a thickness ranging from 16.30 to 17.72 m in the Xisha Islands and 17.30 m in the Well Nanyong-1 in the Nansha Islands (Wang Chongyou eta/., 1979; Zhu Yuanzhi & Sha Qingan, 1992). A survey of the reef-crests of 70 reefs in the South China Sea (Table 2) shows that generally only one atoll developed on an antecedent platform. Only in a few cases can two atolls develop along the trend of an antecedent platform. The morphology and char acteristics of modern reefs are closely related to the extent and width of their antecedent platforms. Purdy ( 1974) and Davies (1983) suggested that the morphology of modern reefs mimics that of the antecedent platforms on which they grew. The extent to which the antecedent morphology has been maintained and later masked by Holocene reef growth offers the basic framework for the classifica tion of shelf reefs (Hopley, 1983). The principle of this classification is also suitable for continental slope coral reefs, and in combination with the extent of the reef-crest, the continuity of the reef wall, and the existence of a lagoon, modern conti nental slope coral reefs can be classified into three classes, detailed below. The boundary values of reef-crest widths for the various classes are 5.5 and 2.5 km for the slope atolls in the South China Sea, in contrast to 3.5 and 1. 75 km for those in the Great Barrier Reef(Table 3; Wang Guozhong eta/., 1988). 1 Large atolls with separate table reefs. Numerous separate table reefs or islets surround an open central lagoon and form a large atoll with a diame ter ranging from 5.5 to 62 km; the average for 26
80
-- - - Average depth
100 m
Fig. 4. Depth of atoll lagoons in the South China Sea.
atolls is 17.61 km. The depth of the lagoons of these atolls ranges from 9 to 100 m, but 62 % of them have depths between 40 and 69 m (Fig. 4). 2 Medium atolls with a continuous reef-wall are surrounded by a narrow, continuous reef-wall rim and form a single relatively closed ring lagoon. The width of the reef-crest varies from 5. 1 to 2. 6 km, and the average width for 18 atolls is 3.91 km. The length of the reef-crest varies from 4.0 to 33 km. The depth of these lagoons is 9-32 m with an average value of 18.61 m (Fig. 4). 3 Isolated small atoll or table reefs grow separately on a narrow antecedent platform with a reef-crest less than 2.5 km wide. The only exception is Jiny indao Table Reef in Xisha Islands, which is 3 km wide. According to their morphology and stage of evolution, these small slope reefs may be further subdivided into three groups: (a) Small closed atolls-a single lagoon is sur rounded by a near-continuous reef-wall with, locally, a lagoon channel connected with the open sea; Haikou Atoll is 1.8 km wide and its lagoon has a channel of 3-5-m depth. The depth of these small atoll lagoons varies from a few metres to 33 m, but half of the lagoons have depths less than 10 m (Fig. 4). (b) Table-like atolls are a group between small closed atolls and table reefs. An example is Hejiao Atoll, which has a closed shallow lagoon of 3-4-m depth; the areal extent of the lagoon is about one-fifth that of the reef-crest.
242
Wang Guozhong
Table 2. Characteristics and location of 70 continental slope atolls in the South China Sea Location
Size of atolls Longitude (E)
Width (km)
Length (km)
Lagoon depth (m)
Trend
18 60 49 60 30 100 65 45 90 42 34 67 53 53 46 50 43 82 27 51 40 30 9 45 65 82
EW90• NNE16. NE57" NE35. NE69• NNE59. NNE28• NE39• NNE31• NNE32• E8o· NE38. NE39• NNE15• NNW347" NE46• NE44• NEE75• E9o· NEE58. NW314• NE52• NW30l• NEE74• NE49• NNE20·
Number
Name
Latitude (N)
LARGE ATOLLS I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Dongsha A. Xuande A. Yongle A. Dongdao A. Huaguang A. Zhongsha A. Yongdeng A. Shuangzi A. Lilu A. Lesi A. West Zhongye A. Changtan A. Daoming A. Haima A. Mahuan A. Nanfangtan A. Jiesheng A. Zhenghe A. Meiji A. Quyuantan A. Xianbin A. Yongshu A. Lisheng A. Yuya A. Andu A. Nanwei A.
20.42' 16.51' 16.31' 16.28' 16.13' 15.50' 11.28' 11.25' 11.21' 11.21' 11.03' 11.02' 10.48' 10.46' 10.45' 10.30' 10.30' 10.18' 9.55' 9•01' 9.44' 9•37' 8·5o' 8·o8' 7.45' 7•43'
116.48' 112.18' 111.40' 112.35' 111.42' 114.19' 114.41' 114.21' 116.51' 114.36' 114.15' 114.44' 114.29' 117.47' 115.52' 116.41' 115.45' 114.28' 115.31' 114.27' 116.29' 113.00' 114.00' 114.41' 114.14' 111.39'
24.0 20.0 17.0 28.0 8.2 60.0 10.0 8.0 62.0 5.5 6.0 6.0 13.0 8.3 10.0 28.0 10.0 23.0 6.0 15.5 7.0 6.0 7.4 16.0 28.0 27.0
27.0 30.0 22.0 55.0 27.5 140.0 19.0 15.0 143.0 7.4 13.0 33.4 39.0 15.8 18.0 67.0 12.0 58.5 9.0 55.0 23.0 25.0 20.0 37.0 74.0 54.0
MEDIUM ATOLLS 1 Beijiao A. 2 Yuzhuo A. 3 Langhua A. 4 Panshiyu A. 5 East Zhongya A. 6 Zhubi A. 7 Libai A. Renai A. 8 9 Xiane A. 10 Jianzhang A. 11 Banyue A. 12 Tianlan A. 13 East Yinging A. 14 Nanhua A. 15 Siling A. 16 Baijiao A. 17 Nanhai A. 18 Danwan A.
17"06' 16.21' 16.03' 16.03' 11.04' 10.54' 10.10' 9.44' 9.22' 9.02' 8·54' 8.52' 8·51' 8.47' 8.22' 8·12' 7•57' 7•23'
111.04' 112.02' 112.31' 111.48' 114.21' 114.06' 115.20' 115.52' 115.27' 116.40' 116.18' 114.39' 112.32' 114.12' 115.13' 113.18' 113.57' 113.48'
4.4 3.6 5.1 3.9 3.3 5.0 4.5 5.0 3.5 2.6 3.0 2.8 5.0 3.0 3.0 4.0 5.0 3.7
12.0 14.7 17.6 8.2 8.3 6.5 5.5 16.5 7.5 4.4 6.8 4.0 13.5 17.0 15.8 33.0 12.0 7.0
SMALL ATOLLS (a) Atolls I 2 3 4 5 6
10.53' to·o3' 9.35' 9•27' 9.20' 9•Jj'
114.56' 113.44' 116.09' 116.55' 115.57' 116.27'
2.0 2.0 0.8 1.7 1.5 1.8
6.6 13.0 2.0 2.5 7.0 2.6
Cailun A. Daxian A. Niuchelun A. Pengbuo A. Xinyi A. Haikou A.
20.0 16.8 14.8 10.8 27.0 22.0 17.0 27.0 20.0 30.0 32.0 14.6 9.0 14.5 3.7
5.0 33.0 12.0 20.0
NEE65• NEE76• NEE7o· NEE64• NEE64• NE56• NEE64• NNE13• N2• NW315• NE44• NE52• W28o· N354• NWW283• NE37" NW318• NEE74.
NNE25• N6• NNE29• NNEt6•. NEE77• NEE69•
243
Controls on co ral atolls in South China Sea Table 2. Continued
Size of atolls
Location Number
Name
7 8 9 10 II 12 13 14 15
Bisheng A. Yinqing A. West Yinqing A. Riji A. Boji A. Guangxingzi A. Guangxing A. Huanglu A. Nantong A.
(b) Table-like 16 17 18
atolls Hejiao T.A. Bantu T.A. Huayang T.A.
( c) Table reefs 19 20 21 22 23 24 25 26
Jinyindao T.R. Zhongjiandao T.R. Xiyue T.R. Fulusi T.R. Xiaoxian T.R.. Mayuan T.R. Nanweidao T.R.. Anbuo T.R.
Lagoon depth (m)
Longitude (E)
Width (km)
Length (km)
s·58' 8"55' 8"52' 8"38' 8"06' 7"37' 7"37' 6"56' 6"20'
113"42' 112"21' 112"14' 111"40' 114"07' 113"55' 113"48' 113"34' 113"16'
2.0 0.7 2.3 1.8 1.9 2.0 2.0 1.5 0.9
10.7 0.9 8.3 6.0 2.8 2.8 9.3 2.0 1.8
27.0 14.6 18.3
NEE66" E90" NEE62" NEE68" E90" E90" NEE75" E84" NESS"
10"15' 10"08' s· s1'
115"21' 116"08' 112"50'
1.3 0.5 1.5
2.0 1.1 5.5
4.0
NEss· NW325" NESI"
16"27' 15"47' 11"10' 10"15' 10"02' 9"12' 8"39' 7"54'
111"31' 111"13' 115"00' 113"12' 114"05' 113"40' 111"55' 112"55'
3.0 2.4 1.9 0.9 0.6 1.2 0.9 1.0
6.0 4.2 3.7 2.8 0.6 1.2 1.9 2.5
Latitude (N)
Trend
NEE77" NE54" NESO" NE45" NE45" NE45" NNE30" NNE23"
A, Atoll; T.A., table-like atoll; T.R., table reef.
(c) Isolated table reefs grow on a narrow anteced ent platform and there is no lagoon in the reef crest because of the small size of its roof, or because the small pre-existing shallow lagoon has been filled with reefal sediments. The morphology, construction and size of conti nental slope atolls are controlled by the morphology and extent of their antecedent platforms. The slope atolls of the various classes were formed by the differential dissolution of previous reefal lime stones. When carbonate platforms of various size were exposed by lowered sea-level during Pleis tocene glacial periods, the rates of dissolution were different. Therefore, a variety of karst features were developed and subsequently different reefs formed in the post-glacial period: 1 Isolated atolls or table reefs may have formed in the Holocene period on small platforms with mound-like or planar topography. 2 The surfaces of the medium antecedent platforms may have been modified in the glacial period, giving rise to a shallow karst depression with a continuous solution rim, leading to the formation of modern closed atolls with continuous reef-walls.
3 The extensive dissolution of larger exposed car bonate platforms during the glacial period resulted in the formation of deep depressions in the centres of platforms and of some drainage channels. Large atolls with separate table reefs may have formed on positive features of these antecedent platforms dur ing the Holocene.
BI-DIRECTIONAL FORMATION MODEL AND MONSOONS
The South China Sea lies between the Pacific and Indian Oceans, climatically in a tropical and sub tropical area ( 23"N to 4"N). Sea surface tempera tures on the continental slope range between 28 and 30"C. The South China Sea area is the main field of East Asian monsoons. From October to April, north-easterly winds predominate, and from June to August south-westerly winds prevail. Depending on latitude, the predominant wind may differ in time and intensity. The north-easterly monsoon may prevail for 8 months in the north and 6 months in the south. The mean monthly velocity of surface
244
Wang Guozhong
winds ranges from 8-1 1 m s-1 in the north to 6 -9 m s-1 in the south. The frequency of north easterly and northerly winds reaches 80-90 %. The south-westerly monsoon lasts 3 months in the north and 5 months in the south. The mean monthly wind velocity is 6 -7 m s-1. The SW monsoon is weaker than the NE, and frequently breaks off. The fre quency of south-westerly and southerly winds is 70 % in the south and 50 % in the north (Sun Xiangping et al., 1981). The period from June to October is the season of typhoons; these influence the South China Sea area more than nine times every year and they are always south-westerly. The wind velocity in the centre of the typhoons fre quently reaches 60-6 5 m s-1, a force 10- 12 wind on the Beaufort scale, and 110 m s-1 at a maximum. Surface wave intensity is directly controlled by monsoons and also has an obvious seasonal pattern. In the period of the NE monsoon, the mean monthly wave height is 1. 2 -2. 7 m in the north and 0. 5-2. 0 m in the south. In summer, the mean monthly wave height is 0.6-1.4 m in the south and 1. 0-2. 2 m in the north. The swell is higher than the surface waves caused by the wind. Wave height typically varies from 1 to 3 m with a maximum of 10 m. The wave height of typhoons is usually higher than 5 m and up to 1 5 m (Sun Xiangping et al., 1981). Surface currents are mainly wind-driven. The mean velocity of the wind-drift driven by NE monsoons is greater than that of SW monsoons. Thus, the variables of winds, waves and currents are all related to the East Asian monsoons. The monthly major and secondary wind frequency dia gram (Fig. 5) indicates the intensity and anisotropy of these dynamic parameters. NE and SW mon soons alternate in the South China Sea, but the wind period of the NE monsoon is longer than that of the SW. The NE monsoon is also the stronger. The extraordinary position and topography of the continental slopes, and the circulation of deep-sea currents which promote upwelling provide rich nutrients for reef-building organisms and create favourable conditions for the formation of coral reefs on the submarine platforms. The growth rate of reef-building organisms and the accumulation of carbonate sediments depend on the sedimentary environment, including the influence of winds, waves and currents. Because of the alternation of the NE and SW monsoons, the maximum growth of continental slope atolls takes place on the well oxygenated NE and SW margins of the antecedent platforms, with slower growth elsewhere. Thus,
20N
15N
ION
5N
I I 0
10%
.� IIOE
115E
120E
Fig. 5. Wind-rose diagram of the monthly major and
secondary wind frequency in the South China Sea. (A) Southern part of Hainan Island (18 "N); (B) Dongsha Islands (20"N); (C) Xisha Islands (16"N); (D) Nansha Islands (7 "N).
crescentic coral reefs developed mainly on the NE and SW margins and reef detrital sediments were transported to the centre, forming the relatively symmetrical morphology and sedimentary facies distribution of the atolls. Because the intensity of the NE monsoon is higher in the north than in the south, the NE margins of the atolls display a greater development than the SW ones; an algal ridge may form on the NE reef flat of some atolls (Zhuang Qiqian et al., 1983); this feature is widely reported on Pacific coral atolls (Milliman, 1974). The slope atolls become more symmetrical south wards in parallel with the change in monsoon intensity. The morphology of the atolls is basically symmetrical, but is less so in north. This bi directional formation model, controlled by the al ternating monsoons, is characteristic of continental slope atolls in the South China Sea. The rose diagram of long dimensions of continen tal slope atolls (Fig. 3) clearly shows that the NE direction is predominant. This results from the
Controls on co ral atolls in South China Sea
interaction of regional tectonics and the East Asian monsoons and reflects the wind-rose diagram (Fig. 5). The sedimentary facies maps of Yongle and Panshiyu atolls in the Xisha Islands (Fig. 6) are relatively symmetrical, but the NE parts of the atolls are more developed than the SW. Based on the investigations of the main coral atolls in the South China Sea, a model of the sedimentary facies of continental slope atolls has been proposed by Wang Guozhong et a!., (1982, 1990). From the open marine fore-reef to the protected lagoon, modern coral atolls have been subdivided into seven sedimentary facies: (i) fore reef sand -gravel facies; (ii) fore-reef colluvial facies; (iii) autochthonous reef facies; (iv) reef-crest rubble (or boundstone) facies; (v) reef flat gravel-sand with bushy coral facies; (vi) sand cay gravel-sand facies; and (vii) lagoon sand with lush sponge-coral facies.
Hm
(b)
0-J-Km
Fig. 6. Sedimentary facies map of Yongle Atoll (a) and Panshiyu Atoll (b); I, sand cay; 2, autochthonous coral reef facies; 3, reef-crest boundstone facies; 4, reef flat sand-rubble and lagoon patch reef facies; 5, depth of lagoon floor; 6, lagoon channel.
245
These facies are distributed around the atolls. The sedimentary facies map and profile of Huaguang Atoll (Fig. 7) may serve as an example (Wang Guozhong eta!., 1988). It shows that the shape and distribution of the sedimentary facies of this atoll are relatively symmetrical and that corals on the NE side display a better development than those on the SW; the NE reef-crest is wider than the SW one.
DISCUSSION
Oceanic atolls and continental shelf atolls have long been recognized in tropical oceans and some other seas (Stoddart, 196 5; Maxwell, 1968). Ladd (1977) subdivided atolls into two groups: (i) deep-sea atolls, which rise from the deep-sea floor; and (ii) shelf atolls, which develop on the continental shelf. Deep-sea atolls are isolated structures and are found in the Pacific and Indian Oceans. Most deep-water atolls grew on submarine volcanic seamounts. Shelf atolls are found in many parts of the world. Some small atolls rising from the slopes of larger island pedestals also belong to this category (Ladd, 1977). In the South China Sea, a third group of atolls has been recognized, the continental slope atolls, which develop on continental slope highs (Nansha Islands, Zhongsha Islands, Xisha Islands and Dongsha Is lands in the South China Sea, Wang Guozhong et a!., 1988, 1990). They are based on Tertiary terres trial formations or on residual relics of Precambrian gneiss (Wang Chongyou et a!., 1979; Taylor & Hayes, 1983). These three types make up a com plete system of coral atolls, which coincides with major domains of submarine geomorphology. The distribution and trends of continental slope atolls in the South China Sea are seemingly con trolled by NE -SW regional tectonics; submarine geomorphology and continental slopes are con trolled by these trends and the organic reefs are developed on the positive relief of slopes. In the South China Sea, the continental slopes are cut by submarine troughs and canyons. The geo graphical locations of the continental slopes be tween the deep-sea basin and continental shelves is favourable for upwelling and strong hydrodynamic conditions, which provide rich nutrients for reef building organisms. For this reason, there is a great development of continental slope atolls in the South China Sea. These may be further subdivided. Hopley (1983) suggested a threefold division for continental shelf reefs in the Great Barrier Reef
246
Wang Guozhong c
N
l
A
0 tl
·10
s·w·
NC:
4 �IV �� · ··.·
V
. ;c
-50 m
1 Zkm
_,III
��
�\
�
�
. .. . . ·
· · · · "' ' -:� .
.
;1
�: .
VIa
-·
. ·
VIb
VIc
VIc ')"'==•"--.Jkm
..
1
VIb VIa II,
�3
Fig. 7. Sedimentary facies map and profile of Huaguang Atoll. I, Sand cay; 2, autochthonous coral reef facies; 3,
reef-crest boundstone facies; 4, reef flat sand-rubble and lagoon patch reef facies; 5, depth of lagoon floor; 6, lagoon channel; II, fore-reef colluvial facies; III, autochthonous reef facies; IV, reef-crest rubble facies; V, reef flat gravel-sand with bushy coral facies; VIa, lagoon slope with growing coral facies; VIb, lagoon patch reef facies; VIc, lagoon floor sponge-coral facies.
based on the sizes of their antecedent platforms; 1.75 and 3.5 km of platform width are the bound ary values of the three classes of shelf reefs. The principle of Hopley's classification can be applied to the South China Sea atolls. Because the continental slope atolls here are characterized by the larger sizes of the antecedent platforms or reef-crests, the boundary values for classification are 2.5 and 5.5 km (Table 3). The continental slope in the South China Sea is influenced by East Asian monsoons, alternating seasonally from the NE to the SW. On the well oxygenated windward NE and SW margins, where hydrodynamic conditions are especially favourable for reef growth, reef-building organisms flourished, resulting in hard-line crescentic coral reefs. Later coral reefs developed continuously on the other margins of carbonate platforms and formed rela-
tively symmetrical atolls. This bi-directional forma tion and relatively symmetrical morphology of atolls are specific characteristics of continental slope atolls in the South China Sea. They differ from the monodirectional formation and asymmet rical morphology of shelf atolls in the Great Barrier Reef and oceanic atolls in the Pacific and Indian Oceans. Maxwell (1968) and Hopley (1983) suggested a monodirectional model for the formation of conti nental shelf atolls in which maximum reef growth is expected on the .windward reef margins. Thus, a crescentic reef would develop first, slowly changing into an open ring and then a closed ring or atoll. Such an asymmetrical form of shelf atolls is re ported in the Great Barrier Reef and the Coral Sea. This monodirectional formation model is con trolled by the trade winds, which are predominantly
247
Contr ols on co ral atolls in South China Sea Table 3. Classification of continental slope atolls in the South China Sea
Sample size
Average width (km)
Min. width (km)
Max. width (km)
Example (width)
Large atolls with separate table reefs
26
17.61
5.5
62
Yongle Atoll (17.0 km)
Medium atolls with continuous reef walls
18
3.91
2.6
5.1
Panshiyu Atoll (3.9 km)
Small atolls (a) Atoll (b) Table-like atoll (c) Table reef
26
1.54
0.5
2.4
Class of slope atolls
southeasterly in the Coral Sea (Hopley, 1982). An asymmetrical morphology is common in oceanic atolls (e.g. Bikini in the Pacific-Tracey eta!., 1948; Hogsty Atoll in the Atlantic-Milliman, 1974), which developed in the trade wind domain.
CONCLUSIONS
In the South China Sea there are three types of atolls: continental shelf atolls, continental slope atolls and oceanic atolls. These are controlled by submarine geomorphology and form a complete sequence of coral atolls, not reported in other seas and oceans. Continental slope atolls are the most developed coral reefs in the South China Sea and are walled reef complexes with a rigid organic framework, which grew on exposed and eroded Pleistocene carbonate platforms and form a veneer of only about 20-m thickness of Holocene reef carbonate sediments. Relationships between reef-crest width, the continuity of a reef wall and the occurrence of a lagoon, allow continental slope atolls to be grouped into three classes. The boundary values of reef-crest width for these various classes are 5.5 and 2.5 km. Thus, the morphology, construction and sizes of continental slope atolls are controlled by the mor phology and extent of Pleistocene antecedent plat forms. Continental slope atolls in the South China Sea are larger than those reported in the Coral Sea. Hydrodynamic conditions in the South China Sea are influenced by the NE and SW East Asian monsoons, resulting in a relatively symmetrical morphology and distribution of sedimentary facies. This bi-directional formation model differs from the trade-wind-controlled monodirectional forma-
Haikou Atoll (1.8 km) Hejiao Atoll (1.3 km) Zhongjiandao Table Reef (2.4 km)
tion model of the shelf atolls in the Coral Sea and oceanic atolls.
ACKNOWLEDGEMENTS
The author thanks the National Natural Science Foundation of China for financial support, and Lu Bingquan, Quan Songing and others for their co operative study on the coral reefs and carbonate sediments in the South China Sea. Robert Gins burg, Bernard Thomassin, Colin Braithwaite and Gilbert Camoin are thanked for reviewing the manuscript. Many thanks are also due to Wei Keqin and Wang Junda for revising the English text, and to Wang Xiuya for drawing figures. REFERENCES
CHEN SHIJIAN & ZHONG JINLIANG (1990) The BriefHis tory of Is lands in the South China Sea. Hainan People's Publisher, Haikou (in Chinese). DAVIES, P.J. (1983). Reef growth. In: Pros pective on Coral Reefs. Austral. Inst. mar. Sci. Contrib., 200, pp. 69-106. Brian Clouston, Manuka. FENG WENKE, XUE WANJUN & YANG DAYUAN (1988). The Geological Environment of Late Quaternary in the Northern South China Sea. Guangdong Science and Technology Publishing House, Guangzhou (in Chinese). HoPLEY, D. (1982) The Geomorphology of the Great Barrier Reef Quaternary Development of Coral Reefs. John Wiley, New York. HoPLEY, D. (1983) Morphological classifications of shelf reefs: a critique with special reference to the Great Barrier Reef. In: Pros pective on Coral Reefs. Austral. Inst. mar. Sci. Contrib., 200, pp. 180-199. Brian Claus ton, Manuka. KAnu, J. A. (1981) Geology of Southeast Asia with special reference to the South China Basin. Energy, 11, 10771091.
248
Wang Guozhong
LADD, H.S. (1977) Types of coral reefs and their distribu tion. In: Biology and Geology of Coral Reefs, IV-2 (EJs Jones, O.A. & Endean, R.), pp. 1-19, Academic Press, London. LIU YIXUAN & MAO SHUZHEN (1989) Submarine relief of the Nansha Islands. In: Report of the Multidis ciplinary Oceanographic Expedition Team to the Nans ha Is lands , I. Science Press, Beijing (in Chinese). LU BINGQUAN, WANG GUOZHONG & QUAN SONGQI'IG (1984) The characteristics of fringing reefs of Hainan Island. Geogr. Res ., 3, 1-16 (in Chinese, with Engli>h abstract). LUDWIG, W.J., KUMAR, N. & HOUTZ, R.E. (1979) Profi es on buoy measurements in the South China Sea Basin. J. geophys . Res ., B7, 3505-3518. MAXWELL, J.D.H. (1968) Atlas of the Great Barrier Reqf Elsevier, Amsterdam. MILLIMAN, J.D. (1974) Marine Carbonates . Springer Verlag, Berlin. PuRDY, E.G. (1974) Reef configurations: cause and effeet. In: Reefs in Time and Space (Ed. Laporte, LF.), Sp
The sedimentary facies zones of the Luhuitou fringing reefs, Hainan Island. J. Tongji Vniv., 4, 70-89 (in Chinese). WANG GUOZHONG, LU BINGQUAN & QUAN SONGQING (1982) The fundamental sedimentary facies of modern coral reefs in Hainan and Xisha Islands. Oil Uas Geol., 3, 211-222 (in Chinese). WANG GUOZHONG, Lu BINGQUAN & QUAN SONGQING (1986) The sedimentary environments and characteris tics of the coral reefs of the Yongxing Island. Oceano/. Limnol. Sin., I, 36-44 (in Chinese, with English ab stract). WANG GUOZHONG, LU BINGQUAN & QUAN SONGQING (1988) Sedimentary facies model and evolution of a continental slope atoll-Huaguang Coral Atoll in the South China Sea. J. Tongji Vniv., I6, 145--158 (in Chinese, with English abstract). WANG GuozHONG, Lu BINGQUAN & QuAN SoNGQING (1990) The evolution of continental slope atolls in the South China Sea. In: Proceed ings of the Firs t Interna tional Conference on As ian Marine Geology, pp. 291306, Shanghai China Ocean Press, Beijing. WANG GUOZHONG, LU BINGQUAN & QUAN SONGQING (1994) Modern coral reefs of the South Chin1 Sea. In: Sed imentology of China (Eds Feng Zengzhao, Wang Yinghua, et a/.), pp. 541-565. Petroleum Industry Press, Beijing (in Chinese). Xu ZHONG FAN & ZHONG JINLIANG (1980) The geomorpho logic characteristics of Huangyan Island. S;ud . Mar. Sin., I, 11-16 (in Chinese, with English abstract). ZHu YUANZHI & SHA QINGAN (1992) Stratigraphy of core Nanyong-1. In: Quaternary Coral Reef Geology ofYong s hu Reef Nansha Is lands (Ed. the Multidi:>ciplinary Oceanographic Expedition Team of Chinese Academy of Sciences to The Nansha Islands) pp. 47-1:5. China Ocean Press, Beijing (in Chinese, with English abstract). ZHUANG QIQIAN, TANG ZHICAN & LI CHUNSHE'IG (1983) Ecological investigations of the Jinyin Island and Dongdao Island, Xisha Islands, Guangdong Province, China. Stud . Mar. Sin., 20, pp. 1-50 (in Chinese, with English abstract).
Spec.
Pubis
int.
Ass.
Sediment. (1998) 25, 249-258
Steady-state interstitial circulations in an idealized atoll reef and tidal transients in a deep borehole by computer simulation
A.-M. L E CL ERC*, D. BRO Ct, P h . J E AN-BA P T I S T E * and J. RAN C HER* *CEA/SACLAY/DSM/LMCE, Orme des Merisiers, Batiment 709, 91191 Gif-sur-Yvette Cedex, France; t CEA/SACLAY/DRN/DMT/SEMT/TTMF, 91191 Gif-sur-Yvette Cedex, France; and *CEA. BJIDSRIDTDS/SARI, B. P. 12, 91680 Brugeres-le-Chatel, France
ABSTRACT
Hydraulic circulations induced by thermal convection inside the carbonate platform of atolls are still not well known. As part of a project aimed at modelling these circulations using the computer code TRIO/CASTEM, developed at the CEA/SEMT, we present preliminary simulations with simplifying assumptions. This numerical study is carried out with the objective of clarifying the basic physical phenomena responsible for pore-water motion within reefs and in deep boreholes. In an entirely homogeneous carbonate body, inward and upward thermo-convective circulations are observed as a result of the oceanic thermal gradient and geothermal heating from the volcanic bedrock. However, the simulated thermal field inside the reef differs significantly from the field observations reported in the literature. Including a highly permeable zone at the base of the carbonate body, as is usually observed in atolls at the transition between carbonate and basalt, modifies the circulations. The velocities and the flow rate are greater, and the temperature gradients are in closer agreement with field data, pointing to the major role played by heterogeneities in shaping the velocity and temperature fields. The numerical study of water motion in a deep borehole raises the question of interpreting the measurements made in such deep wells. If the borehole intersects the basal highly permeable zone, the system becomes an open hydraulic circuit in which the flow is mainly independent of the flow in the carbonate body. Our calculation of the borehole response to a simple sinusoidal tidal forcing shows that considerable hydraulic circulations develop in the borehole, leading to extensive temperature and velocity oscillations. These oscillations are comparable with those reported for the P7 well in Tahiti.
INTRODUCTION
work has focused on these circulations: Rougerie & Wauthy ( 1993) propounded the endo-upwelling concept to explain the supply of nutrients to corals. Wilson et al. (1990) suggested that high Rayleigh number convection was the only hydrological model likely to dolomitize the Triassic Latemar build-up within the constrained time-frame. Samaden et al. (1985) performed the first numerical simulations of the thermo-hydraulic behaviour of an atoll. We have studied these circulations from a theoretical point of view using a computer code. To build first a generic model of behaviour, we modelled an atoll with sim plifying assumptions regarding the geometry, perme ability, rock structure and boundary conditions. As a first step, we looked at the steady-state circulation in
Atoll carbonate platforms are time evolutive struc tures built on top of a basaltic substrate. They result from coral reef growth, sediment accumulation and their secondary transformation through diagene netic processes. Many physico-chemical field data on reef interstitial waters (hydraulic, thermal, geochem ical, geological, etc.) show a link between the pore water inside the carbonate and the ocean water. For instance, the temperature profiles in the reef and under the lagoon tend to follow the temperature pro file of the ocean: the temperature decreases with depth whereas it usually increases in geological sub strates. This suggests that water of oceanic origin circulates within the rock, influencing its tempera ture field and its chemical composition. Previous
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
249
250
A. -M Leclerc et a!.
the case of a homogeneous platform. In reality, the rock is highly anisotropic and heterogeneous, be cause of stratification of the substrates and occur rence of joints, fractures or karsts. To take these heterogeneities into account, we studied, as a prior ity, the role of a transition zone, i.e. a zone of higher permeability located at the base of the carbonate platform. Because of its special position, this layer has a strong influence on the hydraulic circulations. Finally, we focused on a more specific problem, i.e. the hydraulic behaviour of a borehole and its re sponse to transient stresses which include tidal ef fects.
EQUATIONS The groundwater flow in the carbonate platform is governed by four classic porous media equations: 1 hydraulic equation, Darcy's law, giving the veloc ity: U
=
K
-
(
)
p Pe V -+-Vz ; Pog Po
2
the condition of incompressibility with Bouss inesq's approximation (i.e. we neglect the compress ibility of water except in the gravity term): V U ·
3
=
O;
a thermal equation with three terms: Pm Cpm
��
Af:.T
=
'-------v--
transient
'-----v-----'
-
Pe Cp,U VT; ·
'-------v---
diffusion convection
4 an equation of state: sea-water density follows a first-order law in terms of temperature given by
P= Po(l - fJT).
The notation is as follows: U is Darcy's velocity or filtration velocity (this is a macroscopic velocity that allows flow rate calcula tions, in contrast to the microscopic velocity of the fluid particles inside the pores; in the Darcy's law equation, the porosity of the medium intervenes indirectly through the permeability K and it is not a variable of the problem); p is the pressure of the interstitial fluid; g is the gravitational acceleration; p0 is the water density at the reference temperature (constant); Pe is the water density (a function of T);
K kp0g!f.1 is the permeability used in hydrogeology, expressed in m s-1; the intrinsic permeability of the medium k, expressed in m2, depends only on the solid matrix whereas K takes the whole medium into account, not only the matrix but also the fluid, with viscosity Jl and density p0, throughout the matrix; T is the temperature·of the medium; A is the thermal conductivity; this is an equivalent macroscopic conductivity of the medium (matrix + interstitial fluid); Pm CPm is the heat capacity of the medium; Peep, is the heat capacity of water; f3 is the coefficient of expansion of water; =
( ( rx; ;�; ;�)} ) ( ( ) --:2
V is gradient Vf
=
. . V· Is divergence
V
a
aax
+
aay
+
aaz
;
ax ay az a2J a2J a2 t:. is the laplacian t:.f - + + -2 • ax 2 ay az ·
=-
-
-
=
In Polynesian atolls, ocean salinity varies from c. 36o/oo at the surface to 34.5o/oo at 800-m depth, and the temperature varies from c. 27 o C to 6 o C (Kessler & Monbet, 1984). Using the international one atmosphere equation of state of sea water (Millero & Poisson, 1981), we evaluate that neglecting salinity variations in our simulations leads to c. 30o/o error in the density gradient between the surface and the base of the system. Thus the effect of salinity on water density is significant, and it will be included at a later stage. However, as a first approximation, we consid ered only thermally induced density gradients, which are the basic driving force of the system. The equation of state introduces the thermo hydraulic coupling by making the temperature ap pear in the hydraulic equation. Hence Darcy's law becomes:
[v _.E.._+
]
(1- /JT)Vz . Pog When written in dimensionless form, these equa tions show that the coupling depends only on one parameter, namely, the Rayleigh number Ra: U
=
-K
V*.U*
=
0
U*= -V*h* + T*V*z* Pm Cpm aT * -Pe Cp, at* __
Ra=
=
t:.*T* - RaU*.V*T*
KfJToLPe Cp,
A
,
Computer simulation of interstitial water circulation
where the asterisk refers to dimensionless quanti ties, L is a reference length, characterizing the size of the system T0 is a reference temperature and h* is the dimensionless hydraulic head. The Rayleigh number can be defined in different ways, depending on the type of problem that has to be solved. For example, to study convection in a rectangular box with isothermal bottom and top boundaries, T0 is usually ( T,op - Tbottom). If the bottom boundary condition is a thermal flux ¢, the Rayleigh number can be defined as: Ra
=
KfJ¢L2peCpJA2.
Our simplified atoll can be characterized by two Rayleigh numbers corresponding to the two forces driving the flow: one for the temperature gradient in the ocean: Ra
=
KfJ(Tsurface- T bottom)LpeCpJ).
and one for the geothermal flux ¢: Ra
=
KfJ¢L2peCpJA2
The Rayleigh number is dimensionless and charac terizes the ratio between convection and diffusion: for a low Rayleigh number, diffusion is predomi nant and the coupling is low; for a high Rayleigh number, convection prevails and the coupling is strong. In some cases, cells can develop beyond critical values of this parameter. These critical values, when they exist, cannot always be expressed analytically but nevertheless numerical approxima tions can be found. NUMERICAL ANALYSIS OF NATURAL CONVECTION
The computer code used to solve these equations is TRIO/CASTEM. This finite element simulation was developed by the Department of Mechanics and Technology (DMT) at the CEA (Commissariat a l'Energie Atomique) for a wide range of applica tions ranging from nuclear reactor hydraulics to groundwater studies. As far as porous media are concerned, it is used extensively to study the dis posal of radioactive waste. As an illustration of the capability of the simulation to treat natural convec tion, we first simulated simple convection boxes for which analytical solutions can be obtained rela tively easily. In all the simulations presented in this paper, the equation system was solved by a finite
251
element method with an upwind scheme for the discretization of the convective term in the heat equation. We considered a rectangular homogeneous do main, governed by the above equations. As bound ary conditions, we assign the following: (i) tempera tures at the bottom and top boundaries; in dimensionless calculations, the bottom temperature is unity and the top temperature is zero; and (ii) hydraulic fluxes are set to zero along all boundaries. Depending on the Rayleigh number and the initial conditions, different configurations can be obtained. Each one can only exist beyond a specific critical Rayleigh number. The critical Rayleigh number can be expressed analytically as a function of the convection cell dimensions:
where H is the height and M the width of the cell. Figure I shows some of the possible configurations; for instance, the smallest critical Rayleigh number (4n2) corresponds to a square convection cell (H M). Using a 50 x 50 mesh for the domain, the relative error between the analytical and numerical values of the critical Rayleigh number is about 2%, thus confirming the ability of the code to model natural convection. =
MODELLING
The carbonate platform lies upon a basaltic bedrock (Fig. 2). Both are permeable media but the basalt permeability is between three and four orders of magnitude lower than that of the carbonate (Guille et al. , 1993; Samaden et al. , 1985). In the study by Samaden et al. (1985), both the carbonate and the basaltic edifices were modelled but the purpose was different: these researchers wanted to evaluate the capability of the volcanic bedrock to contain radio active wastes from nuclear tests and to act as a barrier protecting the environment. They pointed out that, in fact, water flow occurred essentially inside the very permeable carbonate layer. Hence, although there are hydraulic circulations inside the basaltic bedrock, these have little effect on the upper circulations in the carbonate medium com pared with those induced by the ocean thermal gradient and geothermal flux. The basalt can be considered impervious. Besides, the thermal flux between the basalt and the carbonate medium is
A.-M Leclerc
252
et al.
Fig.
Ra Atmosphere
c
=
1. Different convection modes in a rectangular box. The square convection cell requires the lowest Rayleigh number (4n2). Rectangular cells such that H 2M can be obtained beyond a higher critical value ( n2 ).
25 2 7r 4
=
-
/
¥
Lagoon
cean
�0 Transition zone
regular because the heat transfer in the basalt is diffusive. This allows us to easily set the thermal boundary condition at the interface between the two domains. Thus it is not necessary to include the basaltic domain in our thermo-hydraulic study. In the following, only the carbonate platform is con sidered. Our mesh, represented in Fig. 3, contains a large low-permeability carbonate body and a transition zone at its base with a varying permeability ratio. The transition zone and the borehole will be con sidered as homogeneous and isotropic porous me dia. We should rigorously solve the Navier-Stokes equations inside the borehole and also presumably inside the transition zone in the case of a karstic structure with large interconnected voids as ob served in Tahiti (R. Fichez, personal communica tion). This would require coupling the solutions of two different problems in two distinct domains. If
Fig. 2. Typical geological cross-section of an atoll.
the flow is laminar, however, the homogenization process is simple, leading to equations similar to Darcy's law, and the Navier-Stokes domains can be treated, as a first approximation, as equivalent high-permeability porous media. The Reynolds number determines whether a flow is laminar or turbulent. In a pipe, it is expressed as: Re= UDp//1,
where U is the longitudinal velocity, D the diameter, p the density of the circulating fluid and 11 its viscos ity. The flow will be purely laminar if Re � 2000 (Roy, 1988, p. 149). The critical zone covers the range 2000 < Re < 4000, and the transition zone has the range 4000 < Re < 6000. The flow is turbulent when Re > 6000. With a borehole diameter of 10 cm, p � 1000 kg m-3 and11 � 10-3 kg m-1 s-1, we obtain Re < 2000 for velocities < 2 X 1o-2 m s-1• Both recorded and calculated velocities are lower
Fig. 3. Two-dimensional mesh of
the domain (carbonate + transition zone).
253
Computer simulation of interstitial water circulation
than this value. In the transition zone, typical calcu lated velocities are of the order of 4 em day-1 � 4.6 X I Q-7 m S-l, Which yields: Re � (4.6 X 10-7 X 50 X 1000)110-3 � 25.
This indicates a perfectly laminar state. Thus the flow can be approximated everywhere in the system by Darcy's law with equivalent permeabilities. Ho mogenizing a laminar flow in a medium containing parallel cracks (De Marsily, 1981) leads to:
K
=
wilpgl12f.1,
where w is the porosity, e is the thickness of the crack, f.1 is the viscosity of the circulating fluid and p is its density. This implies
K � ( I x 502 X 1000 x 10)112 x I Q-3 �2x 109m s-1 for the transition zone and
K �(I X (10 X I Q-2)2 X 1000 x 10)112 X I Q-3 � 1.25 x 104m s-1 for the borehole, whereas real permeabilities vary between I Q-9m s-1 and 10-2 m s-1• In the following, we shall use permeabilities up to I 04m s-1• Higher permeabilities are unnecessary because the system reaches asymptotic behaviour. We assign the following thermal and hydraulic boundary conditions: I Thermal boundary conditions: fixed temperature at the top free surface (Tsurface = 25 ·q and at the interface between the ocean and the carbonate according to the vertical temperature profile in the tropical ocean. At the basalt-carbonate interface, the volcanic basement induces a geothermal flux of 0.1 W m-2• 2 Hydraulic boundary conditions: at the free sur-
face, we assign the atmospheric pressure. In the ocean, the hydrostatic pressure is determined by the temperature profile and the equation of state. At the interface between the basalt and the carbonate, the hydraulic fluxes are set to zero, implying tangential velocities only. At the centre of the atoll, the flux is also set to zero because of the axial symmetry of the system. This hypothesis, already proposed by Sa maden et a!. (1985), implies that there is no convec tion cell across the symmetry axis of the atoll. In the same way, studying the system in two dimensions requires the assumption that no three-dimensional circulation occurs. The values of the variables of the equations are: Pm CPm = P eep,= 4.19 x 106 W s m-3 ·c-1; ). = 5.028 w m-1 ·c-1; and p 1.5 X 10-4 ·c-1. The carbonate platform is 200 m thick at the centre and 800 m thick at the periphery, with a radius of 7600 m. The standard permeability (of the low permeability body) used in the study is K = I o-4m s-1• These characteristics correspond to the Mururoa morphology. =
STUDY OF A HOMOGENEOUS ATOLL PLATFORM
Previous work performed in our laboratory with the same computer code (Leclerc et a!., 1995) has highlighted some features of a homogeneous car bonate body, i.e. without the transition zone. Up ward hydraulic circulations from the ocean towards the centre of the atoll were shown to develop, as illustrated in Fig. 4a. At the periphery of the atoll, the isotherms are perturbed and the temperature gradient is negative (Fig. 4b). On the other hand, in the centre of the atoll, the isotherms are essentially
A.
6.0
c
9. 9
B
�
E·
f.! w: �: �j:
il
i:
Fig. 4. Velocity (a) and temperature (b) fields in a homogeneous platform.
� ft m
7
9
12. 14.
16.
18.
20.
22.
23 . 25.
27.
29
31.
A.-M Lec lerc et
254
al. A
6.1
B
8. 2
c
10.
()_
13.
t:::
;: [: w:: ±:: J]
*�� £::
� I li
15. 17. 19. 21. 23 26 26 30.
32
34
Fig. 5. Purely diffusive temperature field in a homogeneous platform. horizontal with a positive vertical temperature gra dient. This thermal distribution is the force driving the flow: the centre of the platform is warmed by the thermal boundary condition at the free surface and by the geothermal flux below, whereas the pore water is cooled by the ocean on the side. As it is heavier, the cold water tends to slip underneath the warm water, leading to an inward groundwater flow. With the numerical values given above, the Ray leigh number defined by the thermal gradient in the ocean is close to 200. The domain with a negative temperature gradient widens continuously as the Rayleigh number increases (there is no critical value below which no flow occurs). Indeed, for low Ray leigh numbers, the temperature field drives the flow but as diffusion, rather than convection, is predom inant, the velocity field does not reshape the tem perature field, which remains diffusive (Fig. 5). For higher Rayleigh numbers, the flow is able to signif icantly modify the temperature field and the veloc ities are greater. (a)
The geothermal flux contributes to pore-water motion but the main cause of these circulations is the vertical temperature structure of the ocean. Water flow occurs even when the geothermal flux is suppressed (Fig. 6) and the velocity remains the same order of magnitude. This work confirms previous studies (Samaden et a!. , 1985) showing that hydraulic circulations do occur within the carbonate platform. It also shows, however, that the hypothesis of a homogeneous body is inadequate to obtain the negative tempera ture gradient observed in boreholes, not only at the periphery but also at the centre of the atoll (Guille et a!. , 1993). INFLUENCE OF
THE TRANSITION ZONE
One of the major heterogeneities influencing the: exchange between the interstitial water and the: ocean is the occurrence of the transition zone,,
L�-----====-2��
(bliL_
�
�
___ ____
Fig. 6. Velocity (a) and temperature (b) fields in a homogeneous platform with zero geothermal flux.
A
5.7
8
7
c
8.6
D
10.
i�
13.
!
J[ i.l ii �· i m i! I
1
11.
14 16. 17. 19. 20. 21. "23
24.
255
Computer simulation of interstitial water circulation
A
6 1
�· t:
9. 2
c
(blf.�
Fig. 7. Velocity (a) and temperature (b) fields with K 10°m s-' in the transition zone.
4.5
B
J.[ �f if ¥ ;. i � I I
7.6 11.
12.
14
15.
17
19. 20. 22.
23. 25.
=
which is modelled as a 50-m-thick highly permeable layer at the base of the carbonate body. The Ray leigh number, as defined above, cannot be used here; we should look for a new expression that takes the heterogeneity of the new system into account. The high permeability of this zone allows the cold oceanic water to penetrate easily deep into the platform (Fig. 7a). The model shows that when the permeability of the transition zone increases, the isotherms are perturbed much further inwards, as can be seen in Fig. 7b. With an equivalent perme ability greater than w-1 m s-1, the isotherms are horizontal throughout the carbonate body, which means that the temperature gradient is negative not only at the periphery of the atoll, but also in the centre (under the lagoon), a feature that is clearly
observed in field temperature measurements (Guille et a!., 1993). Buddemeier & Holladay (1977) had already noticed that the occurrence of more permeable layers influenced the hydrology of Enewetak Atoll. These layers transmitted tidal sig nals horizontally, so that the transient response in boreholes was almost independent of their distance from shore. The velocity in the transition zone increases with the permeability, as long as the transmissivity (product of the permeability and the width of the flow zone) of the transition zone is smaller than that of the whole body. Beyond this threshold, the flow throughout the system is driven by the low permeability body and the velocity in the transition zone tends towards an asymptotic value (Fig. 8).
4e-07 3.5e-07 3e-07
� 2.5e-07 ;g�
>
2e-07 1.5e-07 1e-07 Se-08
Fig. 8. Velocity in the transition
zone in relation to the permeability.
4
log (Ktrans/Kbody)
. 7
A.-M. Leclerc et al.
256
The average Darcy velocity in the transition zone obtained in the present calculations is around 4 em day-1 and its value in the overlying platform is of the order of 0.2 mm day-1• These simulations show that the transition zone is indeed a key feature which qualitatively and quantitatively modifies the hydraulics of the sys tem, and leads to more realistic results. Conse quently, it will be taken into account in all the following simulations.
p �--� I
=0
2H
-T I
BOREHOLE RESPONSE TO TIDAL FORCING
All the physico-chemical data for atolls come from measurements in boreholes. However, these mea surements may not always be representative of the hydraulics of the atoll, as the flow circuit is modified by the presence of the borehole itself. As an exam ple, the measurements in the P7 borehole in Tahiti (F. Rougerie, personal communication) show large temperature oscillations which, in view of their time dependence, appear to be related to the tide. To study this specific behaviour, as a first approx imation, the load produced by the tidal range and the thermal internal waves on the flank of the structure was simulated by a sinusoidal overpres sure applied on the atoll flanks. The signal ampli tude is ± 15 em, a value representative of the real tidal signal in Tahiti. The deep borehole and the transition zone constitute an open hydraulic circuit between the ocean and the atmosphere through the carbonate platform, in which the water can flow very easily. Therefore, on the time-scale of the tidal oscillations, this circuit is a one-dimensional sys tem. The long-term effect of the net perturbation inside the carbonate body (typical time frame c. 1010 s), which could arise from the tidal forcing, would require a specific treatment and will not be addressed here. The tide induces velocity and temperature oscil lations inside the well with amplitudes and phase lags (in relation to the tidal pressure) depending on the permeability of the highly permeable zones. As a first step, these zones (borehole+ transition zone) were simulated by a vertical porous tube consisting of two layers at different temperatures (Fig. 9), the lower temperature -T representing the oceanic 1 water circulating in the warmer borehole at (tem perature + T ). The motion is governed by a differ 1 ential equation:
1'1P
=
1'1P0
sin (w t)
Fig. 9. Schematic representation of a porous tube.
dx
dt
Kf3T1 KMo . sm wt � x. 2H -
=
The solution is a sine function that can be obtained analytically: X= X0sin (wt+¢) with
Hw K-ao 0. tan¢ = Kf3T1 -
The amplitude X0 and the phase ¢ depend on the permeability: when the permeability increases and tends towards infinity, the amplitude of these oscil lations increases and tends towards an asymptotic finite value; the phase lag tends to zero. Thus for high permeability ratios, the motion is in phase with the tidal pressure and the velocity presents a n/2 phase lag. For low ratios, the opposite is true: the velocity is in phase and the motion is out of phase. We checked the validity of this sequence by running the two-dimensional atoll model (simulat ing the tide) for permeability ratios ranging from l to 109. Figure 10 represents the numerical average velocity in the borehole during one period (12 h}. For low ratios, the oscillations are very small and in phase with the tidal signal. This characterizes the oscillations inside the carbonate body. When· the permeability ratio increases, the amplitude and the phase lag increase. The system tends towards an
257
Computer simulation of interstitial water circulation xl.E-2 Velocity (m/s) 2.50 ,----.-----.---�--,
Fig. 10. Evolution of the average 2.15
2.20
2.25
2.30
2.35
2.40
2.45
asymptotic state, with a lag of around n/2. In this case, at flood tide, cold oceanic water penetrates the transition zone and then the borehole, cooling it down. These velocities are great enough to bend the isotherms upwards in the well (Fig. 11). At ebb tide, the opposite is true: the water flows downwards inside the borehole, bending the iso thermic lines downwards.
2.50
2.60 2.55 xl. E5
velocity (from numerical calculations) in the borehole for various permeability ratios.
The temperature oscillations obtained with high permeability ratios have amplitudes similar to the oscillations observed in the P7 borehole in Tahiti (Fig. 12). For instance, for a 107 permeability ratio,
24.5
(a)
9
24
� 23.5 ::J
'§
23
� 22.5
E
Ql 1-
22 21.5 '-----22/07/93 23/07/93 24/07/93 25/07/93 26/07/93 12:00 12:00 12:00 12:00 12:00 Date/hour
(b)
21.5 ,-----...,
G
21
�
20.5
::J
�
20
�
19.5
�
E
�
19 18.5 18�-�-�--�-�-��� 0 0.5 1.0 1.5 2.0 2.5 3.0 x1.E5 Time(s)
Fig.
11. Temperature field at high tide (zoom on the
borehole).
Fig. 12. Temperature oscillations in the borehole (a) ORSTOM P7 data in Tahiti (F. Rougerie, personal communication); (b) numerical calculation.
258
A.-M. Leclerc et al.
i.e. K = 103 m s-1 in the permeable zones, we obtain thermal oscillations of 3 oc. CONCLUSION
Several conclusions can be drawn from this prelim inary study of a simplified atoll platform: 1 As already shown by Samaden et a!. (1985}, atolls are characterized by upward convective circula tions. Oceanic water penetrates the atoll through its flanks. These circulations are mostly due to the temperature gradient in the ocean. The geothermaL flux, which intuitively would appear to be the main driving force of the thermoconvective circulation, is in fact of lesser importance. 2 The system is very sensitive to heterogeneities. Taking a basal highly permeable transition zone into account modifies the results both qualitatively and quantitatively and leads to a thermal field which is in closer agreement with field thermal data (Guille et a!., 1993). 3 Measurements in deep wells may be biased be cause of perturbation of the system by the borehole itself, which reacts as a specific hydraulic system. The long-term goal of our work is to simulate the migration of various chemical elements and to quantify the material transfer between the ocean and the carbonate structure. It will require a de tailed description of both terrain heterogeneity and external forcings. The main motivations are twofold: 1 The first is to characterize the role of the convec tive circulations with respect to the nutrient balance of coral reefs. The circulations could supply nitrates and phosphates from the semi-deep layers of the ocean up to the atoll surface through the porous platform, as proposed by the endo-upwelling con-
cept (Rougerie & Wauthy, 1993). The net efficiency of this supply remains to be determined. 2 The second is related to the use of carbonate as a palaeo-indicator: hydraulic flows not only influence the formation of the rocks but also their later transformation (dolomitization, etc.). Modelling these circulations should allow progress in under standing these transformations and their influence over the preservation of palaeo-climatic signals.
REFERENCES BUDDEMEIER, R.W. & HOLLADAY, G. (1977) Atoll hydrol
ogy: island ground-water characteristics and their rela tionship to diagenesis. In: Proceedings of the 3rd International Coral Reef Symposium, Miami, pp. 167173. DE MARSILY, G. (1981) Hydrogeologie quantitative. Mas
son, Paris.
GuiLLE, G., GounERE, G. & SoRNEIN, J. (1993) Les Atolls de Mururoa et de Fangatauja, Vol. I. Masson, Paris. KESSLER, M. & MONBET, Y. (1984) Centrale ETM Avant projet Tahiti. Resultats des etudes de site, Vol. I, pp. 439 and 454. Technical report IFREMER/DIT. LECLERC, A.M., BROC, D., JEAN-BAPTISTE, P. & RANCHER, J. (1995) Modelisation des circulations hydrauliques dans un atoll. Technical Report 5712 CEA, Soclay. MILLERO, F.J. & POISSON, A. (1981) International one-· atmosphere equation of state of seawater. Deep-Sea Res., 28A(6), 625-629. ROUGERIE, F. & WAUTHY, B. (1993) The endo-upwelling
concept: from geothermal convection to reef construc tion. Coral Reefs, 12, 19-30. RoY, D.N. (1988) Applied Fluid Mechanics. Ellis Hor wood, Chichester. SAMADEN, G., DALLOT, P. & ROCHE, R. (1985) Atoll d'Eniwetok. Systeme geothermique insulaire a l'etat nature!. La Houille Blanche, 2, 143-151. WILSON, E.N., HARDIE, L.A. & PHILLIPS, O.M. (1990) Dolomitization front geometry, fluid flow patterns, and the origin of massive dolomite: the Triassic Latemar buildup, Northern Italy. Am. J Sci., 290, 741-796.
Active Margins
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
Spec. Pubis int. Ass. Sediment. (1998) 25, 261-277
Environmental and tectonic influence on growth and internal structure of a fringing reef at Tasmaloum (SW Espiritu Santo, New Hebrides island arc, SW Pacific) G. C A B I O C H*, F. W. TAY L O Rt, J. Rf:CY*, R. LAWRENCE E DWA R DSt, S. C. G RA Y§, G. FAUREII, G. S. B U R R� and T. C O R REGE**
*ORSTOM, UMR GEOSCIENCES AZUR, B.P. A5, Naimea Cedex, New Caledonia; tinstitute of Geophysics, University ofTexas-Austin, 8701 North Mopac EXPY, Austin, TX 78759, USA; tDepartment of Geology and Geophysics, Newton Horace Winchell School of Earth Sciences, 421 Space Science Building, 100 Union Street, SE Minneapolis, MN 55455, USA; §Marine and Environmental Studies, University of San Diego, Alcala Park, San Diego, CA 92110, USA; IILaboratoire Hydrobiologie Marine et Continentale, USTL, Place E. Bataillon, 34095 Montpellier Cedex 5, France; �University of Arizona, Department ofPhysics, AMS Facility, Tucson, AZ 85721, USA; and **ORSTOM, LFS, 32 avenue Varagnat, 93143 Bondy Cedex, France
ABSTRACT
Subduction of the Australian Plate has caused rapid uplift of the central New Hebrides island arc (15"S, SW Pacific). The d'Entrecasteaux ridge system, a prominent bathymetric feature on the downgoing plate, is underthrusting the central part of the New Hebrides arc. The coastlines of most islands are characterized by emerged Holocene coral reef terraces. A maximum uplift rate of 6 mm yr-1 occurs along the south-west coast of Espiritu Santo, near the plate boundary. To investigate the Late Quaternary neotectonic and environmental evolution of the uplifted fringing reefs, we drilled the emerged Holocene reef at Tasmaloum (SW Espiritu Santo) to depths as great as 40-45 m. Coral samples from various levels were dated by 23<>-fh and 14C, as described elsewhere, and the internal structure of the reefs was studied. Preliminary palaeoecological and sedimentological data indicate the following. First, the coral reef colonized a substrate and began to grow by 24 ka on weakly indurated calcareous sand beds, which probably formed during the Late Pleistocene and are possibly as old as 30 ka or more. These sand levels could represent the deep fore-reef area of an older reef, at present behind the uplifted Holocene terraces. Second, the biofacies and coral ages from the reef sequence, which is continuous from the last glacial maximum (LGM), provide a view of the internal structure: between 24 and 12-10 ka, coral levels, composed mainly of Acropora gr. hyacinthus and gr. cytherea (accompanied by a few Galaxea gr. fascicularis), constitute medium- to high-energy assemblages, reflecting relatively deeper and more protected environments. However, at a few levels, acroporid build-ups of Acropora gr. danailrobusta indicate high-energy environment alternation. Between 12-10 ka and the present, assemblages of acroporids (Acropora gr. danailrobusta), scarce poritids, numerous encrusting coralline algae, verrnetid gastropods and encrusting foraminiferids indicate high-energy environments, probably corresponding to the upper part of the exposed reef slopes. This biofacies succession indicates changes in wave energy, related to fluctuations in local bathymetry controlled by the net effect of variable rates of sea-level rise, reef growth and tectonic uplift. After 6 ka, the replacement of coral, coralline algal and stromatolite assemblages by coral and coralline algal frameworks implies the establishment of a new hydrological and oceanographic regime. The biofacies and age structure of the reef show that the uplift rate has varied since 24 ka.
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
261
262
G. Cabioch et al. INTRODUCTION
Located in the tropical zone of the SW Pacific Ocean, the Vanuatu (New Hebrides) archipelago comprises the narrow volcanic island chain which stretches between 13•s (Torres Group in the north) and 200S (Anatom in the South) and between 166 ·E and 172 ·E (Fig. 1). These islands are a part of the New Hebrides Island Arc (NHIA) at the convergent boundary of the Australian and Pacific Plates. This arc shows neotectonic features related to the subduction of ridges and seamounts. In particular, in its central part, there is no physio graphical trench along the plate boundary west of Espiritu Santo, where the d'Entrecasteaux zone
intersects the arc (Fig. 1), emerged islands in the forearc (Espiritu Santo and Malekula) occur anom alously close to the plate boundary and the backarc islands of Pentecost and Maewo have uplifted rapidly (Taylor eta!., 1987). The maximum Late Quaternary uplift rates of Espiritu Santo (3-4 to 6 mm yr-1 during the Ho locene in its SW part) and North Malekula (maxi mum of 3.4 mm yr-1 since 6 ka) coincide with the western topographical axis of these islands, which is subparallel to the trend of the trench (Jouannic et a!., 1980; Taylor et a!., 1980, 1985). Previous studies of the coral reefs in Vanuatu
(a)
(b) Torres Islands
I ', .
..
Espirit u Santo
'-� Banks Islands ,.
·
\..... Aob ;\ Maewo � \ Pentecost
15�
·� 4:. Ambrym
Malekula
t.... Epi-.
�
Vate
Erroman go
'
Tanna \ .
2o•s
-'--
Anatom (c)
Q ueiros Peninsula
Fig. 1. (a) Geodynamic setting of the SW Pacific, showing the location of Espiritu Santo, Vanuatu, in the New Hebrides island arc; (b) the main islands of the Vanuatu archipelago; (c) location of the Tasmaloum site (square) and the two main geomorphological divisions of Espiritu Santo: I, the eastern plateaux (coral limestone terraces and plateaux); 2, the western mountains (volcanic and sedimentary rocks).
263
Influences on fringing reefgrowth and structure have examined neotectonics, sedimentological fea tures, palaeoenvironments and palaeoclimate. Guilcher ( 1974) studied the structure of reef types in Vanuatu, and identified fringing and open-sea reefs as primary types. Veron ( 1990) recognized about 296 species (belonging to 62 genera) of hermatypic corals in the archipelago. Done & Navin ( 1990) studied habitats of the shallow water communities of many coral reefs. Our studies have focused on neotectonics and palaeoclimate (Beck et al., 1992; Gray et al., 1993; Recy et al., 1993; Taylor et al., 1993). In the Tasmaloum area (Fig. 1), a high uplift rate ( 6 mm yr-1) was recorded along the coast at least for the last 6 ka (Taylor et a/., 1980; Gilpin, 1982; Bloom & Yonekura, 1985) and perhaps even higher for the last 12 ka (Taylor, 1992). Although this rate is high, we may assume that reef growth was continuous during a large part of Quaternary times, in the same way as on Huon Peninsula (Chappell & Polach, 1991; Edwards et a/., 1993). If so, we can obtain corals representing nearly all of the last 20 ka of reef growth by drilling relatively shallow cores. We calculated that drilling to depths less than 20-30 m would be sufficient. For example, if the uplift rate has been 6 mm yr-1, then the 20-ka shoreline has been raised 120 m from its original level. Combining the uplift rate and total postglacial sea-level rise (e.g. Fairbanks, 1989), the net trans gression at SW Espiritu Santo was hypothesized to
have been only about 10 m. Based on this reason ing, we began a coring programme on the emerged part of the Tasmaloum fringing reef in 1990. We drilled several sites in 1990, 1992 and 1994 to sample a great number of pristine corals represent ing the past 20 ka. Before our reef drilling programme, the growth history of Vanuatu fringing reefs was unknown. However, the internal structure of reef sites is documented in the SW Pacific, especially in Austra lia (Davies & Marshall, 1979; Hopley, 1982; Mar shall & Davies, 1982; Collins et al., 1993) and in New Caledonia (Cabioch et al., 1995) and serves as a basis for comparison with the Vaimatu reefs. The objective of this paper is to present the first palaeoecological and sedimentological results on the internal structure of a particular fringing reef (Tasmaloum, SW Espiritu Santo, Fig. 1), as,well as preliminary geomorphological data concerning reef settlement and growth patterns. In addition, we interpret the data in the context of the environmen tal and tectonic features of the central NHIA.
MATERIALS AND METHODS
Fourteen cores, obtained by vertical (9A to -G, -I, -M & -N) and oblique (9H, -J, -K & -L) drilling (Fig. 2), were recovered using a Jacro and, later, a Sedidrill coring system. The total depth to which we
North
South
Drill holes (Recent upliftedcoral reef )
Reef Terrace 2
Reef Terrace 4 Reef Terrace 3 (+36m) + ( 55m) (6,600 yrat + 33/37 m after Gilpin, 1982and Bloom & Yonekura , 1985) marine notch
"'.
-�-
?
Fig. 2. Schematic cross-section of the uplifted Tasmaloum fringing reef, showing the main reef terraces and the two
units recognized by drilling: (I) the Late Pleistocene (24 ka) to Holocene upper unit (coral debris and build-ups); (2) the Late Pleistocene (from, at least, 30 ka to 24 ka) lower unit (bioclastic sands). The deepest cores (vertical or inclined) are, at each drilling site, 9C, 9E, 9F, 9H and 9N. The inclined cores are 9H, 9J, 9K and 9L. The dates have been obtained through TIMS (U/Th dates) and AMS e4C dates calibrated in calendar years, in italics).
264
G. Cabioch et a!.
drilled ranged from 15.25 to 41.35 m (Fig. 2), and the core diameters were 4.6-3.5 em. Recovery ranged from 10 to 80%. Sections of poor recovery are assumed to be due to cavities or caverns, or may represent the occurrence of intervals of fine sand or unconsolidated sediments (Figs 2 and 3). During drilling, core sections were recovered every 1.50 m, thus permitting a depth accuracy of ± 0.5 m. The depths (m) are given by reference (a indicates above; b indicates below) to the precisely levelled determinations of the highest living Pocillopora and Goniastrea retiformis level, which corresponds to the highest level of survival (HLS) of corals (e.g. Taylor et a!., 1987). Core studies include geochemical, sedimentologi cal and mineralogical analysis focusing on evaluat ing diagenesis, and biological aspects (fauna and flora components and associations). Samples were examined using standard petrographical binocular microscopy, scanning electron microscopy (SEM), X-ray diffraction and specific staining techniques. Semi-quantitative estimates were made from thin sections to determine the composition of the biolog ical assemblages associated with the coral frame work. Coral samples for isotopic dating were most commonly of Porites sp. or acroporids. We selected only corals which appeared pristine, containing more than 98% aragonite and showing minimal evidence of aragonite cement or dissolution. 230Th/ 234U dates were obtained by thermal ionization mass spectrometry (TIMS) in the Minnesota labo ratory, using the method developed by Edwards et a!. ( 198 7). The same coral samples were also dated by accelerator mass spectrometry (AMS) 14C (AMS facility, Tucson), to provide a calibration between the two methods (Gray et a!. , 1993). Dates only obtained by 14C were converted to calendar years, using the calibration methods (Bard et a!., 1990, 1993; Edwards et a!. , 1993; Stuiver & Reimer, 1993).
REEF GEOMORPHOLOGY AT TASMALOUM
The coral reefs surrounding Espiritu Santo are generally fringing reefs. They are typically narrow, with reef fronts very close to the shores (Guilcher, 1974). They are not protected by barrier reefs or any other natural barrier, except near SE Espiritu Santo where small islands protect a few sites
(Fig. 1 ). As a result, reefs are occasionally subjected to large waves, especially along the open SW coast where the Tasmaloum area is located (Fig. I). The shore morphology immediately around the drilling area exhibits both narrow and broad terraces, in which gentle to steep slopes are observed (Fig. 2). On the uplifted reef, a succession of minor scarps are distinguishable. Several broad terraces on the emerged Holocene reef have been formed successively between the present sea-level and the Holocene terrace level at about 36 m. The 36-m terrace level terminates against a scarp which rises to an upper terrace of older coral limestone at +55 m (Fig. 2). The +36-m reef terrace is the highest and broadest Holocene terrace, dated at 6.6 ka (Gilpin, 1982). A deep notch in the palaeo-sea cliff at the back of the 36-rn terrace is observed. Except against this cliff, notches are generally missing from the scarps of the Ho locene reef. The present uplifting reef flat near the drilling sites shows typical reef zonation: an outer reef zone occu pied by small coral colonies (generally acroporid forms) with encrustations of coralline algae; a me dian reef zone with many branching corals and somewhat indurated fine bioclastic sediments; and a back-reef zone having few large coral colonies and weakly indurated bioclastic sediments. Behind, the supratidal zone consists of an ancient and very gently sloping uplifted reef surface. This zonation is similar to that typically encountered in the south-west Pa cific, i.e. in modern Vanuatu coral reefs (Done & Navin, 1990}, in many neighbouring New Cale donian (Cabioch et at., 1995) and Solomon (Morton & Challis, 1969) fringing reefs, and in the Australian Great Barrier Reef (Partain & Hopley, 1989). The modern fringing reef is narrow, and living corals are restricted to the reef front, the seaward part of the reef flat and the reef slope. Robust, wave resistant organisms, well adapted to strong water movement, and coralline algal veneers are fre quently present in this high-energy environment. Bioeroders are common and constitute good mark ers of sea-level, especially the Echinometra level near mean sea-level. The modern fore-reef slope is steep.
TASMALOUM INTERNAL REEF STRUCTURE
Cores drilled at Tasmaloum penetrated the Ho locene, Late Pleistocene and LGM reef sequence
265
Influences on fringing reefgrowth and structure back to 24 ka (Fig. 2), and recovered the underlying reef substrate. Coral reefs that grew during the postglacial transgression colonized a substrate of calcareous sands, gravel and pebbles (Fig. 2).
cially by Amphistegina spp., accompanied by A. les soni and A. radiata (the most abundant), Alveolinella quoyii or Borelis schlumbergeri, Calcarina sp. at -35 m bHLS and-37.50 m bHLS. Green algae, Hali meda sp., are particularly abundant at around -36.50 m bHLS.
Reef substratum
The substrate upon which the Tasmaloum fringing reef is established consists of sand layers, found only in the deepest cores, i.e. 9E, 9F, 9G and 9H (Fig. 2). This sand substrate (older than 24 ka and probably as old as, at least, 30 ka) consists of fragments, mainly of molluscs (bivalves and gastropods), Hal imeda and benthic Foraminifera, represented espe-
Reef biofacies s�quences
Corals recovered in the cores are dominated by acroporids and poritids (Figs 2 & 3). Coralline algae belong usually to Porolithon cf. onkodes, Porolithon sp., Neogoniolithon spp., Neogoniolithon fosliei, Dermatolithon tesselatum, Lithophyllum sp.,
DRILLING 9C
DRILLING 9D DRILLING 9E (+7. 42 m) 0 0 DRILLING 98 0
(+3.57 m) -
2
DRILLING 9A ...... -4,413-3,996
2
4
_9,928
8
12 -13,207
most characteristic vertical cores. Italic type indicates the 14C dates, calibrated in calendar years, and bold type indicates the TIMS U/Th dates (2cr precision in parentheses). The top of each drilling is given by reference to HLS.
_10,924
10
10
Fig. 3. Lithological features of the
(±13)
. _8,242-7,947
8
14
_4,228
4
6
-
(±52)
2
0
11,843
12
- (±35)
14
- (±30)
12,405
-12,468
(±46)
4
4
6
6
8
8
10
10
12
14 16 18
1
h���
12,061
(±54)
12 14
16
12,791
24
18
16
20
18
22
20
24
26 13,959
(±69)
15,780
- (±51) • coral head & coralline crusts dominated framework = coral rubbles & detritus = laminar micritic crusts dominated framework (stromatolites s.l.) l:8l fine sands & cavities Core depth : m.
26 28 30 32
(±63)
266
G. Cabioch et a!.
Tenarea sp., and various other forms. The main coral-facies of the 24 ka Tasmaloum reef sequence are distributed within two distinct units (Fig. 4): I A lower unit, from 24 to 12- 10 ka (between -26 and -5 m bHLS), is composed of branching coral facies with Acropora sp. gr. hyacinthus, Acropora sp. gr. cytherea, various other acroporids, and of coral heads with faviids such as Montipora digitata and sometimes Galaxea gr.fascicularis. Nevertheless, in several cores, Acropora sp. gr. danai and various acroporids are found. In this lower unit, the coral line algae (generally Porolithon onkodes and/or Der-
DRILLING 9C
0
(+14.25 m)
4
DRILLING 9E (+7.42 m) 0
Porolithon onkodes
Vermetids
DRILLING 9F (+3.57 m) 0
2 4
2
6
4
8
Favia cf steffigera Acropora gr danai & Porites sp Porites sp Leptastrea sp
Porites sp
Faviids
6 8
12
10
14_ Platygyra sp
12
6 8 10
Porites sp Galaxea sp Galaxea sp Acropora gr hyacinthus 22 14,670
12- 10 ka (between - 14 and -5 m bHLS), acro porids are abundant, composed usually of Acropora gr. danai and branches of Acropora sp., accompa-· nied by scarce faviids. 2 An upper unit, from 12- 10 ka to the present, consists of in situ massive Porites facies and a few faviids, alternating with beds of Acropora fragments (i.e. coral rubble facies) or Acropora sp. gr. danai, or various acroporids. Several coralline algal levels are: composed of Porolithon cf. onkodes and/or Neogo-· niolithon spp. (or Neogoniolithon fosliei}, and/or
Porites sp & Acropora gr danai Porites sp 4,857-4,453 Poritessp 458-4,871 t-'orites sp ·& Acropora gr danai Diploastrea heliopora 6,724-6,412
5_,_
Acropora gr danai Acropora sp Porites sp
12
10,040
14
Neogoniolithon cf fosliei
10,964
Leptastrea sp Porolithon onkodes
16
Dennatolithon cf tesseffatum
Acropora sp
18
Acropora gr danai
Faviids &
Porolithon cf onkodes 20 Ve rm etid s
&
Montastrea curta & Stylophora pistil/ala & 22 Gomastrea cf retiformis & Porolithon sp
24
14
matolithon cf. tesselatum) are scarce. From 15 to
Acropora sp Neogoniolithon cf fosliei & Favia cf steffigera & 26 Goniastrea cf retiformis Porolithon onkodes &
Vermetids
Dermatolithon cf tessel/atum Acropora sp
12,791 Dermatolithon cf tesseffatum Montastrea curta Acropora gr danai
13,959
Acropora gr danai
Pocil/opora cf verrucosa a r c , ris Acropora sp
�C:,���� � %;;;�t� 15,780
Porites sp
- coral and Ior coralline algae dominatedframeworks G coral r ubbles &detrit us arm laminar micritic cr usts dominatedframeworks (stromatolitess.l.)and detrital accum ulation c:J coarse to medi um sands B83il pebblesand gravels I::2J cavitiesandfine sands D� Unconformities Core depth m ( )
Fig. 4. Main corals (faviid, poritid
and acroporid forms), coralline algae (species of Porolithon, Neogoniolithon, Dermatolithon) and vermetid gastropods recognized in cores 9B, 9C, 9E and 9F. The dates in bold type have been obtained through TIMS and those in bold italic type through AMS calibrated in calendar years. The top of each drilling is given by reference to HLS.
Influences on fringing reefgrowth and structure
Fig. 5. Thin-section photomicrograph of Dermatolithon cf. tesselatum, Porolithon onkodes (coralline algae) and vermetid assemblage at -15.50-m core depth (i.e. 11.93 m bHLS) in core 98, typical of a high-energy environment (scale bar represents 500 Jlm).
Dermatolithon tesselatum in which vermetid gastro pod borings (generally Dendropoma) are usually intermingled (Fig. 5). Several layers of encrusting coralline algae are commonly observed around coral debris. The detritus, generally uncommon except for the cavity infillings, consists of corals, green algal parti cles (Halimeda sp.), benthic foraminiferids, red algae and molluscs. This biofacies usually occurs near the top of the cores. Lithification
Micritic infilling and cements Thin-section studies reveal numerous borings in filled by micrite and internal sediments, particularly abundant in the upper metres of the cores. The bioclasts are usually cemented by high-magnesium calcite, formed generally of peloidal micrite (Fig. 6), a type described in reefs by Macintyre (1977) and Marshall ( 1986). Coral framework, coralline algal veneers and most bioclasts (molluscs, etc.) are bored by various organisms, especially fungi, sponges (clionids) and algae. In most cases, these borings and most other cavities (in particular, the primary pores) are infilled by micritic sediments, admixed with very fine debris of corals, algae, molluscs, and sometimes by geopetal pelleted high magnesian calcite. Isopachous high-magnesium fi brous cement occurs occasionally, indicating cements of marine origin (Macintyre, 1977; Ais-
267
Fig. 6. Thin-section photomicrograph of a peloidal high-magnesian calcite cement partially infilling a cavity in core 98 at -18.60-m core depth (i.e. 15.03 m bHLS). Peloids are rimmed by a high-magnesian calcite fibrous to bladed cement. The dotted texture of the peloidal cement is less dense at the top than at the bottom of the cavity (scale bar represents 160 Jlm).
saoui & Purser, 1986; Marshall, 1986; Aissaoui, 1988). In a few cases, botryoidal aragonite cements link the coralline algae layers. A particular sedimentological feature is the par tial micritization of mollusc and coralline algal fragments, especially in the upper metres of the cores. The complete replacement of the fragments by micrite is usually not observed, except in some red algal layers. These micritizations are probably due to abundant borings observed in these pieces (Bathurst, 1966). Although the degree of lithifica tion depends on the skeletal framework (Aissaoui & Purser, 1986; Macintyre & Marshall, 1988), a great extent of submarine cementation is observed from the bottom to the top of cores, independent of the reef biofacies. This is probably due to the external position of the reef build-ups in a high-energy water environment, as reported elsewhere. A few dark minerals, infilling cavities, are widely observed throughout the cores, indicating inputs of terrestrial sediments from the island.
Laminar micritic crusts In most of the cores, from around 5-6 ka (at + 3 m aHLS) to around 20 ka (at -23 m bHLS), high magnesian micritic calcite laminar crusts, clotted or peloidal in aspect, are observed (Fig. 7). These laminar micritic crusts are especially abundant from -6.50 m bHLS (at 1 1-10 ka) to - 18.50 m
268
G. Cabioch et al.
Fig. 7. Thin-section photomicrograph of a micritic laminar crust at -11.80-m core depth (i.e. 8.23 m bHLS) in core 98. The voids and micritic reticulate network are typical of an organic edification (scale bar represents 160 !!m).
bHLS (at 16 ka) (Figs 3 and 4), and seem to disappear after 6 ka. The laminar crusts are usually smooth and generally occur intermixed with corals and coralline algae, forming a typical biological succession. Corals in growth position, usually Acro pora gr. danai!robusta (or various acroporids) indi cating high-energy conditions, are coated by coralline algae. In most cases, these encrusting red algae are associated with vermetid gastropods (Fig. 5), and seldom with encrusting foraminiferids. Above these encrustations, micritic plan-laminar
Fig. 9. SEM view of a micritic plan-laminar crust at
-10-m core depth (i.e. 5.34 m bHLS) in core 9A, showing high-magnesian calcite cement and grape-shaped forms attributed to microbial remains (scale bar represents 2 !!m).
coatings are found, usually up to 5 em thick. These laminar micritic coatings consist of peloidal and dense micrite (Fig. 7), forming wavy laminae and, sometimes, a dense micritic network is observed (Fig. 8). Occasionally, many well-defined stratified levels of bioclastic fragments are interbedded in these laminae. Grape-shaped aggregates, similar to bacterial remains (Figs 9 & 10), are attributed to organic constructions similar to the stromatolites described in Polynesian high-energy barrier reef
Fig. 8. Thin-section photomicrograph of dense micritic
tubes in a plan-laminar micritic crust at -5.40-m core depth (i.e. 0. 74 m bHLS) in core 9A. In this dense micritic crust, a polygonal form is observed. This form could represent filament remains characteristic of an organic activity (scale bar represents 80 !!m).
Fig. 10. SEM view of a micritic plan-laminar crust in which many grape-shaped forms are recognized, similar to bodies of microbial remains (at -I 7.20-m core depth, i.e. 2.95 m bHLS, in core 9C) (scale bar represents 2 !!m).
269
Injluences on fringing reefgrowth and structure Table I. o180 and o13C values of laminated stromatolitic crusts of cores 9A, 98 and 9C at various depths
0180 (%o PDB)
013C (%o PDB)
-4.85 -5.35 -6.85
0.05 -0.02 0.08
3.77 3.78 3.67
129 165 206
-7.50 -10.20 -14.30
0.28 0.23 0.12
3.61 3.66 3.11
143 151 176
-3.0 -5.55 -10.95
-0.13 0.11 -0.71
3.44 3.41 2.15
Core number
Sample number
9A
202 201 va 32
98
9C
Depth* (m)
*Depth below the highest living Pocillopora and Goniastrea retiformis level.
environments (Montaggioni & Camoin, 1993; Camoin & Montaggioni, 1994). Values of 81 80, from -0.02 to 0.28, and 81 3C, from 2. 15 to 3.78 (Table 1), are typical of organic fractionation and confirm this assumption. The Vanuatu crust values are close to the values found in the Polynesian stromatolitic crusts (Fig. 11). Nevertheless, a shift toward positive values in 8180 is observed. In summary, morphology, structure and isotopic data of the laminar micritic crusts indicate an organic
origin, as pointed out in Tahiti (Camoin & Montag gioni, 1994). Freshwater cementation and diagenetic changes in bioclastic fragments
In a few cases, oxidized, grey-brown silt-size infill ings and partial dissolution of aragonitic compo nents are observed from the bottom to the top of the cores, indicating emergence events and meteoric water throughflow. Also, matrix dissolution occurs around the bioclasts, especially at the top of the cores (Fig. 12). Coral heads or coral branches are not much affected by freshwater leaching. Sometimes, although corals seem to be pristine (conservation of aragonitic fabrics), SEM observations reveal solu tion features {Fig. 13) in which the leaching of sclerodermite centres is observed. Dissolution of
Vanuatu stromatolitic crusts
fossil corals from Tahiti -3.0
modern corals from Tahiti -6.0
-5.0
-4.0
-3.0
-2.0
-1.0
0.0
+1.0
+2.0
o18o %o PDB Fig. 1 1 . Isotopic data values of o180 measured in the laminar micritic veneers (•, see values in Table 1). For comparison, the o180 and o13C isotopic values of Polynesian stromatolitic crusts and Polynesian modern and fossil Acropora branches are taken from Camoin & Montaggioni (1994).
Fig. 12. Thin-section photomicrograph of micritic matrix dissolution around a gastropod at -0.80-m core depth (i.e. 2. 72 m aHLS), in core 98 (scale bar represents 500 Jlm).
G. Cabioch et al.
270
Fig. 13. SEM view of a Porites sample from the upper metres of core 9A, showing partial leaching of the aragonitic needles. The centre of the sclerodermite is clearly affected by dissolution (scale bar represents 3 J.. lm).
aragonite needles is also found in the corals (Fig. 13), at both the bottom and top of cores. Most rarely, low-magnesian cement fills voids in coral (Fig. 14). Such a leaching sequence is interpreted as the first stage of progressive freshwater diagenesis (James, 1974; Gvirtzman & Friedman, 1977; Hen rich & Wefer, 1986).
DISCUSSION
Morphology
The narrow Holocene terraces along the Tas maloum peninsula, separated by gentle slopes with out notches (except at the back of the 36-m terrace), may indicate continual rapid uplift movements with only brief intervening pauses. Likewise, the pronounced steepness of the slope that is observed at present along the Tasmaloum shore indicates a continued process of frequent successive uplift movements resulting in a rapid mean uplift rate. Several researchers have assumed that the small terraces and scarps could represent coseismic uplift events (Ota et al. , 1993; Chappell et al. , 1994; Ota & Chappell, 1996). However, the observations previ ously made in Vanuatu after the 1965 uplift of North Malakula, which raised corals by as much as 1.2 m (Benoit & Dubois, 1971; Taylor et al. , 1980,
Fig. 14. SEM view of a Porites sample (see Fig. 8) from the upper metres of core 9A, in which we can recognize some low-magnesian calcite crystals (euhedral form), indicating freshwater diagenesis: the aragonitic sclerodermite leaching is followed by a partial low-magnesian calcite cement replacement (scale bar represents 4 J..lm).
1987), did not show any kind of distinctive terrace or scarp building. The same observation has been made in the Tasmaloum area, where the same 1965 seismic event uplifted the coast by about 0.25 m (Taylor et al., 1980, 1987). Building of small ter races and scarps could be due to simultaneous combination of coral reef growth especially when the reefs arrive in the intertidal zone because of uplift, bioerosion and solution processes, sea-level rise and succession of more or less numerous tectonic uplift events, such as those previously reported in other parts of the archipelago (Benoit & Dubois, 197 1; Jouannic et al. , 1980; Taylor et a/., 1980, 1987, 1990), or in Papua New Guinea (Bloom et al., 1974; Chappell, 1974; Chappell & Polach, 199 1). Formation of the deep notch at the back of the 36-m terrace, dated at c. 6 ka, is directly related to growth of the broad 36-m reef flat and the relative stability of sea-level at a time when the rates of uplift and rising sea-level equalized for some time (Fig. 2). Absence of other representative notches from the scarps of the Holocene reef below the 36-m terrace (Fig. 2) indicates that, since 6 ka, the posi tion of the shoreline has regressed nearly continu ously, so that dissolution processes and endolithic organism borings were never at one level long enough to form other notches (Pirazzoli, 1986).
27 1
Influences on fringing reefgrowth and structure Pre-reef substrate, reef colonization and initial growth
Colonization of the Tasmaloum fringing reef oc curred upon a pre-reef substrate, encountered in the cores at depths of -20 to -25 m bHLS (Figs 2 & 15). This substrate is composed of a huge bioclastic sand formation accumulated during a part of the last glacial period. The faunistic (especially the benthic foraminiferids) and floristic assemblages of these sand levels are typical of an algal environment. The apparent absence of corals (colonies or debris) seems to be in accordance with the fact that the sand beds could represent the fore-reef of an older reef just behind the present Holocene terraces. A combination of sea-level fall or rise and tectonic uplift eventually produced a substrate with a depth suitable for colonization and reef development. However, the substrate became shallow so fast that coral growth was apparently prevented; this is consistent with the lack of coral before 24 ka. Above the basal sand formation, pebbles and gravel conglomerates indicate a palaeo-coastline (Fig. 15). The reef growth initiation was almost instantaneous, as shown by similar dates obtained in core 9F at -22.65 m bHLS (20 96 1 cal yr BP) and in core 9J at -28.25 m bHLS (2 1 303 cal yr BP). Nevertheless, in core 9£, a date of 24 390 cal yr BP was obtained at -22.70 m bHLS. The earliest stages of Tasmaloum reef initiation seem to have been
contemporaneous with the late stages of the LGM and the earliest stages of the sea-level rise, at around 20 ka. Previous to this study, Holocene reef initia tions were observed to occur in the Pacific at around 9-7 ka (Easton & Olson, 1976; Davies & Montaggioni, 1985; Davies et al., 1985, Montag gioni, 1988; Cabioch et a!. , 1995; Kan et al. , 1995; Larcombe et a!., 1995), and in the barrier reef in Tahiti at around 14 ka (Bard et al., 1996). Varia tions in reef initiation ages of these various areas could be explained by sea surface temperature (SST) differences unfavourable to reef development or the unavailability of suitable foundations for reef establishment (Davies et al., 1985; Partain & Hop ley, 1989). Near Tasmaloum, located at 15°S, at around 24 ka, SSTs were probably high enough for the reef growth. Also, the gravels and pebbles pro vided a favourable substrate for reef initiation, as previously observed in Australia (Johnson & Risk, 1987; Partain & Hopley, 1989). In New Caledonia, we know that the time lag for Holocene coral colo nization between reef areas is related to differences of substrate type, fringing reefs colonizing preferen tially karstified limestone foundations (or other similar surfaces), the roughness of karstic surfaces (or similar) being favourable to recruitment and attachment of coral larvae (Cabioch et a!., 1995). At Tasmaloum, the morphology of the underlying pre-reefal substrate (top of the sand accumulations, Figs 2 & 15) influenced the earliest stages of reef
6 ka reef flat
40 30 20
Drill holes
10
-10 __ _____ ___ _
-20
')
___ _
volcanic substrate ?
-30
Fig. 15. Cross-section of the uplifted Tasmaloum fringing reef, showing the time-lines of 6, 8,
I 0, 12, 14, 16, 18 and 20 ka. Only the most characteristic dates are reported (those in italics are 14C dates converted to calendar years).
272
G. Cabioch et al.
development, as indicated by the regularity of the time-lines between 24 ka and 12-10 ka (Fig. 15), which is also usually observed in the younger reefs (Marshall & Davies, 1982; Walbran, 1994). After 12-10 ka, the tectonic behaviour of Espiritu Santo constrained the modern structure. This gives an explanation for the relatively narrow reef and steep reef slope (Bloom & Yonekura, 1985), similar to the situation in Papua New Guinea (Chappell, 1974; Bloom et al. , 1974). Palaeoecological significance
The environmental factors which influence reef growth are directly related to water depth, light intensity, wave energy, nutrients, sediment and freshwater inputs: characteristic assemblages of or ganisms represent well-defined zones, as pointed out elsewhere in the south-west Pacific (Morton, 1973, 1974; Guilcher, 1974; Done & Navin, 1990). Consequently, coral and organism assemblages found in the cores commonly have a well-defined palaeoecological significance (biozonal partitioning) and we can use these to reconstitute the various stages of reef development. The coral assemblage of the lower unit (from 24 to 12-10 ka) is principally composed of branching coral facies and various acroporids, Acropora SP,p., as generally observed in the Pacific (Davies et al. , 1985; Cabioch, 1988; Montaggioni, 1988; Veron, 1990). The most frequent Acropora, A. gr. hya cinthus and A. gr. cytherea, are common in environ ments having heavy waves and surges and in the central part of reef slopes (Morton, 1974; Faure, 1982; Veron, 1990). This coral assemblage, accom panied by scarce coralline algae, indicates medium to high-energy conditions developed predominantly in relatively deep or protected environments (Done, 1982; Faure, 1982; Veron, 1990), perhaps related to the 5-15-m reef slopes. The textural analysis con firms this interpretation. Furthermore, from 15 to 12 ka, a A. gr. danai!robusta and Porolithon cf. onkodes community, sometimes found mixed with the previous acroporid assemblage, is typical of high wave energy (Faure, 1982; Veron, 1990). The second most frequent forms, poritids, are abundant, but are found everywhere, and thus are not good bathymetric indicators. Faviids are common, but never abundant, except in core 9B, where a large colony of Diploastrea heliopora is observed. The upper unit (between 12-10 ka and the present) is composed principally of Acropora sp. gr.
danai, or various other strong branching acro porids, typical of high-energy conditions, probably related to the reef front and the upper reef slope (Done, 1982; Faure, 1982; Veron, 1990). In situ massive Porites sp. are more common. We note that in core 9D, at around -1 m HLS, the coral assem blage is made of Favites cf. pentagona, Seriatopora sp. and Pachyseris rugosa, which indicates a moderate-energy coral reef environment. In these cores, the occurrence at many levels of heavy coralline red algal crusts, especially in the upper unit (generally Porolithon cf. onkodes), is very closely linked to local environmental conditions, especially wave energy and light intensity (Littler & Doty, 1975; Bosence, 1984, 1985). In these thick coralline algal crusts, vermetid gastropods (gener ally Dendropoma maxima) and encrusting foramin iferids are common (Fig. 4). In the Pacific, this specific assemblage is typical of high-wave-energy environments on hard substrates, generally at depths less than 2-6 m below mean sea-level (i.e. near the reef front or at the top of the upper reef slopes). In particular, Porolithon onkodes requires intense light and continuous disturbance (Morton, 1973, 1974; Littler & Doty, 1975; Richard, 1982; Adey, 1986; Nunn, 1993). Sometimes, Porolithon cf. onkodes is found to be associated with Neogoni olithonfosliei or Dermatolithon cf. tesselatum in the uppermost metres of the cores (Fig. 4), these species being sharply restricted to depths less than 10 m (Adey et al. , 1982). Biofacies succession, reef development and neotectonic behaviour
During the postglacial sea-level rise, the vertical succession of reef assemblages generally reflects environment and bathymetry variations. These changes are related to antecedent topography and variations in wave energy (Davies & Hopley, 1983; Montaggioni, 1988; Collins et al. , 1993). Similar vertical variations occurred in the Tasmaloum fringing reefs during the last 24 kyr. From 24 to 15 ka, the water-energy conditions suggest that reef growth at Tasmaloum did not keep up with sea-level rise. However, from 15 to 1210 ka, the intermingled high-wave-energy assem blage (A. gr. danailrobusta community) seems to indicate an alternation between catch-up and keep-up biofacies in the sense of Neumann & Macintyre (1985). These community variations are probably related to uplift events.
Influences on fringing reefgrowth and structure From 12-10 ka to the present, the high-wave energy assemblages demonstrate that the reef re mained close to sea-level. Thus, the Tasmaloum fringing reef can be regarded as a keep-up reef between 12-1 0 ka and the present, because reef growth was able to keep pace with rising sea-level, or, alternatively, because successive uplifts permit ted it to keep pace with sea-level rise, ranging from 0- to 6-m water depth. At about 5-6 ka, the sea level stabilized, as reported elsewhere in the Pacific (Lambeck & Nakiboglu, 1986), but the uplift con tinued (Jouannic et al., 1980; Taylor et al., 1980, 1987) and thus terraces, scarps and notches formed (Fig. 2). The bathymetry variation, observed in the inter nal structure of the Tasmaloum fringing reef, shows also that the uplift rate may not have been constant during the last 24 ka. Changes in reef growth pat terns, particularly well characterized from 15 to 12-10 ka, imply an increase in uplift rate. In SW Espiritu Santo, this increase in uplift rate is proba bly related to the collision of the d'Entrecasteaux Zone with the New Hebrides arc. Lithification
Physico-chemical lithification and freshwater diagenesis The Tasmaloum fringing reef is well lithified by silt-size sediment infillings and, more importantly, by pelleted micritic high-magnesian calcite, espe cially in the uppermost metres of the cores. As reported by numerous workers, these well-lithified materials are restricted to outer reef environments, exposed to heavy wave action (Aissaoui & Purser, 1986; Cabioch, 1988; Macintyre & Marshall, 1988). This agrees with the implications of biological assemblages typical of medium- to high-energy wa ter conditions, which represent the reef front (upper unit) or the reef slopes (upper and lower units). Sometimes, freshwater diagenesis, revealed by discrete dissolution features, low-magnesian calcite cements (Fig. 14) and rare dissolution moulds, clearly shows that reef growth was disturbed by short periods of emergence. These are accompanied by freshwater inputs from episodic and successive uplift movements or pauses in sea-level rise related to deglaciation. Although leaching is observed, these reef materials are usually consolidated by freshwater cementation with low-magnesian calcite spar cement or micrite infillings.
273
Organic lithification and stromatolites
In the Tasmaloum fringing reef, organic lithification is one of the most interesting sedimentological features, and numerous stromatolitic coatings are observed. The Tasmaloum reef is exposed to fre quent heavy surf, similar to the Polynesian high energy barrier reef (Montaggioni & Camoin, 1993). In neighbouring New Caledonia, such micritic coat ings are rare or absent. Nevertheless, we observe similar thin laminations in the Holocene fabric of a lagoonal reef located near the barrier reef, where high-wave-energy conditions prevail. These data suggest that high-wave-energy conditions are re quired for reef stromatolitic development. In the Tasmaloum reef cores, the stromatolites are found in abundance from 16-15 to 12-1 0 ka, a period of very rapid sea-level rise (Fairbanks, 1989; Bard et al., 1990, 1996; Edwards et al., 1993), in which palaeoceanographic parameters (SSTs and perhaps nutrient inputs) varied significantly. An other interesting fact is the disappearance of these laminar crusts after 6 ka. This corresponds to the time of sea-level stabilization (Lambeck & Nakibo glu, 1986), which constituted a generally favourable period for reef growth (Davies et a!., 19 85; Montag giani, 1988). Sea-level stabilization caused the establishment of a new hydrological and oceano graphic regime, accompanied by, in particular, warming of tropical waters (Beck et al., 1992, 1997). After 6 ka, conditions became more favour able for coral growth, but unfavourable for stroma tolite growth. This is particularly well explained by changes in the SSTs recorded by isotopic methods in many Porites from Tasmaloum (Beck et al., 1992; Recy et al., 1993; Beck et al., 1997). The Sr/Ca palaeo-SSTs, measured by TIMS, are very close to present-day values in the youngest corals ( < 5 ka), but several degrees lower than present-day values in the other corals (between 5 and 10.3 ka). These recent results suggest also that the tropical belt was compressed toward the Equator before 10 ka (Beck et al., 1992). Another explanation of these ecologi cal changes may be related to palaeoceanographic changes, such as those pointed out recently by McCulloch et al. (1996).
CONCLUSIONS
At Tasmaloum, morphological, palaeoecological and sedimentological data show the initiation of a
274
G. Cabioch et a!.
fringing reef along an uplifted coast and an example of unusual reef growth on a rapidly uplifting sub strate. 1 Although SSTs before 10 ka were probably lower than at present, as inferred by Beck et al. ( 1992), and coral build-ups less numerous than they are today, the reef colonized a more-or-less indurated substrate (calcareous sands) by 24 ka (Gray et al., 1993) and continued to grow until now. 2 We have cored a reef sequence as complete as the Barbados section (Fairbanks, 1989; Bard et al., 1990) and more complete than the Huon Peninsula (only the last 13 ka were recovered in that case; Chappell & Polach, 199 1; Edwards et at., 1993) and French Polynesian (only the last 14 ka; Bard et a!., 1996) sections. 3 The biofacies analysis reveals two distinct units: (a) From 24 to 12- 10 ka, rare coralline algae and poritid and acroporid build-ups of Acropora gr. hyacinthus and A. gr. cytherea are considered to represent a medium- to high-wave-energy facies, related to 5- 15-m water depth. Nevertheless, from 15 to 12-10 ka, this biofacies alternates with a high-wave-energy facies, composed mainly of Acropora gr. danailrobusta, implying several uplift events. (b) From 12- 10 ka to the present, assemblages of acroporids (Acropora gr. danailrobusta), rare poritids, numerous encrusting coralline algae (Po rolithon cf. onkodes, Neogoniolithon spp. , Neogo niolithon fosliei and Dermatolithon tesselatum), vermetid gastropods, and sometimes Acervulina foraminiferids reveal a high-wave-energy facies, corresponding to the upper part of the exposed reef slope or the outer part of the fringing reef flat, i.e. ranging from 0- to 6-m water depth. From 6-5 ka to the present, reef growth is marked by a succession of emergences resulting from the com bination of sea-level stabilization and incremen tal uplift movements. 4 Patterns of Tasmaloum reef growth appear to depend on the tectonic behaviour of Espiritu Santo in relation to the uplift movements and the more or-less rapid rises or stillstands of postglacial sea level. Thus, it is inferred that the uplift rate has varied during the last 24 ka (from 24 ka to 4.5 ka). After 6 ka, the vertical biofacies change, marked by the disappearance of the laminar crusts (or stroma tolites sensu lato), reflects the establishment of a new hydrological and oceanographic regime. 5 Once sea-level stabilized, the wide 6-7-ka reef flat at +36 m quickly emerged by uplift. The reef
then had only the steep transgressive reef-front as substrate. The rapid emergence prevented it from growing wide.
ACKNOW LEDGEMENTS
The field-work required extensive logistical sup port, and many people participated in the drilling operations. We are particularly grateful for the contributions of Yvan Join, Jean-Louis Laurent, Claude Ihilly (ORSTOM Noumea), Paul and Ray mond Aroug, Christian Livo, Edwin Tae (Tas maloum) and Bernard Labrousse (ORSTOM Villefranche). We also thank the crew of ORSTOM R. V. Alis for assistance in transporting equipment. We also wish to thank the Vanuatu Government for permits to drill and for assistance, and particularly thank the Public Works Department of Espiritu Santo. We also acknowledge Claude Reichenfeld and Michel Lardy (ORSTOM) for assistance in the preparation for field-work, and also Michel Noel (Luganville, Espiritu Santo) for help in logistical support. Our thanks are extended to the people of Espiritu Santo, in particular to the people of Pa kataora, Vounapissu and Vimala (SW Espiritu Santo). We also acknowledge Dr Christophe Chev illon (ORSTOM Noumea) for examinations and analysis of the bioclastic sands. Thanks are due to Roger Notonier and Christine Castellaro (Univer site de Provence, Marseille) for their assistance in SEM. Gratitude is expressed to Dr Christian Jouan nic for his critical and helpful comments, and to Professor Edouard Bard for discussing calibration methods. We also thank Professor David Hopley and Dr Gilbert Camoin for constructive reviews. This work is supported jointly by ORSTOM (l'In stitut Francais de Recherche Scientifique pour le Developpement en Cooperation), National Science Foundation Grants ATM-8922 1 14 and EAR8904987 (F.W. Taylor) and National Science Foun dation Grant OCE-950 1580 (G. Burr). UMR Geosciences Azur Contribution 93 and University of Texas Institute for Geophysics Contribution 1238. REFERENCES ADEY, W.H. (1986) Coralline algae as indicators of sea
level. In: Sea-level Research: a Manualfor the Collection and Evaluation of Data (Ed. van de Plassche, 0.), pp. 229-279. Geobooks, Norwich.
Influences on fringing reefgrowth and structure ADEY, W.H., TOWNSEND, R.A & BOYKINS, W.T. ( 1982) The crustose coralline algae (Rhodophyta: Corallinaceae) of the Hawaiian Islands. Smithson. Contrib. mar. Sci., 15, 1-74. AISSAOUI, D.M. ( 1988) Magnesian calcite cements and their diagenesis: dissolution and dolomitization, Muru roa atoll. Sedimentology, 35, 82 1-841. AISSAOUI, D.M. & PURSER, B.H. (1986) La cimentation dans les recifs: principe de cimentation maximale. C. R. Acad. Sci. Paris, 303, 301-303. BARD, E., HAMELIN, B., FAIRBANKS, R.G. & ZINDLER, A. ( 1990) Calibration of the 14C timescale over the past 30 000 years using mass spectrometric U-Th ages from Barbados corals. Nature, 345, 405-410. BARD, E., ARNOLD, M., FAIRBANKS, R.G. & HAMELIN, B. (1993) 230Th-234U and 14C ages obtained by mass spectrometry on corals. Radiocarbon, 35, 191-199. BARD, E., HAMELIN, B., ARNOLD, M. et a/. ( 1996) Deglacial sea level record from Tahiti corals and the timing of global meltwater discharge. Nature, 382, 241-244. BATHURST, R.G.C. ( 1966) Boring algae, micrite envelopes and lithification of molluscan biosparites. Geol. J. , 5, 15-32. BECK, J.W., EDWARDS, R.L., ITO, E. et a/. (1992) Sea surface temperature from coral skeletal strontium/ calcium ratios. Science, 257, 644-647. BECK, J.W., RtCY, J., TAYLOR, F.W., EDWARDS, R.L. & CABIOCH, G. (1997) Holocene sea surface temperature variations in the SW Pacific Ocean derived from coral Sr/Ca and oxygen isotope thermometry. Nature, 385, 705-707. BENOIT, M. & DUBOIS, J. ( 1971) The earthquake swarm in the New Hebrides archipelago, August 1965. R. Soc. N. z., 9, 141-148. BLOOM, A.L. & YONEKURA, N. ( 1985) Coastal terraces generated by sea-level change and tectonic uplift. In: Models in Geomorphology (Ed. Woldenberg, M.J.), pp. 139- 154. Allen and Unwin, Winchester, MA. BLOOM, A.L., BROECKER, W.S., CHAPPELL, J.M.A., MAT THEWS, R.K. & MESOLELLA, K.J. (1974) Quaternary sea-level fluctuations on a tectonic coast: new Th230/ U234 dates from the Huon Peninsula, New Guinea. Quat. Res., 4, 185-205. BosENCE, D. (1984) Construction and preservation of two modern coralline algal reefs, Saint-Croix, Caribbean. Palaeontology, 27, 549-574. BosENCE, D. (1985) Preservation of coralline-algal reef frameworks. In: Proceedings of the 5th International Coral Reef Congr., Tahiti, 6, pp. 623-628. CABIOCH, G. (1988) Recifs frangeants de Nouvelle Caledonie (Pacifique sud-ouest). Structure interne et in fluences de l 'eustatisme et de Ia neotectonique. These Doct., Universite Aix-Marseille I. CABIOCH, G., MONTAGGIONI, L.F. & FAURE, G. (1995) Holocene initiation and development of New Cale donian fringing reefs, South-West Pacific. Coral Reefs, 14, 131-140. CAMOIN, G.F. & MONTAGGIONI, L.F. (1994) High energy coralgal-stromatolite frameworks from Holocene reefs (Tahiti, French Polynesia). Sedimentology, 4 1 , 655676. CHAPPELL, J. (1974) Geology of coral terraces, Huon Peninsula, New Guinea: a study of Quaternary tectonic
275
movements and sea-level changes. Geol. Soc. Am. Bull., 85, 553-570.
CHAPPELL, J. & POLACH, H. (1991) Post-glacial sea-level rise from a coral record at Huon Peninsula, Papua New Guinea. Nature, 349, 147-149. CHAPPELL, J., 0TA, Y. & BERRYMAN, K.R. (1994) Holocene and Late Pleistocene coseismic uplift of Huon Penin sula, Papua New Guinea. In: Study on Coral Reef Terraces of the Huon Peninsula, Papua New Guinea (Ed. Ota, Y.), pp. 35-62. Department of Geography, Senshu University, Kawasaki, Japan. COLLINS, L.B., ZHU, Z.R., WYRWOLL, K.-H. et a/. (1993) Holocene growth history of a reef complex on a cool water carbonate margin: Easter Group of the Houtman Abrolhos, Eastern Indian Ocean. Mar. Geol., 1 1 5, 2946. DAVIES, P.J. & HOPLEY, D. ( 1983) Growth facies and growth rates of Holocene reefs in the Great Barrier Reef. Bur. Min. Res. J. Aust. Geol. Geophys., 8, 237251. DAVIES, P.J. & MARSHALL, J.F. (1979) Aspects of Holocene reef growth-substrate age and accretion rate. Search, 10, 276-279. DAVIES, P.J. & MONTAGGIONI, L.F. ( 1985) Reef growth and sea-level change: the environmental signature. In: Pro ceedings of the 5th International Coral Reef Congr., Tahiti, 3, pp. 477-51 1. DAVIES, P.J., MARSHALL, J.F. & HOPLEY, D. (1985) Rela tionships between reef growth and sea level in the Great Barrier Reef. In: Proceedings of the 5th International Coral Reef Congr., Tahiti, 3, pp. 95-103. DONE, T.J. (1982) Patterns in the distribution of coral communities across the central Great Barrier Reef. Coral Reefs, 1 , 95-107. DONE, T.J. & NAVIN, K.F. ( 1990) Shallow-water benthic communities on coral reefs. In: Vanuatu Marine Re sources: Report of a Biological Survey (Ed. Done, T.J. & Navin, K.F.), pp. 10-36. Australian Institute of Marine Science, Townsville, Qld. EASTON, W.H. & OLSON, E.A (1976) Radiocarbon profile of Hanauma Reef, Oahu, Hawaii. Geol. Soc. Am. Bull., 87, 71 1-719. EDWARDS, R.L., CHEN, J.H. & WASSERBURG, G.J. (1987) 23sU-234U-230Th-232Th systematics and the precise measurement of time over the past 500 000 years. Earth planet. Sci. Lett., 81, 175-192. EDWARDS, R.L., BECK, W.J., BuRR, G.S. et a/. (1993) A large drop in atmospheric 14C/12C reduced melting in the Younger Dryas, documented with 230Th ages of corals. Science, 260, 962-968. FAIRBANKS, R.G. ( 1989) A 17 000-year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Na ture, 342, 6 37-642. FAURE, G. (1982) Recherche sur les peuplements des scler actiniaires des recifs cora/liens de l 'archipel des Mascar eignes (Ocean Indien occidental). These Doct. es Sci., Universite Aix-Marseille II. GILPIN, L. ( 1982) Tectonic geomorphology of Santo Island, Vanuatu. MS thesis, Cornell University, Ithaca, NY. GRAY, S.C., BURR, G.S., DONAHUE, D.J. et a/. ( 1993) · Comparison of TIMS 230Th and AMS 14C ages of Pacific corals for the last deglaciation: evidence for a
276
G. Cabioch et a!.
rapid oscillation of delta 14C. EOS Trans., Am. geophys. Union, 1 993 Fall meeting, 74, pp. 326. GuiLCHER, A. ( 1974) Coral reefs of the New Hebrides, Melanesia, with particular reference to open-sea, not fringing, reefs. In: Proceedings of the 2nd International Coral Reef Symposium, Brisbane, Qld, 2, pp. 523-535. GviRTZMAN, G. & FRIEDMAN, G.M. ( 1977) Sequence of progressive diagenesis in coral reefs. Am. Assoc. petrol. Geol., Stud. Geol., 4, 357-380. HENRICH, R. & WEFER, G. ( 1986) Dissolution of biogenic carbonates: effects of skeletal structures. Mar. Geol., 71,
6th International Coral Reef Symposium, Townsville, Qld, 3, pp. 455-460. MONTAGGIONI, L.F. & CAMOIN, G.F. (1993) Stromatolites
529. JOUANNIC, C., TAYLOR, F.W., BLOOM, A.L. & BERNAT, M. (1980) Late Quaternary uplift history from emerged reef
associated with coralgal communities in Holocene high energy reefs. Geology, 2 1 , 149-152. MoRTON, J.E. ( 1973) The intertidal ecology of the British Solomon Islands. I . Zonation patterns of the weather coasts. Phil. Trans. R. Soc. London, 265, 491-542. MoRTON, J.E. (1974) The coral reefs of the British So lomon Islands: a comparative study of their composi tion and ecology. In: Proceedings of the 2nd International Coral Reef Symposium, Brisbane, Qld, 2, pp. 31-53. MORTON, J.E. & CHALLIS, D.A. ( 1969) The biomorphology of Solomon Islands shores with a discussion of zoning patterns and ecological terminology. Phil. Trans. R. Soc. London, Ser.B, 255, 459-516. NEUMANN, A.C. & MACINTYRE, I. (1985) Reef response to sea level rise: keep-up, catch-up or give-up. In: Proceed ings of the 5th International Coral Reef Congr., Tahiti, 3, pp. 105-110. NuNN, P.D. (1993) Role of Porolithon algal-ridge growth in the development of the windward coast of Tongatapu island, Tonga, South Pacific. Earth Surf Process.
terraces on Santo and Malekula, Central New Hebrides island arc. UN ESCAP, CCOP/SOPAC, Tech. Bull., 3,
Land!, 18, 427-439. 0TA, Y. & CHAPPELL, J. (1996) Late Quaternary coseismic
34 1-362. HOPLEY, D. ( 1982) The Geomorphology of the Great Barrier Reef Quaternary Development of Coral Reefs. John Wiley, New York.
JAMES, N.P. ( 1974) Diagenesis of scleractinian corals in the subaerial vadose environment. J. Paleontol., 48, 785799. JOHNSON, D.P. & RISK, M.J. (1987) Fringing reef se quences and geochemistry, Britomart Reef, central Great Barrier Reef, Australia. Sedimentology, 3 1, 5 15-
9 1- 108. KAN, H., NOR!, N., NAKASHIMA, Y. & ICHIKAWA, K. (1995)
uplift events on the Huon Peninsula, Papua New Guinea, deduced from coral terrace data. J. geophys.
The evolution of narrow reef flats at high-latitude in the Ryukyu Islands. Coral Reefs, 14, 123-130. LAMBECK, K. & NAKIBOGLU, S.M. (1986) Recent sea level changes in the south-west Pacific. R. Soc. N. Z. Bull., 24,
Res., 101, 6071-6082. 0TA, Y., CHAPPELL, J., KELLEY, R. et a/. (1993) Holocene
485-491. LARCOMBE, P., CARTER, R.M., DYE, J., GAGAN, M.K. & JOHNSON, D.P. (1995) New evidence for episodic post glacial sea-level rise, central Great Barrier Reef, Austra lia. Mar. Geol., 1 27, 1-44. LITTLER, M.M. & DoTY, M.S. ( 1975) Ecological compo nents structuring the seaward edges of tropical Pacific reefs: the distribution communities and productivity of Porolithon. J. Ecol., 63, 117-129. MACINTYRE, I.G. ( 1977) Distribution of submarine ce ments in a modem Caribbean fringing reef, Galeta Point, Panama. J. sediment. Petrol., 47, 503-516. MACINTYRE, I.G. & MARSHALL, J.F. (1988) Submarine lithification in coral reefs: some facts and misconcep tions. In: Proceedings of 6th International Coral Reef Symposium, Townsville, Qld, 1, pp. 263-272. MARSHALL, J.F. ( 1986) Regional distribution of submarine cements within an epicontinental reef system: central Great Barrier Reef, Australia. In: Reef Diagenesis (Eds Schroeder, J.H. & Purser, B.H.), pp. 8-26. Springer Verlag, Berlin. MARSHALL, J.F. & DAVIES , P.J. ( 1982) Internal structure and Holocene evolution of One Tree Reef, Southern Great Barrier Reef. Coral Reefs, 1, 2 1 -28. McCULLOCH, M., MORTIMER, G., ESAT, T., XIANHUA, L., PILLANS, B. & CHAPPELL, J. (1996) High resolution windows into early Holocene climate: Sr/Ca coral records from the Huon Peninsula. Earth planet. Sci. Lett., 138, 169-178. MONTAGGIONI, L.F. ( 1988) Holocene reef growth history in mid-plate high volcanic islands. In: Proceedings of the
coral reef terraces and coseismic uplift of Huon Penin sula, Papua New Guinea. Quat. Res., 40, 177-188. PARTAIN, B.R. & HOPLEY, D. (1989) Morphology and
development of the Cape Tribulation/ringing reefs, Great Barrier Reef Australia. Great Barrier Reef Marine Park Authority Tech. Mem., 21 , 1-45. PIRAZZOLI, P.A. (1986) Marine notches. In: Sea-level Re search: a Manual for the Collection and Evaluation of Data (Ed. van de Plassche, 0) , pp. 361-400. Geobooks,
Norwich. RECY, J., BECK, J.W., TAYLOR, F.W., EDWARDS, R.L. & CABIOCH, G. (1993) Variations holocenes de Ia temper ature de surface de Ia mer dans le Sud-Ouest Pacifique. Reun. spec. Soc. geol. Fr. Soc. geol. Fr., Paris, 90 pp. RICHARD, G. ( 1982) Mollusques lagunaires et recifaux de
Polynesie fam;aise. Inventaire faunistique, bionomie, bi lan quantitatif croissance, production. These Doct. es Sci., Universite P. & M. Curie, Paris. STUIVER, M. & REIMER, P.J. (1993) Extended 1 4C data base and revised calib 3.0 1 4C age calibration program. Radiocarbon, 35, 2 15-230. TAYLOR, F.W. (1992) Quaternary vertical movements of the central New Hebrides island arc. In: Proceedings of the Ocean Drilling Program, Initial Reports, 134 (Eds Collot, J.-Y, Greene, H. G. & Stokking, L.B.), pp. 33-42. Ocean Drilling Program, College Station, TX.
TAYLOR, F.W., !SACKS, B.L., JOUANNIC, C., BLOOM, A.L. & DUBOIS, J. (1980) Coseismic and Quaternary vertical tectonic movements, Santo and Malekula islands, New Hebrides island arc. J. geophys. Res., 85, 5367-5381. TAYLOR, F.W., JOUANNIC, C. & BLOOM , A.L. (1985) Qua ternary uplift of the Torres Islands, northern New Hebrides frontal arc: comparison with Santo and
Influences on fringing reefgrowth and structure Malekula islands, central New Hebrides frontal arc. J Geol., 93, 4 1 9-438. TAYLOR, F.W., FROHLICH, C., LECOLLE, J. & STRECKER, M. ( 1 987) Analysis of partially emerged corals and reef terraces in the central Vanuatu arc: comparison of con temporary coseismic and nonseismic with Quaternary vertical movements. J. geophys. Res., 92, 4905-4933. TAYLOR, F.W., EDWARDS, R.L. & WASSERBURG, G.J. ( 1 990) Seismic recurrence intervals and timing of aseismic subduction inferred from emerged corals and reefs of the central Vanuatu (New Hebrides) frontal arc. J. geophys. Res., 95, 393-408.
277
TAYLOR, F.W., MANN, P., LAGOE, M. et a!. ( 1 993) Mecha nisms for rapid reversals of vertical motion in the New Hebrides and Solomon arcs related to colliding features.
EOS Trans., Am. geophys. Union, 1 993 Fall Meeting, 74, 545 pp.
VERON, J.E.N. ( 1 990) Checklist of the hermatypic corals of Vanuatu. Pacific Sci. , 44, 5 1 -70. WALBRAN, P.D. ( 1 994) The nature of the pre-Holocene surface, John Brewer Reef, with implications for the interpretation of Holocene reef development. Mar. Geol. , 122, 63-79.
Passive Margins
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
Spec. Pubis int. Ass. Sediment. ( 1998) 25, 281-294
Lagoonal sedimentation and reef development on Heron Reef, southern Great Barrier Reef Province B. T. S M I TH*1, E. FRANKEL* and J. S. J ELLt *Department ofApplied Geology, University ofTechnology, Sydney, PO Box 123, Broadway, NSW 2007, Australia; and
tDepartment ofEarth Sciences, University of Queensland, Brisbane, Qld 4072, Australia
ABSTRACT
Most studies of coral reef lagoonal sediments have been confined to surface materials. A three dimensional appreciation of materials and processes in this environment has therefore not been fully provided. The combined use of vibrocoring and seismic profiling on Heron Reef provides the basis for a more thorough understanding of sedimentation processes, and allows them to be related to the overall development of the reef system. Heron Reef is a lagoonal platform reef in the Capricorn Group at the southern end of the Great Barrier Reef Province. The lagoon is complex with both shallow (0-2 m) and deep (3-5 m) areas. The sediment accumulations in these areas are compositionally similar, but texturally different. Both areas have gravelly muddy sands (Facies 3) as basal fill, overlain by coarse sand in the shallow lagoon (Facies I) and muddy sand in the deep lagoon (Facies 2). Radiocarbon ages of these sediments provide a minimum date of 4200 yr BP for the beginning of lagoonal filling. It is believed that Facies 3 was deposited before any of the reef reached sea-level. The windward margin reached sea-level first (possibly around 2700 yr BP) creating a different hydrodynamic environment across the platform. Subsequent sedimentation involved sorting behind the windward margin into the coarse sand of Facies I to windward and muddy sand of Facies 2 to leeward. Seismic profiles across Heron Reef indicate a planar foundation surface dipping at 0.03 to the east, which correlates with the position of the Pleistocene antecedent platform. The lagoon has not mimicked the substrate morphology. •
INTRODUCTION
The processes of coral reef sedimentation are broadly understood, with the general concept of coarse fore-reef debris, medium-grained reef-top material, and finer sediment accumulation in the back-reef and lagoons. However, most investiga tions of lagoonal sedimentary processes have only dealt with surface sediments, with little information on the development of lagoon sediments through time. On Heron Reef the lagoon surface sediments have been investigated by Maxwell et al. (1961, 1964), Maxwell (1973), Jell & Flood (1978) and
Groves (1993). The present study discusses the nature of lagoon sediment accumulations through time, and relates this to the overall growth history of the reef. The methods used in this study were vibrocoring to collect relatively undisturbed cores, and high resolution seismic profiling across the reef to pro vide subsurface information from which a clearer understanding of reef development may be ob tained. These methods have not previously been used together on individual reefs.
1Present address: Australian Antarctic Division, Depart ment of Environment, GPO Box 252-80, Hobart, Tas. 700 I, Australia.
HERON REEF MORPHOLOGY
Heron Reef lies within the Capricorn Group c.
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
281
282
B. T Smith, E. Frankel & J. S. Jell
150 km east of Rockhampton, Queensland (Fig. 1). This group, together with the Bunker Group, form the southernmost development of the Great Barrier Reef Province (GBRP). It is an enclosed elongate lagoonal platform reef system (Maxwell, 1968) c. 9. 5 km by 4. 5 km (Fig. 2). It has a well-developed windward rim to the south, and a less well developed irregular leeward rim. A vegetated sand cay (800 m by 300 m) is situated at the western margin. The reef exhibits concentric zoning from reef-slope to back-reef environments as described by Maiklem ( 1968). The lagoonal area is a complex region of varying water depth and sediment thick ness, with two main subdivisions-the deep and shallow areas (Fig. 2). The deep lagoon is characterized by a water depth of 3 m below mean low water (bmlw), and sediment thickness of at least 5 m as demonstrated by vibro coring. It has numerous large patch reefs forming individual structures. A small lagoon exists in the north-east area (herein referred to as the north-east lagoon}, separated from the deep lagoon by a coral-algal ridge running east to west, perpendicu lar to the reef edge (Fig. 2). To the north-west there
CORAL SEA
Tropic of
is an extension of the deep lagoon. The shallow lagoon is typically 1 m deep with sediment accumu lations up to 4 m thick. It is further divided into areas of sporadic coral cover and areas with no coral growth. The outer reef slope has two distinct terraces: one at 4-6 m bmlw and another between 15 and 20 m. The latter marks the base of the reef slope, where there is an abrupt step from the steep reef structure to a gently sloping sediment floor. This is thought to be a Pleistocene foundation, as it corresponds to the depth of the solution unconformity in the Heron Island Borehole described by Davies (1974). The terrace at 4-6 m has not been addressed, although it: was previously noted by Maiklem (1968) and Jell&: Flood ( 1978).
METHODS OF INVESTIGATION
Seismic profiling
Seismic profiling across Heron Reef was undertaken using a high-resolution boomer system. One com-
North TryonI Broomfield North ..,. Wreck West Wils6n Wistari H�t Sykes s n eOne Tree Mas��:J :; -...._: Polmaise 1 • Fitzroy Lamont 41' Llewellyn e Boult Hoskyn • Fairfax Lady Bunker Musgrave
Capricorn Group
•
•
•
•
Capricorn
Group
QUEENSLAND
0
Km
500
Fig. 1. Location of Heron Reef, within the Capricorn Group of the southern Great Barrier Reef Province.
Lady • Elliot
283
Lagoonal sedimentation and reef development
Northeast lagoon
C:;;:���=7Z----J-/- Remnant
coral-algal rim Coral-algal reef rim
Legend
lagoon with numerous patch reefs D - Deep lagoon with occasional patch reefs - Deep D
A Shallow lagoon with occasional patch reefs Shallow lagoon with no patch reefs
D � Hll
Reef flat Vegetated sand clay Vibrocore site
Fig. 2. Heron Reef map indicating reef zones and vibrocore sites, with the position of cross-section (A-A') marked (refer to Fig. 5). Note the position of the north-east lagoon, separated from the deep lagoon by a coral-algal ridge located perpendicular to the present reef rim. The lagoon system has deep and shallow areas with varying amounts of patch reef cover.
plete west-east profile was obtained (Fig. 3), with several partial profiles from most areas of the reef. These records have not been reprocessed in any way. Reef traversing was invariably hampered by choppy sea conditions and time constraints im posed by the tides. All data were collected during the spring tides of November 1992-1994 (inclu sive) to allow maximum time within the lagoon for sediment sampling and geophysical investigation. Sediment sampling
Twelve vibrocores of 75-mm diameter were taken from the shallow lagoon, two from the north-east lagoon, and 24 from the deep lagoon (Fig. 2). At three sites multiple cores were taken: two at H1 and H2, and eight at H55 (the reasons for this are contained in a separate study). Compaction was calculated as a percentage using the difference in
lagoon floor level to the level of sediment within the tube before extraction. Compaction values ranged from 3 to 30%. Penetration depths ranged from 580 mm to 5440 mm. The cores were sampled at 200-mm or 500-mm intervals, depending on the nature of the material, giving a total of 328 samples. These materials were wet-sieved into gravel, sand and mud fractions, and classified according to the method of Folk ( 19 54). Selected samples were radiocarbon dated to give some indication of the timing of lagoonal filling.
RESULTS
Seismic profiles
Seismic profiling across Heron Reef shows two distinct reflectors (Fig. 3). The upper reflector is the
284
B. T. Smith, E. Frankel & J. S. Jell
(a)
w
E
/Water
surface
0 �
�
� Water - sedlffient interface
-
Extent of Heron Lagoon
/
Pleistocene foundatio
-10 -15
I
I 0
-5
2000
1000
Horizontal scale (m)
Traverse location sketch Heron Reef
water-sediment interface, and accurately delineates the position of the lagoon between 0 and 3 m bmlw (Fig. 3a). The lower reflector has a constant appar ent dip of c. 0.03° to the east at around 14 m bmlw (corrected for non-vertical ray paths) near the cen tre of the deep lagoon (Fig. 3a). This angle is likely to be the true dip, as profiles from south to north across the reef are horizontal, indicating strike direction. This reflector is interpreted to be the Pleistocene surface on which the reef has grown, as it coincides with the solution unconformity de-
I
-
_ 20
-
_25
-
_ 30
-
_3
3000
5
Fig. 3. Seismic profiles across Heron Reef showing the water-sediment interface and the Pleistocene foundation. (a) West-east profile across Heron Lagoon. The extent of the lagoon is indicated by the upper reflector and the Pleistocene foundation by the lower reflector. It should be noted that multiples have not been removed. (b) (Opposite) West-east profile of the eastern reef slope of Heron Reef and the western reef slope of Sykes Reef. The upper reflector is the water-sediment interface of the inter-reef area, and the lower reflector is the Pleistocene foundation, dipping constantly across the profile.
scribed by Davies (1974). Harvey (1986) used seismic refraction techniques around Heron Island to locate the pre-Holocene surface. His results confirmed those of Davies (1974) that this surface is at c. 12 m bmlw in the vicinity of the cay. The present investigation confirms both of these previ ous studies; the platform is at 12 m bmlw at the western margin of the reef, and at 16 m at th<e eastern margin. At a scale of a few tens of metres this surface has an undulating form (Harvey, 1986); however, across the 9.5 km width of the foundation
285
Lagoonal sedimentation and reef development (b)
w
E
Water Depth (m) 0
Water surface
-5
Water - sediment interface
-10
-15 Pleistocene foundation
-20 -
25
-30
2000
1000
3000
Horizontal scale (m)
Traverse location sketch Heron Reef
Fig. 3. (Continued).
there is an overall dip of 0.03 The seismic data lack the resolution to define any internal textural discontinuities within the Holocene sediments. • .
Sediment composition
Almost all of the sediment within the lagoon system is calcareous skeletal material. Trace amounts of siliceous sponge debris represent the only non calcareous material present. Jell & Flood (1978) used multivariate analytical techniques on surface sediments at Heron Reef in an attempt to create mappable component facies. They concluded that this was not possible because
Sykes Reef
of the compositional similarity of the sediments. Other penetrative lagoonal sediment studies within the GBRP include those of Kiene ( 1983), on One Tree Reef in the Capricorn Group, and Tudhope (1983), on Davies Reef in the central GBRP. Neither study revealed any mappable composi tional units. Grain counts of the gravels and 1/z<j> sand fractions larger than 1. 5 <1> show that the major components in all samples are fragments of corals, molluscs and calcareous algae. Minor amounts of Halimeda, Foraminifera, crustaceans, agglutinated worm bur rows, echinoderms, sponge spicules, fine-grained aggregates (probably faecal pellets), fish bones and
286
B. T. Smith, E. Frankel & J. S. Jell Gravel t:::o,/'"%
( a)
Sand
25
50
75
Mud.
1oo
Sand
25
H4
Sand
"' ����wu�.-.-r.-�.-.-r�.-r.-�o 25
5o
50
75
H5
75
1oo
Mud
Sand
25
50
1oo
75
10o
75
100
Mud
Mud
H9
H8
Gravel
o....._...-q,
Sand
25
50
Hll
75
1oo
Mud
Sand
25
50
H56
Mud
287
Lagoonal sedimentation and reef development
worm tubes are also present. This study confirms that the composition of all sediments is similar throughout the lagoonal system; however, prelimi nary investigation indicates that there may be broad component variation over sediment depth. This aspect of the study is at present under investi gation. Sediment texture
Previous studies on surface sediments from Heron Reef show the relationship between sediment size and the energy regime across the reef. Coarse material accumulates in the high-energy zones around the reef edge and across the reef flat envi ronments, and finer material is deposited in the lower-energy lagoons. The material from the cores confirms these relationships, with coarse sand accu mulations in the shallow lagoon and fine sands and muds in the deep lagoon. However, these relation ships do not remain constant throughout the sedi ment pile. The gravel/sand/mud ratios for each sample have been plotted on textural ternary diagrams (exam ples are presented in Fig. 4a). Each diagram repre sents all the samples from a particular core, with depth labels downhole (uncorrected for compac tion) connected with a single line to indicate the progressive change in texture with depth. There are three main observations on these data:
Fig. 4. (a) (Opposite) Representative core samples showing textural changes down-hole: H4, 8 and H9 are examples of deep lagoon vibrocores, and H5, HI! and H56 are examples of shallow lagoon vibrocores. (Refer to text for explanation of distribution patterns.) (b) Ternary textural classification system with respect to the gravel, sand and mud ratios based on the method of Folk ( 1954). The groupings observed in the plots of data points for each hole have led to classification within three facies: Facies !-sand and gravelly sand; Facies 2-muddy sand and sand; Facies 3-gravelly muddy sands, muddy sandy gravels, with occasional muddy gravels and gravelly muds.
1 There are one or two distinct groups of data points in every core. 2 Where there is one group present, the data points are concentrated along or close to the zero mud axis, and are sand dominated. 3 Where there are two groups present, the group of data points representing the upper core samples always plots on the edge of the diagram, either towards zero gravel, or zero mud, and is always sand dominated. The second group represents sam ples towards the base of the cores and is a less well sorted group which plots towards the centre of the diagram; usually above 20% for each component. For ease of classification these groups have been defined as facies within the method of Folk (1954) (Fig. 4b, Table 1). Facies 1 represents those samples which group along or close to the zero mud axis within the sand (S), gravelly sand (gS) and sandy gravel (sG) areas (Fig. 4a, cores H5, H l l and H56). Facies 2 represents the samples which group along or close to the zero gravel axis within the muddy sand (mS) or sand (S) areas (Fig. 4a, cores H4, H8 and H9). Facies 3 represents the samples which loosely group towards the centre of the diagram, within the areas of gravelly muddy sand (gmS), muddy sandy gravel (msG), muddy gravel (mG) or gravelly mud (gM) (Fig. 4a, cores H4, H8, H9 and H56). No samples plotted within the sandy mud (sM), mud (M), or gravel (G) areas of the diagram. Having designated each group of data points to a
Gravel
(b)
G- gravel S - sand M- mud
IB
Facies 1
1111!11
Facies 2
g - gravelly s - sandy m- muddy
�
Facies 3
CD
Facies 1
gM
Sand
Mud
or
2
Table 1. Textural classification of all sediment samples using the method of Folk ( 1954) (Fig. 4), giving core numbers, sample numbers, depth of core represented (corrected for compaction), ternary textural classification and facies allocation Texture
Texture
Hole
Sample
Depth down-
(Folk,
Textural
Hole
Sample
Depth down-
(Fol k,
Textural
number
number
core (mm)
1954)
facies
number
number
core (mm)
1954)
facies
I or 2
1-2
0-387
s
3
387-748
gS
4-5
748-1109
sG msG
HIa
1-5
840
Hlb
1-6
1050
H2a
1-8
0-1923
s s s
9-10
1923-2242
gS
11-14
2242-3935
H2b
H3 H4
HS
H6
1-5
0-940
s s
6-7
940-1524
gS
8-18
1524-4318
s
I or 2 I I or 2 I or 2 I
19-20
4318-5440
gS
I
1-18
0-3605
mS
2
19
3605-4315
gmS
3
1-9
0-1665
mS
2
10-17
1665-4320
gmS
1-3
0-715
4-5 1-5
gmS
3-8
366-1952
mS 3
1-14
0-3870
mS
2
H27
1-4
0-735
mS
2
5
735-945
gmS
3
sG
6-8
945-1575
mS
3
715-1058
gS
9-10
1575-3010
gmS
3
0-1342
s
2
I
0-146
gS
2
146-878
sG
3
878-1300
I
0-297
gS
2
297-976
sG
3-4
976-2261
gS
5-6
2261-3225
gmS
H53
I
0-270
s
2
270-580
sG
H54
1-4
0-2196
gS
5-6
2196-3230
gmS
I
0-540
s
2
2
540-1040
gmS
3
3
1040-1620
msG
3
4
1620-2309
mG
H55b
1-5
0-2542
gmS
2
H55c
I
0-456
mS
2
2-5
456-2679
gmS
3
6
2679-3215
msG
2
8-10
1708-2400
gmS
3
H8
1-6
0-1512
mS
2
7-9
1512-2205
gmS
3
10
2205-2394
mG
3
HIS
0-244 244-366
H26
3
Hl7
mS
I 2
gmS
gmS
HIS
sG
msG
mS
Hl4
132-930
2440-2940
1342-4400
Hl3
gS
2-4
1952-2440
0-1708
H12
1109-1350 1-132
9-11
1-7
H11
H25
6 I
12
6-17
HIO
H24
I or 2
H7
H9
H22
I or 2
11
2394-2671
gM
3
12
2671-2797
mG
3
13
2797-3130
msG
1-6
0-1298
s
2
7-8
1298-1700
gmS
2
9-10
1700-2600
msG
3
I
0-126
mS
2
126-378
s
2
HSI
H52
H55a
2
3
378-430
sM
3
4
430-882
mS
3
5
882-1273
gmS
6-7
1273-1764
msG
8
1764-1950
gmS
H55d
I
0-74
gS
2-3
74-713
sG
4-10
713-2226
gS
1-7
0-1395
mS
2
8-10
1395-3080
gmS
3
1-8
0-1695
mS
2
9-16
1695-3503
gmS
3
17-18
3503-4240
msG
3
1-7
0-1417
mS
2
8-9
1417-1853
gmS
3
10
1853-2050
msG
3
I
0-134
s
2-10
134-2500
sG
1-2
0-268
sG
3
268-710
gS
I
H55e
s
I
0-590
mS
2
2
590-1180
gmS
3
3
1180-1345
msG
3
I
0-836
mS
2 3
2-4
836-1815
gmS
H55f
1-3
0-1391
mS
4-7
1391-4080
mG
H55g
1-4
0-1856
mS
1856-2390
msG
3
2390-3364
gM
3 3
H55h
H56
3364-3770
mG
I
0-270
msG
3
2-4
270-1879
gmS
3 3
5-6
1879-2808
7
2808-3402
g)\1 mG
1-2
0-848
sG
3-4
848-1740
msG
5-6
1740-3028
gmS
7-8
3028-4040
msG
1-7
0-1504
mS
2
8
1504-1744
gmS
3
9-11
1744-2343
mS
2
12
2343-2442
msG
3
I
0-335
gS
13
2442-2703
mS
2
2-3
335-2500
sG
14-16
2703-3430
msG
3
4-5
2500-3040
msG
H57
2
3
289
Lagoonal sedimentation and reef development
facies type, it is then possible to observe a distribu tion pattern of the facies both laterally and vertically through the lagoonal sediment pile. The shallow la goon sediments always contain Facies 1 within the uppermost 1800 mm. This unit has no internal stratigraphy and generally coarsens with depth. Six cores from the shallow lagoon did not penetrate be yond this unit. The remaining six cores penetrated up to 3500 mm and intersected Facies 3. The tran sition from Facies 1 to 3 is gradational over a few centimetres at around 1800 mm. The deep lagoon sediments usually contain Facies 2 in the uppermost 1500 mm of core. At this depth there is an abrupt transition in most cores to Facies 3. At the base of several cores penetration was halted by large lumps of coral rubble. It is unlikely therefore that the base of Facies 3 was reached. The sediments of the north-east lagoon, and north-west extension of the deep lagoon, do not follow the general sediment accumulation patterns. In these areas the sediments are sandy throughout and are ascribed to Facies 1 or 2. Sediment ages
Ten samples have been radiocarbon dated (Table 2). Seven of these were taken from the base of the cores, and three from intervals above the base. The base of Facies 3 in the deep lagoon cores (H4, H6, H8 and H18) is dated between 3020 and 3900 yr BP. The age of the base of Facies 3 in the shallow lagoon is 4 180 yr BP. Samples from the base of Facies 1 and 2 yield ages between 2380 and 3100 yr BP. All ages are reported as conventional uncalibrated carbon years before present (BP).
REEF GROWTH HISTORY
Initial development of Heron Reef occurred in response to inundation across the planar substrate detected by seismic investigation to be between 12 and 16 m bmlw. This is estimated to have occurred around 9000 yr BP (Davies, 1977; Hopley, 1982; Marshall & Davies, 1982; Davies & Hopley, 1983) at a rate of c. 6-7 mm yc1• Sea-level continued to rise until about 6000 yr BP (Thorn & Chappell, 1975; Hopley, 1982), when it may have been 1 m higher than its present level (Flood, 1983; Thorn & Roy, 1983). Stabilization at the present level oc curred between 5000 and 2000 yr BP (Flood, 1983; Larcombe et at., 1995). Direct evidence for the sea-level history on Heron Reef has not yet been obtained. The detailed nature of reef growth behaviour at Heron Reef between 9000 and 5000 yr BP remains unclear. However, it has been established by Hopley (1982) and others that reef growth commonly lagged behind sea-level rise. Given the present morphology of Heron Reef, it is suggested that the windward side accreted before or more rapidly than the leeward side, thereby reaching sea-level first. Evidence for this lies in the presence of the coral algal ridge towards the north-east corner of the reef, which runs perpendicular to the present coral-algal reef rim (Fig. 2). This ridge may be the remnant of a previous leeward reef rim position, before the rest of the leeward side reached sea-level. Distribution and timing of lagoonal sedimentation
The interval between the Pleistocene foundation
Table 2. Radiocarbon ages in conventional radiocarbon years before present (BP); the rate of sediment accumulation is calculated from the depth of core penetration (corrected for compaction) and radiocarbon ages
Sample number (hole numbersample number)
Penetration depth (mm)
Carbon dated yr BP
General lagoonal position of hole
Rate of sediment accumulation (mm yr-1)
Facies
H2b-20 H3-16 H4-17 H6-17 H8-13 H15-IO H18-16 H56-4 H56-6 H56-8
5440 4315 4230 4400 3130 2500 3430 1695 2825 3921
2860 3100 3850 3020 3890 2400 3900 2380 3580 4180
NE lagoon Deep Deep Deep Deep Shallow Deep Shallow Shallow Shallow
2.10 1.47 1.1 1.55 0.8 1.21 0.88 0.71 0.94 1.83
2 2 3 3 3 I 3 I 3 3
290
B. T. Smith, E. Frankel & J. S. Jell
and the base of the cores is c. 8 m (between 8 and 16 m bmlw). The nature of sedimentation within this interval is uncertain at present; it is possibly unconsolidated detrital lagoonal infill, cemented framework-like material, or a combination of these. The depth of the lagoon before sediment infill began is therefore unknown; however, minimum depths and ages have been established. Vibrocore H56 (Figs 2 & 4a) is of particular significance because of its position close to the windward margin within the shallow lagoon area. It yielded some of the deepest and oldest sediments recovered from the lagoon; at 4040 mm the age is c. 4200 yr BP (Table 2). Thus lagoon infilling began before this time. There are two distinct phases of lagoon filling: one which began before 4200 yr BP when Facies 3 began to accumulate, the other starting around 2700 yr BP with concurrent deposition of Facies I to the windward and Facies 2 to the leeward (Fig. 5). It is suggested that the lagoon was an open system until around 2700 yr BP, as there appears to be no sorting of sediments within Facies 3 to suggest that the reef had created a barrier to the open water. When the windward margin reached sea-level at the beginning of the second phase, the hydrodynamic regime within the lagoon system changed; the mar-
gin formed a barrier to the open ocean and pro vided sediment which was transported and sorted in the back-reef environment. Facies 1 is restricted to the sand sheet which forms the floor of the shallow lagoon to the wind ward side of the reef. The lack of coral communities and coarse texture of this unit indicates that this is an active surface apparently unsuitable for biotic colonization or bioturbation. This unit is currently prograding from the windward margin towards the centre of the reef, which is a common feature of lagoonal platform reefs for this area (refer to Orme et a!., 1974; Davies et a!., 1976). Marshall & Davies (1982) defined this feature as a 'prograding sand facies'. Facies 2 is located in the uppermost sediments of the deep lagoon and in the north-east lagoon. This facies is restricted to the leeward side of the reef and is considered to be the lower-energy equivalent of Facies I. Marshall & Davies (1982) referred to this style of sedimentation as a 'muddy sand facies'. It is heavily bioturbated and laterally interrupted by patch reefs. Facies 3 is distributed across most of the lagoonal system. It predominantly contains gravelly muddy sand, or muddy sandy gravel. There does not
Leeward
Windward (a)
m
A'
A sea level
/
£'f!:!·.!j·��J;Jr,;KWA
...----
(b)
m
--
A'
A
Key
�
Facies!
reef framework
-
Facies2
Patch Reef
-
Facies3
Indefinite boundary
Km
[2] �
Holocene
Fig. 5. Cross-sections from windward to leeward (A-A' marked in Fig. 2) across the lagoon of Heron Reef. (a) Facies 3 was deposited in the lagoon before the margins reached sea-level. The windward side reached sea-level around 2700 yr BP. (b) The coincidence of sea-level with the windward margin changed the sedimentation style within the lagoon. This resulted in the sorting of sediments into the coarse sand of Facies I behind the windward margin and the muddy sand of Facies 2 towards the leeward margin.
Lagoonal sedimentation and reef development
appear to be a clear textural differentiation within Facies 3 which defines variation in texture from windward to leeward, although its absence is noted within the isolated lagoonal systems. It is therefore assumed that the reef margin was below sea-level during this period of deposition.
DISCUSSION
Purdy ( 1974a,b), Hopley (1982), Harvey (1986) and others have suggested that present lagoonal morphology of reefs such as Heron Reef is the direct result of mimicking of a karstic Pleistocene surface. It is thought that the lagoons mimic depressions either from pre-existing lagoons in a Pleistocene reef or from a centrally eroded limestone basin within a platform. Patch reefs are believed to have grown on high features within these basins. There appears to be some evidence to support this on One Tree Reef (Davies & Hopley, 1983). The west-east profile presented in Fig. 3 does not show this relationship; the Pleistocene reflector is seen to remain planar below the irregular reflector of the lagoon floor. The present lagoonal morphol ogy in this case is therefore not related to an irregular substrate. Patch reef development is also not explained in terms of initiation on topographi cal highs of the antecedent platform. There may be irregularities apparent in the seismic profiles be neath patch reefs (Fig. 3a); however, these are most likely to be pull-up features caused by the patch reef relief. The thickness and style of initiation of patch reefs are therefore unknown. Vibrocore penetration within the shallow lagoon area provided two types of core: some penetrated several metres and intersected Facies 3, whereas others terminated abruptly on a hard substrate, and only intersected Facies 1. It is believed this was caused by interruption by buried patch reefs. The prograding sand sheet of Facies 1 has covered any patch reefs which may have developed on the windward side, and is in the process of burying those in the transitional area between the sandy shallow lagoon and the deep lagoon. The deeper cores of the shallow lagoon and all cores from the deep lagoon intersected Facies 3. It is concluded that patch reef development began before deposi tion of Facies 3, and patch reefs were established well enough not to be buried by this sedimentation. Three-dimensional lagoon sedimentation in the central GBRP has been studied in some detail by
29 1
Tudhope ( 1983) and Scoffi.n & Tudhope (1988). The degree of correlation between the central and southern regions remains to be ascertained; how ever, a preliminary comparison indicates that the patterns of lagoonal sedimentation are similar. Scoffi.n & Tudhope (1988) suggested a model for complete lagoonal sedimentation based on predic tions from their study on Davies Reef (Fig. 6). The sediments of Heron lagoon correlate with the lower subtidal deposits of their model. This investigation, however, has defined a clearer differentiation be tween the leeward and windward lagoonal sedi ments. It is probable that each individual reef responds to its local environment with different growth styles and sedimentation rates. The differences observed between Davies and Heron Reefs implies the exist ence of individual sedimentation styles within broadly similar models. The differences are as sumed to develop as a result of the timing of marginal reef growth and its relationship across the reef to the height of sea-level during the transgres sion. Marshall & Davies (1982) established that the windward margin of One Tree Reef reached sea level c. 5000 yr BP. This is significantly earlier than the estimations for Heron Reef. Confirmation with radiocarbon ages of reef-top materials is required to clarify the timing of reef growth events. A comparison can also be made between the growth history of Heron Reef and that of Lady Elliot Reef at the southern end of the Bunker Group. Shingle ramparts on this reef have been dated to 3200 yr BP (Chivas et a!., 1986). This age correlates with the accumulation period for Facies 3 on Heron Reef. As there are no similar deposits on Heron Reef, they cannot be directly compared; however, the age of the ramparts confirms that the reef structure reached its current elevation by at least 3200 yr BP. Each individual reef appears to have developed with slightly different rates and styles depending, for example, on the foundation morphology, proximity to the shelf edge, water circulation and nutrient supply. This would pro duce slightly different growth curves for each reef.
CONCLUSIONS
The reefs of the southern GBRP in its present form are growing on a Pleistocene foundation (Davies, 1974; Marshall, 1983a,b). The nature of the surface
292
B. T. Smith, E. Frankel & J. S. Jell
LEEWARD
(a)
WINDWARD
Phosphatic crust Horizontally bedded
clay s ands
(grainstone)
Soil horizons with rhizoids
Bassett edge and phytokarst truncated surfaces
�������!
Coral shingle with Mg calcite micritic
cements
Cross-bedded beaches and beachrocks with aragonite cement plus intraclasts and encrustations (grainstone)
Graded, imbricated, cross-bedded coral s hingle (rudstone) Local in-situ framework esp. microatolls (framework)
Bioturbated (wackestones) and locally ripple and cross-bedded sands
Increase in mean grain size
(grainstone)
Stormlayers
Increase in reef rim
(floatstone)
components
Increase in sorting Bioturbated, poorly sorted, gravelly, muddy sands (wackestone and floatstone)
o .I I.
• D
� t) '
L
�
t
Heron Lagoon sediments
Basal lag of coarse epilithic gravel (rudstone) -
In-situ fram ework (framestone) - -
;: :: Ho
e
�'7-.,---�.---.--i
�
Fig. 6. (a) Lagoon sedimentation model derived from a study of Davies Reef, central GBRP (after Scoffi.n & Tudhope, 1988). Heron lagoon sediments correlate with the lower subtidal deposits. (b) (Opposite) Lagoon sedimentation model for Heron Reef, indicating the distribution and texture of Facies 3 across the entire lagoon system, Facies 2 to the leeward side, and Facies I to the windward side (G, gravel; S, sand; M, mud).
beneath Heron Reef is planar and dips gently to the east. Sea-level stabilized near its present position c. 6000 yr BP (Hopley, 1982). Lagoonal deposition began before the reef margins reached sea-level, with deposition of Facies 3 from before 4200 yr BP. Facies 1 and 2 were deposited once the windward margin reached sea level c. 2700 yr BP. The leeward
margin reached sea-level during deposition of Facies 1 and 2.
ACKNOWLEDGEMENTS
This project was funded by an Australian Research Council Grant (ARC A3921123) awarded to E.
293
Lagoonal sedimentation and reef development (b)
LEEWARD s
WINDWARD
M
-
Facies 2
0
Facies I
0 ==
-------------
Heron Lagoon
0 --
0
Facies 3
-0 <=::; Fig. 6. (Continued).
Frankel and J. Jell during 1992-1994, and an internal research grant from the University of Tech nology, Sydney to E. Frankel. B. Smith is supported by an Australian Postgraduate Award. Deepest thanks are due to the numerous field assistants who have helped with the sample collection over three major field trips. We thank also Steve Hearn and Nick Sibbald (University of Queensland) for the collection of geophysical data, and the Geological Survey of Queensland for loan of the geophysical equipment. The staff of Heron Island Research Sta tion are thanked for their co-operation and support. Appreciation is extended to various reviewers of this manuscript, particularly John Marshall of the Aus tralian Geological Survey Organisation.
REFERENCES
CHIVAS, A., CHAPPELL, J., POLACH, H., PILLANS, B., & FLOOD, P. (1986) Radiocarbon evidence for the timing and rate of island development, beach rock formation, and phosphatization at Lady Elliot Island, Queensland, Australia. Mar. Geo/., 69, 273-287. DAVIES, P.J. (1974) Subsurface solution unconformities at Heron Island, Great Barrier Reef. Proceedings of the 2nd International Coral Reef Symposium, Brisbane, Qld, pp. 573-578. DAVIES, P.J. (1977) Modern reef growth-Great Barrier Reef. Proceedings of the 3rd International Coral Reef Symposium, Miami, FL, pp. 325-330. DAVIES, P.J. & HOPLEY, D. (1983) Growth fabrics and growth rates of Holocene reefs in the Great Barrier Reef. Bur. Min. Res. J Aust. Geo/. Geophys., 8, 237251. DAVIES, P.J., RADKE, B.M. & ROBISON, C.R. (1976) The evolution of One Tree Reef southern Great Barrier
0
t
sediments
0 Facies 3 �
� � -
�
Reef, Queensland. Bur. Min. Res. J. Aust. Geo/. Geo phys., I, 231-240. FLOOD, P.J. (1983) Holocene sea level data from the southern Great Barrier Reef and southeastern Queensland-a review. In: Australian Sea Levels in the Last 15 000 years: a Review (Ed. Hopley, D.), James Cook University of Northern Queenland, Townsville, Monogr. Ser. Occas. Pap., 3, 85-92. FOLK, R.L. (1954) The distinction between grain size and mineral composition in sedimentary rock nomencla ture. J Geol., 62, 344-359. GROVES, C.D. (1993) Morphological zonation and sedi ment distribution of the eastern part of Heron Reef BSc honours thesis, University of Queensland, Brisbane. HARVEY, N. (1986) The Great Barrier Reef Shallow Seis mic Investigations. Department of Geography, James Cook University of Northern Queensland, Townsville, Monogr. Ser., 14. HoPLEY, D. (1982) The Geomorphology of the Great Barrier Reef Quaternary Development of Coral Reefs. John Wiley, Brisbane, Qld. JELL, J.S & FLOOD, P.J. (1978) Guide to the geology of reefs of the Capricorn and Bunker Groups, Great Barrier Reef Province, with special reference to Heron Reef Austra lian Sedimentologists Group, Department of Geology, University of Queensland, Brisbane, 8. KIENE, W.E. (1983) Spatial and temporal aspects of la goonal sedimentation at One Tree Reef southern Great Barrier Reef Australia. MSc thesis, University of Syd ney, NSW. LARCOMBE, P., CARTER, R.M., DYE J., GAGAN, M.K. & JOHNSON, D.P. (1995) New evidence for episodic post glacial sea level rise, central Great Barrier Reef, Austra lia. Mar. Geo/., 127, 1-44. MAIKLEM, W.R. (1968) The Capricorn Reef complex, Great Barrier Reef, Australia. J sediment Petrol., 38, 785-798. MARSHALL, J.F. (1983a) Submarine cementation in a . high-energy platform reef: One Tree Reef, Southern GBR. J sediment Petrol., 53, 1133-1149. MARSHALL, J.F. (1983b) Lithology and diagenesis of the
294
B. T Smith, E. Frankel & J. S. Jell
carbonate foundations of modern reefs in the southern GBR. Bur. Min. Res. J. Aust. Geol. Geophys. 8, 253-265. MARSHALL, J.F. & DAVIES, P.J. (1982) Internal structure and Holocene evolution of One Tree Reef, southern Great Barrier Reef. Coral Reefs, 1, 21-28. MAXWELL, W.G.H. (1968) Atlas of the Great Barrier Reef Elsevier, Amsterdam. MAXWELL, W.G.H. (1973) Sediments of the Great Barrier Reef. In: Biology and Geology of Coral Reefs, Vol. 1 (Eds Jones, O.A. & Endean, R.), pp. 299-345. Academic Press, New York. MAXWELL, W.G.H., DAY, R.W. & FLEMING, P.J.G. (1961) Carbonate sedimentation on the Heron Island Reef, GBR. J. sediment Petrol., 31, 215-230. MAXWELL, W.G.H., JELL, J. & McKELLAR, R.G. (1964) Differentiation of carbonate sediments in the Heron Island Reef. J. sediment. petrol., 34, 294-308. 0RME, G.P., FLOOD, P.G. & EwART, A. (1974) An investi gation of the sediments and physiography of Lady Musgrave Reef-a preliminary account. Proceedings of the 2nd International Coral Reef Symposium, Brisbane, Qld, pp. 371-386.
PURDY, E.G. (1974a) Reef configurations: cause and effect. In: Reefs in Time and Space (Ed. Laporte, L.F.), Spec. Pub!, Soc. econ Paleont. Miner., 18, 9-76 PuRDY, E.G. (1974b) Karst-determined facies patterns in British Honduras: Holocene carbonate sedimentation model. Bull. Am. Assoc. petrol. Geol., 58, 825-855. SCOFFIN, T.P. & TUDHOPE A.W. (1988) Shallowing upward sequences in reef lagoon sediments: examples from the Holocene of the Great Barrier Reef of Australia and the Silurian of Much Wenlock, Shropshire, England. Pro ceedings of the 6th International Coral Reef Symposium, Townsville, Qld, 3, 479-484. THOM, B. G. & CHAPPELL, J. (1975) Holocene sea levels relative to Australia. Search, 6, 90-93. THOM, B.G. & RoY, P.S., (1983) Sea level change in New South Wales in the last I 5,000 years. In: Australian Sea Levels in the Last 15,000 years: a Review (Ed. Hopley, D.), James Cook University of Northern Queensland, Townsville, Monogr. Ser. Occas. Pap., 3, 64-84. TUDHOPE, A.W. (1983) Processes of lagoonal sedimentation and patch reef development, Davies Reef, Great Barrier Reef of Australia. PhD thesis, University of Edinburgh. _
Spec. Pubis int. Ass. Sediment. ( 1998) 25, 295-310
Terrigenous sediment accumulation as a regional control on the distribution of reef carbonates K. J. W O OLFE and P. L A R C O M B E Marine Geophys ical Laboratory, School ofEarth Sciences, James Cook University, Townsville, Qld 4811, Australia
ABSTRACT
Sediment supply to coastal oceans is a primary factor in controlling the occurrence of offshore and nearshore carbonate provinces, particularly coral reefs. Sediment supply to the inner shelf of the central Great Barrier Reef and to the Gulf of Papua is significant, and varies temporally and spatially by several orders of magnitude, as do sediment fluxes of material on the shelf itself. However, coral reefs are common in turbid waters along the inner shelf of the Great Barrier Reef, located where suitable substrates are available. In Halifax Bay, coastal turbid-zone reefs have developed on stable gravel substrates, in water depths of less than 4-5 m. These environments are subject to high rates of coastal sediment transport. Near-bed turbidities of up to 140 nephelometric turbidity units (NTU; equivalent to c. 140 mg 1-1) occur in a wave-dominated coastal boundary layer. In such locations, reef initiation and survival appears to be more strongly controlled by sediment accumulation, and hence substrate availability, than by water quality or sediment flux past the reefs. Sediment accumulation is, in tum, partly controlled by local oceanography and sediment transport.
INTRODUCTION
detrimental for coral reef commumties (Adey, 1 978; Rogers, 1 990; Stafford-Smith et a!., 1994; Anonymous, 1995). The global river-derived terrigenous sediment supply to the worlds' oceans has been estimated as 1 3.5 x 109 t yr-1 (Milliman & Meade, 1 98 3). If this sediment were uniformly distributed throughout the global ocean, the background terrigenous (river derived) flux (BTF) of material to the sea-floor would be c. 45 g m-2 yc1 ( 1 BTF). This globally averaged value provides a convenient measure against which to compare the terrigenous sediment fluxes of different regions. Truly oceanic corals, such as those living on coral atolls, probably ex perience terrigenous sediment accumulation rates significantly below 1 BTF, whereas corals on conti nental shelves may receive tens to hundreds of BTFs because of the input of riverine sediment to the shelf and its distribution by oceanographic processes. We review some recent data from the Great
Coral reefs are generally associated with clear blue water and pristine white sandy beaches, driving public perceptions of some environmental issues, and probably also influencing interpretation of the geological record. However, many modern coral reefs in inner-shelf locations experience prolonged, if not permanent, inundation by highly turbid water (> 1 00 mg 1-1 of total suspended solids, Larcombe et a!., 1995b; Larcombe & Woolfe, 1 997). In this paper we discuss some sedimentary controls on such shelf-reef carbonates. We emphasize the role of 'coastal turbid-zone reefs' (Larcombe & Woolfe, 1 997), not because they necessarily dominate the modern distribution of coral reefs, or the geological record, but because the historical paucity of their biological and geological documentation, in combi nation with their potential significance as pioneer ing reefs under rising sea-levels, requires that they receive an increased emphasis. Our observations temper the widely publicized view that high levels of sediment supply to the coastline are necessarily
=
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
295
·
296
K. J. Woolfe & P. Larcombe
Barrier Reef (GBR) region and conclude that sedi ment accumulation is a major limiting factor on the development of some coral reefs. Water turbidity alone is, in most cases, more likely to influence the particular species composition of the coral assem blage (Done, 1982; Veron, 1 995) than it is to determine the presence or absence of a coral reef.
GBR shelf (Belperio, 1978, 1983, 1 988; Johnson & Searle, 1 984;0hlenbusch, 1 99 1 ;Cartereta/., 1 99 3; Larcombe et a!., 1 995a; Larcombe & Woolfe, 1 997). Some of these data have been presented in the proceedings of a workshop held in Townsville (Larcombe & Woolfe, 1 996).
DATABASE AND
CENTRAL GREAT BARRIER REEF
THE INNER SHELF, METHODS
There are substantial sedimentological and oceano graphic datasets for the inner shelf of the central GBR province between Bowen and Ingham, from most sedimentary environments. The inner shelf is a location of high rates of sediment resuspension and, particularly in the presence of wind-driven or tidal currents, sediment transport (Belperio, 1983; Larcombe et a!., 1 995b). Sediments supplied to the coast by the Burdekin River enter Upstart Bay. From here, sediments are transported northwards along the inner shelf mostly by the effects of the SE trade winds, and material is accumulated in Bowl ing Green, Cleveland and Halifax bays. The coast parallel sediment transport path is thus more than 200 km long (Fig. 1 ).
We draw upon a database of over 1 000 h of current and wave data, 30 000 nephelometer hours of turbidity data, 3000 line-km of shallow-penetration marine seismic records (3.5 kHz and Uniboom), 250 sediment trap retrievals, 350 water samples, 500 grab samples, 1 50 short cores ( < 4 m long) and 1 200 high-resolution grain-size determinations, from the GBR shelf north of Mackay, and from the Gulf of Papua. These data have been collected for a range of completed and continuing projects, includ ing: (i) regional oceanographic, sedimentological and geological mapping of the sea-floor (e.g. Brun skill et al., 1995; Woolfe et al., 1 995); and (ii) examination of the post-glacial evolution of the
147'00' 'b .,
<£;
0
19'S
20"S
km 25
Fig. 1. The central GBR coastline, showing the Burdekin River delta, the series of northward-facing embayments to its north which form sediment traps for terrigenous sediments supplied to the coast, and the locations of . places mentioned in the text. In Halifax Bay, three coastal turbid-zone reefs are labelled.
297
Sediment accumulation and reef distribution
Regional sedimentation
The Burdekin River is by far the largest single point source of water and sediment in the central GBR province. It has a mean annual discharge of 10 x 106 Ml (Queensland Government, 1 99 3) with a maximum daily freshwater discharge of over 3.4 x 1 06 Ml day-1• Belperio ( 1 979) estimated that the river supplies 3.5 Mt of sediment annually to the shelf, over 85% as washload. More recent estimates are 2.7 Mt (Moss et al., 1 99 3) and 8.5 Mt (Neil & Yu, 1 995). These estimates are clearly very dependent upon the characteristics of the years measured, and upon the discharge rating curves used, especially because large-scale single events are so important in transporting sediment. Using seismic and core data, Way ( 1987) calcu lated the Holocene sediment volume in Upstart Bay. She demonstrated that, at the mouth of the Burdekin River, sediment has accumulated through the Holocene at an average of 0.45 mm yc1 (Table 1), or 0.2 Mt yc1• Using Belperio's ( 1 979, 1 983) figure (above) for sediment supply to the coast by the Burdekin River, this indicates that c. 95% of the material is transported northward along the coast line. On a similar basis, Carter et al. ( 1 993) esti mated that 50-70% of the sediment (mud) supplied to Cleveland Bay from the Burdekin River bypasses
the embayment, and is transported northwards alongshore. The three main embayments north wards from the Burdekin (Bowling Green Bay, Cleveland Bay, Halifax Bay) show successively de creasing rates of terrigenous infill (Table 1 ), indicat ing that their sediments are, in the long term, probably derived largely from the Burdekin River. Significant coastal progradation has occurred along many parts of the coastline since the Ho locene sea-level highstand (Hopley & Murtha, 1975; Belperio, 1983;Carteret al., 1 99 3). Around Towns ville, coastal progradation rates have been 0.53.0 m yc1• Progradation has been attributed to a slowed overall rate of sea-level rise before the highstand, and to the highstand itself (Harris et al., 1990). Progradation may also have been enhanced by the subsequent slight fall in sea-level. Further, Shulmeister ( 199 5) has speculated that the SE trade winds may also have intensified after c. 5 kyr BP, and would have contributed to embayment infilling and development of the coastal wedge at this time. The data of Belperio ( 1 98 3) demonstrate that, for the last 30 yr, the sediment accumulation rate of intertidal sediments (0.5-8.0 mm yr-1) is 5- 1 0 times that o f the inner shelf, whereas net accumu lation rates are only rarely above 0.2 mm yc1• Sediment accumulation rates for the sediment wedge in Halifax Bay are all below 0.2 mm yc1• It
Table 1. Comparison of long-term accumulation rates of sedimentary environments between the mouth of the
Burdekin River and Halifax Bay (arranged in order of distance from source), including the resulting flux of terrigenous sediment to the bed
Upstart Bay 2500-yr mean*
Sediment accumulation rate (mm yr-1)
Terrigenous fraction of sediment* (%)
Terrigenous sediment flux to the bedt (BTF)
0.7
70
20
Upstart Bay 6500-yr mean*
0.45
85
15.8
Upstart Bay and Bowling Green Bay, intertidal sediments§
0.5-8.0
95
19-300
Cleveland Bay 30-yr mean§
<0.2
85
< 6.8
<0.25
85
<8.5
Cleveland Bay 6000-yr meanll Halifax Bay 7000-yr mean§
<0.1
75
<3
Mid-shelf off Townsville 6000-yr mean�
0.03-0.2
10-30
0.12-2.4
* From data of Belperio (1978), Way (1987) and Ohlenbusch (1991). t BTF calculated assuming particle density of 2.6 and porosity of 30%, and scaled to remove biogenic carbonate. 1 BTF 45 g m-2 yr-1• *Data calculated from core and seismic data of Way (1987), using an area of 330 x 106m2. §Data from Belperio (1983). II Recalculated from Carter et a!. (1993) and P. Larcombe & R.M. Carter, unpublished data. � Calculated from core and seismic data of Ohlenbusch (1991). =
298
K J. Woolfe & P. Larcombe
is noteworthy that the mid-shelf off Townsville has accumulated only 0. 1 2-2.4 times the global average BTF over the last 6000 yr. Along the coastline, similar physical environ ments tend to have sedimentary regimes formed of common combinations of oceanographic and sedi ment transport processes, leading to similar turbid ity and sedimentation regimes (Table 2). The 4-month data sets of Larcombe et al. ( l 995b) show that, measured at 30 em above the bed, the shallow reef sites of Magnetic Island (sites 8- 1 4 of Larcombe et al., 1 99 5b) have long-term average suspended sediment concentrations (SSCs) of 7.5 mg 1-1 and gross sedimentation rates (i.e. trap capture) of 1 5. 1 mg cm-2 day-1. These are only 36% and 26% respectively of the values for the deeper soft-bottom sites on the adjacent inner shelf of Cleveland Bay. The fringing reefs of Magnetic Island show elevated turbidities as a result of resuspension by wind waves, generated mostly by the afternoon sea breeze, and the measured sedimentation rates are almost purely due to resettling of resuspended material on the reef flats. Indeed, to our knowledge, modern accumulation of mainland-derived terri genous sediment has not been documented at any inner-shelf reef sites. In contrast, wind-waves have little effect upon the sedimentation regimes on the inner shelf itself, where the action of swell waves is by far the dominant process causing resuspension. These waves are the main influence upon sedimen tation rates measured using sediment traps. Applying the concept of BTF to both long-term core data (Table 1) and measurements of modern processes (Table 2) allows estimation of the relative importance of sediment resuspension and transport as compared with long-term accumulation of terrig enous material. Over the last 30 yr, sediment has accumulated in Cleveland Bay at c. 0.2 mm yc1, at c. 7 BTF, and the Holocene accumulation rate is similar {Table 1 ). Sediments that accumulate in sediment traps do not necessarily represent material that ultimately falls to the bed. However, if the sediment trap data given above are taken as a reasonable estimate of bed accumulation, they would equate to a potential terrigenous sediment flux to the bed of c. 4000 BTF for Cleveland Bay. This indicates that, as a first approximation, only up to c. 0.2% of the material resuspended in Cleveland Bay each year is actually accumulated there on time-scales of a decade and longer. This emphasizes the highly dynamic and turbid nature of the inner shelf, and the distinction of sediment
resuspension and transport from net sediment ac cumulation. We also note that the fringing reefs receive 1 00- 1 000 BTF in the short term, emphasiz ing the importance of sediment resuspension and removal for their presence. On the basis of the measured turbidities, available fine sediments, and the principal control upon sediment transport {Table 2), we would expect Paluma Shoals and other Halifax Bay coastal reefs to experience even higher fluxes of terrigenous sediment to the bed. Corals and sedimentation in Halifax Bay
Until recently, corals on the inner shelf in Halifax Bay (Figs 1 & 2) had received only minor scientific attention (an exception being the study by Done ( 1 982)), but the widespread presence of a pioneer ing coral association has now been documented in the shallow subtidal zone adjacent to the shoreline (Veron, 1 995; Larcombe & Woolfe, 1 997). On Paluma Shoals {Plate 1 , facing p. 304) the main corals on the reefs are faviids and mussids ('dome' types), small Acropora, and large stands of Galaxea, with coral cover of 40-60%. The coral heads on the reef shoals are devoid of modern terrigenous sedi ment, and sediments between corals on the reef are clean quartz sand. The inner-shelf plains of Cleve land and Halifax bays are formed of a terrigenous muddy sediment wedge (Belperio, 1 978; Carter et al., 1 99 3). The sea-bed adjacent to the Paluma Shoals reef is a muddy sand bottom with scattered Turbinaria and small Por ites. Environmental monitoring of dredging opera tions for the Port of Townsville, together with more recent work, has demonstrated that a highly turbid coastal boundary layer is frequently formed in Cleveland and Halifax bays, close to fringing and nearshore reefs (Larcombe & Ridd, 1 994; Lar combe et al., 1994, 1 995b; Stafford-Smith et at., 1 994; Larcombe & Woolfe, 1 997). In Halifax Bay the boundary layer is generated during periods of regional SE winds, when waves resuspend muddy terrigenous sediments on the inner shelf (Fig. 3). In Halifax Bay, reef establishment is apparent within the coastal boundary layer. Cattle Creek Shoals and Lady Elliot Reef are examples of other coastal turbid-zone reefs (Fig. 1 ). Young corals, at and below the lower water mark, grow on relict (pre-highstand) alluvial gravels, which in places are probably slightly reworked, presumably by storin waves (Plates 1 and 2). The size and morphology of Paluma Shoals, and the presence of spur and groove
299
Sediment accumulation and reef distribution
18'45' s
I 'C,
¥ �o.
•
=
coral reefs
I
\
\ HALl FAX BAY o···-...
'
\
'
\--�------
19'00' s
Approximate outer boundary of terrigenous
20m ,
',
..., ..,._
•"''
Paluma Shoals
� :� : � � ;x ',,�:_:. . ....
:: . . ::: : c_.::::::::::: ,_ ::
146'30' E
146'45' E
Island
� � � j 1 � j j::::: (� ..
146'15'E
.
/"-- .... ----
..:
147'00'E
Fig. 2. Halifax Bay, near Townsville, central Great Barrier Reef. Fringing coral reefs (marked in black) occur on most of the continental islands of the inner shelf, such as Magnetic, Great Palm and Orpheus islands. Coastal turbid-zone reefs (such as Paluma Shoals) and other coral accumulations also occur close to much of the Halifax Bay shoreline, in turbid water of the shallow subtidal zone, influenced by resuspension of the inner-shelf terrigenous sedimentary wedge. The whole shoreline and nearshore zone has not yet been surveyed in detail, and it is likely that more coral reefs and coral accumulations occur than are shown here (after Larcombe & Woolfe, 1997).
structures, indicates that these reefs are long-term (Holocene) features, are well established and may be forming a coral framework. An alternative pos sibility is that their form may be inherited from Pleistocene reefs, and that their spur and groove morphology is an erosional feature, but similar coastal fringing reefs along the GBR coastline are underlain by terrigenous gravels (Johnson & Carter, 1 987;Hopley, 1995), and no Pleistocene reefs have been documented along the central GBR coastline (Murray-Wallace & Belperio, 1 9 9 1). Finally, their morphology may be a result of erosion of their reef framework having occured in slightly deeper water, at the peak Holocene highstand at c. 5000 yr BP. From their elevation, and using the sea-level curve of Larcombe et a!. ( 1 99 Sa, their Fig. 7), these reefs have a maximum possible age of c. 7000 yr BP. The
nearby Magnetic Island reefs have been dated as forming as much as c. 6500 yr BP (Hopley, 1983), and are thus of similar age to other inner-shelf fringing reefs along the GBR shoreline (Johnson & Carter, 1987; Hopley, 1 995). It is difficult to determine the short- and long term 'health' of the modern 'coastal turbid-zone reefs' at Paluma Shoals, but it is likely that their long-term survival and growth is disrupted episod ically by cyclonic waves, or by prolonged periods of low coastal salinity resulting from high river dis charges. Relict gravels beneath or adjacent to shoreface sands are widespread along the Halifax Bay coastline, indicating that net sediment accumu lation over much of the intertidal and shallow subtidal zone has been zero or less during the last 6000-7000 yr. This period of non-deposition or
w 0 0
Table 2. Summary of oceanographic characteristics, turbidity data and sediment-trap data for some reef and inner-shelf sites at Cleveland and Halifax bays
Site
Site depth and aspect (dominant wave trains approach from SE)
Con tinen tal island fringing reefs Magnetic Island 4-month meanll
Potential terrigenous sediment flux to the bed§ (BTF)
Principal processcontrols upon turbidity
Turbidities/ SSCs* (mg ,-I)
Proportion of the time IOmg l-1 was exceeded at 30 em above the bed (%)
-
-
-
7.5
15.1
10-60
120-720
-
2-23**
10-60
16-1090
Sedimentation rates (measured Time-averaged by sediment traps)t sse (30 em above bed) (mg cm-2 day-1)
Terrigenous fraction of sedimentt (%)
Geoffrey Bay (site 9)�
3.5 m, S
Locally induced wind-waves
B <50
17
Arthur Bay (site I 0)
7.5 m, SE
Locally induced wind-waves
B <45
16
Middle Reef (site 13)
7 m, NW and SE
Swell and wind-waves
s 23-39 M 19-41 B < 70
24
-
8.8-61.8
40-80
280-3900
Rattlesnake Island Halifax Bay (site 14)
8 m, S
Swell and Spring tides
B <10
I
-
5. 7-39.6
10-50
45-1565
4 m, open
Waves, at low water
B <140
Coastal turbid-zone reefs Paluma Shoals Halifax ·Baytt
�
�
� 0
s:; "' R<> � t-< >::>
� 0
;:s
<:)"'
Table 2. (Continued)
Turbidities/ SSCs* (mg l-1)
Proportion of the time 10 mg l-1 was exceeded at 30 em above the bed (%)
Sedimentation rates (measured Time-averaged by sediment sse (30 cm traps)t (mg cm-2 day-1) above bed)
-
-
-
20.9
11 m, open
Swell waves
M 5-35 B 9 to> 250
35
Outer Cleveland Bay (site 4)
15 m, open
Swell waves
B <250
Eastern Cleveland Bay (site 1)
4.5 m, NW
Waves
<220
Site depth and aspect (dominant wave trains approach from SE)
Principal processcontrols upon turbidity
-
Outer Cleveland Bay (site 3)
Site Inner-shelfsediment wedge Cleveland Bay 4-month meanH
Inner to mid-shelf transition zone Cape Cleveland§§ 20 m, open
Terrigenous fraction of sediment; (%)
Potential terrigenous sediment flux to the bed§ (BTF)
58.5
85
4000
-
16.5-106.4
85
1110-7160
34
-
16.7-72.9
85
1120-4900
47
-
14-129
90
1000-9180
�;:s � ..,. $::)
� Swell waves and flood tide
;:s
<25
l::
§: (3• ;:s
* In Cleveland Bay and Halifax Bay, turbidities measured in NTU are approximately numerically equivalent to SSC in mg l-1• S, Surface; M, mid-depth; B, <1 m above bed. t 'Sedimentation rates' here refer purely to data measured by sediment traps. Therefore the data do not distinguish between resuspension and resedimentation as opposed to net sediment accumulation. ; From data of Belperio (1978), Smith (1978), Way (1987) and Ohlenbusch (1991). § BTF calculated assuming particle density of 2.6 and porosity of 30%, and scaled to remove biogenic carbonate. 1 BTF 45 g m-2 yr-1• II Calculated from full datasets for sites 8-12 of Larcombe et a !. (1995b). II Site numbers are those of Larcombe & Ridd (1994) and Larcombe et a/. (1995b). **Data from Stafford-Smith et a!. (1994). ttData from Larcombe & Woolfe (1997). H Calculated from full datasets for sites 1-7 of Larcombe et a/. (1995b). §§ A. Orpin, personal communication. The potential terrigenous sediment flux to the bed is the result assuming that all the sediment accumulated in sediment traps had fallen to the bed. Most data cover a range from fair-weather to rough sea conditions, although they do not cover cyclonic conditions. The periods of wind-waves and swell waves are < 7 s and> 7 s, respectively, as used by Larcombe et a/. (1995b). Over 350 water samples, 250 sediment-trap retrievals, 1000 h of current and wave data, and 30 000 nephelometer hours of data are summarized here. =
$::) ;:,:
$::>...
;;; �
fr q
5: � c:;· ;:s
w 0
302
K. J. Woolfe & P. Larcombe 200
NTU)
Turbidity (
100
0
Cur
t
ren direction (degs)
300
Wind-driven coast-parallel current moving towards the NW, fully suppressing the normal tidal currents
100 0 ••
Wind direction (degs)
•
•• •
300
• •
•
1-l --·�
N winds 100 0
• • • •
•
Period of persistent SE winds
•• •
•
•
• • ••
•
50
_____..,._
• •
.
••
.... .••... 150
100 Time- Hours
erosion has coincided with reef growth at Paluma Shoals and elsewhere on the inner shelf. Sediment supply to the coast
Three main methods have been used to estimate terrigenous sediment supply to the central GBR shelf: 1 Extrapolation of measurements of sediment transport in rivers and estuaries to an annual or longer time-scale (e.g. Belperio, 1 979, 1 983, 1 988); 2 Modelling of the sediment yields from soil ero sion and runoff in the river catchments (Moss et al., 1993; Neil & Yu, 1 995); 3 Estimation of the volume of Holocene sediment on the shelf, to give a long-term average of sediment accumulation rate (Belperio, 1 98 3; Harris et al., 1990; Carteret al., 1 99 3). Application of the first two techniques to the rivers of northern Queensland can result in considerable uncertainty in the calculated rate of sediment sup ply. This is a result of the rivers' characteristics, which include relatively low mean annual runoff, large inter-annual variability of discharge, heteroge neous vegetation patterns in their catchments (in cluding altered catchments), and unusual channel morphologies. Sediment yield models developed overseas are not applicable to most Australian catchments (Neil & Yu, 1 995). Back-calculating long-term sediment input from the Holocene sedi ments involves uncertainties about the stratigraphy, dating and the extent of the deposits, but appears to be the most reliable indicator. Few Holocene terri genous sediments now occur on the middle and outer shelf of the GBR; most are confined to the
•• •
• ••• •
200
Fig. 3. Turbidity (measured 0.3 m above bed), current direction (measured I m above bed) and wind direction at Paluma Shoals, Halifax Bay. Note that turbidities increase dramatically following the commencement of SE winds (70 h onwards) because of resuspension by waves. The SE winds also suppress the normal tidal currents to produce a constant coast-parallel current, towards the NW (simplified from Larcombe & Woolfe, 1997).
inner-shelf terrigenous sedimentary wedge. It is assumed that terrigenous accumulation in these areas has been negligible since rising post-glacial sea-levels approached modern levels. We conclude that sediment supply to the central GBR coastline and coast-parallel sediment trans port, although small by world standards, have not been insignificant over the Holocene. Terrigenous sediment accumulation rates are highest inshore and near the Burdekin River mouth. Yet, except close to the mouths of the large rivers, inner-shelf reefs appear to occur in most places where suitable substrates are available. A significant control on the distribution of coral reefs on the inner shelf of the central GBR is thus a lack of net sediment accumu lation at reef sites, rather than the actual terrigenous sediment supply to those areas or flux past them.
THE GULF OF PAPUA
Having established the general model, we comment upon another example. The Gulf of Papua (Fig. 4) is fringed in the west and north by sandy beaches, tidal mudflats and extensive mangroves (Thorn & Wright, 1 98 3). Fringing reefs occur along the north eastern coast of the gulf, north of Port Moresby, and in the Torres Strait, which extends south and west from the gulf. To the south is the northernmost section of the GBR (Harris et a!., 1 993). A large (> 1 0 000 km2) partially relict carbonate platform has recently been described immediately seaward of the Fly River mouth (Fig. 5) (Alongi & Robertson, 1 995; Brunskill et al., 1 995). This platform occurs in 40-70 m of water and is dominated by calcare-
.... • 146'E .
·:.
• 1195 1171 • .
1197
0 224
1198
Tauri River .1 .. :::::
'ij
. •
V:J
0227,210
� §'
� >::.
R ::::
�
::::
i:l ..... c;·
•1167
;:s
• tt66
>::.
04 �2::.,:.
.
..j
. /'
j•
1 0
'Cc······
�errie
� � �
-····
?
�
England
Shoals
c;
�
1500m
,. -·
..
�
·..
;:s
Papua Plateau
'/� Torr-es . -�
,.,,_<;;; ·
:';;�:\¢ �;:· , >·.:s;;.;\ ':; �z;?�;�:/ ·
..
(?,
v
·
Fringing reefs
.. J,
r;-;·_:::: ·..(..'
_
__
_
· :/·Y(r�f-:�;���:':};- · :;
__
0
25
km
50
'·
. - .::�.
!O'S
Fig. 4. Geomorphological map of the Gulf of Papua, showing an extensive carbonate platform occurring directly seaward of the Fly River mouth. The location and extent of the fringing reefs in the eastern gulf (black) are estimated from Admiralty Chart Aus. 379. The reefs of the Torres Strait at the northern extreme of the Great Barrier Reef are marked in a pale stipple. 0, core locations; e, grab sample locations; lines show 12.5-kHz echo-sounder track lines collected during cruise 93/5 on R.V. Franklin ; •, grab sample sites taken from R.V. Harry Messel.
'-'-' 0 '-'-'
304
K J Woolfe & P. Larcombe
Key to terrigenous sediments
c::::J �
Laminated mud
Sand
Mottled mud
Interbedded mud and sand
Organic·rich laminae
Fig. 5. Distribution of sedimentary facies on the Fly-Gulf of Papua shelf. Hatchures mark terrigenous sediments of the Fly River delta and inner Gulf of Papua (modified after Alongi & Robertson, 1995).
ous red algal crusts (Brunskill et al., 1 995). Despite its proximity to the Fly River the surface of the platform is mud free. The principal riverine sediment sources for the Gulf of Papua are the Fly River, which contributes 80- 1 20 Mt yc1 of sediment (Ok Tedi Mining Lim ited, 1 988;Harris eta!., 199 3), and the Purari River (80 Mt yc1: Milliman & Meade, 1 983; Thorn & Wright, 1 98 3). It has been estimated that the total sediment supply to the gulf from all its rivers is 365 Mt yc1 (Milliman, 1 995), which is nearly six
times the sediment yield of the entire continent of Australia (62 Mt yr-1: Milliman & Meade, 1 98 3). Given this large regional supply of terrigenous sediment into the coastal ocean (58 times the BTF), it might conventionally have been viewed as sur prising that 90 km of fringing reef occurs between Kerema and Port Moresby (Admiralty Chart Aus. 379), and also that many small reefs are present in the highly turbid waters immediately south-west of the Fly River mouth, including Merrie England Shoals 18 km south of the mouth (Fig. 4). Unfortunately, compared with the GBR shelf, limited data are available on these reef communi ties, which have only been documented by AIDAB ( 1 994). However, their existence indicates that, over a time period of centuries to millennia, signif icant local accumulations of terrigenous sediment have not formed. Further, water quality has been maintained at a level which has allowed reef forma tion and growth. Local, nearshore sediment trans port processes are undocumented, but both these factors probably depend upon local and regional oceanographic processes. Regionally, there is a sharp decline in the abundance of terrestrial sedi ment immediately seaward of the Fly River mouth, indicated by both textural (Figs 4 & 5) and geochemical studies (Fig. 6). When viewed together with the oceanographic work of Ok Tedi Mining Limited ( 1 988), Wolanski & Alongi ( 1 995) and Wolanski et al. ( 199 5), and with the previous sedimentary work of Harris et al. ( 1 993), these studies indicate a northward deflection of the Fly River sediment plume. This deflection is (probably) in response to the persistent SE trade winds, which produce a strong northward drift in the coastal boundary layer, as far east as the eastern margin of
8S
Fig. 6. Carbon isotope
Torres Strait
.....
) I .
-.
,
. ,.
1
,•·
�;{r ·
Great
� Barrier � Reef .
10 s 146 E
composition of sediments in the Gulf of Papua. Contours indicate percentage of terrestrial particulate organic carbon (which decreases sharply offshore to less than 8%) and the measured carbon isotope composition of the sediments (modified after Bird et a!., 1995).
Sediment accumulation and reef distribution
the Gulf of Papua. Combined with the effects of waves the result is a strong landward partitioning of terrigenous sediment, and a focus of rapid sediment accumulation at the head of the gulf between Kerema and the Fly River mouth, an area devoid of documented reefs (Admiralty Chart Aus. 378). Fringing reefs do occur south-east of Kerema along the coast to beyond Port Moresby. We recognize that insufficient oceanographic and water-quality data are available to determine causal relationships between sediment supply to the gulf, sediment flux within it, turbidity, sediment accu mulation, and coral reef distribution within the Gulf of Papua. The spatially averaged terrigenous supply into the Gulf of Papua is 58 BTF, compared with only 3 BTF for the whole GBR lagoon (data from Milliman & Meade, 1 98 3; Moss et a/., 1 99 3; Milliman, 199 5) and c. 8 BTF for the inner GBR shelf alone. However, despite the large sediment supply to the gulf, hydrodynamic conditions still result in regions where offshore carbonate tracts and coastal fringing reefs persist.
DISCUSSION
Reefal limestones, containing and associated with terrigenous mud, occur throughout the geological record (e.g. Hayward, 1 982; James & Bourque, 1992), so the presence of coastal turbid-zone reefs along modern reef-forming coastlines should not be surprising. It is widely known that some coral species and communities are more tolerant of tur bidity than others (Done, 1 982; Rogers, 1 990; Veron, 1 995), yet 'threshold turbidities' have proved difficult to determine. This is perhaps unsur prising, partly because the temporal and spatial variations in the hydrodynamic, sedimentary and geological environments of many reefs have not been sufficiently well considered (Larcombe et a/., 1 99 5b). A simple 'threshold condition' might be defined purely in terms of turbidity (e.g. a maxi mum turbidity which a coral may be subjected to without death). However, this would not take into account the hydrodynamic and sediment dynamic factors, and the significance of such a threshold would therefore be questionable. Moreover, as the examples summarized above indicate, net sediment supply into the coastal ocean cannot be directly correlated with the regional presence or absence of coral carbonate provinces. Noting examples of quasi-independence between
305
carbonate and terrigenous sedimentation, Woolfe & Larcombe ( 1 997) have described conceptual boundary conditions within which reefs can de velop (Fig. 7). Their model is based upon the separation of three variables-terrigenous sediment accumulation, erosion and water turbidity-and summarizes the likely relationships amongst these parameters with respect to coral reef growth or carbonate removal. A series of existence fields is predicted for coral reefs for various combinations of sedimentary, hydrodynamic and sediment dynamic conditions. Corals tend to colonize hard stable substrates (Hopley, 1 982; Veron, 1995) and, by definition, the rate of carbonate production must exceed the rate of terrigenous sediment accumula tion for a carbonate province to develop. It is therefore implicit that, at the time of reef initiation, the rate of terrigenous sediment accumulation must be zero, otherwise a hard (or at least stable) sub strate would not exist (Fig. 7). However, whereas it is intuitively correct to link turbidity and sedimen tation, the two are not mutually dependent; most marine terrigenous sediment is deposited from relatively turbid water, but not all turbid water deposits sediment. This is exemplified by the near shore corals in Halifax Bay, which occur within the coastal boundary layer (Plate 1) but in a region of zero net sediment accumulation. It is also impor tant to note that the presence of high turbidities does not necessarily reflect terrigenous sediment, because in some erosional environments, turbidity may be generated by transport of carbonate muds. The presence of reefs in turbid waters highlights two other potential problems regarding sedimenta tion at coral reefs: 1 Terrigenous sediments are commonly incorpo rated into the framework of a coral reef. One or both of two different mechanisms may take place. First, assuming potential new 'reefs' (as described above) are successful in establishing a growth framework (i.e. assuming they do become, in fact, true reefs as opposed to muddy carbonate accumu lations), the baffling effect of the corals may increase as the reef develops, and trap increasing amounts of muddy terrigenous sediment within the reef frame work. Second, given their location, the reefs could ultimately be swamped by sediment, especially during landward migration of the marine mud wedge (for instance, during the Holocene sea-level rise;see below). Burial by sediment could also occur by seaward progradation of a shore-connected sedi ment wedge, under conditions of high sedimenta-
306
K J Woolfe & P. Larcombe Terrigenous accumulation T
Carbonate removal
Reefal accumulation
Terrigenous erosion
Detail of reef fields
R
+
Fig. 7. Diagram showing the conceptual existence fields for coral reefs in terms of net carbonate production and
terrigenous sedimentation. Fields of erosion and sedimentation along with zones of terrigenous- and carbonate dominance are defined by a plot of T (rate of terrigenous sediment accumulation) against R (rate of reefal carbonate accumulation). Coral reefs are restricted to the central right portions of the diagram, and some examples of other
Sediment accumulation and reef distr ibution
tion rate and/or falling sea-level. In either case, the preserved reefal deposit could reasonably be inter preted in the geological record as recording evi dence for increased sediment supply through sea level change, yet the terrigenous sediment content is, in part, autocyclic (that is, trapping efficiency is a function of framework development). James & Bourque ( 1992) have previously noted the difficulty in distinguishing autostratigraphy from allostratig raphy in the geological record. In our examples, distinguishing increased sediment supply produced by sea-level change from increased trapping effi ciency would not be easily achieved from preserved deposits. 2 Sediment traps have been used in a number of monitoring studies of coral reef environments, to document turbidity and sedimentary characteristics (e.g. Dodge & Vaisnys, 1977; Rogers, 1 982, 1983, Cortes & Risk, 1985; Pastorak & Hilyard, 1985). Where deployed in regions of wave and current resuspension, these traps record a gross sediment flux in the downward direction, towards the bed. However, great care has to be taken in interpreting sediment-trap data, because both the sign and magnitude of sedimentation are likely to be in error. In regions of sediment resuspension, sediment traps will always overestimate net sediment accumula tion (compare Tables 1 & 2). In areas of net erosion, sediment caught within traps might erroneously indicate net accumulation of sediment, unless used with a range of other information. Moreover, the relationship of sediment-trap data to water turbid ity is likely to be extremely complex, temporally variable and probably locally unique. The use of sediment traps in combination with recording cur rent meters and nephelometers greatly increases the chances of understanding sedimentary processes.
(Con tinued from opposite page) depositional environments are given. Reef initiation can occur in any field provided that T and R are near zero and that environmental conditions allow the juvenile reef to move rapidly into the reef stability fields. Turbidity increases away from the x-axis (T 0) through sediment accumulation from turbid water (for T> 0), or from erosion of existing terrigenous substrate (T < 0), and 'turbid reefs' result. With an increasing absolute magnitude of terrigenous sedimentation, reef death occurs through a combination of light attenuation, burial or erosion. Continued divergence from the x-axis leads to the formation of siliciclastic shelf sediments or carbonate lags (ITI> R). (Modified after Woolfe & Larcombe, 1997 .) =
307
This understanding is likely to improve studies of the physiological responses of corals to various aspects of sedimentation.
SEA-LEVEL CYCLES
The evidence from Halifax Bay indicates that the availability of a suitable substrate is a greater control on reef initiation and development than is water clarity, at least for initial colonization. Substrate character is controlled by sediment accumulation rather than by turbidity. Cycles of transgression and regression imposed by eustatic and tectonic factors will thus act in combination with local oceano graphic and sediment transport regimes, and will be of far greater importance for the establishment of reefs than sediment supply per se. During transgressive phases, shelf sediments are commonly transported landward, increasing the volume of sediment in the coastal zone. Within the central GBR lagoon, the isobaths in the 40-20-m depth range are relatively linear and evenly spaced, indicating that during lower (and rising) post-glacial and Holocene sea-levels, the coastline would have been relatively linear (Harris et a/., 1990;Larcombe & Woolfe, 1997). With the likely presence of SE trade winds, substantial northward movement of coastal sediment would have occurred, limiting the availability of stable, clean substrates. In concert with a rapidly rising sea-level, few coastal or inner shelf reefs would have been initiated. However, as sea-level approached the level of the modern 101 5-m isobath, a highly indented coastline evolved (Belperio, 1978). The embayments provided loci for sedimentation and attenuated the coastal boundary layer, and coral reefs were able to initiate in areas where sediment accumulation was mini mal, including exposed headlands (Fig. 8).
CONCLUSIONS
Coral-dominated carbonate provinces are not re stricted to clear-water conditions. 'Coastal turbid zone reefs' occur throughout the central GBR, and in the Gulf of Papua. Although near-bed turbidity is necessary for fine-grained terrigenous sediment ac cumulation, the presence of highly turbid water does not necessarily lead to net sediment accumu lation. Indeed, reef establishment on the GBR shelf is currently occurring in some of the most turbid
308
K. J. Woolfe & P. Larcombe
"Coastal turbid-zone reefs" on gravel patches of Pleistocene substrate Shore-detached sediment wedge Patch reef
Depth >20m Inner-shelf sediment wedge " Turbid inner-shelf waters
""'
Mid-shelf lag surface
"
Clear mid-shelf waters
Prevailing SE trade winds
20-40 km
Fig. 8. Perspective view of a sedimentary model of the central GBR coastline, showing terrigenous sediment bodies and geomorphical features (partly after Belperio. 1978), sites of coral reef growth and simplified oceanographic features. The terrigenous sediment wedges are attached to the shoreline near the major sediment sources and in the depositional embayments, but lie offshore further downdrift (and) on eroding (linear) coastlines. Where the sediment wedge lies offshore, coarse substrates are exposed in the nearshore zone, forming sites of coral growth and potential reef formation (from Larcombe & Woolfe, 1997).
parts of the inner shelf, where resuspension and erosion are exposing new hard substrate. In noting the relationships of some corals to terrigenous sediment bodies, we stress the distinc tion between sediment supply to the coast, sedi ment flux on the shelf, and sediment accumulation. The sediment accumulation rate is a more funda mental control on reef development than is sedi ment supply. However, the occurrence of terrigenous sediment within developing reefs is probably not, by itself, an indication of increased sediment supply. Increasing terrigenous compo nents within the upper portion of coral cores may reflect increased sediment trapping rather than increased sediment accumulation, supply or water turbidity. Substrate availability is more important as a
control upon reef initiation and growth than is water clarity. Substrate availability is controlled by local oceanography and sediment transport, super .. imposed upon regional sedimentation patterns, and in the longer term, tectonic and global eustatic: change. High sediment fluxes through an area do not necessarily limit substrate availability.
ACKNOWLEDGE MENTS
This paper draws on data collected for a variety of projects. Financial and logistical support was pro.. vided by the Cooperative Research Centre for Sustainable Development of the Great Barrier Reef, the Australian Institute of Marine Science, the Australian Research Council, James Cook Univer..
Sediment accumulation and reef distribution
sity, Ok Tedi Mining Ltd and the Townsville Port Authority. We thank our colleagues Bob Carter and Tim Naish for their editorial comments, and Alan Orpin for helpful advice. Colin Braithwaite pro vided inspired suggestions upon the manuscript, and comments were also received from an anony mous reviewer. Arnstein Prytz processed the turbid ity datasets, and, finally, we thank Richard Purdon for his skilful and patient drafting assistance with Figs 1 , 2, 4 & 8. REFERENCES
W.H. (1978) Coral reef morphogenesis: a multidi mensional model. Science, 202, 831-857. AIDAB (1994) Papua New Guinea and Gulf Coastal Zone Management Plan, pre-feasibility study, Final Report. Australian International Development Assistance Bu reau, Canberra, ACT. ALONGI, D.M. & ROBERTSON, A.l. (1995) Factors regulat ing benthic food chains in tropical river deltas and adjacent shelf areas. Geomar. Lett., 15, 145-152. ANONYMOUS ( 199 5) Old photos chart destruction of Aus tralia's reef. New Sci. , 4, 7. BELPERIO, A.P. (1978) An inner shelfsedimentation model for the Townsville region, Great Barrier Reef province. PhD thesis, James Cook University, Townsville, Qld. BELPERIO, A.P. (1979) The combined use of wash load and the bed material load rating curves for the calculation of total load: an example from the Burdekin River, Aus tralia. Catena, 6, 317-329. BELPERIO, A.P. (1983) Late Quaternary terrigenous sedi mentation in the Great Barrier Reef lagoon. In: Proceed ings of the Great Barrier Reef Conference (Eds Baker, J.T., Carter, R.M., Sammarco, P.W. & Stark, K.P.), pp. 71-76. James Cook University, Townsville, Qld. BELPERIO, A.P. (1988) Terrigenous and carbonate sedimen tation in the Great Barrier Reef province. In: Carbonate Clastic Transitions (Eds Doyle, L.J. & Roberts, H.H.), Developments in Sedimentology, 42, pp. 143-174. Elsevier, Amsterdam. BIRD, M.l., BRUNSKILL, G.J. & CHIVAS, A.R. (1995) Carbon-isotope composition of sediments from the Gulf of Papua. Geomar. Lett. , 15, 153-159. BRUNSKILL, G., WOOLFE. K. & ZAGORSKJS, I. (1995) Dis tribution of riverine sediment chemistry on the shelf, slope and rise of the Gulf of Papua. Geomar. Lett. , 15, 160-165. CARTER, R.M., JOHNSON, D.P. & HOOPER, K.G. (1993) Episodic post-glacial sea-level rise and the sedimentary evolution of a tropical embayment (Cleveland Bay, Great Barrier Reef shelf, Australia). Aust. J. Earth Sci. , 40, 229-255. CoRTES, J. & RISK, M.J. (1985) A reef under siltation stress: Cahuita, Costa Rica. Bull. mar. Sci. , 36, 339-356. DODGE, R.E. & VAISNYS, J.R. (1977) Coral populations and growth patterns: responses to dredging and turbidity associated with dredging. J. mar. Res. , 35, 715-730. DoNE, T. ( 1982) Patterns in the distribution of corals ADEY,
309
commumties across the central Great Barrier Reef. Coral Reefs, 1, 95-107. HARRIS, P.T., DAVIES, P.J. & MARSHALL, J.F. (1990) Late Quaternary sedimentation on the Great Barrier Reef continental shelf and slope east of Townsville, Australia. Mar. Geol. , 94, 55-77. HARRIS, P.T., BAKER, E.K., CoLE, A.R. & SHORT, S.A. (1993). A preliminary study of sedimentation in the tidally dominated Fly River Delta, Gulf of Papua. Continent. ShelfRes., 13, 441-472. HAYWARD, A.B. (1982) Coral reefs in a clastic sedimentary environment: fossil (Miocene, SW Turkey) and modern (Recent, Red Sea) analogues. Coral Reefs, 1, 109-114. HoPLEY, D. ( 1982) The Geomorphology of the Great Barrier Reef Quaternary Development of Coral Reefs. Wiley-Interscience, New York. HOPLEY, D. (1983) Evidence of 15000 years of sea-level change in tropical Queensland. In: Australian Sea-levels in the Last 15000 Years: a Review (Ed. Hopley, D.), De partment of Geogaphy, James Cook University, Towns ville, Qld, Monogr. Ser, Occas. Pap., 3, pp. 93-104. HoPLEY, D. (1995) Continental shelf reef systems. In: Coastal Evolution: Late Quaternary Shoreline Morpho dynamics (Eds Carter, R.W.G. & Woodrolfe, C.D.), pp. 303-340. Cambridge University Press, Cambridge. HOPLEY, D. & MURTHA, G.G. (1975) The Quaternary Deposits of the Townsville Coastal Plain. Geography Department, James Cook University, Townsville, Qld, Monogr. Ser, 8. JAMES, N.P. & BouRQUE, P.A. (1992) Reefs and mounds. In: Facies Models: Response to Sea-level Change (Eds Walker, R.G. & James, N.P.), pp. 323-347. Geological Association of Canada. JOHNSON, D.P. & CARTER, R.M. (1987) Sedimentary framework of mainland fringing reefdevelopment, Cape Tribulation area. Great Barrier Reef Marine Park Au thority, Townsville, Qld, Technical Memorandum, 14. JoHNSON, D.P. & SEARLE, D.E (1984) Post-glacial seismic stratigraphy, central Great Barrier Reef, Australia. Sedi mentology, 31, 335-352. LARCOMBE, P. & Rmo, P.V. (1994)Data interpretation. In: Townsville Port Authority Capital Dredging Works 1 993: Environment Monitoring Program (Eds Benson, L.J., Goldsworthy, P.M., Butler, I.R. & Oliver, J.), pp. 165194. Townsville Port Authority. LARCOMBE, P. & WOOLFE, K. (Eds) (1996) Great Barrier Reef Terrigenous Sediment Flux and Human Impacts, 2nd edn. CRC Reef Research Centre, Research Sympo sium Proceedings, Townsville, Qld. LARCOMBE, P. & WOOLFE, K. (1997) sedimentary and sea-level controls on the distribution of Holocene inner shelf coral reefs, Great Barrier Reef, Australia. Mar. Geol. (submitted). LARCOMBE, P., RIDD, P.V., WILSON, B. & PRYTZ, A. (1994) Sediment data collection. In: Townsville Port Authority Capital Dredging Works 1 993: Environment Monitoring Program (Eds Benson, L.J., Goldsworthy, P.M., Butler, I.R. & Oliver, J.), pp. 149-164. Townsville Port Author ity. LARCOMBE, P., CARTER, R.M., DYE, J., GAGAN, M.K. & JOHNSON, D.P. (1995a) New evidence for episodic post glacial sea-level rise, central Great Barrier Reef, Austra lia. Mar. Geol. , 1 27, 1-44.
310
K J Woolfe & P. Larcombe
P., RIDD, P.V., WILSON, B. & PRYTZ, A. (1995b) Factors controlling suspended sediment on inner-shelf coral reefs, Townsville, Australia. Coral Reefs, 14, 163171. MILLIMAN, J.D. (1995) Sediment discharge to the ocean from small mountain rivers: the New Guinea example. Geomar. Lett., 15, 127-133. MILLIMAN, J.D. & MEADE, R.H. (1983) World-wide deliv ery of river sediments to the oceans. J. Geol. , 91, 1-21. Moss, A.J., RAYMENT, G.E., REILLY, N. & BEST, E.K. (1993) A preliminary assessment of sediment and nutrient exports from Queensland coastal catchments. QueenslandDepartment of Environment and Heritage, Brisbane, Environment Technical Report 5. MURRAY-WALLACE, C.V. & BELPERIO, A.P. (1991) The last interglacial shoreline in Australia-a review. Quat. Sci. Rev., 10, 441-461. NEIL, D.T. & Yu, B. (1995) Simple climate-driven models for estimating sediment input to the Great Barrier Reef lagoon. In: Great Barrier Reef Terrigenous Sediment Flux and Human Impacts (Eds Larcombe, P. & Woolfe, K.), pp. 67-73. CRC Reef Research Centre, Research Symposium Proceedings, Townsville, Qld. 0HLENBUSCH, R. (1991) Post-glacial sequence stratigraphy and sedimentary development of the continental shelf off Townsville, central Great Barrier Reefprovince. Honours thesis, Geology Dept., James CookUniversity,Towns ville, Qld. OK TED! MINING LIMITED (1988) Sixth supplemental agree ment environmental study, 1 986-1 988, final draft report, Vols I, II and III, unpublished. PASTORAK, R.A. & BILYARD, G.R. (1985) Effects of sewage pollution on coral reef communities. Mar. Ecol. Progr. Ser. , 21, 175-189. QuEENSLAND GovERNMENT (1993) The Condition of River Catchments in Queensland. Department of Primary Industries, Brisbane. RoGERS, C.S. (1982) The marine environments ofBrewers Bay, Perseverance Bay, Flat Cay and Saba Island, St. Thomas, U.S VI., with emphasis on coral reefs and seagrass beds (November 1978-July 1981). Department of Conservation and Cultural Affairs, Government of the Virgin Islands. RoGERS, C.S. (1983) Sublethal and lethal effects of sedi ments applied to common Caribbean reef corals in the LARCOMBE,
·
field. Mar. Pollut. Bull., 14, 378-382. C.S. (1990) Responses of coral reefs and reef organisms to sedimentation. Mar. Ecol. Progr. Ser. , 62, 185-202. SHULMEISTER, J. (1995) Holocene climate change in Queen sland: implications for sea-level change and coastal sedimentation. In: Great Barrier Reef Terrigenous Sed iment Flux and Human Impacts (Eds Larcombe, P. & Woolfe, K.), p. 91. CRC Reef Research Centre, Research Symposium Proceedings, Townsville, Qld. SMITH, A. (1978) Case study: Magnetic Island and its fringing reels. In: Geographical Studies of the Townsville Area (Ed. Hopley, D.), Department of Geography, James Cook University, Townsville, Monogr. Ser. Oc cas. Pap., 2, pp. 59-64. STAFFORD-SMITH, M.G. , KALY, U.L. & CHOAT, J.H. (1994) Reactive monitoring (short-term responses) of coral species. In: Townsville Port Authority Capital Dredging Works 1 993: Environmental Monitoring Program (Eds Benson, L.J., Goldsworthy, P.M., Butler, I.R. & Oliver, J.), pp. 23-53. Townsville Port Authority. THOM, B. G. & WRIGHT, L.D. (1983) Geomorphology of the PurariDelta. In: The Purari-Tropical Environment ofa High Rainfall Basin (Ed. Petr, T.), pp. 47-65. Dr W. Junk, The Hague. VERON, J.E.N. (1995) Corals in Space and Time: the Biogeography and Evolution of the Scleractinia. Univer sity of New South Wales Press, Sydney. WAY, A.J. (1987) Post-glacial stratigraphy of Upstart Bay off the Burdekin River, north Queensland. MSc thesis, James Cook University, Townsville, Qld. WOLANSKJ, E. & ALONGI,D.M. (1995) A hypothesis for the formation of a mud bank in the Gulf of Papua. Geomar. Lett. , 15, 166-171. WOLANSKJ, E., KING, B. & GALLOWAY, D. (1995) Water circulation in the Gulf of Papua. Continent. Shelf Res., 15, 185-212. WOOLFE, K. & LARCOMBE, P. (1997) Terrigenous sedimen tation and coral reef growth: a conceptual framework. Mar. Geol. (submitted). WOOLFE, K.J., DALE, P.J. & BRUNSKJLL, G.l. (1995) Sedi mentary CIS relationships in a large tropical estuary: evidence for refractory carbon inputs from mangroves. Geomar. Lett. , 15, 140-144.
RoGERS,
Spec. Pubis int. Ass. Sediment. (1998) 25, 311-323
Comparison between subtropical and temperate carbonate elemental composition: examples from the Great Barrier Reef, Shark Bay, Tasmania (Australia) and the Persian Gulf (United Arab Emirates) C. P. RA O*, Z. Z. A M I N I* and J. F E R G U SONt *Department of Geology, University of Tasmania, GPO Box 252C, Hobart, Tas. 7001, Australia; and tAustralian Geological Survey Organisation, GPO Box 378, Canberra, ACT 2601, Australia
ABSTRACT
The Persian Gulf (United Arab Emirates) is renowned for subtropical carbonates. Extensive subtropical (Great Barrier Reef and Shark Bay) to temperate (Tasmania) carbonates are also forming in shallow seas around Australia. These carbonates differ in the types and proportions of skeletal to non-skeletal grains and cements, and are forming in normal to hypersaline shallow- marine environments. The elemental composition of subtropical carbonates differs from that of their temperate counter parts mainly because of differences in seawater temperature, carbonate mineralogy, salinity, rate of precipitation and the proportion of skeletal to non- skeletal grain composition. Differences in the Mg concentrations of the bulk carbonates result from variations in the temperature of sea water and carbonate mineralogy. Sr concentrations are higher in subtropical carbonates relative to temperate ones because of a higher proportion of aragonite in the tropical carbonate and calcite mineralogy. Na values increase with increases in salinity and rate of precipitation. Under reducing conditions appreciably higher Mn and Fe concentrations enter the calcite lattice compared with aragonite. The results from this study demonstrate that modern subtropical carbonate elemental composition differs distinctly from that of temperate carbonates. Thus, these differences can be used in the recognition of the ancient spectrum of subtropical to temperate carbonates based on the relative concentrations of elements and their ratios.
INTRODUCTION
Extensive subtropical to temperate carbonates are forming in shallow seas in many areas, particularly around Australia (Rao, 1996a). A similar spectrum of ancient subtropical to temperate carbonates may exist in the stratigraphical record but only a few ancient non-tropical carbonates have been identi fied (e.g. Nelson, 1978; Rao, 1981a, 1988; Brook field, 1988; Draper, 1988; Boreen & James, 1995). To fill this gap in our understanding, we present here a comparison of the elemental composition of subtropical Great Barrier Reef and Shark Bay car bonates, Australia, and subtropical Persian Gulf carbonates, United Arab Emirates, with temperate Tasmanian carbonates, Australia. Temperate carbonates differ from tropical carbon ates in the proportions of skeletal and nonskeletal-
grains (Lees, 1975), the mineralogy and products of diagenesis (e.g. Rao, 1981b; Reeckman, 1988; Rao & Adabi, 1992), oxygen and carbon isotopic compo sition (Rao & Green, 1983; Rao & Nelson, 1992), and in the concentrations of major and minor ele ments (Rao, 1990a; 1996b; Rao & Jayawardane, 1994). Elements that are mainly used in understand ing the origin of carbonates are Ca, Mg, Sr, Na, Mn and Fe (Veizer, 1983a,b). The Mg contents are sen sitive to variations in seawater temperature (e.g. Bur ton & Walter, 1991), and aragonite and calcite min eralogy (Milliman, 1974). The Sr contents vary with the proportion of aragonite and calcite, seawater temperature (Morse & Mackenzie, 1990) and rate of precipitation (Carpenter & Lohmann, 1992). The Na contents depend on salinity (Land & Hoops,
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
311
312
C. P.
Rao, Z. Z. Amini & J. Ferguson
197 3), biochemical fractionation and rate of precip itation (Busenberg & Plummer, 1985). The Mn and Fe concentrations are sensititive to changes in oxi dizing and reducing conditions (Veizer, 1983a,b; Morrison & Brand, 1987). The results from this study demonstrate that tropical carbonates differ distinctively from temperate carbonates. These re sults can be used to recognize ancient subtropical and temperate carbonates.
METHODS OF STUDY
Bulk sample powders were dissolved in l N HCI and analysed for Ca, Mg, Sr, Na, Fe and Mn by atomic absorption spectrophotometry (AAS). These values were normalized for 37% Ca to facilitate comparison of elemental composition in pure carbonate frac tions. The detection limits are ± 1o/o for Ca and Mg and ± 5 p.p.m. for Sr, Na, Mn and Fe (Robinson, 1980). Data of subtropical carbonates are listed in Table 1. The Tasmanian temperate carbonate ele mental data are from previous publications (Rao, 198 l b; Rao & Adabi, 1992; Rao & Jayawardane, 1993, 1994; Rao & Amini, 1995) and unpublished data.
SEDIMENTOLOGICAL FEATURES
Petrographical examination of samples analysed in this study showed that the bulk samples vary con siderably in skeletal and non-skeletal grain compo sition, micrite, spar and terrigenous constituents.
high carbonate facies range from reef to reworked reef debris and algae. Most of the reefs are shelf, lagoonal or elongate platform reefs (Maxwell, 1968). The core samples studied from the Great Barrier Reef are from a transect off north-east Australia (Fig. 1). Petrographical examination of these samples revealed abundant biotic constitu ents, particularly algae (Halimeda), foraminifers, molluscs, echinoderms and rare bryozoans, corals, worm tubes and crustaceans, and mud, debris and terrigenous material. Non-skeletal grains are rare in the Great Barrier Reef carbonates. Shark Bay
Modern carbonates at Shark Bay, Western Australia (Fig. 2), are forming in an arid to semi-arid climate, at a subtropical latitude of 26 S. The samples stud ied here are from Hamelin Pool (Fig. 2), which is a landlocked embayment in Shark Bay that contains carbonates. Seawater temperatures range from 15 to 29oC. The salinity of seawater varies considerably from oceanic (35-40o/oo), to metahaline (40-53o/oo) and hypersaline (56-70o/oo). The salinity variation is due to strong evaporation and restriction of seawater inflow by the Faure sill (Fig. 2). Organic communi ties change with salinity. In the metahaline phase, seagrass communities dominate. In the hypersaline phase, the biota is restricted to bivalves and algae with non-skeletal grains, such as ooids, intraclasts and pellets. The Shark Bay samples are stromatolitic or microbial limestones with foraminifers, molluscs o
144 143 142
176
Great Barrier Reef
The Great Barrier Reef is the largest modern reef province in the world. It is about 2000 km long, 23-290 km wide and is composed of about 2500 reefs. It is situated off the north-east of Australia in latitudes ranging from 9 to 24oS. Open seawater temperatures range from 21 to 29 C. The salinity variation is small and is between 35 and 36.6o/oo because of strong ocean water circulation. Three major sediment facies were recognized from the modern northern Great Barrier Reef (Flood & Orme, 1988). These are coastal terrigenous sand facies, transitional facies and impure carbonate algal facies. The impure carbonate facies are ex tremely heterogeneous mixtures of terrigenous mud and debris of corals, algae and other skeletons. The
"-., "-.. 122cm "-., "' 100cm""-....._ ""20km
__j
o
137 Scm
139 22cm 63cm
1 03cm
176cm
63cm /
230cm
242cm "-..
140 212
/
/
/
320cm
Fig. 1. Core samples from Great Barrier Reef, north-east
Australia.
Subtropical and temperate carbonate elemental composition
313
Table 1. Elemental composition of subtropical bulk carbonates from the Great Barrier Reef, Shark Bay and Persian
Gulf Sample no.
Mg(%)
Sr(p.p.m.)
Na(p.p.m.)
Mn(p.p.m.)
Fe(p.p.m.)
Great Barrier Reef Top 83/16 30 em top 83/13 330 cm 83/13 200 em 83/13 200 em 83/20 200 em 83/20 100 em 83/2 137 em Core 140 down 22 em Core 140 down 85 em Core 140 down 175 em Core 140 down 230 em Core 140 down 320 em Core 137 top 5 em Core 137 down 103 em Core 137 down 242 em Core 144 top Core 144 down 122 em Core 212 top Core 212 down 63 em Williamson No. 4 Mynidon3-186(27.1-28.61) Boulder! 7A(19.65-20.75) RibbonS No. 37A OTI No. 5(15.05-16.73) Coral No. I Coral No. 2
1.71 2.35 1.72 1.58 1.93 1.51 2.25 2.21 0.73 0.87 0.81 1.11 0.74 1.47 1.10 1.15 2.66 3.17 2.22 2.27 0.25 0.36 0.11 0.31 0.11 0.24 0.28
3592 3651 5687 5925 4874 6188 3239 3390 7049 7184 7247 6790 7499 5757 7409 7097 2542 2484 3177 3051 860 1775 813 1543 7135 1474 3695
9853 21 661 12 439 12 141 12 355 12 241 14 189 14 478 5968 7442 7718 8269 6885 8459 II 510 9059 17 790 33 036 4933 4877 444 522 534 1085 4219 896 1962
155 436 94 85 95 95 333 452 12 16 15 21 13 47 56 51 299 159 13 14 29 30 17 12 7 13 8
3871 10 735 3818 3221 3178 3953 6414 7507 261 556 455 680 355 2123 2832 2595 17 595 26 772 226 167 245 67 175 132 46 168 95
0.30 0.34 0.29 0.40 0.46 0.50 0.66 0.65 0.35 0.29 0.32 0.33 0.37 0.38 0.40 0.58 0.44 0.42 0.43 0.43 0.37 0.34 0.46 0.49 0.38 0.40 0.39
9039 8935 8654 8176 7369 7390 7099 6944 8127 8404 8228 8048 7996 7989 8664 8291 8211 8160 8194 7986 8488 8144 8617 8454 8365 8238 7789
5656 5518 5037 4884 4658 4766 5078 4562 4791 4799 4819 4703 4662 4457 5504 5037 4359 4278 4343 4428 5021 4464 6568 5807 5593 6273 7432
5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
333 315 269 330 263 261 269 228 262 254 312 282 298 298 734 829 568 527 543 526 422 593 1880 1813 1294 1330 1104
Shark Bay lA IB IC ID IE IF 2A 2B 2C 2D 2E 2F 2G 2H 3A 3B 3C 3D 3E 3F 3G 3H 4A 4B 4C 4D 4E
Continued on p. 314.
314
C. P.
Rao, Z. Z. Amini & J. Ferguson
Table 1. Continued.
Sample no.
Mg(%)
Sr(p.p.m.)
Shark Bay (Continued) SA 58 5C 50 SE
0.40 0.31 0.30 0.37 0.43
8781 8615 8398 8235 8289
5688 4639 5136 4583 6067
5 5 5 5 5
861 1193 1412 1456 1249
0.66 1.16 0.54 0.45 0.73 0.99
5955 6925 7191 6989 6584 3740
4029 14 295 4436 3176 3347 2811
21 75 18 18 32
219 1626 232 285 363 679
Na(p.p.m.)
Mn(p.p.m.)
Fe(p.p.m.)
Persian Gulf I
2 3 4 5 6
and micrite (Logan et al., 1974; Burne & Moore, 1987). Persian Gulf
The Persian Gulf is a landlocked sea that covers a 2 large area (c. 226 000 km ). The average water depth is about 35 m, and it is virtually cut off from open
Fig. 2. Samples from Hamelin Pool, Shark Bay, Western Australia.
Ill
ocean circulation. The average Gulf water tempera tures fluctuate mostly between 4o·c in summer and 15·c in winter. The salinity is in the range 40-50%o in open shallow waters, and 60-70%o in remote la goons, rising to 60 to > 1 OO%o in coastal embay ments. The facies patterns along the Trucial coast of the United Arab Emirates of the Persian Gulf de pend on three major factors (Purser & Evans, 1973): orientation of the shoreline with respect to 'shamal' winds, proximity to the Qatar Peninsula (an up-wind barrier) and the presence of the Great Pearl Bank coastal barrier. The western regiori. is protected lat erally by the Qatar Peninsula from 'shamal' winds. The subtidal facies are mainly carbonate muds with molluscs. The intertidal flat facies are pelletal sands with imperforate foraminifers. The protection de creases rapidly to the east, where the sabkha embay ments contain fringing reefs, and oolitic and mollus can sands. The central region is protected from 'shamal' winds by the Great Pearl Bank, which acts as a coastal barrier with tidal deltas and channels and passes laterally to coastal lagoons, intertidal flats and wide coastal sabkha (supratidal flat). On the seaward side of the barrier, spectacular coral reefs and oolitic tidal deltas have formed. The barrier is covered by molluscan (bivalve) sand; whereas lagoons contain mud with imperforate foraminifers, gastropods, pel lets and gastropods. Intertidal flats have abundant algal mats. Supratidal flats contain algae, mud, evaporites and dolomite. The north-eastern region is unprotected from 'shamal' winds, and a linear coast line has developed. The sediments here are muddy sands and clean sands (with oolites, pellets and skel etons); these suffer the effects of maximum wave fetch, which has led to the development of major
Subtropical and temperate carbonate elemental composition longshore spit systems. The Persian Gulf samples studied are from beaches on the Trucial coast. These contain abundant ooids (up to 95%) with intraclasts and various amounts of biota, such as molluscs, for aminifers, echinoderms, skeletal debris and micrite. Tasmania
Carbonate is the predominant sediment on the tem perate shelf off Tasmania (Fig. 3). Siliciclastics occur in water shallower than about 30 m and carbonates in deeper water (greater than 30 m). The modern carbonates are mixed with, and grade into, pre Holocene glacial carbonates on the outer continental shelf between 130 and 200 m. The average winter and summer surface-water temperatures on the shelf around Tasmania are about 11 C and 16 C, respec tively. During the maximum of the last glacial period (c. 18 000 yr BP), sea-level dropped about 130 m and shallow-marine carbonates formed around Tasma nia; surface-water temperatures were about 4 C lower than they are now. The salinity of the deep to shallow Tasman Sea ranges from 34.4o/oo to 35. 7o/oo because of mixing of water masses (Rao & Huston, 1995). In summer, seawater temperatures and salino
315
ity are higher than winter ones, because of the influx of the Eastern Australian Current. In winter, seawa ter temperatures and salinity decrease as a result of introduction of subantarctic water. Through the year, the Tasman Sea water is mixed with low salinity and -temperature deep Antarctic intermedi ate water. The temperate Tasmanian carbonates comprise mainly skeletal grains, intragranular ce ments and rare non-skeletal grains (Rao, 1981b, 1990b). In Tasmanian carbonates bryozoans are the dominant fauna in all samples, along with some mol luscs, foraminifers, echinoderms, brachiopods, algae and coccoliths. The samples studied here are from eastern and western Tasmania, collected by dredging on a grid of 10 nautical miles.
o
o
ELEMENTAL COMPOSITION
Subtropical bulk carbonates
Elemental composition of subtropical bulk carbon ates presented here is from the Great Barrier Reef, Shark Bay and Persian Gulf samples. Magnesium
=-========= 0
•
•
•
•
0
:.:.:.:.:. .
: :.: ������
. . . . .. .. .
- - --. .-:. :-:-:-,:-:-:-:-···
In these samples, the concentration of Mg ranges from 0.11 to 3.2%. Aragonite contains < 1o/o Mg whereas high-Mg calcite contains 1-3.2% Mg (Fig. 4). These ranges are similar to equilibrium Mg concentrations observed in other tropical world wide carbonates (Veizer, 1983a,b) .
.:J
.�
[] Quartz sand i!=P.I Quartz/ 1t:ti1 carbonate sand
l8l Mud jgj Carbonate 44°8 Fig. 3. Distribution of shelf carbonates and siliciclastics
off Tasmania.
Strontium The Sr concentrations range from 813 to 9039 p.p.m. in examples presented here (Fig. 4) and are similar to those in other tropical carbonates (Morse & Mackenzie, 1990). The variation of Sr in the sam ples studied here is mainly related to relative propor tions of aragonite containing around 9000 p.p.m. Sr and calcite with Sr values;;;. 813 p.p.m.. Sodium The Na content in the bulk carbonates studied here ranges from 444 to 33 036 p.p.m. with a mean value of 6780 p.p.m.. These Na values are posi tively correlated with Mg concentrations, with a significant r2 value of 0.67 (Fig. 5). The increase in Na with Mg values results from the increase in the
316
C. P.
�
Aragonite
Rao, Z. Z. Amini & J. Ferguson 500
+ ppmSr: P.G. • ppmSr: GBR.
•
o ppmSr: SB
'/+� )
4000
•
+
•
....
....
.
-5-u 0
....t..i.
.
+
•
.
•
..
1.5
.5
_....
2.5
2
3
3.5
Mg%
Fig. 4. Variation of Sr and Mg in bulk carbonates from the Great Barrier Reef(GBR), Persian Gulf (PG) and Shark Bay (SB). In most samples Sr concentrations decrease with increasing Mg values because of increasing amounts of high- Mg calcite and decreasing amounts of aragonite. The biotic calcite line is after Carpenter & Lohmann(1992).
proportions of calcite in the samples. These Na values are much higher than those in other world wide carbonates (Veizer, 1983a,b). As explained below, these anomalously high concentrations of Na are related t'<�-high calcitic biota content in the Great Barrier Reef samples, high rate of carbonate formation and high-salinity sea water in Shark Bay and the Persian Gulf.
3500Qt--��---r--�-�-��-�.....____,
25000
•
•
Shark E\ay
50 • 0
Mg%
30000
Aragonite
100
......
tliotic calcite
7 o���--�--4 0 .5 1.5 2 2.5 3 3.5
1000
High-Mg calcite
0..150 •
• •
25 [2o0
3000 2000
o p p mMn: SB
..5 300 � 0
§.. 5000 0..
•
• pp mMn: GBR.
350
+ +
J) 6000
+ p pmMn: P.G. •
450 40 0
+ ppmNa P.G.
Fig. 6. Variation of Mn and Mg in bulk carbonates from
the Great Barrier Reef(GBR), Persian Gulf (PG) and Shark Bay(SB). It should be noted that Mn concentrations are low in aragonitic samples and high in high- Mg calcitic samples.
Manganese and iron The Mn concentrations (Fig. 6; 5-452 p.p.m.; mean 47 p.p.m.) are small in aragonite samples with < 1o/o Mg and high in calcite samples with > 1 o/o Mg. Sim ilarly, Fe concentrations (Fig. 7; 46-26 772 p.p.m.; mean 1,907 p.p.m.) are small in aragonite samples with < 1o/o Mg, whereas Fe concentrations are high in calcite samples with > 1o/o Mg. These Mn and Fe concentrations in the bulk samples studied here (Figs. 3, 6 & 7) are higher than those in other world wide calcitic and aragonitic carbonates (Morrison &
• ppmNa GBR. o ppmNa SB
High-Mg calcite
30000
•
.,20000 z E 15000 Aragonite 0.. 0..10000
+ p pmFe: P.G.
25000
o ppmFe: SB
20000
r.1: 15000 E
..
y 5992.987x 1917.828, r2 .667 -5000,+--��---+���-�---+ .5 1.5 0 3.5 2 2.5 3 =
+
=
Mg%
Fig. 5. Variation of Na and Mg in bulk carbonates from
the Great Barrier Reef(GBR), Persian Gulf(PG) and Shark Bay(SB). It should be noted that Na values increase with increasing Mg contents because of the abundance of high- Mg calcite biota in the Great Barrier Reef samples. The Shark Bay and Persian Gulf samples have Na values higher than 2700 p.p.m. because of higher salinity of sea water.(See text for details.)
8:woo0
•
-500v 0
High-Mg calcite
Aragonite
•
�
5000 0
•
A ppmFe: GBR.
mil1
.5
' :··. 1.5
2
2.5
3
3.5
Mg%
Fig. 7. Variation of Fe and Mg in bulk carbonates from
the Great Barrier Reef(GBR), Persian Gulf(PG) and. Shark Bay(SB). It should be noted that Fe concentrations are low in aragonitic samples and high in high- Mg calcitic samples.
317
Subtropical and temperate carbonate elemental composition 30000
ppmFe: P.G. • ppmFe: GBR. o ppmFe: SB
+
25000 20000 �15000 E
2::10000 5000 0
=.79
-5000 -5000
0
5000 10000 15000 20000 25000 30000 35000 ppmNa
aragonite (Veizer, 1983a,b; Rao, 1996a). Therefore, low Mg values in the Shark Bay and Persian Gulf samples are due to high aragonite content, very low to high Mg values in the Great Barrier Reef samples are related to calcite to aragonite mixed mineralogy, and mainly intermediate Mg values in temperate Tasmanian carbonates result from the occurrence of low- to intermediate-Mg calcite. As Mg content in calcite increases with rising temperature (Mucci, 1987; Burton & Walter, 1991), subtropical Great Barrier Reef samples contain higher Mg values compared with temperate Tasmanian carbonates (Fig. 9).
Fig. 8. Variation of Fe and Na in bulk carbonates from
the Great Barrier Reef(GBR), Persian Gulf (PG) and Shark Bay(SB). It should be noted that Fe and Na concentrations are positively correlated.
Brand, 1987). The Fe and Na values are positively correlated, with a significant? value of 0.8 (Fig. 8). Comparison with temperate carbonates
Magnesium The Mg contents in carbonates are low in Shark Bay, moderate in the Persian Gulf, high in Tasma nia, and range from lowest to highest in the Great Barrier Reef (Fig. 9). The Mg concentrations are high and range up to a few per cent in calcite, whereas Mg contents are small and are < 1o/o in
8
0.. 0.. bO
�
3.5 3 2.5 2 1.5
+
The Sr concentrations in bulk carbonates are low in the temperate Tasmanian carbonates, moderate in the Great Barrier Reef samples, high in the Persian Gulf samples and highest in the Shark Bay samples (Fig. 10). The amount of Sr in aragonite is between 8000 and 10 000 p.p.m., whereas in calcite Sr ranges from 900 to 1800 p.p.m. The highest Sr concentrations of around 8000 p.p.m. in Shark Bay are related to the predominance of aragonite in a mainly aragonite-calcite mixture. Low to moderate Sr concentrations in the Great Barrier Reef result from the mixture of calcite and aragonite, whereas the lowest Sr concentrations in temperate Tasma nian carbonates are due to the predominance of intermediate-Mg calcite and some aragonite (Rao & Adabi, 1992).
14000r'-�-�--���--�-�......
MgppmGBR
+
• MgppmPG o
Strontium
12000
MgppmSB
10000
eMgppm TAS
SrppmGBR
• SrppmPG o
SrppmSB
ll
Srppm TAS
1
-.5-'--r---�--�----.-�--,--L 0
20
40
60
80
100
Percentile
0
20
40
60
80
100
Percentile
Fig. 9. Percentile distribution of Mg in bulk carbonates
Fig. 10. Percentile distribution of Sr in bulk carbonates
from the tropical Great Barrier Reef (GBR), Persian Gulf(PG) and Shark Bay(SB), and temperate carbonates from Tasmania(TAS).(See text for details.)
from the tropical Great Barrier Reef(GBR), Persian Gulf(PG) and Shark Bay(SB), and temperate carbonates from Tasmania(TAS).(See text for details.)
318
C. P.
+
25000
Rao, Z. Z. Amini & J. Ferguson
300�----�----��--y,
NappmGBR
+
•NappmPG
250
o NappmSB
20000
o
!!. Nappm TAS
MnppmGBR
•MnppmPG MnppmSB
!!. MnppmTAS
200 8
fusooo
0.. 0..
0.. "'
150
�
'Soooo
100 50 0
20
40
60
80
100
0
Percentile
20
40
60
80
100
Percentile
Fig. 11. Percentile distribution of Na in bulk carbonates from the tropical Great Barrier Reef (GBR), Persian Gulf(PG) and Shark Bay(SB), and temperate carbonates from Tasmania(TAS).(See text details.)
Sodium The Na concentrations in bulk carbonates are low in the temperate Tasmanian. carbonates, low to moderate in the Persian Gulf samples, high in the Shark Bay samples, and range from lowest to highest in the Great Barrier Reef samples (Fig. 1 1). Na contents in the samples studied are related to the combination of salinity, biochemical fractionation and growth rate. The salinity of sea water around Tasmania and the Great Barrier Reef is 35o/oo, whereas salinity ranges up to 300o/oo in tidal flats and lagoons in Shark Bay and the Persian Gulf (Rao, 1996a). The increase in Na from the temper ate Tasmanian to the Persian Gulf and Shark Bay bulk carbonates is due to increasing salinity. The lowest Na values observed in the Great Barrier Reef samples (Fig. 11) represent normal seawater salinity of 35o/oo, whereas the highest Na values in the Barrier Reef samples are due to a combination of biochemical fractionation and growth rate. The biotic calcites have higher Na values relative to abiotic calcites as a result of biochemical fraction ation (Morrison & Brand, 1987; Rao, 1996a). The Na values increase with increasing rates of crystal growth (Busenberg & Plummer, 1985). The rate of formation of tropical carbonates is greater than that of temperate ones (Rao, 1994, 1996a).
Fig. 12. Percentile distribution of Mn in bulk carbonates
from the tropical Great Barrier Reef (GBR), Persian Gulf(PG) and Shark Bay (SB), and temperate carbonates from Tasmania(TAS).(See text for details.)
samples, moderate in the Persian Gulf carbonates, high in the temperate Tasmanian carbonates, and lowest to highest in the Great Barrier Reef samples (Fig. 12). The Fe contents are similar in the Persian Gulf and Shark Bay samples, moderate in the temperate Tasmanian carbonates, and lowest to highest in the Great Barrier Reef samples (Fig. 13). The vari�tion of Mn and Fe in modern carbonates is related to carbonate mineralogy, oxidizing and reducing conditions, and availability of Mn and Fe released from terrigenous sediments. The concen-
10000 �----��-�--�-----'--t--r-'-T + FeppmGBR 9000 • FeppmPG 8000 o FeppmSB 7000 !!. Feppm TAS §_ 6000 �5000 p.. 4000 3000 2000 100 �G������==:_�_j
�1
0
20
40
60
80
100
Percentile
Fig. 13. Percentile distribution of Fe in bulk carbonates
Manganese and iron The Mn contents are lowest
m
the Shark Bay
from the tropical Great Barrier Reef (GBR), Persian Gulf(PG) and Shark Bay(SB), and temperate carbonates from Tasmania(TAS).(See text for details.)
Subtropical and temperate carbonate elemental composition trations of Mn and Fe are low in aragonite (< 20 p.p.m.), whereas calcite can take up to a few per cent (Mucci, 1987; Rao & Jayawardane, 1994). Thus, aragonitic carbonates from Shark Bay and the Persian Gulf contain low Mn and Fe concentrations relative to calcitic carbonates from temperate Tas mania. Low Mn and Fe concentrations enter calcite in oxidizing conditions, whereas high concentra tions of these elements can occur in calcite in reducing environments. Tasmanian carbonates studied are from water depths from 30 to 200 m, where conditions are predominantly reducing. In contrast, the Shark Bay, Persian Gulf and Great Barrier Reef carbonates are from water depths less than 50 m, where conditions are predominantly oxidizing. The availability of Mn and Fe released from terrigenous sediments is low in Shark Bay and the Persian Gulf, because in these areas mostly pure carbonates are forming. In contrast, the temperate Tasmanian carbonates and the Great Barrier Reef carbonates grade into coastal siliciclastics deposits that release Mn and Fe into shallow seas in these regions.
DISCUSSION
The vanatJon of major and minor elements in subtropical to temperate carbonates is mainly re lated to: (i) carbonate mineralogy; (ii) seawater temperatures; (iii) seawater composition; (iv) frac tionation of elements; (v) rate of precipitation; (vi) oxidizing and reducing conditions; (vii) pC02level, and (viii) salinity. Carbonate mineralogy
Carbonate mineralogy is a major control on Mg, Sr, Na, Mn and Fe values in modern carbonates. Abiotic aragonite contains low concentrations of Mg ( < 1%), Mn (mostly <5 p.p.m.) and Fe (< 50 p.p.m.), moderate Na (c. 2700 p.p.m.) and high Sr (c. 9000 p.p.m.). Biotic aragonite contains lower concentrations of Sr ( < 9000 p.p.m.) higher amounts of Mn (> 5 p.p.m.) and Fe (> 50 p.p.m.), and variable amounts of Mg and Na relative to abiotic aragonite (Morrison & Brand, 1987). Abi otic calcite and biotic calcite have a similar range of Mg concentrations, because of the occurrence of similar types of low-Mg to high-Mg calcites. Mg and Sr (Fig. 4), and Mg and Na (Fig. 5) concentrations are positively correlated in bulk carbonates because
319
of variable amounts of aragonite and calcite com ponents, biotic content and salinity. Seawater temperature
Seawater temperature determines the carbonate mineralogy and thus affects the concentrations of elements in carbonates. Experimental studies dem onstrate that entirely abiotic aragonite forms at > 30·c, mixtures of abiotic aragonite and high-Mg calcite precipitate between 16 and 30·c, whereas low-Mg calcite forms at < 3·c (Kinsman & Hol land, 1969). Biotic carbonate mineralogy also varies with seawater temperature (Lowenstam, 1954; Morse & Mackenzie 1990). The mol o/o MgC03 in abiotic calcite decreases with lower seawater temperature because of changes in calcite mineralogy from high-Mg to low-Mg calcite (Mucci, 1987; Burton & Walter, 1991). Similarly, mol o/o MgC03 in many biotic calcites decreases with lower seawater temperature because of changes in calcite mineralogy from high-Mg to low-Mg calcite (Chave, 1954; Morse & Mackenzie, 1990). The biota that show the temperature depen dence in calcite mineralogy are benthic foramin ifera, bryozoans, echinoids and barnacles, and these are important constituents in temperate carbonates. Higher Mg contents in the Great Barrier Reef samples relative to temperate Tasmanian ones are due to higher seawater temperatures in tropical regions. Seawater composition
In sea water, the concentrations of Mg, Sr and Na increase with increasing salinity, and Fe and Mn contents depend on input of terrigenous material that contains high concentrations of Fe and Mn. The Mg concentrations in marine calcite are not related to the rate of precipitation or saturation of CaC03 (Mucci, 1987; Burton & Walter, 1991). Therefore, the relationships between Mg, Sr, Fe and Mn values observed in modern carbonates are due to seawater temperature, seawater composition, oxidizing and reducing conditions, and pC02 in seawater. Similar slopes of regression lines of Sr and Mg between biotic and abiotic calcites are due to uniform composition of Mg and Sr in normal sea water (Fig. 4; Carpenter & Lohmann, 1992). Sr and Na concentrations in sea water increase with in creasing salinity. High Na concentrations in carbon ate from Shark Bay and the Persian Gulf are due to
320
C. P.
Rao, Z. Z. Amini & J Ferguson
an increase in salinity. Sr concentrations in arago nite from Shark Bay and the Persian Gulf are similar despite the increase in salinity. This is because the Sr concentrations in carbonates are not significantly affected by increasing concentrations of Sr in sea water. Higher concentrations of Mn and Fe in bulk carbonates that contain > 1% Mg than in those samples that contain < 1% Mg indicate that Mn and Fe enter calcite preferentially compared with aragonite. Fractionation of elements
The incorporation of a trace element into a CaC03 lattice is governed by the distribution coefficient (D) in the following equation (Mcintire, 1963; Kins man, 1969): (mMelmCa)S
=
D (mMelmCa) W,
where m is molar concentration, Me is trace ele ment, Ca is calcium, and S and W indicate solid phase (i.e CaC03) and water, respectively. This equation is valid only when the system is at com plete equilibrium and the water and solid phase do not show any concentration gradients in Me during precipitation (homogeneous distribution law; Gor don et a!., 1959). The distribution coefficient D is related to the equilibrium constant K by several equations (Veizer, 1983b). These show that the dis tribution coefficient is a complex function of tem perature, pressure and the chemical compositions of the liquid and solid phases. Distribution coefficients involving various ele ments incorporated into CaC03 during the precipi tation of aragonite and calcite have been summa rized by Veizer (1983a,b). The distribution of the minor and trace elements is governed by the follow ing conditions of the distribution coefficient (Veizer, 1983a,b; Morrison & Brand, 1987): 1 When D 1, the precipitated carbonate contains similar amounts of Me relative to the carrier in both liquid and solid; 2 When D > 1, there is an enrichment of the Me concentration in the precipitated solid phase rela tive to its proportion in the liquid phase; 3 When D < 1, there is a proportional depletion of the minor and trace elements in the solid phase relative to their proportions in the liquid phase. In general, cations that are larger than Ca (e.g. Sr, Na) are preferentially incorporated into the open orthorhombic structure of aragonite. Cations that are smaller than Ca (e.g., Mg, Fe, Mn) are preferen=
tially incorporated into the tighter rhombohedral structure of calcite. The precise calculation of dis tribution coefficients is only necessary to determine the exact water chemistry. The magnitude and sign of distribution coefficient are sufficient to provide information on the seawater composition and the effect of diagenesis (Veizer, 1983a,b; Morrison & Brand, 1987). Rate of precipitation
The rate of precipitation affects the concentrations of Mg, Sr, Na and Mn in calcitic carbonate, partic ularly in biotic calcite. Kolesar ( 1978) observed that the Mg concentration in the coralline alga Calliar t hron are related to its growth rate. In summer months the growth of the alga is slow, which means that the demand for Mg during calcification is matched by the supply from sea water. In winter months the growth of the alga is so fast that the Mg in solution cannot keep pace with the demand. Thus, the growth rate of algae is also a function of water temperature. The Sr distribution coefficient is correlated with changes in calcite precipitation rate (Lorens, 1981; Mucci & Morse, 1983; Mucci, 1988). Slow rates correspond to equilibrium conditions and high rates to kinetic effects. Thus relatively high Sr concentrations in biotic calcite result from rapid precipitation associated with shell growth of marine organisms (Fig. 4; Carpenter & Lohmann, 1992). Experimental studies indicate that the amounts of Na and S04 in abiotic calcite increase linearly with increasing crystal growth rate (Busenberg & Plummer, 1985). The number of crystal defects increases with increasing crystal growth rate. The amount of Na that can be incorporated in the calcite depends on the number of crystal defects. The large variation of Na values in biotic calcite (Land & Hoops, 1973; Busenberg & Plummer, 1985) is related to salinity, crystal growth rates, ionic substitution and other causes. Experimental studies indicate that Mn concentra tions decrease with increasing crystal growth rate (Lorens, 1981; Mucci, 1988; Pingitore et a!. 1988), in contrast to Mg, Sr and Na values for which concentrations increase with increasing crystal growth rate. Lower concentrations of Mn relative to Mg, Sr and Na in bulk carbonates and biotic calcite are due to the fast rate of formation of tropical carbonates and biota.
Subtropical and temperate carbonate elemental composition Oxidizing and reducing conditions
Mn and Fe concentrations are sensitive to oxidizing and reducing conditions (Eh). In oxidizing waters Mn and Fe rapidly precipitate as highly insoluble ferric and manganoan oxyhydroxides, and the wa ter will contain very small concentrations of these elements. In reducing sea waters (anaerobic waters) Mn and Fe can enter the calcite lattice in apprecia ble amounts (Mucci, 1988). pC02 in sea water
The Mg concentrations in marine calcite vary with seawater temperature, pC02 levels and sulphate concentrations in sea water (Burton & Walter, 1991 ). Regardless of temperature, Mg concentra tions increase linearly with decreasing pC02 levels. Therefore, variations in Mg concentrations other than those attributable to a given temperature are due to changes in pC02 level. Salinity
Na concentrations in abiotic aragonite and calcite indicate the salinity of sea water (Land & Hoops, 1973). In biotic calcite and aragonite Na concentra tions are due to both salinity and rates of crystal growth (Busenberg & Plummer, 1985; Morrison & Brand, 1987). The positive correlation observed between Na and Mg concentrations in bulk carbon ates from the Great Barrier Reef, Persian Gulf and Shark Bay (Fig. 5) is due to a combination of salinity and growth rate of biota. Na values in the Great Barrier Reef bulk carbonate, which has formed in normal seawater salinity of 35o/oo, are much higher than those of the Persian Gulf and Shark Bay because of rapid growth of biota, which increases the amount of crystal defects into which Na is incorporated. Na values in marine abiotic aragonite forming at normal salinity of 35o/oo are around 2700 p.p.m. Na values in the Persian Gulf bulk carbonates (2500- 13 500 p.p.m.) and in the Shark Bay bulk carbonates (4000-7000 p.p.m.) are much higher than 2700 p.p.m. partly because of higher salinities of about 40-60o/oo. Preservation of original 'elemental' signal in ancient carbonates
Ancient carbonates are affected by meteoric and/or burial diagenesis. This normally leads to a decrease
32 1
in Mg, Sr and Na, and an increase in Mn and Fe concentrations relative to modern carbonates (Veizer, 1983a,b). Despite these changes, the rela tive concentrations and ratios of elements depict original mineralogy and elemental concentrations (Rao, 1981a, 1990b, 1991). For example, tropical Ordovician Tasmanian carbonates, interpreted to be originally aragonitic on the basis of petrograph ical features, contain high Sr, moderate Na and low Mn, similar to relative elemental concentrations in modern tropical aragonitic carbonates (Rao, 1981a, 1990b). In contrast, subpolar calcitic Permian polar carbonates, which contain glacial dropstones, con tain equal amounts of Sr and Na and higher amounts of Mn and Fe, similar to modern calcitic carbonates (Rao, 1991). Similarly, originally cal citic temperate Cainozoic limestones from New Zealand were found to contain lower concentra tions of Mg and Sr and higher concentrations of Na, Fe and Mn than tropical carbonates (Winefield et al., 1996). In these calcitic non-tropical carbonates, the ratio of Sr/Na is about unity or less, in contrast to higher Sr/Na ratios greater than three in origi nally aragonitic tropical carbonates (Rao, 1991; Winefield et al., 1996).
CONCLUSIONS
The major features of elemental composition in subtropical and temperate carbonates are as follows: 1 The Mg and Sr concentrations in the bulk carbonates from the Great Barrier Reef, Persian Gulf and Shark Bay are similar to those of other tropical carbonates, whereas Na, Mn and Fe con centrations in these carbonate localities are higher than in other subtropical carbonates because of high salinity in the Persian Gulf and Shark Bay areas and abundant biota in the Great Barrier Reef area. The Mg contents in temperate Tasmanian carbonates are higher than those of aragonitic subtropical carbonates. Sr concentrations in temperate carbon ates are lower than in subtropical counterparts because of calcitic mineralogy of temperate carbon ates. The Na contents are lower and Mn and Fe concentrations are higher in temperate carbonates relative to tropical aragonitic carbonates because of normal salinity of sea water, higher rate of precipi tation and the formation of temperate carbonates in a reducing (dysaerobic) environment. 2 Inorganic components have a similar range of Mg, higher Sr, and lower Na, Mn and Fe concentra-
322
C. P.
Rao, Z. Z. Amini & J. Ferguson
tions relative to biotic equivalents. Non-skeletal grains are absent in temperate carbonates. Cements occur in temperate carbonates but these were not chemically analysed because of the fine crystal size of cements. 3 Carbonate mineralogy is a major control on Mg, Sr, Na, Mn and Fe concentrations in tropical and temperate carbonates. Subtropical carbonates are entirely aragonite or mixtures of aragonite and calcite, whereas temperate carbonates are mainly calcite with some aragonite. 4 Seawater temperatures determine carbonate mineralogy and thus affect the concentrations of elements in carbonates. Mg and Sr concentrations decrease with lower seawater temperature. 5 In sea water, the concentrations of Mg, Sr and Na increase with increasing salinity, and Mn and Fe contents depend on input of terrigenous material. 6 The concentration of elements in carbonates is determined by distribution coefficients. Cations larger than Ca are preferentially incorporated into the open orthorhombic structure of aragonite, whereas cations that are smaller than Ca are prefer entially incorporated into the tighter rhombohedral structure of calcite. 7 Mg, Sr and Na values increase and Mn values decrease with increasing rate of crystal growth. 8 Appreciable Mn and Fe concentrations enter the calcite lattice in preference to aragonite in reducing conditions. 9 Mg concentrations other than those attributable to a given temperature are due to changes in pC02 levels. 10 Na values in abiotic and biotic carbonates in crease with increasing salinity and rate of precipita tion. 1 1 Relative cocentrations of elements and their ratios can be used to differentiate originally tropical aragonitic limestones from calcitic non-tropical limestones.
ACKNOWLEDGEMENTS
Financial assistance was provided by a grant from the University of Tasmania. We thank P. Robinson for supervision of chemical analysis, and Richard Orme and John Marshall for providing the Great Barrier Reef samples. We also thank Lucien Mon taggioni and Gilbert Camoin for their valuable comments, which improved an earlier version of the manuscript.
REFERENCES
T.D. & JAMES, N.P. (1995) Holocene sediment dynamics on a cool-water shelf: Otway, southeastern Australia. J. sediment. Petrol., 63, 574-588. BROOKFIELD, M.E. (1988) A mid-Ordovician temperate carbonate shelf-the Black River and Trenton Lime stone .Groups of southern Ontario, Canada. In: Non tropical Shelf Carbonates-Modern and Ancient (Ed. Nelson, C.S.). Sediment. Geol., 60, 137-153. BURNE, R.V. & MOORE, L.S. (1987) Microbialites: orga nosedimentary deposits of benthic microbial communi ties. Palaios, 2, 241-254. BURTON, E.A. & WALTER, L.M. (1991) The effects of pC02 and temperature on magnesium incorporation in calcite in seawater and MgC12-CaC12 solutions. Geochim. Cos mochim. Acta, 55, 777-785. BUSENBERG, E. & PLUMMER, N.L. (1985) Kinetic and thermodynamic factors controlling the distribution of S04 and Na in calcites and selected aragonites. Geochim. Cosmochim. Acta, 49, 713-725. CARPENTER, S.J. & LOHMANN, K.C. (1992) Sr/Mg ratios of modern marine calcite: empirical indicators of ocean chemistry and precipitation rate. Geochim. Cosmochim. Acta, 56, 1817-1849. CHAVE, K.E. (1954) Aspects of the biogeochemistry of magnesium 2: Calcareous sediments and rocks. J. Geol., 62, 587-599. DRAPER J.J. (1988) Permian limestone in the southeastern Bowen Basin, Queensland: an example of temperate car bonate deposition. In: Non-tropical Shelf Carbonates Modern and Ancient (Ed. Nelson, C.S.). Sediment. Geol., 60, 155-162. FLOOD, P.G. & 0RME, G.R. (1988) Mixed siliciclastic/ carbonate sediments of the northern Great Barrier Reef province, Australia. In: Carbonate-Clastic Transitions (Eds Doyle, L.J. & Roberts, H.H.), pp.175-205. Elsevier, Amsterdam. GORDON, L., SALUTSKY, M.L. & WILLARD, H.H. (1959) Precipitation from Homogeneous Solution. John Wiley, New York. 2 KINSMAN, D.J.J. (1969) Interpretation of Sr + concentra tion in carbonate minerals and rocks. J. sediment. Petrol., 49, 937-944. KINSMAN, D.J.J. & HOLLAND, H.D. (1969) The coprecipi tation of cations with CaC03• IV, The coprecipitation of 2 Sr + with aragonite between 16 to 96'C. Geochim. Cosmochim. Acta, 33, 1-17. KOLESAR, P.J. (1978) Magnesium in calcite from the coralline alga. J. sediment. Petrol., 48, 815-820. LAND, L.S & HooPs, G.K.(l973) Sodium in carbonate sediments and rocks: a possible index to salinity of diagenetic solutions. J. sediment. Petrol., 43, 614-617. LEES, A. (1975) Possible influence of salinity and temper ature on modern shelf carbonate sedimentation. Mar. Geol., 19, 159-198. LOGAN, B.W., HOFFMAN, P. & GEBELEIN, C.D. (1974) Algal mats, cryptalgal fabrics and structures, Hamelin Pool, Western Australia. In: Evolution and Diagenesis of Qua ternary Carbonate Sequences, Shark Bay, Western Aus tralia. (Eds Logan, B.W., Read, J.F., Hagan, G.M. et a!.), Mem. Am. Assoc. petrol. Geol., Tulsa, 22, 140-194. LORENS, R.B. (1981) Sr, Cd, Mn and Co distribution BOREEN,
Subtropical and temperate carbonate elemental composition coefficients in calcite as a function of calcite precipita tion rate. Geochim. Cosmochim. Acta, 45, 553-561. LowENSTAM, H.A. (1954) Factors affecting the aragonite: calcite ratios in carbonate-secreting marine organisms. J Geol. , 62, 284-322. MAXWELL, W.G.H. (1968) Atlas of the Great Barrier Reef Elsevier, Amsterdam. MciNTIRE, W.L. (1963) Trace element partition coeffi cients-a review of theory and applications in geology. Geochim. Cosmochim. Acta, 27, 1209-1264. MILLIMAN, J.D. (1974) Recent Sedimentary Carbonates I, Marine Carbonates. Springer-Verlag, Berlin. MORRISON, J.O. & BRAND, U. (1987) Geochemistry of Re cent marine invertebrates. Geosci. Can., 13, 237-254. MORSE, J.W. & MACKENZIE, F.T. (1990) Geochemistry of Sedimentary Carbonates, Developments in Sedimentol ogy, 48. Elsevier, Amsterdam. Muccr, A. (1987) Influence of temperature on composi tion of magnesian calcite overgrowths precipitated from seawater. Geochim. Cosmochim. Acta, 51, 19771984. Mucci, A. (1988) Manganese uptake during calcite pre cipitation from seawater: conditions leading to the formation of a pseudokutnahorite. Geochim. Cosmo chim. Acta, 52, 1859-1868. MuCCI, A. & MORSE, J.W. (1983) The incorporation of 2 Mg2 + and Sr + into calcite overgrowths: influences of growth rate and solution composition. Geochim. Cos mochim. Acta, 47, 217-233. NELSON, C.S. (1978) Temperate shelf carbonate sediments in the Cenozoic of New Zealand. Sedimentology, 25, 737-771. PINGITORE, N.R., EASTMAN, M.P., SANDIDGE, M., 0DEN, K. & FREIHA, B. (1988) The co-precipitation of manga nese(II) with calcite: an experimental study. Mar. Chern., 25, 107-120. PuRSER, B.H. & EVANS, G. (1973) Regional sedimentation along the Trucial Coast, SE Persian Gulf. In: The Persian G ulf (Ed. Purser, B.H.), pp. 211-231. Springer Verlag, Berlin. RAo, C.P. (198la) Geochemical differences between trop ical (Ordovician) and subpolar (Permian) carbonates, Tasmania, Australia. Geology, 9, 205-209. RAo, C.P. (1981b) Cementation in cold-water bryozoan sand, Tasmania, Australia. Mar. Geol., 40, M23-M33. RAo, C.P. (1990a) Geochemical characteristics of cool temperate carbonates, Tasmania, Australia. Carbonates and Evaporites, 5, 209-221. RAo, C.P. (1990b) Petrography, trace elements and oxygen and carbon isotopes of Gordon Group carbonates (Ordovician), Florentine Valley, Tasmania, Australia. Sediment. Geol., 66, 83-97. RAo, C.P. (1991) Geochemical differences between sub tropical (Ordovician), cool temperate (Recent and Pleis tocene) and subpolar (Permian) carbonates, Tasmania, Australia. Carbonates and Evaporites, 6, 83-106. RAo, C.P. (1994) Implications of isotopic fractionation and temperature on rate of formation of temperate
323
shelf carbonates, eastern Tasmania, Australia. Carbon ates and Evaporites, 9, 33-41. RAo, C.P. (l996a) Modern Carbonates: Tropical, Tem perate and Polar. Carbonates, Hobart, Tasmania, Aus tralia. RAo, CP. (1996b) Elemental composition of marine cal cite from modern temperate shelf brachiopods, bryozo ans and bulk carbonates, Eastern Tasmania, Australia. Carbonates and Evaporites, 11, 1-18. RAo, C.P. & ADABI, M.H. (1992) Carbonate minerals, major and minor elements and oxygen and carbon isotopes and their variation with water depth in cool, temperate carbonates, western Tasmania, Australia. Mar. Geol., 103, 249-272. RAo, C.P. & AMINI, Z.Z. (1995) Faunal relationship to grain-size, mineralogy and geochemistry in Recent tem perate shelf carbonates, Western Tasmania, Australia. Carbonates and Evaporites, 10, 114-123. RAo, C.P. & GREEN, D.C. (1983) Oxygen- and carbon isotope composition of cold shallow-marine carbonates of Tasmania, Australia. Mar. Geol. , 53, 117-129. RAo, C.P. & HusTON (1995) Temperate shelf carbonates reflect mixing of distinct water masses, eastern Tasma nia, Australia. Carbonates and Evaporites, 10, 105-113. RAo, C.P. & JAYAWARDANE, M.P.J. (1993) Mineralogy and geochemistry of modern temperate carbonates from King Island, Tasmania, Australia. Carbonates and Evaporites, 8, 170-180. RAo, C.P. & JAYAWARDANE, M.P.J. (1994) Major minerals, elemental and isotopic composition in modern temper ate shelf carbonates, Eastern Tasmania, Australia: im plication for the occurrence of extensive ancient non tropical carbonates. Palaeogeogr. Palaeoclimatol. Palaeoecol. , 107, 49-63. RAo, C.P. & NELSON, C.S. (1992) Oxygen and carbon isotope fields for temperate shelf carbonates from Tas mania and New Zealand. Mar. Geol., 103, 273-286. REECKMAN, S.A. (1988) Diagenetic alterations in temper ate shelf carbonates from southeastern Australia. Sedi ment. Geol. , 60, 209-219. RoBINSON, P. (1980) Determination of calcium, magne sium, manganese, strontium, sodium and iron in the carbonate fraction of limestones and dolomites. Chern. Geol., 28, 135-146. VEIZER, J. (1983a) Chemical diagenesis of carbonates: theory and application of trace element technique. In: Stable Isotopes in Sedimentary Geology (Ed. Arthur, M.A.). Soc. econ. Paleont. Miner., Short Course, 10, 3/1-3/100. VEIZER, J. (1983b) Trace element and stable isotopes in sedimentary carbonates. In: Carbonates: Mineralogy and Chemistry (Ed. Reeder, R.J.), Rev. Mineral., Min eral. Soc. Am. Washington, DC, 1 1, 265-299. WINEFIELD, P.R., NELSON, C.S. & HODDER, A.P.W. (1996) Discriminating temperate carbonates and their diage netic environments using bulk elemental geochemistry: a reconnaissance study based on New Zealand Cenozoic limestones. Carbonates and Evaporites, 11, 19-31.
Index
Page numbers in italics refer to figures; those in bold refer to tables.
active margins, fringing reef growth and structure, New Hebrides Island Arc 261-77 Allison Guyot, Mid-Pacific Mountains 79-90 age-depth model l l 0-12 Albian sequences 130-3 Aptian sequences 129-30 comparisons 109 location maps 80, 96, 138 summary 107 Alps, French, Cretaceous, Vercors Platform 13, 14 Alps, Southern, Triassic Schlern Platform 7 Alps, Swiss, Cretaceous, Helvetic Domain l l , 13 ammonites, time-scale calibration
113
Aptian-Albian, Pacific Albian sequence boundaries
115-16 Aptian sequence boundaries
117
and European sequences 112-16 eustatic sea-levels 116-18 Japanese seamounts 60 time-scale calibration 113 Arkansas, Mississippian, Burlington platform, exposure and drowning unconformities, close superposition 18 atolls see coral atolls Aube, Folkestone, correlation with Aptian-Albian, Pacific 114 Australia development of subtropical shelf carbonates 188-9 see also Great Barrier Reef Australia, E climate and oceanography 169 distribution of carbonate sediments 165 Fraser Island and Platform
166-89 Australia, NE, Marion Plateau
11
Boussinesq's approximation of incompressibility 250 Bowling Green Bay, Queensland, sediment accumulation
296-302 bryozoan-foraminifera-mollusc assemblage, and coral Halimeda assemblage 187 Burdekin River delta, Queensland, sediment accumulation
8,
145-61 Australia, W, Shark Bay, geographical and geological setting 312-14
61-2 porous media equations 250-l post-drowning evolution 59 sedimentology 4-5 seismic expression of exposure and drowning events 15-16 sequence stratigraphy 7-9 subtropical vs temperate, four examples 311-23 subtropical development, Quaternary and Tertiary
296-302 Burlington platform, exposure and drowning unconformities, close superposition 18
Cainozoic carbonate platform, Marion Plateau, NE Australia, computer simulation 145-61 Campanian, initiation of guyots 57 Campanian-Maastrichtian, development of platforms 57,
58
112-15 composite sequence biostratigraphy-isotope scale
without exposure 9-15 as global markers 17 nutrient excess 62-4 sea-level changes 62 as sequence boundaries 16-17 topography 6 erosion during drowning 4 exposure before drowning, seismic evidence 17-18 groundwater flow, porous media equations 250-l insights into demise 59-61 origin of morphological features
B0lling meltwater pulse 231-3 borehole response, tidal transients, computer simulation 256-58 Borneo, Oligocene of Kalimantan 9,
Canada, Western Canada Basin, Devonian reefs and platforms
15
carbon dioxide levels, sea-water 321 carbon reservoir function of coral reefs 74 carbonate ancient, preservation of elemental signal 321 composition, subtropical and temperate 311-23 distribution control by terrigenous sediment
295-310 terrigenous sediment supply
295-310 fractionation of elements 320 lithofacies, Brazil 188 production vs water depth 4 and rate of precipitation 320 carbonate platforms computer simulation 145-61 correlation with European continental margins and Haq curve 110, 112-16 development 57-9 drowning events 62-4 defined 3, 59 with exposure 7-9
163-95 unconformities, scenarios
interglaciation to present 70 Cleveland Bay, Queensland, sediment accumulation 296-302 climate change origin of Great Barrier Reef 32-6 see also sea-level changes coastal turbid-zone reefs, central GBR and Gulf of Papua
295-310 Comanchean platforms, southern USA, Cretaceous 12 Comoro Islands, Mayotte, morphology and sediments 219-36 computer simulation borehole response, tidal transients
256-58 groundwater flow, porous media equations 250-l Marion Plateau, Cainozoic carbonate platform, N E Australia 145-61 steady-state interstitial water circulation, coral atolls 249-56 coral atolls cross section 252 homogeneous atoll platform 253-4 porous media equations 250-l
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9 325
5
see also named locations chloralgal lithofacies 188 chlorozoan lithofacies 188 clam, Tridacna spp., last
Index
326 coral atolls (cont.) South China Sea, tectonic and monsoonal controls 237-48 steady-state interstitial water circulation, computer simulation 249-56 transition zone 254-6 coral reefs carbon reservoir function 74 coastal turbid-zone reefs, central GBR and Gulf of Papua
295-310 conceptual existence fields, carbonate production and terrigenous sedimentation 306 development and lagoonal sedimentation, Heron Reef, S Great Barrier Reef Province
281-94 extinction and contraction during Last Glaciation 74 internal structure, Tasmaloum, SW Pacific 261-277 limiting factors 72 coral-Halimeda assembly, and bryozoan-foraminifera-mollusc assembly 187 Cretaceous Comanchean platforms, southern USA 12 Helvetic Domain, Swiss Alps 11,
13
Maracaibo platform, Venezuela Pacific guyots as recorders of
12
sea-level 95-100 polar ice caps 78 Pyrenees, S 9, 12 Shuaiba platform, Persian Gulf 12 Vercors Platform, French Alps 13,
14 Cycloclypeus-Operculina limestone, features 20 l
Darcy's law 250 d'Entrecasteaux Zone, New Hebrides Island Arc, location map 262 Devonian reefs and platforms, Western Canada Basin 15 dolomite, Barremian 140 dolomite, white sucrosic, origin in shallow water 137-44 dolomitization Queensland shelf 190 role of sea water, strontium isotopes
137-44 . drowned reefs 223 drowning events, carbonate platforms 3
Espiritu Santo, SW Pacific, fringing reef biofacies sequences 265-7, 272-3 o180 and o13C, laminated stromatolitic crusts 269 geomorphology and lithology 264-5
lithification 267-70, 273 location map 262 schematic cross-sections 263, 271 tectonic and environmental growth influences 261-77 European margin, sea-level curve, correlation with Aptian-Albian, Pacific 114 EXXON global sea-level chart comparison with Marion Plateau data 159 correlation with carbonate platform sequences 110, 112-16 as a working model 147-8, 156,
157, 158-60
Fieberling Guyot, oscillatory currents 5 Fly River, Gulf of Papua, sediment accumulation 302-5 foraminifera, time-scale calibration
113
Fourier analysis, sedimentary cycles at ODP Sites Leg 143, geophysical logs 77-92 Fraser Island and Platform 166-9 dolomite, comparison with Marion Plateau 190 location maps 166-8 Miocene subtropical carbonate buildups 189 Quaternary subtropical carbonate buildups 189 sedimentology 172-87 rhodoliths, radiocarbon age 178 seismic line 170 seismic stratigraphy 171-2 sub-reef environment, comparison with Great Barrier Reef
34-6 fringing reef, Tasmaloum, SW Pacific, tectonic and environmental growth influences 261-77
Gardner Banks 167, 187 location maps 168, 174-5 seismic data 170 Great Bahama Bank, platform history 5-6 Great Barrier Reef central coastline inner shelf, sediment accumulation 296-302 sedimentary model 308 turbid-zone reefs 295-310 geographical and geological setting
312 location map 24-5 origins 32-6 reconciliation of seismic, sedimentological and isotopic data 30-2 Great Barrier Reef Province, Heron Reef, lagoonal sedimentation
281-94
Gulf of Papua, turbid-zone reefs, sediment accumulation 302-5 guyots age-depth models l l 0-12 correlation of sequences 107-110 defined 39 drowning events 116-18 typical sectional view 98
Halifax Bay, Queensland, sediment accumulation 296-302 Haq curve see EXXON Haynesville platform, Jurassic, Texas 14 Helvetic Domain, Cretaceous, Swiss Alps 11, 13 Heron Reef, Great Barrier Reef Province lagoonal sedimentation 281-94 location map 282 radiocarbon dating 289 ryyf growth history 289-91 seisq�ic profiling and sediment si'mpling 282-9 Huevo Guyot see Resolution Guyot
Indian Ocean, W, last glacial extinctions 70-2 Indo-West Pacific habitat fragmentation 73 molluscan assemblage extinctions
69-76 see also Pacific, NW
JOIDES Resolution see ODP, route location maps Jurassic, Haynesville platform, Texas 14
karst processes
4-5
Lady Elliot Reef, Bunker Group, reef growth history 291 lagoonal sedimentation and reef development, Heron Reef, S Great Barrier Reef Province
281-94 last glacial maximum 231 logging data, spectral analysis
78-9
Maastrichtian cooling of sea-surface 63 drowning of platforms 59 magnesium, subtropical vs temperate shelf carbonates 315, 317 manganese and iron, subtropical vs temperate shelf carbonates
316, 318 Maracaibo platform, Cretaceous, Venezuela 12 Marion Plateau, NE Australia
Index chronostratigraphical diagram 150 comparison with EXXON global sea-level chart 153, 159 computer simulation 145-61 carbonate parameters 154 diagenetic sequence, 3 stages 190 dolomite, comparison with Fraser shelf 190 mega-seismic sequences and events in cover sequence 149 Marshall Islands, Wodejebato Guyot, NW Pacific, development and demise model 39-68 Maryborough Basin 167 Mayotte, Comoro Islands geological setting 220-1 location map 220 morphology and sediments 219-36 U-Th geochemistry and ages
223-4
Mid Pacific Mountains see Allison and Resolution Guyots mid-oceanic carbonate platforms, NW Pacific, development and demise model 57-64 Milankovitch periodicities, ODP Leg
143
83-4, 86
Miocene Luahua platform, South China Sea 9, 10 Mississippian, Arkansas, Burlington platform, exposure and drowning unconformities, close superposition 18 MIT Guyot, NW Pacific 96, 107 age-depth model 110-12 Albian sequences 127-9 Aptian sequences 126-7 comparisons 109 molechfor lithofacies 188 molluscan assemblage extinciions, Indo-West Pacific 69-76 monsoonal controls, coral atolls, South China Sea 237-48
327
8180, laminated stromatolitic crusts
269
normalized 8180 curve
230
71 Pacific, NW guyots as recorders of Cretaceous eustatic sea-level changes
95-136 seamount morphology 61-4 Wodejebato Guyot, development and demise model 39-68 see also Indo-West Pacific Pacific, SW, Tasmaloum, fringing reef tectonic and environmental growth influences 261-77 Paluma Shoals, Halifax Bay, Queensland, corals and sedimentation 298-300, 302 Papua, Gulf of, sediment accumulation 302-5 passive margins Heron Reef, Great Barrier Reef Province 281-94 terrigenous sediment, control on reef carbonate distribution
295-310 patch reefs, origins 291 Persian Gulf geographical setting 314-15 Shuaiba platform, Cretaceous 12 Pleistocene reef complex deposits, Central Ryukyus, SW Japan
197-213 porous media equations 250-1 Purari River, Gulf of Papua, sediment accumulation 302-4 Pyrenees, S, Cretaceous 9, 12
Quaternary subtropical carbonate platform development, S Queensland
163-95 Navier-Stokes domains 252 New Hebrides Island Arc, active margins, fringing reef growth and structure 261-77
ODP Leg 133 Great Barrier Reef 23-38 route location map 25 ODP Leg 143 route location maps 80, 96, 138 see also Allison Guyot; Resolution Guyot ODP Leg 144 route location map 40 see also MIT, Takuyo-Daisan and Wodejebato Guyots . Oligocene of Kalimantan, Borneo 9,
11
One Tree Reef, reef growth history
291 oxygen isotopes
Red Sea last glacial extinctions 70-2 small-scale mass extinction 71-2 sibling species in Indo-West Pacific
Tasmaloum, SW Pacific, fringing reef, tectonic and environmental growth influences 261-77 Queensland climate and oceanography 169 East Australia Current 169 Plateau 24 shelf, comparison with Ryukyus, SW Japan 190-2 subtropical carbonate platform development, Quaternary and Tertiary 163-95
radiocarbon dating Heron Reef, Great Barrier Reef Province 289 rhodoliths, Fraser Island and Platform 176, 178 Tasmaloum, SW Pacific, fringing reef, laminated stromatolitic crusts 269 Rayleigh numbers 251, 252
tropic stability, last interglaciation to present 69-76 Resolution Guyot, Mid-Pacific Mountains age-depth model 110-12 Albian sequences 122-4 Aptian sequences 119-22 comparisons 109 location maps 80, 96, 138 origin of white sucrosic dolomite
137-44 reconstruction of emersion horizons 101 strontium isotopes, evidence for role of sea water in dolomitization
137-44 summary 103-5, 104-5 rhodalgal lithofacies 188 rhodolith limestone, features 20 I rhodoliths, fore-reef, Ryukyu Group
210-11 rhodoliths, radiocarbon age, Fraser Platform 176, 178 Ryukyu Group, SW Japan carbonate and siliciclastic rocks
199
comparison with S Queensland shelf
190-2 corals and corallines 202 location maps 191, 197 Pleistocene reef complex deposits
197-213 sedimentary facies 190-1 stratigraphy 203-7
sea-level changes Aptian-Albian 95-136 carbonate platforms drowning events 62 theory 97-100 and practice I 00-3 European margin, correlation with Aptian-Albian, Pacific 114 limiting factor of coral reefs 72 meltwater pulses 231-3 normalized 8'80 curve 230 schematic eustatic curve, Aptian-Albian, Pacific 116-18 and sediment accumulation 307-8 sea-water carbon dioxide levels 321 composition 319-20 density equations 250 salinity, tolerances 70 temperature carbonate mineralogy 319 limiting factor of coral reefs 72 sediment accumulation control on reef carbonate distribution 295-310 coral reefs, potential problems 305
Index
328 sediment accumulation (cont.) methods of measurement 302 sedimentary cycles, carbonate platform facies, ODP Sites 865 and 866 logs 77-92 SEDPAK platform modelling 156 setup and input variables used in modelling 155 Seiko Guyot see Takuyo-Daisan Guyot seismic data Gardner Banks 170 Great Barrier Reef 30-2 Heron Reef, Queensland 282-3 mega-seismic sequences, Marion Plateau, NE Australia 8, 149 Wodejebato Guyot, NW Pacific 40-1 sequence boundaries 3 Shark Bay, W Australia geographical and geological setting 312-14 location map 314 Shuaiba platform, Cretaceous, Persian Gulf 12 sodium subtropical shelf carbonates 315 temperate shelf carbonates 318 South China Sea coral atolls characteristics and location
242-3
classification 247 location map 239 tectonic and monsoonal controls 237-48 types 240 Miocene Luahua platform 9, IO tectonic map 238 spectral analysis, logging data 78-9 strontium subtropical shelf carbonates 315 temperate shelf carbonates 317 strontium isotopes, evidence for role of sea water in dolomitization 137-44
subtropical shelf carbonates vs temperate, composition, four examples 311-23 composition, comparisons 313-14 development, Australia 188-9 environmental significance 187-8 Quaternary and Tertiary, S Queensland 163-95 subtropical and temperate carbonate platforms, diagenetic features 189-90
Tahiti, P7 borehole 256-8 Takuyo-Daisan Guyot, NW Pacific 96, 105-6 age-depth model 110-12 Albian sequences 126 Aptian sequences 124-6
comparisons I09 Tasmaloum see Espiritu Santo Tasmania, shelf carbonates and siliciclastics 3 I 5 tectonic controls, coral atolls, South China Sea 237-48 tectonic growth influences, Tasmaloum, Espiritu Santo, SW Pacific 261-77 temperate shelf carbonates, vs subtropical 311-23 terrigenous sediment control on reef carbonate distribution 295-310 world total annual 29 5 Tertiary subtropical carbonate platform development, S Queensland 163-95 Texas, Haynesville platform, Jurassic 14 thermal convection, inducing steady state hydraulic water circulation 249-56 tidal forcing, borehole response, computer simulation 256-8 tidal transients, borehole response, computer simulation 256-58 time-scale calibration 113
Triassic Schlern Platform, Alps, Southern 7 TRIO/CASTEM, simulation of thermal convection systems 251 tropic stability, Red Sea and Western Indian Ocean 69-76 turbidites, threshold condition 305
U-Th geochemistry and ages, Mayotte 223-4 Upstart Bay, Queensland, sedimP.nt accumulation 296-302 USA, Comanchean platforms, Cretaceous 12
Vanuatu (New Hebrides Island Arc), active margins, fringing reef growth and structure 261-77 Venezuela, Maracaibo platform, Cretaceous 12 Vocontian Basin, correlation with Aptian-Albian, Pacific II4
Walther's Law of Succession 23-38 water depth, vs carbonate production 4 Wodejebato Guyot, NW Pacific depositional history 57-9, 60 development and demise model 39-68 diagenesis 49-57 morphology and seismic data 40-1 paragenetic sequence and interpreted relative sea-level changes 56 post-drowning facies 57 shallow-water platform carbonates 44-56 stratigraphy 41-3 weathering profiles 44
Younger Dryas meltwater pulse 231-3
Reefs and Carbonate Platforms in the Pacific and Indian Oceans Edited by G. F. Camoin and P. J. Davies © 1998 International Association of Sedimentologists ISBN: 978-0-632-04778-9
0.0
20.0
8
40.0
60.0
0.0
100.0
200.0 row
0.00 -:;; g �·
.,
�
0.10
0.20
0.30
0.40
amplitude Plate 1 Spectrogram for Hole 865A. The vertical scale is in cycles per 80 m. Each row represents a single Fourier transform. Relative amplitude (percentage variance) has been converted to a colour scale as indicated by the colour bar below the image.
0.0
20.0
8
40.0
60.0
200.0
0.0
0.00
400.0
row
0.25
0.50
amplitude Plate 2 Spectrogram for Hole 866A. The vertical scale is in cycles per 80 m. Each row represents a single Fourier transform. Relative amplitude (percentage variance) has been converted to a colour scale as indicated by the colour bar below the image. .
(a) Sequence Plot
0
c
Sea level H (m) L
A
Distance (km)
� .n .. _- �-� <;J3 00 0-3 00
[Facing p. 160]
�--
-=
--
�
-
17 5 Ma .
MP3
-15 Ma
MP3
-12.5 Ma
MP3
Plate 2 Sequential SEDPAK simulation sequence plots from 25 Ma to present using the Marion Plateau sea-level curve showing the evolution of the Marion Plateau platform.
Plate 3 (a) Close-view of sequence sea-level plot of the MP2 platform from SEDPAK simulation showing the four platform-building phases (a-d) and their relationships with the MP3 platform and the Pliocene Holocene sediments. Note that the progradation rate of the MP2 platform changed dramatically from a to d (see Table 4). Triangles indicate platform progradation (inclined) and aggradation (vertical). Arrows illustrate the onlap nature between the MP3 coeval sediments and the MP2 platform. (b) Close-view of sequence-sea-level plot of the MP3 platform from SEDPAK simulation showing the four platform-building phases (a-d) and their relationships with the MP2 platform basinal facies and the Pliocene-Holocene sediments. Vertical triangles indicate platform aggradation, and arrows illustrate the on lap relationship between MP3a and its underlying sediments.
(b)
Sealevel H (m) L -5
d c
(a)
Sequence Plot
Distance (km)
(b)
245
-180
Sequence Plot
I c: 0
�w
Distance (km)
Plate 5 Sequential SEDPAK simulation sequence plots from 20 Ma to 5 Ma using the EXXON sea-level curve. Note that the sea-level fall in the Mid-Late Miocene (c. 10 Ma) was not great enough to initiate the MP3 platform.
150
-150
Plate 4 Sequence-sea-level plot from SEDPAK simulation using the EXXON sea-level curve (a), and using a stochastic sea-level curve (b). Note that, in both cases, the simulations failed to reconstruct the gross architecture and geometry of the platform.
Plate 1 Aerial photograph of Paluma Shoals and the adjacent shoreline, taken with sea-level 15 em above lowest astronomical tide (10 August 1991). North is to the top right hand corner, and the width of field of view is c. 2.8 km. The reef flats of Paluma Shoals are exposed, and on the northern reef, the reef flat clearly extends to the low water mark. Spur and groove structures are clearly visible on the seaward side on the reefs, with breaking waves slightly further seaward. The coastal water is very turbid (chocolate brown), but small patches of less turbid water (green) are visible landward of the reefs, particularly at the southern reef. The muddy intertidal zone (dark brown) has superposed sand bodies formed by wave reworking of fluvial outwash. (Photograph reproduced with permission of Department of Natural Resources, Queensland.)
(b)
(a)
Plate 2 (a) View looking seawards of the low-gradient lower intertidal and shallow subtidal zone. Note the highly turbid water generated by waves. Lower part of photo shows coral colonies present upon exposed pebbles and cobbles (photo taken at low water). (b) Goniastrea rediformis in turbid shallow water in the lower intertidal zone. (c) A colony of G. rediformis (diameter ca. 20 em) in the lower intertidal zone, surrounded by rippled silty sand with sparse seagrass. The colony is probably based upon a buried hard substrate.
(c) [Facing p. 304]
