Correlation of the early Paleogene in northwest Europe

Correlation of the Early Paleogene in Northwest Europe

Geological Society Special Publications Series Editor

A. J. FLEET

GEOLOGICAL SOCIETY SPECIAL PUBLICATION NO. 101

Correlation of the Early Paleogene in Northwest Europe EDITED BY

R. W. O'B. KNOX British Geological Survey Nottingham, UK

R. M. CORFIELD University of Oxford Oxford, UK and

R. E. DUNAY Mobil, North Sea Ltd London, UK

Stratigraphy Commission Petroleum Group

1996 Published by The Geological Society London

THE GEOLOGICAL SOCIETY The Society was founded in 1807 as the Geological Society of London and is the oldest geological society in the world. It received its Royal Charter in 1825 for the purpose of 'investigating the mineral structure of the Earth'. The Society is Britain's national society for geology with a membership of 8000. It has countrywide coverage and approximately 1000 members reside overseas. The Society is responsible for all aspects of the geological sciences including professional matters. The Society has its own publishing house, which produces the Society's international journals, books and maps, and which acts as the European distributor for publications of the American Association of Petroleum Geologists, SEPM and the Geological Society of America. Fellowship is open to those holding a recognized honours degree in geology or cognate subject and who have at least two years' relevant postgraduate experience, or who have not less than six years' experience in geology or a cognate subject. A Fellow who has not less than five years' relevant postgraduate experience in the practice of geology may apply for validation and, subject to approval, may be able to use the designatory letters C Geol (Chartered Geologist). Further information about the Society is available from the Membership Manager, The Geological Society, Burlington House, Piccadilly, London W1V 0JU, UK. The Society is a Registered Charity, No. 210161. Published by the Geological Society from: The Geological Society Publishing House Unit 7 Brassmill Enterprise Centre Brassmill Lane Bath BA1 3JN UK (Orders: Tel 01225 445046 Fax 01225 442836)

Distributors USA AAPG Bookstore PO Box 979 Tulsa OK 74101-0979 USA (Orders: Tel (918) 584-2555 Fax (918) 548-0469)

First published 1996

Australia Australian Mineral Foundation 63 Conyngham Street Glenside South Australia 5065 Australia (Orders: Tel (08) 379-0444 Fax (08) 379-4634)

The publisher makes no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility for any errors or omissions that may be made. 9 The Geological Society 1996. All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with the provisions of the Copyright Licensing Agency, 90 Tottenham Court Road, London W1P 9HE. Users registered with the Copyright Clearance Center, 27 Congress Street, Salem, MA 01970, USA: the item-fee code for this publication is 0305-8719/96/$07.00.

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN 1-897799-47-0 Typeset by EJS Chemical Composition, Midsomer Norton, Bath BA3 4BQ, UK Printed by The Alden Press, Osney Mead, Oxford, UK

India Affiliated East-West Press PVT Ltd G-I/16 Ansari Road New Delhi 110 002 India (Orders: Tel (11) 327-9113 Fax (11) 326-0538) Japan Kanda Book Trading Co. Tanikawa Building 3-2 Kanda Surugadai Chiyoda-Ku Tokyo 101 Japan (Orders: Tel (03) 3255-3497 Fax (03) 3255-3495)

Contents PREFACE KNOX, R. W. O'B. Correlation of the early Paleogene in northwest Europe: an overview

vii

1

Regional studies: stratigraphy, tectonics and volcanism NEAL, J. E. A summary of Paleogene sequence stratigraphy in northwest Europe and the North Sea

15

NADIN, P. A. & KUSZNm,N. J. Forward and reverse stratigraphic modelling of CretaceousTertiary post-rift subsidence and Paleogene uplift in the Outer Moray Firth Basin, central North Sea

43

RITC~E, J. D. & HITCHEN,K. Early Paleogene offshore igneous activity to the northwest of the UK margin and its relationship to the North Atlantic Igneous Province

63

JoY, A. M. Controls on Eocene sedimentation in the central North Sea Basin: results of a basinwide correlation study

79

MUDGE, D. C. & BUJAK,J. P. An integrated stratigraphy for the Paleocene and Eocene of the North Sea

91

THOMAS, J. E. The occurrence of the dinoflagellate cyst Apectodinium (Costa & Downie 1976) Lentin & Williams (1977) in the Moray and Montrose Groups (Danian to Thanetian) of the UK central North Sea

115

WOOD, S. E. & TYSON, R. V. An integrated palynological-palynofacies approach to the zonation of the Paleogene in the Forties-Montrose Ridge area, central North Sea

121

ALl, J. R. & JOLLEY, D. W. Chronostratigraphic framework for the Thanetian and lower Ypresian deposits of southern England

129

POWELL, A. J., BRINKHUIS,H. & BUJAK,J. P. Upper Paleocene-Lower Eocene dinoflagellate cyst sequence biostratigraphy of southeast England

145

ELLISON, R. A., Au, J. R., H1NE,N. M. & JOLLEY,D. W. Recognition of Chron C25n in the upper Paleocene Upnor Formation of the London Basin, UK

185

ALI, J. R., HAtLWOOD,E. A. & KING, C. The 'Oldhaven magnetozone' in East Anglia: a revised interpretation

195

HOOKER, J. J. Mammalian biostratigraphy ascross the Paleocene-Eocene boundary in the Paris, London and Belgian basins

205

JOLLEY, D. W. The earliest Eocene sediments of eastern England: an ultra-high resolution palynological correlation

219

MITLEHNER, A. G. Palaeoenvironmensts in the North Sea Basin around the Paleocene-Eocene boundary: evidence from diatoms and other siliceous microfossils

255

ScnMrrz, B., HEILMANN-CLAUSEN,C., KING, C., STEURBAUT,E., ANDREASSON,F. P., CORFIELD,R. M. t~ CARTLIDGE,J. E. Stable isotope and biotic evolution in the North Sea during the early Eocene: the Alb~ek Hoved section, Denmark

275

Global perspective: geochronology and the oceanic record BERGGREN,W. A. & AUBRY,M.-P. A late Paleocene-early Eocene NW European and North Sea magnetobiochronological correlation network

309

vi

CONTENTS

AUBRY,M.-R, BERGGREN,W. A., STOTr, L. & SINHA,A. The upper Paleocene-lower Eocene stratigraphic record and the Paleocene-Eocene boundary carbon isotope excursion: implications for geochronology

353

STOTr, L. D., SINHA,A., THIRY,M., AUBRY,M.-R & BERGGREN,W. A. Global 813C changes across the Paleocene-Eocene boundary: criteria for terrestrial-marine correlations

381

THOMAS, E. & SHAC~ETON, N. J. The Paleocene-Eocene benthic foraminiferal extinction and stable isotope anomalies

401

CORFIELD,R. M. & NORRIS,R. D. Deep water circulation in the Paleocene Ocean

443

CHAmSI, S. D. & SCHMITZ,B. Early Eocene palaeoceanography and palaeoclimatology of the eastern North Atlantic: stable isotope results for DSDP Hole 550

457

INDEX

473

Preface The early Paleogene of northwest Europe has been the subject of intense investigation over the last quarter century, with important stimulus being provided by the search for oil and gas in the offshore basins and by lUGS-sponsored investigations of the onshore historical stage and system stratotype sections. The Paleogene has long been an exploration target offshore northwest Europe. Giant accumulations, such as the Forties oilfield (UK) and the Ekofisk oilfield and Frigg gasfield (Norway), were discovered in the early days of exploration in the central and northern North Sea. Exploration of the North Sea Paleogene is continuing, with total discoveries now exceeding 12 billion barrels of oil equivalent (BBOE). Paleogene exploration plays are also being actively pursued West of Shetlands, where discoveries reputed to be in excess of 1 BBOE have been made in recent years. The early Paleogene has thus been the focus of major industry interest, and continues to be an attractive exploration target. The onshore sections that fringe the southern margin of the North Sea Basin are home to the historical stratotype successions for most of the Paleogene system and stage boundaries. With the drive towards development of a global standard for the subdivision of Paleogene time, these successions have been the subject of detailed investigation in recent years. Attention is currently focused on the early Paleogene, with international collaboration taking place under the aegis of the lUGS 'Paleocene-Eocene Boundary' and 'Paleocene Stages' working groups and IGCP Project 308 'Paleocene-Eocene Boundary Events in Space and Time'. The igneous province of northwest Britain has also received much attention,with a better understanding of the timing and nature of the volcanism arising from the application of improved analytical techniques and from the acquisition of new information from shallow and deep drilling in the offshore areas. DSDP drilling in the eastern Atlantic has also played a significant part in recent advances in northwest European early Paleogene stratigraphy. Drilling in the Bay of Biscay and the Goban Spur (Legs 48, 80) has provided information on the oceanic succession nearest to northwest Europe, while drilling in the Rockall area (Legs 12, 48, 81) and adjacent parts of the North Atlantic has increased our knowledge of the crustal evolution of the region during the early Paleogene, leading to a better understanding of the history of tectonism and volcanism in northwest Europe. As illustrated by the papers in this volume, the wide range of activities listed above has led to the acquisition of a remarkably diverse dataset, which provides a unique opportunity for the development of a truly comprehensive regional stratigraphy, encompassing terrestrial, epicontinental marine and oceanic successions, and linking these to the tectonic and volcanic events associated with the onset of seafloor spreading between Greenland and Europe. A key element in realizing this potential is the integration of data derived from onshore studies, offshore hydrocarbon exploration activities and ocean drilling programmes. It is hoped that publication of this volume will add further stimulus to the necessary interchange of data and ideas between researchers in these different fields. R. W. O'B. Knox R. M. Corfield R. E. Dunay

Correlation of the early Paleogene in northwest Europe: an overview R. W. O ' B . K N O X British Geological Survey, Keyworth, Nottingham NG12 5GG, UK

The last two decades have seen a major resurgence of interest, both commercial and scientific, in the early Paleogene stratigraphy of northwest Europe. The commercial interest has arisen primarily as a result of major oil and gas finds in the central and Northern North Sea, mostly in deep-water sandstones of late Paleocene to mid Eocene age. Increased interest in the onshore sections has been stimulated partly in response to the offshore hydrocarbon exploration, but largely through the activities of international (IUGS/IGCP) working groups, whose primary concern is the establishment of a globally standardized system of series and stages. The onshore sections of the southern North Sea Basin area are of particular importance in these investigations, because they include the historical stratotypes for the Paleocene and Eocene series and for their constituent stages. Unfortunately, these historical stratotypes are inappropriate as global stratotypes because of their stratigraphic incompleteness, and their limited representation of the standard Paleogene biozones. Only through the fullest understanding of these historical stratotype sections, however, can we ensure that the standard stages are defined in a way that ensures the maximum compatibility with traditional assignments in NW Europe (Knox 1994; Schmitz 1994). For a long time the commercially driven and scientifically driven lines of investigation proceeded more or less independently, partly because of the confidential nature of the offshore investigations and partly because of difficulties in correlating widely separated sections of strongly contrasting lithofacies and biofacies. For these reasons, the earlier stratigraphic compilations for the Paleogene of northwest E u r o p e were concerned almost exclusively with the onshore areas (e.g. Curry et al. 1978; Cavelier & Pomerol 1986; Pomerol 1989). A notable exception is the compilation of data collected in relation to IGCP Project 124 (Vinken et al. 1988), which represents a remarkable achievement in the field of multidisciplinary and multinational stratigraphic collaboration. Other areas that have been the subject of detailed analysis are the oceanic successions of the eastern Atlantic (Fig.l), encountered during DSDP drilling in the Goban Spur area (Legs 48, 80) and in the

Rockall area (Legs 12, 48, 81), and the successions of British Tertiary Igneous (BTIP) and the North Atlantic Igneous (NAIP). Five, largely independent, areas can thus be identified:

volcanic Province Province of study

(1) the onshore sections of the southern margin of the North Sea Basin, restricted to inner shelf, littoral and terrestrial facies; (2) the offshore sections of the North Sea Basin and West of Shetlands area, dominated by outer shelf, slope and basinal facies, but including inner shelf to terrestrial facies around the Scottish landmass; (3) the offshore sections of the Goban Spur area, restricted to bathyal facies (largely calcareous nannofossil oozes); (4) the offshore sections of the Rockall area, representing inner shelf to bathyal facies; (5) the onshore and offshore stratified sections of the BTIP and GFIP, dominated by lavas and tufts, but with intercalations of nonvolcanogenic sediments.

_ !tI Fig.1. Distribution of early Paleogene sedimentary and igneous successions in NW Europe, with locations of the Bay of Biscay, Goban Spur and Rockall DSDP sites (Legs 12, 48, 80, and 81) and central North Sea well 22/10a-4.

From Knox, R. W. O'B., Corfield, R. M. & Dunay, R. E. (eds), 1996, Correlation of the Early Paleogene in Northwest Europe, Geological Society Special Publication No. 101, pp. 1-11.

2

R.W. O'B. KNOX

A broad stratigraphic framework has now been established for each of these areas. For example, extensive studies have now been carried out on all aspects of biostratigraphy, such that regional zonation schemes are firmly established for the more important fossil groups. Similarly, comprehensive magnetic polarity zonations have been established for the onshore areas of the BTIP and the southern North Sea Basin, providing not only a means of correlation, but also a direct link to the geological timescale. It is only now that these comprehensive stratigraphic frameworks have been established that correlation betweeen the individual areas (referred to as 'interregional correlation' in this account) can be realistically attempted. Problems in interregional correlation arise primarily from differences in the type and nature of the available stratigraphic data. Examples of limitation due to data type are seismic data, which are available only for offshore sections, and magnetic polarity data, which can be obtained only from cored sections or outcrops. Even within a single discipline, interpretation can be hampered by different methods of data acquisition. Thus biostratigraphic data for the offshore hydrocarbons basins are largely based on cuttings, and thus dependent on first downhole occurrences (FDOs), whereas for the onshore sections they are based on the standard criteria of first and last appearance datums (FADs, LADs). A more serious limitation on interregional correlation is the effect of facies on the nature and diversity of fossil assemblages. Thus while the Goban Spur sections in the eastern Atlantic possess rich calcareous microfaunas and nannofaunas, allowing assignment to the standard Paleogene biozones, equivalent strata in central parts of the North Sea Basin are commonly devoid of calcareous fossils. Conversely, whereas palynomorphs are ubiquitous in the North Sea Basin, they are reported to be absent from the Goban Spur sections. Under such circumstances, correlation between the two successions must rely on a combination of techniques, with particular emphasis on those that are less facies dependent (e.g. magnetostratigraphy and tephrostratigraphy). Fortunately, not all of the correlation problems are so severe. For example, the early Paleogene succession of Denmark, accessible through outcrop sections and cored boreholes, is of deep-water facies, and provides a valuable insight into the succession of the central North Sea. It is therefore possible to apply techniques such as magnetostratigraphy to a basinal North Sea succession, and thereby assess the relationship between the zonally based biostratigraphic schemes established for onshore sections and the FDO-based schemes established for the offshore hydrocarbons boreholes. Additionally, the Paleocene sections of

SE England provide a more or less continuous transect from sublittoral facies in the historical Thanetian stratotype sections of Kent to outer shelf 'North Sea'facies in northern parts of East Anglia. The major challenges for the future are (i) interregional correlation within northwest Europe, and (ii) calibration of the northwest European sections in terms of the standard biozones and the geological timescale. Because it is clear that no one discipline can solve all these problems, a multidisciplinary approach is paramount. The purpose of this overview is to provide a brief assessment of how the different disciplines have contributed towards interregional correlations, and how the ultimate aim of a chronologically calibrated, integrated stratigraphy for the entire northwest European region might be achieved.

Essential elements of an interregional correlation The geological timescale The publication of a new Paleogene timescale (Cande & Kent 1995) is a welcome development in view of the significant discrepancies between earlier timescales, especially over the late Paleocene to early Eocene interval. The incorporation of a radiometric date obtained from a tephra layer of early NP10 age puts the new timescale on a much firmer footing. This is especially important for assessing the influence of 'Atlantic' tectonism on stratigraphic events, as it allows an improved correlation between the biostratigraphically dated sedimentary successions and the radiometrically dated lava successions of the North Atlantic borderlands. However, as cautioned by Berggren & Aubry (1996), timescales are constructed on a series of assumptions that will be subject to continuous reassessment, and the new Paleogene timescale is no exception. The Cande & Kent (1995) timescale is nevertheless considered to provide a much improved overall chronological calibration of the standard biozones and magnetic polarity zones (see Berggren et al. in press). Its adoption by the lUGS Subcommission on Paleogene Stratigraphy should ensure that it becomes the common standard until such time as sufficient data are accrued to warrant further refinement. The use of such a common standard will greatly facilitate comparison of stratigraphic data of all types and from all parts of the world.

Biostratigraphy Calcareous nannofossils. A more or less complete record of the early Paleogene standard calcareous nannofossil zones is present in the Goban

CORRELATION OF THE EARLY PALEOGENEIN NORTHWEST EUROPE" AN OVERVIEW Spur DSDP Sites 549 and 550, which together may be regarded as providing a composite oceanic reference section for NW Europe. This composite section, which has recently been the subject of a detailed reassessment (Aubry et al. 1996), provides a unique opportunity for linking the epicontinental successions with the global oceanic record. Direct correlation is admittedly hampered by the patchy distribution and in part endemic nature of calcareous nannofossil assemblages in NW Europe, but, as early studies on the onshore sections of the North Sea Basin (e.g. Aubry 1986; Siesser et al. 1987) are augmented by new data (e.g. Steurbaut 1990; Hine 1994; Ellison et al. 1996), a significant part of the NW European succession can now be at least broadly linked to the oceanic calcareous nannofossil record. The Goban Spur sites also provide an important record of the standard planktonic foraminiferal zonation for the eastern Atlantic, which has recently been the subject of a comprehensive review (Berggren & Aubry 1996). However, planktonic foraminifera are of limited value in the epicontinental successions, where faunas are characteristically dominated by endemic assemblages. Only in the early Paleocene and part of the early Eocene is there potential for direct assignment of the standard planktonic zonations. For the remainder of the succession, microfossil zonal schemes have been established largely for the deep-water, basinal areas of NW Europe, either on whole faunas (e.g. King 1989; Mudge & Copestake 1992) or selected elements (e.g. Mitlehner 1996). While these zonal schemes have potential for detailed correlation between the deep-water facies of the North Sea Basin, West of Shetlands basins and the offshore Norwegian basins, they cannot be applied to the shallow-water onshore successions of the southern North Sea Basin margins because of a fundamental change in biofacies. Zonal schemes that apply to the entire region (e.g. Gradstein et al. 1994) are therefore necessarily less detailed than the correlations developed specifically for the basinal areas. However, they have the advantage that they can be directly linked with the calcareous nannoplankton zonation and magnetostratigraphy established for the onshore sections. Microfossils.

Palynomorphs. Following their early use in the correlation of the marginal successions of the southern North Sea Basin, dinoflagellate cysts have proved to be particularly valuable in basinal areas (e.g. Heilmann-Clausen 1985; Powell 1988; Mudge & Bujak 1996). They are now established as perhaps the most effective means of correlating across the broad spectrum of facies encountered in

3

the onshore and offshore basins of NW Europe (see Powell 1992; Heilmann-Clausen 1994). Terrestrial palynomorphs have an even greater potential for interregional correlation, as they occur in both marine and continental facies, including sedimentary intercalations within the lava successions of the Hebridean Province and the Faeroes. Regional stratigraphic analysis of the terrestrial palynomorph assemblages is, however, less straightforward than for marine palynomorphs, because of greater variability in assemblages arising from terrestrial climatic and physiographical controls, and because of the strong geographical control exerted by fluvial catchment, transport and deposition. Nevertheless, the potential for local and regional high-resolution stratigraphy has been demonstrated in studies on the Thanet Formation of southern England (Jolley 1992) and on the Harwich Formation and its Central North Sea equivalents (Jolley 1996). Calibration of the palynomorph zonations against the standard zonal schemes is possible in those parts of the succession where the appropriate calcareous nannoplankton are present, including the early Paleocene, part of the late Paleocene (late NP6 to early NP9), and much of the early Eocene. Calibration over the interval spanning late NP9 to NP10 cannot, however, be achieved within the North Sea Basin, and application of calibrations developed for other areas (e.g. Rockall area: Morton et al. 1983) may be unsafe because of possible diachroneity of dinoflagellate cyst influxes resulting from isolation of the North Sea Basin at this time. Many of the limitations apparent from existing palynological studies will be overcome from the combined use of terrestrial and marine groups in establishing correlations between terrestrial, shallow marine and deep marine facies. Published accounts in which such an approach have been used (e.g. Jolley 1996) are relatively few, but the method is being increasingly used in the study of the offshore hydrocarbons basins. Though the mammal faunas of NW Europe are of very restricted occurrence compared with other zonally significant fossil groups, they provide important information on the evolution of the regional palaeogeography Hooker 1996). The succession of NW European mammal faunas indicates that interchange with North American faunas took place twice during the Paleocene and Eocene, firstly in the early late Paleocene and secondly in the Paleocene/Eocene 'boundary interval' (see below). Both events correspond to periods of maximum lowstand (middle Maureen Formation and basal Sele Formation lowstands of the central North Sea), which presumably resulted in exposure

Mammals.

4

R.W. O'B. KNOX

of the Greenland-Scotland landbridge. The initial highly cosmopolitan nature of the North American and European faunas within the Paleocene-Eocene boundary interval suggests that migration was exceptionally rapid, with favourable palaeogeographical conditions probably being enhanced by short-term climatic amelioration (Hooker 1996, and see below). Integrated biostratigraphic framework. As discussed above, there is a distinct dichotomy between the marine biostratigraphic record in the oceanic Atlantic succession and that in the epicontinental successions of the North Sea Basin and West of Shetlands area. Even within the epicontinental basins, biofacies changes from the centre to the margins of the basins puts severe constraints on long-distance correlations. No one fossil group can be relied on to correlate between all sections, and the key to developing a biostratigraphic correlation framework for the entire NW European region thus lies in determining the interrelationships between bioevents and biozonal schemes established for all fossil groups. Isotope stratigraphy

The standard techniques of carbon and oxygen isotope analysis are carried out on whole-rock carbonate or, for more informative results, on selected species of calcareous benthic and planktonic foraminifera. Analysis is therefore usually restricted to carbonate-rich facies. Knowledge of the Paleogene isotopic record has therefore been built up primarily from oceanic sections encountered in DSDP and ODP drilling (e.g. Charisi & Schmitz 1996; Corfield & Norris 1996; Stott et al. 1996). The long-term trends within the early Paleogene are now well established, and attention is currently focusing on specific aspects, such as the occurrence and age of one or more short-term negative carbon isotope excursions in the late Paleocene (e.g. Corfield & Norris 1996; Stott et al. 1996) and their relationship to other major events, such as the widespread oceanic 'benthic extinction' event (Thomas & Shackleton 1996). Appropriate facies for carbon and oxygen isotope analysis within the lower Paleogene are largely restricted in NW Europe to the lower Eocene. A detailed study on the Danish lower Eocene has revealed both long-term trends and short-term events, as well as demonstrating the influence of fresh water input on isotopic values (Schmitz et al. 1996). No such comprehensive study has been carried out on the NW European Paleocene, because the facies are largely unsuitable. However, Stott et al. (1996) have identified a distinct negative carbon isotope excursion within

early diagenetic (pedogenic) carbonate from the Argile Plastique Bariolre (equivalent to the Reading Formation) in the Paris Basin. This may well correlate with the major negative short-term excursion recorded from oceanic sections in late NP9, in which case it will corroborate the earlier findings of Koch et al. (1992) and constitute an important breakthrough in linking the Paleocene/Eocene succession of NW Europe with the oceanic record. R a d i o m e t r i c dating

Radiometric dating has been carried out on both high temperature minerals and glauconites. With two exceptions, the high temperature dates have been obtained from volcanic rocks in the BTIP and GFIE These dates have contributed to knowledge of the relative timing of volcanic events within and between individual igneous centres, and also to the gross timing of volcanism in the region (e.g. Mussett et al. 1988; Noble et al. 1988; Ritchie & Hitchen 1996). They are also of major significance in assessing the relationship between phases of volcanism and the tectonic and sea-level history of the region as inferred from coeval sedimentary successions. Dating of the sediments themselves has been based on K-At analysis of glauconite, and has played a significant part in the construction of some timescales (e.g. Harland et al. 1989). Single-crystal (sanidine) Ar/Ar dates from early Eocene tephras in Denmark have played a more specific role in the construction of the Cande & Kent (1995) timescale, with the age of ash-layer-17 being used as a tiepoint for the lower part of calcareous nannofossil zone NP10. Magnetostratigraphy

Magnetostratigraphy has also played an important role in the establishment of the igneous history of the region, supplementing the data obtained from superpositional/crosscutting relationships and radiometric dating (Mussett et al. 1988; Ritchie & Hitcben this volume). In the sedimentary successions of the onshore southern North Sea area, a detailed knowledge of the reversal history has now been established for many areas (see Ali et al. 1993; Ali & Jolley 1996), though reassessment and refinement are still possible through the application of improved techniques or through the study of new sections (e.g. Ali et al. 1996). As a means of providing precise chronological correlations, magnetostratigraphic data have unique potential in correlating between igneous and sedimentary successions and in assessing synchroneity of sealevel change both within NW Europe and beyond.

CORRELATION OF THE EARLY PALEOGENEIN NORTHWESTEUROPE: AN OVERVIEW Tephrostratigraphy

Though a detailed tephrostratigraphy had been established in the early Eocene ash-series of Denmark around the turn of the century, the significance of these tephras as regional or interregional correlation tools was not appreciated until their wide geographical extent was revealed by offshore drilling in the North Sea. Equivalents of the Danish ash-series are now known to extend beyond the North Sea Basin, into the West of Shetlands area and into the Goban Spur area, providing a valuable interregional marker for the lower NP10 interval. Several phases of ash deposition have now been recognized, with compositional changes reflecting progressive stages in the volcanic history of the northeast Atlantic rift zone (Morton & Knox 1990). S e q u e n c e stratigraphy a n d crustal history

As with the other methods of stratigraphic analysis, sequence stratigraphy has to a large extent developed separately in the offshore areas (e.g. Armentrout et al. 1993; Den Hartog Jager et al. 1993; Galloway et al. 1993; Mitchell et al. 1993; Jones & Milton 1994) and the onshore areas (e.g. Plint 1988; Grly & Lorenz 1991; Jolley 1992; Hardenbol 1994; Knox et al. 1994; Powell, et al. 1996; Vandenberghe in press). It is only recently that studies have been published that combine the two. Studies of this kind range from broad overviews (e.g. Neal 1996), in which depositional systems are linked to both long-term and short-term sea-level change, to detailed, often biostratigraphically driven, analysis of restricted stratigraphic intervals (e.g. Jolley 1996; Powell et al. 1996). Sequence stratigraphy provides a useful vehicle for the compilation of diverse, multidisciplinary stratigraphic data. However, it is clear from the NW European record that the sequence stratigraphy of the region cannot be assessed independently of its crustal history, which during the early Paleogene as a whole was strongly influenced by both Atlantic and Alpine processes. While the overall aim must be to develop a sequence stratigraphic scheme for the whole of NW Europe, it is clear that proper appreciation must be given to the effect of local tectonics on the relative sea-level curve for different parts of the region. The influence of tectonism during the Paleocene has been amply demonstrated by studies on the Paleogene uplift history of the central and northern North Sea, with an uplift of over 400 m proposed for the Outer Moray Firth area (Nadin & Kusznir 1996). Tectonic control is also proposed as the underlying mechanism for the generation of large-scale sequences in the Eocene (Joy 1996). Such a strong tectonic signal is hardly

5

surprising, considering the complex crustal history that is recorded in the British and North Atlantic igneous provinces, culminating in the opening of the North Atlantic in early Eocene times (Ritchie & Hitchen 1996). Under such circumstances, the erection of an interregional sequence statigraphic scheme will inevitably be hampered by geographical variation in the amount of uplift, and by the interplay between uplift and eustatic sea-level change.

Paleocene/Eoeene boundary events The Paleocene/Eocene boundary has yet to be formally defined, but historical considerations require that it will eventually be placed within the NP9 to early NP10 interval, which, in the meantime, is often referred to as the Paleocene/Eocene boundary 'interval' or 'transition'. This interval includes several events of global significance, of which the following may be considered the most important: (1) a pronounced short-term negative shift in carbon isotope values (Corfield & Norris 1996; Stott et al. 1996); (2) an oceanic 'benthic extinction' event, involving a dramatic reduction in both numbers and diversity of benthic foraminifera in the oceans (Thomas & Shackleton 1996); (3) an influx of kaolinite into both the oceans and shelf seas (Thomas & Shackleton 1996); (4) interchange of mammals between North America and Europe, leading to cosmopolitan faunas with an exceptional level of species in common (Hooker 1996); (5) extensive uplift and volcanism associated with the lead up to opening of the North Atlantic between Greenland and Rockall (Ritchie & Hitchen 1996). The relative timing of these events has yet to be fully established, and is the focus of IGCP Project 308. Until this timing is properly established, the ultimate cause of all these changes will not be known. Explanations for the individual features include: (1) a change in the pattern of oceanic circulation, with warming and/or increased salinity of bottom waters causing mass mortality among the benthic communities; (2) rapid changes in productivity in the oceans; (3) climatic warming, with enhanced chemical weathering leading to an increased production of kaolinite; (4) the development of a land-bridge between North America and Europe as a result of regional, probably plumerelated, uplift, perhaps enhanced by eustatic sealevel fall. The NW European succession, with its detailed, multi-component stratigraphic, tectonic and

6

R.W. O'B. KNOX

volcanic record, provides a unique opportunity to unravel many of these temporal relationships. However, the coverage of this critical interval is somewhat fragmented, and little attempt has been made to piece it together other than in the general context of regional stratigraphic compilations. Attempts at identifying the influence of global events on the NW European stratigraphic record have been largely concerned with the role of eustasy in the history of relative sea-level change. To assess the influence of specific global events, attention has focused largely on the early part of the Paleocene-Eocene boundary interval, during which major changes are known to have taken place in the oceanic environment and in the distribution of terrestrial faunas. Such studies include assessment of the history and mechanisms of mammal migration (Hooker 1996) and the recent search for isotopic signatures in the Paleocene of the Paris Basin (Stott et al. 1996). Unfortunately, both studies have of necessity been concerned primarily with the terrestrial sections of the southern North Sea margin, which suffer from the limitations imposed by incompleteness of the stratigraphic record. There is a real possiblity that some of the events identified in the oceans will be represented in these sections by hiatuses. The offshore sections of the central North Sea (and equivalent deep-water

sections in onshore Denmark) similarly have their limitations. In particular, the absence of calcareous facies precludes both the identification of the standard planktonic biozones and the application of standard techniques of carbon and oxygen isotope analysis. However, in their favour, these sections do provide a more or less continuous record across the Paleocene/Eocene boundary interval and must surely contain some reflection of any major change in the global environment.

The basinal record of the central North Sea In the context of a relatively deep-water, noncalcareous, clastic facies, and bearing in mind the events recorded from oceanic sections, reflections of any change in oceanic circulation or climate might be expected in (i) the benthic community (dominantly agglutinated foraminifera), (ii) the nature and relative dominance of terrestrial palynomorph assemblages,~ and (iii) the composition of the clay mineral assemblages. Evidence from published and unpublished sources indicates that significant environmental changes are recorded by all three of these components. In terms of central North Sea lithostratigraphy, the 'Paleocene/Eocene boundary interval' com-

Fig.2. Stratigraphic events in the early part of the Paleocene-Eocene boundary interval in well 22/10a-4, and possible correlation with the onshore succession of southern England.

CORRELATION OF THE EARLY PALEOGENE IN NORTHWEST EUROPE: AN OVERVIEW prises the uppermost Lista Formation, the Sele Formation, and the Balder Formation. Tephrostratigraphic correlation indicates that much of the Balder Formation and the uppermost Sele Formation (unit $3) is of NP10 age, with the NP9/10 boundary probably occurring in the upper part of unit $2. The remainder of the Sele Formation is probably of NP9 age. A pronounced upward reduction in the abundance and diversity of benthic foraminiferal assemblages has long been recognized to occur near the base of the Paleocene/Eocene boundary intervals offshore, corresponding to Sele/Lista formation boundary as defined by Knox & Holloway (1992). The existence of cores through this interval in a relatively expanded Central North Sea section (well 22/10a-4) has allowed other features associated with this 'benthic extinction' event to be recorded in detail for the first time (O'Connor & Walker 1993). Most significantly, the reduction in the benthic assemblages takes place at the level where pale greyish green, waxy, unbedded claystones (Lista Formation) are replaced upwards by medium to dark grey, crudely laminated mudstone (Sele Formation, base unit S la) (see O'Connor & Walker 1993, fig. 21). The boundary probably marks an increase in sedimentation rate, with greater retention of organic matter. However, the reduction in numbers of the agglutinating foraminifera cannot be ascribed simply to a dilution effect, as it is accompanied by a significant reduction in diversity. Further up the section, a second lithological change takes place, marked by a progressive increase in the number and thickness of turbidite sandstone layers (c.859Y6" in fig. 1 of O'Connor & Walker 1993). An upward increase in fine-scale lamination, and decrease in bioturbation, is observed within this unit, indicating the onset of bottom-water anoxia. The top of the turbidite sandstone unit is marked by a rapid upward transition to delicately laminated mudstone. A marked high gamma-ray wireline-log spike occurs at the top of this transition (here placed at c. 8567' 6" in fig. 1 of O'Connor & Walker 1993). This gamma-ray spike is accompanied by a sharp and sustained increase in uranium content. Benthic foraminifera disappear altogether in the uppermost part of the the turbidite sandstone, and are of only rare occurrence throughout the remainder of the Sele Formation. The changes in microfaunal abundance and diversity are accompanied by changes in the palynomorph assemblages (Thomas 1996), with the upward transition from greenish claystone to grey mudstone being marked by a dramatic increase in the ratio of terrestrial to marine forms. This feature persists to the top of the lower division of the Sele Formation (unit Sla of Knox & Holloway 1992), above which the turbidite sandstone unit is

7

characterized by an increase in the proportion of marine forms and by the incoming of the genus Apectodinium. A peak abundance of Apectodinium occurs in the upper part of the turbidite sandstone unit, immediate below the high-gamma spike. Clay mineral assemblages also display marked changes over this interval. The Lista Formation is characterized by kaolinite-free assemblages, dominated by smectite or chlorite (both probably derived from alteration of fine-grained volcanic material). Similar assemblages initially persist into the lower part of unit S l a, but kaolinite appears about half way up the unit, reaching a low peak within the turbidite sandstone unit. After a slight fall-off in the lower part of unit Slb, kaolinite increases rapidly, paralleling an overall increase in grain size. It is clear that important environmental changes took place in the central North Sea in the early part of the Paleocene/Eocene boundary interval. Two of the features described above, the reduction in benthic faunas and the influx in kaolinite, parallel events described from oceanic sections. The marked increase in the proportion of terrestrial palynomorphs can be interpreted in terms of increased terrestrial run-off, resulting from either sea-level fall or climatic change. According to some interpretations, the increase in relative abundance of Apectodinium might be an indicator of warming, but other factors, such as sea-level change, may also be involved (Thomas 1996). However, an increase in temperature around the S 1a/S l b boundary has previously been inferred by Schrrder (1992) from the composition of the palynomorph assemblages, and fluctuations in the composition of the palynomorph assemblages reported from the remainder of the Slb section (Forties Sandstone) may also be of climatic origin (Wood & Tyson 1996). The occurrence of kaolinite in the lower part of the Sele Formation is almost certainly an indicator of increased humidity in the source areas, as it is virtually absent from the underlying Lista Formation, even in sandier facies lower in the section. However, variations in the relative abundance of kaolinite within the Forties Sandstone are clearly related to the overall grain-size, and cannot be taken as a direct climatic indicator. A combination of palynological data plus clay mineral data thus points to at least a general climatic warming around the Sla/Slb boundary, close to the high gamma-ray wireline-log spike, and it is tempting to think that various events recorded over this interval may in some way be related to the broadly coeval events recorded from ocanic successions. However, many other factors need to be taken into account in view of the progressive restriction and evental isolation of the North Sea

8

R.W. O'B. KNOX

Basin over this period. For example, there is good evidence for substantial sea-level fall at the Lista/Sele formation boundary, leading to an increased land area and a sharp influx of terrestrial palynomorphs. A second sea-level fall may be represented by the base of the turbidite sandstone unit, but several features, such as the slight decrease in the proportion of terrestrial palynomorphs, the abrupt incoming of Apectodinium, and the increase in fine-scale lamination point more to a change in basin/hinterland configuration than to a simple sea-level fall. The succeeding laminated mudstones, which are associated with a reduction in kaolinite percentage, a reduction in Apectodinium abundance, and a sustained reduction in the proportion of terrestrial palynomorphs, are interpreted as representing transgression. The maximum-flooding may be represented by the high-gamma peak. Alternatively, it may occur close to the kaolinite and terrestrial palynomorph minima, in which case the highgamma peak could be regarded as a purely basinal phenomenon, perhaps related to the onset of full anoxia. The fall in sea-level represented by the Lista-Sele facies transition is reflected throughout the UK North Sea area, with an overall fall in sealevel of at least 100 m being inferred for southern England. In the Bradwell section (see Knox et al. 1994) fine-grained mudstones of Rhabdammina biofacies (Thanet Formation, Lista equivalent) are overlain by pedogenically altered lagoonal or shallow marine sandy mudstones (upper division of the Upnor Formation, Sele equivalent). It is difficult to explain a sea-level fall of such magnitude and rapidity other than by tectonic uplift (see also Neal 1996). Precise correlation of the onshore and offshore sections is uncertain. However, it seems likely that the hiatus at the base of the lower division of the Upnor Formation (Ellison et al. 1996), during which the Thanet Formation was subjected to decalcification and other effects of meteoric leaching, equates with the major sea-level fall recorded at the base of the Sele Formation (unit S 1a). In both cases, the overlying sediments appear to have been deposited at a time of continued, if somewhat restricted, connection with oceanic waters (allowing influx of NP9 calcareous nannofloras to southern England). Further restriction of free connection between the North Sea and the eastern Atlantic appears to have occurred prior to deposition of the upper division of the Upnor Formation, which is of restricted marine facies. The onset of continental deposition, represented by the base of the lower leaf of the Reading Beds and by the base of the Argiles plastiques bariolres marks the complete closure of the southwestern oceanic connection.

From these observations, it is plausible to suggest that the Upnor Formation and immediately overlying continental beds are represented in the central North Sea by the transitional, upwardcoarsening turbidite sandstone unit. On this correlation, the turbidite sandstone unit would approximate to the horizon of mammal migration (Hooker 1996) and the negative carbon isotope excursion recorded by Stott et al. (1996). The principal reduction in benthic faunas would, however, not be related to the oceanic 'benthic extinction', however, since it occurs lower in the succession. In the absence of any reliable onshoreoffshore corelation tool, such correlations must be considered speculative, as must any connection between the events recorded from the central North Sea and the global changes that were occurring at about the same time. It is quite possible to explain the central North Sea events purely in terms of regional tectonism, with the progressive elimination of benthic faunas and the onset of basin anoxia resulting from basin isolation and the influx of kaolinite resulting from increased humidity and precipitation in response to changes in palaeogeography. We are thus left with the tantalizing situation that some of the oceanic events recorded from the Paleocene/Eocene boundary interval are paralleled in the broadly coeval sediments of the central North Sea, but that on present evidence it is not possible to say whether some or all of the North Sea events are caused by local (North Atlantic) tectonism or by global climatic change. Of course, it may be wrong to think in terms of these two extremes, as changes in oceanic circulation pattern, global climatic change, and mammal migration could themselves all be explained in terms of plate reorganization prior to the opening of the North Atlantic (Eldholm & Thomas 1993). In conclusion, one of the most intriguing aspects of early Paleogene stratigraphy is the evidence in sediments of late NP9 age for a major but short-term interruption to the long-term trends of oceanic warming and climate change. It must be expected that such a profound change in the world's oceans would have left its mark on the terrestrial and epicontinental marine record, in which case the onshore and offshore successions of northwest Europe must be prime candidates for study. Such study would not be of purely academic interest, since an understanding of the relative roles of regional uplift, eustasy, and global climatic change would not only answer some long-standing scientific questions, but also throw new light on the origin of one of northwest Europe's most productive oil reservoirs, the Forties Sandstone. The elucidation of event stratigraphy within the Paleocene/Eocene

CORRELATION OF THE EARLY PALEOGENE IN NORTHWEST EUROPE: AN OVERVIEW boundary interval in northwest Europe thus provides a prime example of the benefits of interchange of data and ideas across the academic/commercial divide.

9

I am grateful to Richard Corfield, David Jolley and Andy Morton for helpful comments on the manuscript. Publication is with the approval of the Director, British Geological Survey.

References ALl, J. R. & JOLLEY,D. W. 1996. Chronostratigraphic framework for the Thanetian and lower Ypresian deposits of southern England. This volume. , HAILWOOD,E. A. & KING, C. 1996. The 'Oldhaven magnetozone' in East Anglia: a revised interpretation. This volume. , KING, C. & HAILWOOD, E. A. 1993. Magnetostratigraphic calibration of early Eocene depositional sequences in the southern North Sea Basin. In: HAmWOOD, E. A. & KIDD, R. B. (eds) High Resolution Stratigraphy. Geological Society, London, Special Publication, 70, 99-125. ARMENTROUT, J. M, MALACEK, S. J., FEARN, L. B. SHEPPARD, C. E., NAYLOR, P. H., MILES, A. W., DESMARAIS,R. J. & DUNGY,R. E. 1993. Log-motif analysis of Paleogene depositional systems tracts, central and northern North Sea: defined by sequence stratigraphic analysis. In: PARKER, J. R. (ed.) Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference. Geological Society, London, 45-58. AUBRY, M. E 1986. Paleogene calcareous nannoplankton biostratigraphy of Northwestern Europe. Palaeogeography, Palaeoclimatology, Palaeoecology, 55, 267-334. , BERGGREN,W. A., STOTF,L. & SINHA,A. 1996. The upper Paleocene - lower Eocene stratigraphic record and the Paleocene-Eocene boundary carbon isotope excursion: implications for geochronology. This volume. BERGGREN, W. A. & AUBRY, M. -E 1996. A late Paleocene-early Eocene NW European and North Sea magnetobiochronological correlation network. This volume. , KENT, D. V., SWISHER, C. C. I I I & AUBRY,M.-R 1995. A revised Cenozoic geochronology and chronostratigraphy. In: BERGGREN, W. A, KENT, D. V., AUBRY, M.-R & HARDENBOL, J. (eds) Geochronology, Time Scales and Stratigraphic Correlation: Framework for an Historical Geology. Society of Economic Geologists and Paleontologists, Special Publication, 54, Tulsa. CANDE,S. C. & KENT,D. V. 1995. Revised calibration of the geomagnetic polarity timescale for the Late Cretaceous and Cenozoic. Journal of Geophysical Research, 100 (B4), 6093-6095. CAVELIER, C. & POMEROL, C. 1986. Stratigraphy of the Paleogene. Bulletin de la Socidtd Gdologique de France, 8 (II,2), 255-265. CHARISI, S. D. & SCHMrrZ,B. 1996. Early Eocene palaeoceanography and palaeoclimatology of the eastern North Atlantic: stable isotope results for DSDP Hole 550. This volume. CORFrELD, R. M. & NORRIS, R. D. 1996. Deep water circulation in the Paleocene Ocean. This volume. COSTA, L. I., DENISON, C. & DOWNIE, C. 1978. The

Paleocene/Eocene boundary in the Anglo-Paris Basin. Journal of the Geologial Society, London, 135, 261-264. CURRY, D., ADAMS,C. G., BOULTER,M. C., DILLEY,F. C., EAMES, E E., FUNNELL, B. M., & WELLS, M. K. 1978. A Correlation of Tertiary Rocks in the British Isles. Geological Society, London, Special Report, 12, 1-72. DEN HARTOGJAGER,D., GILES, M. R. & GRIFFITHS,G. R. 1993. Evolution of Paleogene submarine fans of the North Sea in space and time. In: PARKER,J. R. (ed.) Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference. Geological Society, London, 59-72. ELDHOLM,O. & THOMAS,E. 1993. Environmental impact of volcanic margin formation. Earth and Planetary Science Letters, 117, 319-329. ELLISON, R. A., ALI, J. R., HINE, N. M. & JOLLEY,D. W. 1996. Recognition of Chron 25n in the upper Paleocene Upnor Formation of the London Basin, UK. This volume. GALLOWAY, W. E., GARBER, J. L., XIJIN, LIU & SLOAN, B. J. 1993. Sequence stratigraphic and depositional framework of the Cenozoic fill, central and northern North Sea Basin. In: PARKER,J. R. (ed.) Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference. Geological Society, London, 33-44. GELY, J. -P. & LORENZ,C. 1991. Analyse s6quentielle de l'Eoc~ne et de l'Oligocbne du bassin Parisien (France). Revue de l'lnstitut Franfais du Pdtrole, 46 (6), 713-747. GRADSTEIN,E M., KAMINSKI,M. A. & BERGGREN,W. A. 1994. Cenozoic biostratigraphy and paleoceanography. In: The North Sea and Labrador Shelf. Micropaleontology, 40, Supplement. HARDENBOL, J. 1994. Sequence stratigraphic calibration of Paleocene and Lower Eocene continental margin deposits in NW Europe and the US Gulf Coast with the oceanic chronostratigraphic record. GFF, 116, 49-51. HARLAND,W. B., ARMSTRONG,R. L., COX, A. V., CRAIG, L. E., SMITH, A. G. & SMITH, D. G. 1989. A Geologic Time Scale. Cambridge University Press. HEILMANN-CLAUSEN,C. 1985. Dinoflagellate stratigraphy of the uppermost Danian to Ypresian in the Viborg 1 borehole, central Jylland, Denmark. Danmarks Geologiske UndersCgelse, Serie A, 7. 1994. Review of Paleocene dinoflagellates from the North Sea region. GFF, 116, 51-53. HINE, N. M. 1994. Calcareous nannoplankton assemblages from the Thanet Formation in Bradwell Borehole, Essex, England. GFF, 116, 54-55. HOOKER, J. J. 1996. Mammalian biostratigraphy across the Paleocene-Eocene boundary in the Paris, London and Belgian Basins. This volume.

10

R. W. O'B. KNOX

JOLLEY, D. W. 1992. Palynofloral association sequence stratigraphy of the Palaeocene Thanet Beds and equivalent sediments in eastern England. Review of Palaeobotany and Palynology, 74, 207-237. 1996. The earliest Eocene sediments of eastern England: an ultra-high resolution palynological correlation. This volume. JONES, R. W. & MILTON, N. J. 1994. Sequence development during uplift: Palaeogene stratigraphy and relative sea-level history of the Outer Moray Firth, UK North Sea. Marine & Petroleum Geology, 11, 157-165. JoY, A. M. 1996. Controls on Eocene sedimentation in the central North Sea Basin: results of a basinwide correlation study. This volume. KING, C. 1989. Cenozoic of the North Sea. In: JENKINS,D. G. & MURRAY, J.W. (eds) Stratigraphical Atlas of Fossil Foraminifera (2nd edition). Ellis Horwood, Chichester, 294-298. KNOX, R. W. O'B. 1994. From regional stage to standard stage: implications for the historical Paleogene stratotypes of NW Europe. GFF, 116, 55-56. - & HOLLOWAY,S. 1992. Paleogene of the Central and Northern North Sea. In: KNox, R. W. O'B. & CORDEY, W. G. (eds) Lithostratigraphic nomenclature of the UK North Sea. British Geological Survey, Nottingham. , HINE, N. &ALI, J. 1994. New information on the age and sequence stratigraphy of the type Thanetian of Southeast England. Newsletters on Stratigraphy, 30 (1), 45--60. KOCH, P. L., ZACHOS, J. C. & GINGERICH, P. D. 1992. Coupled isotopic change in marine and continental carbon reservoirs near the Palaeocene/Eocene boundary. Nature, 358, 319-322. MILTON, N. J., BERTRAM,G. T. & MANN,I. R. 1990. Early Palaeogene tectonics and sedimentation in the Central North Sea. In: HARDMAN, R. F. P. BROOKS, J. (eds) Tectonic Events Responsible for Britain's Oil and Gas Reserves. Geological Society, London, Special Publication, 55, 339-351. MITCHELL, S. M., BEAMISH, G. W. A., WOOD, M. V., MALACEK, S. J., ARMENTROUT,J. A., DAMUTH,J. E. & OLSON, H. C. 1993. Paleogene sequence stratigraphic framework of the Faeroe Basin. In: PARKER, J. R. (ed.) Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference. Geological Society, London, 1011-1023. MITLEHNER, A. 1996. Palaeoenvironments in the North Sea Basin around the Paleocene-Eocene boundary: evidence from diatoms and other siliceous microfossils. This volume. MORTON, A. C. & KNox, R. W. O'B. 1990. Geochemistry of late Palaeocene and early Eocene tephras from the North Sea Basin. Journal of the Geological Society, London, 147, 425-437. --, BACVdVIAN,J. & HARLAND,R. 1983. A reassesment of the stratigraphy of DSDP Hole 117A, Rockall Plateau: Implications for the Paleocene-Eocene boundary in N.W. Europe. Newsletters on Stratigraphy, 12(2), 104-111. MUDGE, D.C. & BUJAK, J.P. 1996. An integrated stratigraphy for the Paleocene and Eocene of the North Sea. This volume.

• COPESTAKE,P. 1992. Revised Lower Palaeogene lithostratigraphy for the Outer Moray Firth, North Sea. Marine & Petroleum Geology, 9, 53-69. MUSSETT, A. E., DAGLEY, P. & SKELHORN, R. R. 1988. Time and duration of activity in the British Tertiary igneous province. In: MORTON, A. C. 8,: PARSON,L. M. (eds) Early Tertiary Volcanism and the Opening of the North East Atlantic. Geological Society, London, Special Publication, 39, 337-348. NADIN, P. A. & KUSZNIR,N. J. 1996. Forward and reverse stratigraphic modelling of Cretaceous-Tertiary post-rift subsidence and Paleogene uplift in the Outer Moray Firth Basin, central North Sea. This volume. NEAL, J. E. 1996. A summary of Paleogene sequence stratigraphy in northwest Europe and the North Sea. This volume. NOBLE, R. H., MCINTYRE, R. M. & BROWN, P. E. 1988. Age constraints on Atlantic evolution: timining of magmatic activity along the E Greenland continental margin. In: MORTON, A. C. & PARSON, n. M. (eds) Early Tertiary Volcanism and the Opening of the North East Atlantic. Geological Society, London, Special Publication, 39, 201-214. O'CONNOR, S. J. & WALKER, D. 1993. Paleocene reservoirs of the Everest Trend. In: PARKER, J. R. (ed.) Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference. Geological Society, London, 145-160. PLINT, A. G. 1988. Global eustasy and the Eocene sequence in the Hampshire Basin, England. Basin Research, 1, 11-22. POMEROL, C. 1989. Stratigraphy of the Palaeogene: hiatuses and transitions. Proceedings of the Geologists' Association, 100, 313-324. POWELL, A. J. 1988. A modified dinoflagellate cyst biozonation for latest Palaeocene and earliest Eocene sediments from the central North Sea. Review of Palaeobotany and Palynology, 56, 327-344. - 1992. Dinoflagellate cysts of the Tertiary System. In: POWELL, A. J. (ed.) A Stratigraphic Index of Dinoflagellate Cysts. Chapman & Hall, London, 155-251. , BRINKHUIS, H. ~:; BUJAK, J. P. 1996. Upper Paleocene - Lower Eocene dinoflagellate cyst sequence biostratigraphy of southeast England. This volume. RITCHIE, J. D. & HITCHEN, K. 1996. Early Paleogene offshore igneous activity to the northwest of the UK margin and its relationship to the North Atlantic Igneous Province. This volume. SCHMITZ, B. 1994. The Paleocene Epoch - stratigraphy, global change and events. GFF, 116, 39-67. - - - , HEILMANN-CLAUSEN,C., KING, C., STEURBAUT,E., COR~ELD, R. M. & CARTLIDGE,J. E. 1996. Stable isotope and biotic evolution in the North Sea during the early Eocene: the Alb~ek Hoved section, Denmark. This volume. SCrmODER, T. 1992. A palynological zonation for the Paleocene of the North Sea Basin. Journal of Micropalaeontology, 11, 113-126. SIESSER, W., WARD,D. J. & LORD A. R. 1987. Calcareous nannoplankton biozonation of the Thanetian stage

--

CORRELATION OF THE EARLY PALEOGENE IN NORTHWEST EUROPE: AN OVERVIEW (Palaeocene) in the type area. Journal of Micropalaeontology, 6, 85-102. STEURBAUT,E. 1990. Ypresian calcareous nannoplankton biostratigraphy and paleogeography of the Belgian Basin. In: DuPuIS, C., DE CONNINCK, J. & STEURSAUT,E. (eds) Bulletin de la Societd Belge de Gdologie, 97 (3-4), 251-285. STEWART,I. J. 1987. A revised stratigraphic interpretation of the early Palaeogene of the central North Sea. In: BROOKS, J. & GLENNIE,K. (eds) Petroleum Geology of North West Europe. Graham & Trotman, London, 557-576. STOTF, L. D., SINHA, A., THIRY, M., AUBRY, M. -P. & BERGGREN,W. A. 1996. Global 513C changes across the Paleocene-Eocene boundary: criteria for terrestrial-marine correlations. This volume. THOMAS, E. & SHACKLETON, N. J. 1996. The Paleocene-Eocene benthic foraminiferal extinction and stable isotope anomalies. This volume. THOMAS, J. E. 1996. The occurrence of the dinoflagellate

11

cyst Apectodinium (Costa & Downie 1976) Lentin & Williams (1977) in the Moray and Montrose Groups (Danian to Thanetian) of the UK central North Sea. This volume. VANDENBERGHE, N., LAGA, P., STEURBAUT, E., HARDENBOL, J. & VAIn, P. R. In press. Sequence stratigraphy of the Tertiary at the southern border of the North Sea Basin in Belgium. In: HARDENSOL,J., VAIL, P., DE GRACIANSKY,P.C. & JACQUIN,T. (eds) Sequence Stratigraphy of Mesozoic and Cenozoic European Basins. CNRS-IFP, Paris. VINKEN, R., VON DANIELS,C. H., GRAMANN,E, KOTHE, A., KNOX,R. W. O'B., KOCKEL,E, MEYER, K. J. & WEISS, W. (eds) 1988. The Northwest European Tertiary Basin. Results of the IGCP Project No. 124. Geologisches Jarbuch, A 100. WOOD, S. E. & TYSON, R. V. 1996. An integrated palynological-palynofacies approach to the zonation of the Paleogene in the Forties-Montrose Ridge area, central North Sea. This volume.

A summary of Paleogene sequence stratigraphy in northwest Europe and the North Sea J. E. N E A L

Rice University, Department of Geology & Geophysics, PO Box 1892, Houston, TX 77251, USA; present address: Exxon Production Research Co., PO Box 2189, Houston, TX 77252, USA Abstract: Sequence stratigraphic analysis of Paleogene central North Sea well-log, seismic and biostratigraphic data recognizes patterns of cyclic sedimentation seen in the physical stratigraphy and biostratigraphy. Numerous authors have documented cyclic sedimentation resulting from relative changes in sea level in northwest Europe, but interregional integration of these observations with North Sea subsurface data is lacking in the literature. Presented here is a chronostratigraphic correlation framework for the Paleogene of northwest Europe, built by integrating subsurface and outcrop data using sequence stratigraphic first principles. Biostratigraphic data from many sources is ordered with the composite standard method. Graphic correlation of this data documents certain correlations and helps suggest previously unrecognized ties. Paleogene North Sea sediments record five major regressions and their intervening major transgressions. Overprinting this low frequency signal are 19 higher frequency sequence cycles that control lithofacies distribution. In northwest Europe, western basins (London-Hampshire, Paris and Belgian) have shallow marine to non-marine settings which reveal basinward and landward facies shifts that indicate sea level changes. The biostratigraphy of these shallow water deposits is linked to deep water central North Sea biostratigraphy by correlating through deeper water deposits outcropping in Denmark that have been tied to western basin stratigraphy. Using this biostratigraphic framework, key bounding surfaces are correlated between basins using sequence stratigraphic principles. Depositional sequences are recognized onshore that are completely sediment starved in the North Sea. The mixing of low and high frequency sea level signals requires that all of northwest Europe be studied to recognize the 'true' signal. Final correlations resolve 30 depositional sequences with five long-term sea-level changes that can vary from one sub-basin to another.

Geologists have recognized transgressions and regressions in northwest European Tertiary sediments since the late 18th century, as Lavoisier noted in 1766 the 'flux and reflux' of the sea represented by map units in the Paris Basin (Rappaport 1969). These sediments challenge us to unravel their stratigraphic equivalence, from shelf to basin and region to region. Sequence stratigraphy is an infant technique compared with traditional correlation procedures of lithostratigraphy, biostratigraphy and even magnetostratigraphy. A geologist using sequence stratigraphy can synthesize results from more traditional correlation methods and suggest ways to solve stratigrapbic contradictions. A sequence stratigrapher uses regional and local stratigraphic and sedimentological observations to reconstruct the relative sea-level history represented within the rock record. Chronostratigraphic charts (Wheeler 1958) represent deposition and lacuna along a given profile through time, graphically identifying depositional sequences. A basin's chronostratigrahic chart will show a record of transgressions and regressions through time.

The northwest European Paleogene basin had many sub-basins, each with its own stratigraphy. Belgium, northern France and southeast England have outcrops of Paleogene shallow marine and estuarine sediments. Boreholes and outcrops from Denmark and northern Germany encounter Paleogene deeper shelf and basinal sedimentary deposits. The central North Sea was a depocentre for sandy submarine fans and prograding deltas identified by interpretation of seismic data and well information. Because all these sub-basins were connected (Ziegler 1990), the rock record of each should show similar relative sea-level histories unless local tectonic uplift obscures the signal. Under the auspices of the CNRS-IFP Sequence Stratigraphy of Mesozoic and Cenozoic European Basins Project, I present below a stratigraphic framework interpreted from a seismic data grid of 7000 line kilometres (Fig. 1) and approximately 250 well-logs (150 of which had detailed biostratigraphic reports). The framework developed from working with experts from European countries that have Paleogene deposits and using

From Knox, R. W. O'B., Corfield, R. M. & Dunay, R. E. (eds), 1996, Correlationof the Early Paleogene in NorthwestEurope, Geological Society Special Publication No. 101, pp. 15-42.

15

16

J.E. NEAL

the ongoing and published biostratigraphic, sedimentological and physical stratigraphic research to correlate a central North Sea sequence stratigraphy with onshore sections. Graphic correlation of biostratigraphic data carried out by J. Stein and J. Gamber of Amoco (Neal et al. 1994) augmented key marker stratigraphy and suggested different and more precise correlations than have been previously published. Sequence stratigraphy can be used in two

different ways depending on the goals of the researcher. The more economically-minded approach focuses mainly on detailed description of the rock record, identifying key stratigraphic bounding surfaces and sedimentary bodies for their fluid flow properties. Regional correlation is less important at this scale. The second approach is more academic, using every available piece of information to construct an internally consistent, documented chronostratigraphic framework. The

Fig. 1. Map of northwestern Europe with the outline of Eocene deposition from Zeigler (1990). Also shown are the locations of North Sea seismic lines and key outcrop and borehole sections. Labelled North Sea Pre-Tertiary structural elements are: E.S.P., East Shetland Platform; W.G.G., Witch Ground Graben; S.V.G., South Viking Graben; EM.H., Forties-Montrose High.

PALEOGENE SEQUENCE STRATIGRAPHY IN NORTHWEST EUROPE second approach is used in this paper. Biostratigraphic data resolution is the limiting factor in the precision, accuracy and detail of a documented chronostratigraphic chart. Each subsurface depositional sequence carried in the framework below has a unique biostratigraphic signature, but a sequence stratigraphic approach integrating sequence stacking patterns within major regression/ transgression facies (R/TF) cycles was necessary when sequences were identified in outcrop section at a higher resolution than individual biostratigraphic schemes could resolve (i.e. two Belgian sequences in N P l l =two English sequences in NP11; an interval largely represented by sediment starvation in the central North Sea). Depositional sequences occur at many timescales (e.g. Mitchum & Van Wagoner 1991; Posamentier et al. 1992) and aliasing sequences with biostratigraphy is a constant pitfall. Employing sequence stacking patterns is one way to enhance the correlation. An important advantage this study has over previous correlation frameworks, as mentioned above, is the composite standard biostratigraphic method and graphic correlation (Shaw 1964; Miller 1977; Stein et al. 1995). Composite standard biostratigraphy synthesizes a detailed ideal section from the most complete individual section available, which is then enhanced with additional data from multiple, stratigraphically overlapping individual sections. Graphic correlation is a technique to graphically represent the completeness of an individual test section relative to the composite standard (Fig. 2). Graphic correlation helps interpret the rock record by identifying and quantifying stratigraphic gaps, which appear as terraces in a line of correlation that relates the completeness of a test section to the composite standard. The line of correlation (LOC) is interpreted from the scatter of individual fossil markers as they appear in the test section and the composite standard. Recognition of data terraces is then used to construct a framework of chronostratigraphic units based on overlapping graphic correlation terraces. If terraces in two or more wells share intervals of time [expressed in terms of composite standard units (CSU)], then the time of overlap can be correlated as a single or composite event (i.e. a lacuna). Terrace-bounded units bracket depositional sequences for correlation purposes, but not all depositional sequences are associated with a terrace (Neal 1994).

Previous work and correlation problems Numerous studies of northwest European Paleogene stratigraphy exist in the literature. The last decade witnessed a revision of Paleogene

17

stratigraphy onshore with the publication of detailed biostratigraphic (Aubry 1986), magnetostratigraphic (Ali et al. 1993) and sedimentological (Plint 1988) frameworks that employ sequence stratigraphic correlation techniques. Other regional papers will arrive with the publication of the Proceedings from the CNRS-IFP Meeting of Dijon 1992 (i.e. Michelsen et al. 1995; Vandenberghe et al. 1995). The central North Sea alone is on a second or third generation of stratigraphy papers, each new generation becomes more complex as data resolution increases. Parker (1975) published the first regional study of Paleogene North Sea deposition that inspired works through to the most recent round of papers presented at the 4th Conference on Petroleum Geology of Northwest Europe (Parker 1993). New aspects of North Sea stratigraphy are continually emerging, creating a need to standardize present published frameworks and document the sea-level signal for the North Sea and northwest Europe. The International Geologic Correlation Project (IGCP) Project 124 (Vinken et al. 1988) synthesized and standardized vast amount of raw stratigraphic information available in northwest Europe, but focused mainly on pure biostratigraphic correlations and emphasized lithostratigraphy. The IGCP Project 124 results provide a control on depositional sequence correlations and allow crosscorrelations of different fossil taxa types. Sequence stratigraphic analysis uses this data source as a framework, but provides additional information not covered the IGCP Project 124 report. The conclusions presented below differ from those of IGCP Project 124 mainly because the depositional sequence-based framework contains more subdivisions and is largely independent of strict lithostratigraphic correlation. This study attempts to standardize the present state of depositional sequence correlations similar to the way IGCP Project 124 ordered biostratigraphic and lithostratigraphic correlations. Standardizing a relative sea level history from subsurface data and published studies has, however, proved difficult. Various groups have created separate biostratigraphic zonation schemes and used different key correlation markers. This study used composite standard biostratigraphy to unify different zonal schemes without losing the ability to recognize individual key correlation markers. This advantage allowed comparisons with the published stratigraphic frameworks and led to a tie with outcrop and borehole sections of northwest Europe. In standardizing relative sea-level histories for northwest Europe, the first step was to evaluate the applicability of the (Haq et al. 1988) eustatic curve. Since northwest Europe was one of the key areas for the development of the curve, its documentation

18

J.E. NEAL

~T~E

.~=~ o,u

do ,..cZ

.~

r

"~

0

09

~~ ,,..a

,~,_qo

PALEOGENE SEQUENCE STRATIGRAPHY IN NORTHWEST EUROPE in this study would seem crucial. The fact that many key biostratigraphic markers in northwest Europe are endemic to the area made correlations to the eustatic curve difficult, but a correlation is made

19

by linking the subsurface framework to outcrop sections containing nannofossil zonations. A possible correlation between the eustatic curve and several North Sea frameworks is shown in

Fig. 3. Comparison of several sequence stratigraphic frameworks published in the literature, corrected to the timescale of Berggren et al. (1995). This figure highlights the different resolutions and internal inconsistencies of various frameworks for the same stratigraphic interval.

20

J.E. NEAL

Fig. 3. This correlation highlights problems of different resolutions and diachroneity in the various published works. The portion of the Haq curve shown is modified to be consistent with the nannofossil stratigraphy of Aubry et al. (1988), tied to the new geomagnetic timescale of Cande & Kent (1992) and Berggren et al. (1995). Armentrout et al. (1993) compared their sequence stratigraphic framework to the Haq curve and recognized fewer events than predicted by the global curve. They based their framework on a series of key fossil markers, which become tie points with the framework presented in this paper. Next is the more detailed framework of Den Hartog Jager et al. (1993), showing a higher resolution than both the Mobil (Armentrout et al. 1993) and Exxon (Haq et al. 1988) frameworks. Den Hartog Jager et al. (1993) based their correlations on the Shell North Sea Paleogene biozones (Schrfder 1992). Stewart (1987) published the first major lower Paleogene sequence stratigraphic framework in the North Sea, displayed at the far right in Fig. 3. This framework is diachronous in places when comparing the type wells and maps of Stewart (1987) to a biostratigraphy-based framework. All these different frameworks, constructed from similar data sets, highlight the problem of finding an agreed relative sea-level signal. The Paleogene North Sea relative sea level signal is easily divided into cycles of two categories: sequence cycles (ranging from 0.3 to 5 million years when tied to the present geomagnetic timescale) and major regressive/transgressive facies (R/TF) cycles (ranging from 3 to c. 13 million years). Sequence cycles of relative sea level form depositional sequences ( s e n s u Mitchum et al. 1977). They, along with sediment supply and the depositional profile, control distribution of lithofacies and seismic stratal patterns. These cycles are correlative within biostratigraphic resolution as stratigraphic events throughout the basin and may represent a eustatic signal. Every location does not record every sequence cycle, though, as composite unconformities or marine hiatal intervals can account for time represented by multiple sequences elsewhere in the basin (Neal et al. 1994). Major R/TF cycles consist of multiple sequence cycles. They reflect a reorganization of the depositional system in a basin or sub-basin and have been related to tectonic mechanisms. These cycles are not always correlative and can be as much as half a cycle out of phase when comparing their effects around northwest Europe. Major R/TF cycles control the facies characteristics of their component sequence cycles. For example, sequence cycles that occur on the falling sea-level limb of a major R/TF cycle will have well developed lowstands and are marked by fan-depositional pulses into distal parts

of the basin. Sequence cycles that occur during the turn-around and rising sea-level portions of a R/TF cycle are characterized by aggradational facies (thick coastal plain deposits). Sequence cycles occurring during the 'highstand' interval of a R/TF cycle may be completely starved in the central North Sea and are recognized only by correlation to onshore sections. Five major R/TF cycles, grouping 18 regionally identifiable sequence cycles, are recognized for the Paleogene central North Sea (Fig. 4). Sequence cycles are named for lithostratigraphic units that typically fall within the sequence (Neal et al. 1994). Various publications have noted major tectonic events in the North Sea area (e.g. Cavelier & Pomerol 1979; Galloway et al. 1993; Japsen 1993; Vandenberghe et al. 1995). Major R/TF cycles are linked to these events, which are recognizable throughout the central North Sea, but their magnitude and precise timing is not correlative throughout northwest Europe. For example, the lower Oligocene is regressive in the central North Sea and Denmark (Michelsen et al. 1995) as a result of the uplift of Norway and possibly the Shetland Platform (Japsen 1993; Galloway et al. 1993). This same time interval corresponds to a major transgression (RupelianStampian) in Belgium (Steurbaut 1992) and France (Cavelier & Pomerol 1979; Gfly & Lorenz 1991). A final complication is that sequence cycles have different apparent magnitudes around northwest Europe and their expression is enhanced or subdued by local tectonic events (Vail et al. 1991). A summary chronostratigraphic chart (Fig. 5) synthesizes all this stratigraphic information to produce a Paleogene North Sea relative sea-level history. Specific formations and fossils are noted below as they document changes in relative sea level from the top Cretaceous to the middle Oligocene.

Danian-Selandian deposition Everywhere in northwest Europe, the Paleogene sequence stratigraphy commences with a major regression that started at various times in the Upper Cretaceous, depending on sub-basin location. In the central North Sea, the Maastrichtian Tor Formation and Danian Ekofisk Formation arc separated by 'a change in sedimentary facies (that) is generally an omission or erosion surface, with the basal Ekofisk a condensed black shale' (Kennedy 1987, p. 477). The duration of this marine hiatus is difficult to determine. The central North Sea biostratigraphic framework relics on ditch cutting samples, and therefore is based on last appearance datums (LAD), which correspond to first downhole

PALEOGENE SEQUENCE STRATIGRAPHY IN NORTHWEST EUROPE

Ma. Central North Sea R\TF Cycles

30 Paleogene (Lower 5 Oligocene)

Major North Sea Subsurface Tectonic Events Sequence Cycles Land ~ " - ~ > S e a Lower L/ark II Uplift oft~

L~

35 ~

Lark l

Brioc

Paleogene 4 (Upper40 Middle Eocene)

21

Skaqerrat:U Katteqat M

Platform]N~ _ Alpine,~ uompressiqn.

~i Alba Undiff.-

~

~

jr ~ "

e'i'umw

Nauchlan

ffaran

45

Upp;r Tay Phases of 1~'

Paleogene 3 SO 55

(Lower Eocene) Paleogene 2

.. (,Upper I~aleocene~ Paleogene I (Lower Paleocene)

~igg Middle Frigg LowerFriggj Upper

Badl"ff-er'uooer Sele,-~

- Middle ~ele

'-'-

.UDDer B~mgPa~ r Forties

Uplift in the Shetland Platform

J

Jt

voc.ncsh Thulea~n ! 1

Lower Balmoral Andrew

/ Maureen /

Ekofisk

/

Selandian Clastics

Dan~ian

Fig. 4. Long-term relative sea-level curve and location on that curve of major regressive/transgressive facies (R/TF) cycles and depositional sequences for the central North Sea. This curve is related to major regional tectonic events from Zeigler (1990), Milton et al. (1990), Galloway et al. (1993) and Vandenberghe et al. (1995).

occurrences (FDO). The most commonly reported fossils in the Ekofisk are planktonic foraminifera such as Globorotalia compressa, Globorotalia pseudobulloides and Eoglobigerina trivialis, which range upwards into the Maureen sequence. The oldest nannofossil marker found in the Ekofisk chalks is the acme occurrence of Prinsius dimor-

phosus, which has been placed near the NP2-3 boundary (Thomsen & Heilmann-Clausen 1985; Thomsen 1992). The oldest correlatable FDO of dinoflagellate cysts in the central North Sea Paleogene is Senoniasphaera inornata, marking the top of dinocyst zone D1 that occurs within NP3 (Vinken et al. 1988).

22

J.E. NEAL

Danian deposits around northwest Europe have a complex and variable stratigraphy. Depositional sequences are difficult to identify within the Ekofisk Formation due to numerous reworking events and allocthonous units (Johnson 1987; Kennedy 1987), therefore the Ekofisk has been considered here as a single sequence. In Denmark, detailed stratigraphic mapping of chalk facies reveals four sequence cycles of transgression and regression (Thomsen 1990, 1992). The oldest of these cycles occurs within the NP1 zone. Whether one or two cycles can be confidently identified and correlated remains a possibility, but at least one cycle can be correlated to other regions. A thin deposit, known as the 'Pa-layer', assigned to NP1 zone, is also reported in Belgium (Verbeek, pers. comm.) and Bignot (1993) placed the Mont AimEVertus Formation within the lowermost Danian. This sequence cycle does not appear on the eustatic curve of Haq e t al. (1988) and may be recognizable only in northwest Europe. The Marnes de Meudon of the Paris Basin are placed in zones NP2-3 by Aubry (1985), but this is contradicted by Bignot (1993) who positioned the base of the Montian stage boundary in NP3, not NP2. Sediments of NP2 age have only been positively identified in northwest Europe in Denmark and possibly the central North Sea. In the London-Hampshire Basin, and parts of the Paris Basin, the lower Danian section is not present and is represented by an unconformity that erodes down to the Campanian and Santonian (Aubry 1985; Pomerol 1989). Deposits of NP3--4 age exist in the Belgium, Paris and Denmark basins and document another sequence cycle. The Mons Basin near the BelgiumFrance border is a local area of subsidence that preserves deposits of a relative sea level cycle

with the Malogne conglomerate, Vroenhoven and Houthem Formations, and the Calcaire de Ciply (Vandenberghe e t al. 1995). Similar-aged deposits are found south and west of Paris in the Calcaire de Meudon (Vinken e t al. 1988; Pomerol 1989). The '61 Ma' flooding event on the Haq curve could be responsible for these deposits in Belgium and France (Aubry 1985; Bignot 1993; Vandenberghe e t al. 1995) and for the most widespread Danian Limestone depositional event (Thomsen 1992). A sequence cycle not identified by Haq e t al. (1988) has been recently identified in the Gulf of Mexico Alabama section (Mancini & Tew 1991) between the planktonic foraminifera zones of Morozovella uncinata (P2) and M o r o z o v e l l a a n g u l a t a (P3A). Vandenberghe e t al. (1995) note evidence for this cycle in the Maasmechlen Formation (P3A - Hooyberghs 1983), which is a thin transgressive calcarenite above the Calcaire de Ciply. The IGCP Project 124 assigned this unit to the NPF2 and B 1 foraminifera zones (Hooyberghs 1988; Laga 1988). These two fossil zones cover broad time intervals, but overlap only during NP4 (Vinken e t al. 1988).

Paleogene 1 R/TF cycle (Danian chalk/ Thanetian clastics) Uplift of northwest Europe continued through the next sequence cycle, bringing major siliciclastic deposits to the basin due to erosional removal of the blanketing Upper Cretaceous chalks. The type Selandian-Danian stage boundary marks the transition from chalk to siliciclastics in Denmark with deposition of the Lellinge Green Sand and Kerteminde Marl (Rosenkrantz 1924; Thomsen & Heilmann-Clausen 1985; Berggren 1994;

Fig. 5. Sequence stratigraphic framework chart for the Paleogene of Northwest Europe. 1 Age scale comes from Berggren et al. (1995) with nannofossil correlations from Aubry et al. (1988). 2 Outcrop correlations compiled primarily from the work of Heilmann-Clausen (1985, 1994), Thomsen & Heilmann-Clausen (1985), Nielsen et al. (1986), Heilmann-Clausen & Costa (1989), Michelsen et al. 0995), Vandenberghe et al. (1992, 1995), De Coninck (1990), Steurbaut (1988), Plint (1983, 1988), Knox (1994), Knox et al. (1994), All et al. (1993), Eaton (1976), J. Riveline & M. Renard (pers. comm.), Cavalier (1988), Cavalier & Pomerol (1986), Aubry (1983), 1986, 1994), Bignot (1993), G61y & Lorenz (1991) and Steurbaut & Nolf (1986). 3 Seismic sequences in the eastern North Sea (Denmark) comes from Michelsen et al. (1995). 4 Hiatuses in the Paris Basin come from Pomerol (1989). I'~-New IO &SS)], position of a sequence boundary from outcrop and subsurface data not picked in Haq et al. (1988); I-~--New (O) I , position of a sequence boundary from outcrop data not picked in Haq et al. (1988). Correlation of the frameworks from Den Hartog Jager et al. (1993) and Armentrout et al. (1993) are based on biostratigraphic calibration points. The (Haq et al. 1988) eustatic curve is modified based on a correlation of nannofossil zone differences between Haq et al. (1988) and Aubry et al. (1988). Lithology and facies key: hemipelagic mud/sediment starved, r----q;tuff, ~ ; dominantly-silt turbidites, I ; dominantly-sand turbidites, I ~ ; highstand silt/sand, Illn ; coal, I ; transgressive silt/sand, ~ ; erosion or missing section, ~ ; pelagic chalk, ~ ; allochthonous chalk debris, ~,m ; graphic correlation data terrace, I ; sequence boundary (sb), ~ ; incisive s b , " ' N J ; transgressive surface,- . . . . ; lowstand prograding, ~ (sand-silt/mud); R/TF cycle regressive phase, 9R/TF aggradational phase, ~ ; R/TF transgressive phase, , f f ~ 9

I

Ag( )-

i~1 st

uence ana Pormau(

I-'UDIisnea ~equence Prame~

uequence ~ii

|

~1~ ,.F J-,,----R/TF for CNS Central North Sea ,,u~,su,acei___._~., ~r eN-~

!il

0 Key B ~ . Darnel C h r o n o s t r a t i g r a p h i c (;;hart " R-'zI ~- = LAD (firstdownhole) S h e l f > Basin 9 -- = ac.me oc.curre~,ce (Moray Firth).. (Central Graben, . J~ ~ - - - - ~ : m l c r o x o S S l l $ ~oum viKinq ~raoen] IZ~J. palynomorphs = - .

n..~ ~" :~eq.

C N S Seq. F r a m e w o r k of S h e l l Mod. C N S Seq. of M o b i l Modified from f r o m Den H a r t o g J a g e r e t al., (1993) ~ ~J ou'~cro~)- - = F~ ~J . . . . Seq. in outcrop o~ ly Armentrout o ~J> 200 m. In Airy backstripping, basement subsidence is estimated at a set of points in time (see Sclater & Christie 1980) by: (a) sequentially removing layers from top to bottom; (b) decompacting remaining

2D STRATIGRAPHICMODELLING IN THE OMF BASIN units; (c) using palaeobathymetric estimates to set the depth to the top of the remaining units; (d) subtracting the loading effects of remaining sediments using Airy isostasy. This procedure generates a curve whose form reflects the basement subsidence through time. Each estimate of basement subsidence at each point in time should carry errors arising from uncertainty in the palaeobathymetry and from the decompaction process. The consequence of using inaccurate estimates of palaeobathymetry in Airy backstripping is illustrated in Fig. 2. Basement subsidence history (Fig. 2a) is usually estimated by fitting a curve through the basement subsidence determinations (e.g. Sclater & Christie 1980). Figure 2a shows a typical subsidence history with an apparently complex history of basement subsidence and uplift. Furthermore,

(a)

~, ,'~, r

.'2.

Base subsidence .1_ determination with error bars

r./3

Time BP (Ma)

: ',T | ~

[ ,/ .L / / /

~,1 -J, ~"" ]" ~ _L"~

Apparent basement | subsidence history .L.

,--

"~ "~ = r.~

(d)

Time BP ( M a ) - -

while some palaeobathymetry estimates carry realistic estimates of bathymetry (solid error bars in Fig. 2a), many Airy backstripping studies use values in which the error in palaeobathymetry is underestimated (dashed error bars in Fig. 2a). It is important that errors in palaeobathymetry are not underestimated, otherwise a meaningless subsidence history is produced by the Airy backstripping process. An alternative approach to estimating basement subsidence is to use only palaeobathymetry estimates with very small errors (Fig. 2b), such as may be obtained from the presence of coals, erosion surfaces and carbonate reefs. These wellconstrained estimates of palaeobathymetry may then be used to calibrate the post-rift subsidence as predicted by the McKenzie (1978) model of

(b)

Time BP (Ma)

45

,

I

Calibrated McKenzie post-dft thermal subsidence

Time BP (Ma)

iii

i

Z i I

i

Departure from McKenzie post-rift thermal subsidence

Fig. 2. Schematic plots of basement subsidence obtained from Airy backstripping. Error bars for individual basement subsidence determinations are shown. Solid lines represent realistic estimates of errors, while dotted lines represent underestimates. (a) Fitting a trend curve (dashed line) through basement subsidence determinations gives an apparently complex history of basement subsidence and uplift. (b) An alternative approach is to use the most accurate palaeobathymetric determinations to calibrate the McKenzie post-rift thermal subsidence model (solid exponential curve) giving a [~-stretching factor. (e) Many of the basement subsidence determinations with realistic errors also fit with the calibrated McKenzie subsidence curve. (d) Those that do not may be used to identify departures (dotted line) from McKenzie post-rift thermal subsidence.

46

P. A. NADIN t~ N. J. KUSZNIR

extensional basin formation and subsidence; calibration of the McKenzie (1978) model taking the form of an estimate of the 13-stretching factor. This approach was used by Bertram & Milton (1989) and Barr (1991). Fundamental to this approach is the assumption that post-rift subsidence between times of reliable palaeobathymetric estimates is due to McKenzie post-rift lithosphere cooling. For some points in time estimates of basement subsidence based on accurate palaeobathymetry may conflict with the calibrated McKenzie post-rift subsidence trend (Fig. 2b). If this occurs, assumptions must be made as to which of the reliable basement subsidence points are due to McKenzie post-rift subsidence alone, and which are influenced by some additional uplift or subsidence process, or further rifting events. Assuming that McKenzie (1978) post-rift subsidence may be adequately calibrated, the basement subsidence estimates based on realistic palaeobathymetry errors should be compatible with the calibrated McKenzie post-rift subsidence curve (Fig. 2c). If palaeobathymetry estimates with realistic low errors do not lie on the calibrated McKenzie subsidence curve, then this implies anomalous subsidence (Fig. 2d). (2) Airy backstripping ignores the flexural strength of the lithosphere; the effective elastic thickness of the lithosphere, Te, for Airy isostasy is zero. The assumption that an Airy response can approximate a flexural response with low effective elastic thickness (Te < 5 km) is invalidated where isostatic loads have a short wavelength component. For example, the subsidence on fault-block highs where many wells are located is influenced by sediment loading in the deeper adjacent halfgrabens and vice versa. This is especially important in the Outer Moray Firth Basin where half-grabens are typically 10-30 km wide, i.e. the Witch Ground and Buchan half-grabens (Fig. 4a), and where the bathymetry at the end of rifting was of the order of 1000 m in the graben axis (Bertram & Milton 1989; Barr 1991). Backstripping stratigraphic wells without allowing for the effects of lateral loading (i.e. Airy isostasy) overestimates the post-rift load due to the cooling lithosphere, and hence overestimates the l-stretching factor (Roberts et al. 1993; Kusznir et al. 1995).

Reverse post-rift stratigraphic modelling: flexural backstripping, decompaction and reverse thermal subsidence modelling In order to overcome the limitations of Airy backstripping, reverse post-rift modelling (Fig. 3) consisting of flexural isostatic backstripping, decompaction and reverse thermal subsidence

modelling has been carried out. The technique has been applied to geological depth sections to produce a sequence of restored sections with time, whose predicted palaeobathymetry and/or emergence can be tested against the observations of well-constrained palaeobathymetric markers such as sea-level erosion surfaces and coals. A full description of the reverse modelling technique can be found in Roberts et al. (1993) and Kusznir et al. (1995). The starting point of the model is a depth-converted present-day crosssection. The procedure involves reverse modelling (i.e. backwards in time) the time-dependent postrift thermal subsidence (defined by 13in 2D) of the extensional sedimentary basin (Fig. 3). Lithosphere cooling, sediment loading and compaction are tracked backwards in time to the end of the syn-rift succession, i.e. the base of the post-rift. Thermal subsidence is defined by a 2D [3-stretching factor according to McKenzie (1978). Isostatic response to thermal, sediment and water-loads are calculated using flexural isostasy. 2D reverse post-rift modelling produces a series of stratigraphic restorations with time, showing predicted palaeobathymetry and/or emergence (Fig. 3). The history of palaeobathymetry and emergence are strongly controlled by the magnitude of 13 used to define the thermal subsidence component of the reverse model. The [3-factor may therefore be calibrated using reverse post-rift modelling constrained by palaeobathymetric estimates. 2D forward syn-rift and post-rift modelling, using the flexural cantilever model of continental lithosphere extension, has also been carried out to constrain estimates of [3-stretching factors and lithosphere flexural strength used in the reverse post-rift modelling.

Quantitative 2D analysis of post-rift subsidence in the Outer Moray Firth Basin using reverse post-rift modelling A regional geological cross-section based on seismic reflection and well data across the Outer Moray Firth Basin has been investigated using reverse post-rift modelling. The profile, shown in Fig. 4a, runs N-S across the Outer Moray Firth from the East Shetland Platform to the Peterhead Ridge and crosses the Halibut Horst, Witch Ground Graben and Buchan Graben which became prominent structures during Mesozoic rifting.

Reverse post-rift stratigraphic modelling (flexural bac kstrippin g ) The present-day depth cross-section (Fig. 4a) has been reverse modelled to the base of the post-rift

2D STRATIGRAPHIC MODELLING IN THE OMF BASIN

47

Fig. 3. A schematic illustration of 2D reverse post-rift modelling consisting of flexural backstripping, decompaction and reverse post-rift thermal subsidence modelling. Starting with a present-day depth section, sediment layers are successively removed down to the base of the post-rift succession. A sequence of restored sections is produced which can be tested against observed palaeobathymetric markers (adapted from Roberts et al. 1993).

succession, i.e. the end of Late Jurassic rifting at 157 Ma BP. First-order eustatic sea-level variations from Haq e t al. (1987) have been incorporated in the modelling (Fig. 40. An important test and constraint of the reverse modelling process is that stratigraphic surfaces once at sea level, i.e. eroded fault-block crests, wave cut platforms and coal bearing sequences, should be restored to sea level at the appropriate point in time. Trial estimates of a constant Late Jurassic [3-stretching factor were used to drive the reverse thermal subsidence model. In addition a 15 of 1.10 was used to describe post-rift thermal subsidence from the earlier Triassic rift event (at 250 Ma he) which also makes a minor contribution to Cretaceous and Tertiary post-rift thermal subsidence. A value of effective elastic thickness, Te = 4 km derived from forward

modelling, was used to describe the flexural strength of the lithosphere in the reverse model. A model using a constant 15= 1.12 for Late Jurassic rifting across the profile was found to restore end-of-rift erosion surfaces across the Shetland Platform margin and Halibut Horst back to or close to sea level at the base of the post-rift sequence (Fig. 4b). A lower value of [5 generated too much end-Jurassic bathymetry, while a greater value of J3 generated too much emergence. The reverse model with a 13 of 1.12 may also be used to interpolate McKenzie (1978) post-rift subsidence and produce restored cross-sections within the early Paleogene. The restored section at the end of the Paleocene (55 Ma) is shown in Fig. 4c. The top Paleocene horizon across this section contains a shallow water coal-bearing

Fig. 4. Reverse post-rift model of the Outer Moray Firth with eustatic sea-level variations but no Paleocene uplift. (a) Present-day depth section (original section from Andrews et al. 1990). (b) Restored cross-section to the endJurassic, modelled using a constant p-factor profile of 1.12. (c) Restored cross-section to the end-Paleocene, modelled using a constant ~-profile of 1.12. (d) Restored cross-section to the end-Jurassic, modelled using a variable [3-factor generated from the preferred forward syn-rift model (Fig. 7). (e) Restored cross-section to the end-Paleocene, modelled using the variable B-factor with ~av = 1.14. (f) First-order eustatic sea-level variations (Haq et al. 1987) used in the reverse model.

2D STRATIGRAPHICMODELLING IN THE OMF BASIN succession (Fig. la), corresponding to the Beauly Member of the Dornoch Formation (Knox & Holloway 1992). However, the restored Paleocene section produced by reverse modelling shows a substantial palaeobathymetry of c. 375 m (Fig. 4c). This restoration is therefore not consistent with the simple 2D McKenzie (1978) post-rift thermal subsidence curve. The discrepancy in water depth at the end of the Paleocene is interpreted as 375 m of additional regional uplift relative to the firstorder McKenzie (1978) post-rift subsidence curve. Increasing the Late Jurassic ~ to 4.0 correctly restored the end-Paleocene coal horizons to sea level, however a 13 of 4.0 (300% extension) is not supported by structural data and gave too much syn-rift thermal uplift and emergence at the end of the Jurassic. 2D forward syn-rift and post-rift modelling, using the flexural cantilever model of continental lithosphere extension, was carried out to constrain estimates of ]3-stretching factors and lithosphere flexural strength. The forward syn-rift model gave a profile of laterally varying 13 (Fig. 7b). This laterally varying ]3-profile, with a ~av-- 1.14, was also used in the reverse post-rift modelling. The reverse post-rift model, with laterally varying ]3, satisfactory restores eroded syn-rift surfaces to sea level (Fig. 4d). The restoration to the endPaleocene (Fig. 4e) using the variable [3-profile is little different from the constant [3-model (Fig. 4c), since both 13estimates are similar (1.12 cf. 1.14).

Reverse post-rift modelling using a regional uplift in the early Paleogene The discrepancy in the water depth of 375 m at the end of the Paleocene in the constant-]3 model (Fig. 4c) implies 375 m of regional Paleocene uplift across the Outer Moray Firth with respect to McKenzie (1978) post-rift thermal subsidence with = 1.12. The present-day section has been reverse post-rift modelled using ~ = 1.12 and the Haq sealevel curve, but also with an additional regional Paleocene uplift of 375 m (Fig. 5). The restored section to the end-Paleocene, as well as to the endJurassic, is in agreement with the observed palaeobathymetric indicators. This reverse post-rift model, with regional Paleocene uplift, may also be used to produce a restoration to the end of the Cretaceous (65 million years). The restored crosssection to the end-Cretaceous (Fig. 5b) predicts substantial palaeobathymetry of 500-800 m. The post-rift subsidence history for two locations in the Outer Moray Firth Basin is shown in Fig. 5d & e, for a reverse model incorporating the Haq sea-level curve augmented by regional Paleocene uplift. The reverse post-rift model (Fig. 5b, d & e) demon-

49

strates that the assumption of little or zero palaeobathymetry for the Late Cretaceous, frequently made in Airy backstripping studies, in the central North Sea is in error.

Reverse post-rift modelling with respect to an end-Paleocene datum The Beauly Member coals at the end of the Paleocene provide a very reliable estimate of zero palaeobathymetry for the Outer Moray Firth profile. The effect of post-Paleocene global eustasy and regional uplift/subsidence can be removed from the reverse modelling subsidence analysis by restoring the section to end-Paleocene time. In order to achieve this, Eocene and younger stratigraphic units must be removed, lower units decompacted and the end-Paleocene surface set to zero bathymetry. The resulting restored section to the end-Paleocene is shown in Fig. 6a. Flexural backstripping and reverse thermal subsidence modelling from the flattened end-Paleocene datum has then be continued back to the end of the Jurassic using the preferred constant ]3-profile of 1.12 and the Haq et al. (1987) sea-level curve (Fig. 4f). However, in this reverse model (Fig. 6b), the eroded fault-block of Halibut Horst and stratigraphy across the West Fladen High are restored too high at the end-Jurassic, with over 300 m of predicted palaeoemergence (Fig. 6b, e & f). Backstripping using a [3 of 1.0 was required to restore the eroded fault-block of Halibut Horst back to sea level (Fig. 6c). However, a [3 of 1.0 implies no stretching, which is clearly inconsistent with the structural evidence. Furthermore, if the section was backstripped from the present-day using a ~ of 1.0, i.e. no syn-rift thermal uplift and no post-rift thermal subsidence, the eroded crest of Halibut Horst was restored 400 m below sea-level (Fig. 6d & g). This is also an unsatisfactory restoration and the present-day stratigraphy clearly requires some post-rift thermal subsidence during the Cretaceous and Tertiary, as shown in Figs 4 & 5. The results of reverse modelling from a zero bathymetry at the end-Paleocene datum show that net basement subsidence at the end-Paleocene with respect to the end-Jurassic was small or zero. A similar result has also been reported by Joy (1992). Substantial extension clearly took place in the Late Jurassic, and should have been followed by post-rift thermal subsidence in the Cretaceous and Tertiary according to the McKenzie (1978) rift basin model. The evidence for deep-water conditions at the end of the Cretaceous (Hancock 1984; Lovell 1986; Bertram & Milton 1989) suggest that this thermal subsidence did take place. The most reasonable conclusion to be drawn from

50

19. A. NADIN ~ N. J. KUSZNIR

2D STRATIGRAPHICMODELLING IN THE OMF BASIN this is that McKenzie post-rift thermal subsidence took place through the Cretaceous and Paleocene, but was offset by regional Paleocene uplift. An alternative suggestion that post-rift thermal subsidence was delayed until after the Cretaceous (Joy 1992) is not compatible with the physics of the McKenzie (1978) model.

Subsidence analysis of the Outer Moray Firth Basin using structural and stratigraphic forward modelling Forward syn-rifl and post-tift structural and stratigraphic modelling using the 2D flexural cantilever model of continental lithosphere extension (Kusznir et al. 1991; Kusznir & Ziegler 1992) has been applied to the Outer Moray Firth profile. In this model extensional deformation is accommodated by extension along planar faults in the brittle upper crust and an equal amount of distributed plastic deformation in the lower crust and mantle. The flexural isostatic response to extensional faulting in the upper crust produces footwall uplift and hanging-wall subsidence. The distribution of plastic deformation in the lower crust and mantle is defined by a 2D pure-shear l-stretching factor. The flexural cantilever model takes into account the geothermal perturbation generated by lithosphere extension (as a function of ~ in 2D) and subsequent post-rift thermal subsidence. Isostatic loads generated by crustal thinning, geothermal perturbation and re-equilibration are compensated flexurally. The flexural isostatic consequences of instantaneous compacted sediment loading as well as erosional unloading are also incorporated in the model.

Best-fit forward syn-rift model Fault throws and positions observed on the seismic reflection profile across the Outer Moray Firth for Late Jurassic extension were input into the flexural cantilever model. The forward syn-rift model (Fig. 7) was allowed to thermally subside by 18 million years to the end of the Jurassic, to enable comparison with, and constraint by, the reverse modelled section to the base Cretaceous

51

(Fig. 4b). Reverse modelled stratigraphy in Fig. 7c is represented by [+]. The best-fit forward syn-rift model and resulting l-factor profile are shown in Fig. 7a-c. The dip of the rotated fault-blocks predicted by the flexural cantilever model is particularly sensitive to Te and gave a value of Te = 4 km. The l-profile produced by forward syn-rift modelling (Fig. 7b) was also used to refine the t-estimate used in reverse post-rift modelling.

Best-fit forward post-rift model The forward syn-rift model (Fig. 7a-c) includes the geothermal perturbation generated during extension and the subsequent post-rift cooling using a 2D adaptation of McKenzie (1978), defined by a ~-stretching factor profile. Post-rift thermal subsidence arising from this l-profile, and resulting stratigraphy, has been forward modelled for the Outer Moray Firth profile from the end-Jurassic to the present-day. The best-fit present-day stratigraphy obtained using forward modelling is shown in Fig. 7d. First-order eustatic sea-level variations from Haq et al. (1987) were applied (Fig. 8a). The supply of sediment was iterated to provide a bestfit with observed present-day stratigraphic thicknesses, taking into account the effects of sediment loading and compaction, and the observation of zero palaeobathymetry at the end-Paleocene, indicated by the Beauly Member coals (Fig. la). Forward modelling of observed early Paleogene thicknesses across the profile required 390 m of regional uplift in the Paleocene, followed by sediment infill to sea level at the end-Paleocene, in order to be consistent with the zero bathymetry implied by the Beauly Member coals. This regional uplift is equivalent to a regional relative sea-level fall of 390 m (Fig. 8a; line 2). In order to generate the accommodation space for the observed thickness of Eocene sediments (c. 100 m in the north to c. 250 m in the south; Andrews et al. 1990), the forward model required a decrease in the regional uplift of the Paleocene by 160 m in the Eocene, equivalent to a regional 160 m relative sea-level rise (Fig. 8a; line 2). Through the remainder of the Oligocene-Recent the Paleocene regional uplift was decreased to zero at the present-day (Fig. 8a; line 2), and the model was infilled to the presentday basin bathymetry of c. 130 m. The forward

Fig. 5. Reverse post-rift model of the Outer Moray Firth including Paleocene uplift and eustatic sea-level variations. (a) Section reverse modelled to end of the Paleocene using a constant ~ of 1.12 and incorporating an additional 375 m of Paleocene uplift (water-loaded). (b) Section reverse modelled to end-Cretaceous using a constant [3 of 1.12. (e) Section reverse modelled to end-Jurassic using a constant [~of 1.12.(d) Subsidence history for a location in the Dutch Bank Basin. (e) Subsidence history for a location on the Halibut Horst. (f) Global eustatic and regional sealevel variations used for this model.

52

P. A. NADIN & N. J. KUSZNIR

Fig, 6. Reverse post-tift model of the Outer Moray Firth with respect to an end-Paleocene datum. (a) Decompacted and flattened section at the end of the Paleocene, consistent with the presence of end-Paleocene coals. (b) Section reverse modelled from the end-Paleocene to the end-Jurassic using a constant ~1of 1.12. (c) Section reverse modelled from the end-Paleocene to the end-Jurassic using a constant ~ of 1.00. (d) Section reverse modelled from the presentday to the end-Jurassic using a constant ~ of 1.00. (e) Subsidence history for (b) for a location on the West Fladen High. (f) Subsidence history for (b) for a location on the Halibut Horst. (g) Subsidence history for (d) for a location on the Halibut Horst. Reverse models incorporate first-order eustatic sea-level variations (Fig. 4f) but no Paleocene uplift.

2D STRATIGRAPHICMODELLING IN THE OMF BASIN post-rift model to the present-day (Fig. 7d) compares well to the observed present-day stratigraphic thicknesses (Fig. 7d; dotted lines).

53

Magnitude of Eocene subsidence (Fig. 7i) The combined thickness of Eocene-Recent stratigraphy which is observed (Fig. 7i) can be forward

Eustatic sea-level, sediment supply and loading (Fig. 7e & f) The forward model may be used to explore the consequences of changing sea level and sediment supply on stratigraphy. The model shown in Fig. 7e incorporated first-order eustatic sea-level variations (Fig. 8b; line 1) and assumed sediment infilling to sea level at all Cretaceous-Tertiary post-rift times. The stratigraphic thicknesses generated (Fig. 7e) showed little resemblance to those observed; the Cretaceous was too thick and the Tertiary was absent. This was expected since the Outer Moray Firth Basin has a semi-starved Cretaceous post-rift succession (Bertram & Milton 1989; Andrews et al. 1990). In the model shown in Fig. 7f, sediment supply through the Cretaceous was iterated to provide a best-fit with observed present-day stratigraphic thicknesses, taking into account the effects of sediment loading and compaction. The eustatic sea level variation is shown in Fig. 8b. The model was then infilled to sea-level at the end-Paleocene to be consistent with the Beauly Member coals. This model without Paleocene uplift generated too much Paleocene and no Eocene-Recent (Fig. 713. Post-rift thermal subsidence, eustasy and sediment supply alone are not able to fully account for observed early Tertiary stratigraphic thicknesses and observed zero end-Paleocene palaeobathymetries in the central North Sea Basin without additional regional uplift in the early Paleocene followed by rapid regional subsidence in the early Eocene.

Magnitude of Paleocene uplift (Fig. 7g & h) The forward post-rift model of present-day stratigraphy is very sensitive to the magnitude of Paleocene uplift. The model shown in Fig. 7g incorporated 550 m of regional Paleocene uplift (Fig. 8b; line 2) and generated insufficient Paleocene stratigraphic thicknesses and too much Eocene-Recent stratigraphy. The effect of reducing Paleocene uplift to 250 m (Fig.8b; line 3) is shown in Fig. 7h, it generated too much Paleocene and no Eocene-Recent. As discussed earlier in this section, the correct thickness of Paleocene sediments can be modelled by incorporating 390 m of regional Paleocene uplift.

modelled assuming a linear transgression from the end-Paleocene through to the present day (Fig. 8b; line 4). However, the linear transgression gives relatively slow regional subsidence through the Eocene (Fig. 8b; line 4), even when combined with post-rift thermal subsidence, and gives little accommodation space for Eocene sedimentation. Much less Eocene sediment (40-75 m) was generated in this model than is observed, while the Oligocene-Recent was too thick (Fig. 7i). A rapid reduction of uplift is required in the early Eocene of 160 m (Fig. 8a; line 2) in order to generate the observed thickness of Eocene (Fig. 7d).

Subsidence history plots Figure 9 shows the computed subsidence histories of the forward post-rift models for a point located in the Witch Ground Graben. The subsidence history shows the development of stratigraphic thicknesses (including the effects of sediment loading and compaction) and palaeobathymetry. The subsidence history for the preferred post-rift subsidence model is shown in Fig. 9a. This model predicts substantial Cretaceous palaeobathymetry followed by rapid shallowing during the Paleocene then deepening through the Eocene. Regional Paleocene uplift and Eocene subsidence is superimposed on post Jurassic-rift thermal subsidence. This 'kick' in early Paleogene basement subsidence is consistent with that reported for the Outer Moray Firth Basin by previous workers (Bertram & Milton 1989; Milton et al. 1990; Jones & Milton 1994). The subsidence histories shown in Fig. 9b-f of the discounted post-rift models (Fig. 7e-i) are not consistent with the observed present-day stratigraphic thicknesses.

Errors in the analysis of CretaceousTertiary post-rift subsidence Errors in the analysis of the post-rift subsidence arise through errors in interpreting and depth convetting stratigraphy from seismic data and from parameters used to define the forward and reverse modelling processes: compaction/decompaction, effective elastic thickness (lithosphere flexural strength) and the [3-stretching factor.

~stretching factor. Reverse post-riff modelling is dependent on the [3-factor profile used to define post-rift thermal subsidence. A [3 of 1.0 implies no extension and therefore no thermal subsidence,

54

F'. A. NADIN & N. J. KUSZNIR

2D STRATIGRAPHIC MODELLING IN THE OMF BASIN while increasing [3 increases extension and thermal subsidence. The preferred constant ~-factor profile across the Outer Moray Firth profile had a value of 1.12. The error on this relatively small amount of extension (12%) is estimated to be _+0.05, which gives an error in the estimation of regional Paleocene uplift of + 25 m.

Effective elastic thickness. Sensitivity tests using different values of Te have shown that determination of 13 by reverse modelling is relatively insensitive to variations in Te within the range 1-25 kin. The preferred Te constrained by forward syn-rift modelling was 4 km. The tilt of faultblocks calculated using forward syn-rift modelling (the flexural cantilever model) is particularly sensitive to Te. When a Te of 0 km (Airy isostasy) is used in backstripping, large errors can arise when isostatic loads have a short wavelength component, i.e. in the early post-rift phase. If Airy isostasy is used in reverse post-rift modelling, a [3 of 1.35 is required to restore the eroded fault-block crest of Halibut Horst back to sea-level. Such a large [3-factor can not be supported by structural data and forward syn-rift modelling. Sensitivity tests have shown that the estimate of Paleocene uplift using a very large Te of 25 km (360 m uplift) is more similar to results using a Te of 4 km (375 m uplift) than those obtained using a Te of 0 km (290 m uplift). Sediment compaction/decompaction. Forward and reverse post-rift modelling are also dependent on the parameter used to define sediment compaction (compaction length, surface porosity and matrix density). However, for the range of feasible values for each compaction parameter, tests have shown, neglecting overpressuring, that the estimates of end-Paleocene palaeobathymetry from reverse modelling was relatively insensitive to the variations in these parameters. Sensitivity tests for compaction parameters for the Outer Moray Firth profile gave small changes in predicted 13; however, the estimates for end-Paleocene palaeobathymetry

55

and inferred Paleocene uplift were relatively stable (+_ 35 m).

Triassic extension. While most of the Cretaceous and Tertiary post-rift thermal subsidence which must be reverse modelled is derived from Late Jurassic rifting, a small component of the Cretaceous and Tertiary post-rift thermal sub sidence was inherited from earlier Triassic rifting at c. 250 Ma. The line of section of profile 1 crosses an area where the Triassic sequence is thin and it is therefore assumed that the profile was relatively unaffected by Triassic rifting. Tests have shown that using a Triassic [3 of 1.00 (i.e. ignoring Triassic rifting), gave a 25% increase in the estimated Late Jurassic [3 from reverse post-rift modelling. Using a Triassic [3 of 1.50 gave a 40% decrease in Late Jurassic [3. However, for variation in Triassic [3 the palaeobathymetrical estimate for the end-Paleocene was stable and showed little change (_ 10 m). The 375 m of regional Paleocene uplift estimated by reverse post-rift modelling is significant compared with the observed thickness of the Cretaceous and Tertiary and cannot be attributed merely to errors in the parameterization of the reverse model, discussed above. The error on the end-Paleocene uplift is estimated to be c. + 60 m. The best estimate, from reverse post-rift modelling, of the magnitude of regional Paleocene uplift across the Outer Moray Firth profile is therefore 375 m +_60 m (water-loaded).

Discussion Long-term global eustasy Changes in sea level, superimposed on the combined effect of infill of accommodation space created by rifting, subsidence from lithosphere cooling, sediment loading and compaction, controlled the development of accommodation space and distribution of sediments in the Outer Moray Firth Basin. It is important to incorporate an

Fig. 7. Forward syn-rift and post-rift stratigraphic models of the Outer Moray Firth using the flexural cantilever model of rift basin formation. All models include long-term global sea level from Haq et al. (1987). (a) Late Jurassic syn-rift crustal structure before erosion. (b) Variable p-profile (solid) and crustal thinning factor profile after erosion (dashed) generated by forward model; ~av = 1.14. (c) End-Jurassic basin geometry after erosion and 18 million years early post-rift sediment infill; ++++, represents backstripped stratigraphy. (d) Preferred forward post-rift model following 390 m of regional Paleocene uplift and 160 m Eocene subsidence superimposed on global sea-level variation; . . . . represents present-day stratigraphy. (e) Model with no Paleocene uplift infilled to sea level at all CretaceousTertiary post-rift times gives too much Cretaceous and no Tertiary. (f) Model with no Paleocene uplift infilled to sea level at end-Paleocene gives too much Paleocene and no Eocene-Recent. (g) As (f) with 550 m Paleocene uplift, gives insufficient Paleocene. (h) As (f) with 250 m Paleocene uplift, gives too much Paleocene. (i) As (f) with 390 m Paleocene uplift and 50 m of Eocene subsidence gives observed Paleocene but insufficient Eocene. See Fig. 8 for sea-level variations applied to models (d)-(i).

(a)

D

0

0

mr-

m c

~o

o~i C~ 6o

f-

m

,,

m

~.~ o~

~

~ 6o

m

oo

o~o

o~

o

~

m

;;0 .-
~ > ~1 o z

g

u.:2

o

ELSE] t~

Lu

~ =z,,, _uJuJ/ -r.j~(n

(/) '1"

i

. . . . . . . . . . . . "Ll. btl.'.'.'," ". . . . . . . I' . ' . ' . 11' . . I. I ' . ' I . . . , . .. . , . ,. .I . 9. . ,

r

1 f I 1 ~ 9 .] F.. i p~l~,tl~ J,llJ* I I,i I/'," I~'"" III I Jll'~JJ]>>:>> ~ I' . ~, J , ' ' - ' l.P l.l i l.' . . . . 9 .I ' . .~ . ~ [. ' . . ' . ' .I ~. ~. ~. .' .~ .' . ' ' ' ' ' ' ' ' ] >

>

>1 >

I


.

I I I I [ ~

"0

I~ ,

. . . . . . . . . . . . . . . . - ..'.'.'.','.','." *

I I I I I I I I I I { I I rl i ill II IlJl [ J IIII

",'.'.'."

~,

.~

I I~> > > >1 I II> >1

0

0 Z

~-

u~o

9 '.'

.'.'

.'.','

.'.'

.1' |~ I II ' - ' - / q "

9'

.......... I ............

:J:J

l:t :1

~ r.,T-1 9>- .~

9 -,--, 0

82

A.M. JOY

dated as the top of the Lower Eocene.) (2) The midEocene high-gamma zone; this distinctive zone is typically 10-40 m thick. The top of the zone is dated as basal to intra-mid Eocene; it is taken to mark the boundary between sequences I and II. It correlates with the H2-H3 boundary of Knox & Holloway (1992), separating their informal Caran Sandstone (H2) and Brodie Sandstone (H3) units. This is also usually a strong seismic reflector, except in areas of closely spaced intraformational normal faulting. (3) The top of the earliest Eocene Balder Formation, a distinctive gamma ray and sonic log marker with an almost invariably strong, unfaulted seismic response. This is taken as the base of sequence I (so Balder Formation sands are not included in sequence I). Where the Balder Formation is absent (i.e. in the extreme west of the study area) the base of sequence I is taken to be the top of the Beauly Formation (see Mudge & Copestake 1992; Knox & Holloway 1992). These three horizons appear to fulfil the criteria for well-log and seismic markers mentioned above, and they may be equated with widespread stratigraphic markers identified in detailed recent studies (e.g. Knox & Holloway 1992). It is considered that no other Eocene marker is as widespread or as distinctive as these; however, over smaller areas it may be possible to subdivide the section more finely using more locally developed well-log and seismic markers backed up by biostratigraphic data.

Distribution and sandstone content of Eocene sequences This subject is addressed using regional maps of seismic isopach and of net sandstone thickness for the two Eocene sequences defined above. Reference is also made to a series of representative schematic well logs illustrating the Eocene stratigraphy of the study area (Fig. 2).

Sequence I isopach in time (Fig. 3) This sequence is thickest in two discrete areas: in the Viking Graben in central Quadrant 9 and on the western edge of the Western Central Graben in Quadrant 29. These two depocentres are both mudstone dominated, but whereas the southern depocentre comprises exclusively mudstone, the section in the northern depocentre has an important sandstone component in its lower part (see Fig. 4). Sequence I sediments in the northern depocentre were supplied from northern Britain; the southern depocentre is part of an arcuate belt of mudstone depocentres, mainly off the mapped area, which occur around the southern margin of the Central Graben (Joy 1993) and which may have been

Fig. 3. Time isopach of sequence I. Contours in ms twt (as a first-order approximation, 100 ms twt is c. 100 m).

sourced by a palaeo-Rhine drainage system from continental Europe. Sequence I thins markedly by onlap, and locally appears to pinch out completely, against the upper Paleocene Moray Delta. This delta therefore appears to have formed a palaeobathymetric high during the early to mid Eocene. The basal onlap relationships at the base of sequence I indicate that this sequence infilled pre-existing bathymetry. Over the basinal areas (i.e. Central Graben, South Viking Graben), away from concentrated sediment supply, sequence I is consistently thin [between 100 and 200 ms two-way time (twt)], reflecting its relatively distal, sandstone-poor nature. The sequence is particularly thin (< 100 ms twt) on the Jaeren High and over parts of the

EOCENE SEDIMENTATION IN THE CENTRAL NORTH SEA BASIN

Western Platform. This suggests that these areas, which were the shoulders of the Mesozoic rift basin, still formed positive features during the early to mid Eocene (Joy 1993).

Sequence I net sandstone isopach (Fig. 4) Sequence I sands include the Frigg Formation sandstones in Quadrant 9 (Deegan & Scull 1977) and the majority of the Tay Formation sandstones in Quadrants 21 and 22. However, they do not include the Gryphon reservoir sandstones and the lower Tay Formation sandstones which belong to the Balder Formation. During sequence I times the main sand supply route to the basinal Frigg Formation was situated in the Quadrant 9 area; this appears from its seismic character to have been a delta. A number of some-

83

what smaller sand fairways in the south of Quadrant 21 supplied sand to the Tay Formation. Two, possibly three, sand supply routes can be distinguished; these may have been associated with a delta visible on seismic data to the west. There is a prominent NW-SE component to the sand isopachs; this reflects deflection of sand-bearing gravity flows by newly rejuvenated fault-controlled structures. This is linked to a widespread, if mild, episode of tectonism of earliest Eocene age, possibly associated with the opening of the Norwegian-Greenland Sea. Sandstones are present throughout sequence I in the central part of the mapped area, but sand-supply routes cannot be mapped with confidence. A relatively thin early Eocene delta formed in Quadrant 14 and western Quadrant 15, but while this may have supplied sand during early sequence I time it cannot be correlated with later sequence I sandstones. The upper part of sequence I in Quadrants 14 and 15 appears to be a thin, condensed mudstone (Figs 2 & 3).

Sequence H isopach in time (Fig. 5) Sequence II is generally 50-200 ms twt thick. Typically, it is less prone to short wavelength fluctuations than sequence I. This is because sequence I was deposited on a rugose bathymetric surface which reflected the positions of the Moray Delta, of newly rejuvenated fault structures and of the underlying Mesozoic rift structures (e.g. the South Viking Graben; Fig. 3). By the end of sequence I time much of this short-wavelength bathymetry had been infilled, and sequence II was deposited upon a much smoother surface. The main fluctuations in sequence II thickness are associated with a major delta centred on the northern Witch Ground Graben and the western Fladen Ground Spur. Within this feature the time isopachs reflect the positions of the two long-lived southeastward-directed sand-supply fairways shown on Fig. 6 (see below).

Sequence H net sandstone isopach (Fig. 6)

Fig. 4. Sequence I net sand isopach. Contours in m.

Sequence II sands include the Nauchlan Member of the Alba Formation (Newton & Flanagan 1993). A major delta developed at this time in the western part of the study area, and two persistent sandsupply routes may be distinguished on the basis of the sandstone isopach. One of these fed a series of gravity flows across the Fladen Ground Spur and into an area centred upon central Quadrant 16, while the other fed the Alba sandstones in southern Quadrant 16. The edge of the major sequence II delta protruded into the basin in the centre of Quadrant 15, marking the progradation of the sand

84

A . M . JOY

Fig. 5. Time isopach of sequence II. Contours in ms twt (as a first-order approximation, 100 ms twt is c. 100 m).

supply fairways; this lobate morphology is typical of fluvially dominated deltas (Elliot 1986). The position of the main fluvial input to the depositional system does not appear to have been closely tectonically controlled. There are thick sequence II sands in the Viking Graben in Quadrant 9 and northern Quadrant 16; these appear to be gravity-flow sandstones. However, because of a scarcity of well data it is not clear how these are related to the body of sequence II progradational clinoforms identified on seismic data to the west. There are thin sequence II sands in southwestern Quadrant 21; from seismic and well data these appear to be gravity-flow sandstones that were deposited on or at the base of the palaeoslope. They do not extend on to the basin floor.

Fig. 6. Sequence II net sand isopach. Contours in m.

Patterns of Eocene sedimentation in the central North Sea Basin This subject is addressed using a series of schematic cross-sections across the northern, central and southern parts of the study area.

Northern area (Fig. 7) During sequence I time gravity-flow sandstones were supplied to the South Viking Graben from a series of deltas to the west. The morphology of the basin floor was strongly controlled by buried Mesozoic rift structures, which may have been rejuvenated during earliest Eocene tectonism. Consequently, the sequence I sandstone isopach is closely related to the positions of the Mesozoic structures.

EOCENE SEDIMENTATION IN THE CENTRAL NORTH SEA BASIN

Fig. 7. Schematic cross-section: northern area. Length of section c. 50 km. No vertical scale implied.

85

86

A.M. JOY

It is believed that the location of the sequence I delta in this area was controlled by the position of the upper Paleocene Moray Delta. There was probably relatively shallow water above the Moray Delta throughout sequence I time, and consequently a large sequence I delta could not develop in the central part of the study area. The main fluvial sand supplies were therefore diverted to the northern [and to the southern (see below)] parts of the study area. No sand was supplied to the basin during the later part of sequence I time; instead a thick sequence of mainly middle Eocene mudstones accumulated. The distribution of these mudstones was strongly controlled by the position of the South Viking Graben, which they partially infilled and the flanks of which they clearly onlap (Fig. 3). Sequence II is dominated by sandstone in the northern area. In numerous wells gravity-flow sandstones with 'box-car' log motifs are overlain by a thick deltaic coarsening-upward sequence, indicating that the delta was prograding over deeper-water sediments during sequence II time. However, the South Viking Graben was a pro-

nounced bathymetric feature at this time and the sequence II delta was unable to completely prograde across this area of deep water.

Southern area (Fig. 8) This area lay to the south of the Moray Delta at the end of the Paleocene, and during sequence I time a major mud-dominated delta developed to the south in Quadrants 28 and 29. It is argued that sand supply to the basin was confined between the bathymetric high formed by the Moray Delta to the north and the developing bathymetric high of the mud-dominated delta to the south. This effect was most pronounced during sequence I time; it may have persisted into sequence II time but by then very little sand was being supplied to this part of the basin. Sequence I sands were supplied from the west; some sand was ponded against newly rejuvenated NW-SE structures while some was able to reach the Central Graben in southern Quadrant 22. Seismic evidence suggests that the bathymetric surface had been smoothed by the end of sequence I

Fig. 8. Schematic cross-section: southern area. Length of section c. 70 km. No vertical scale implied.

EOCENE SEDIMENTATION IN THE CENTRAL NORTH SEA BASIN time. The distribution of the scarce sequence II sands in this area appears to have been controlled by the base of the palaeoslope.

Central area (Fig. 9) Deposition in this area during sequence I time was dominated by the presence of the Moray Delta. The position of this delta was originally controlled by the Moray Firth Basin; sand supplies were concentrated into this structure, which formed a major palaeobathymetric low at the end of the Cretaceous (Joy 1993). However, by the end of the Paleocene the Moray Delta formed a positive palaeobathymetric feature with somewhat deeper water areas both to the north, on the East Shetland Platform/Fladen Ground Spur, and to the south, on the Western Platform. As a result, fluvial sand supplies were diverted both north and south of the Moray Delta during sequence I time, and deltas

87

of this age built up in response to this supply of sediment. However, overlying the Moray Delta itself, sequence I consists of a thin condensed mudstone (Fig. 1), and eastward-thickening sequence I mudstones onlap the front of the Moray Delta (Fig. 9). The sequence I sandstones in the central part of the study area are isolated from their marginal equivalents by a sand bypass zone (Fig. 4), which in turn was controlled by the slope of the Moray Delta. Though sand-supply routes cannot be mapped with confidence in this area, the sand was probably supplied from the west and northwest. The gravity flows from which the sand was deposited appear to have been extremely sensitive to bathymetry. Sequence I sediments infilled a significant portion of the water column in the central part of the study area. The ultimate effect of this reduction in palaeo-water depths was to allow the sequence II

Fig. 9. Schematic cross-section: Central area. Length of section c. 100 km. No vertical scale implied. (a) End Paleocene. Maximum basinward extent of Moray Delta. (b) Early to Middle Eocene. A large delta cannot develop due to limited accommodation space. Sequence I is consequently mudstone-dominated. Thin sequence I gravity flow sands supplied from out of plane section. (c) Middle to late Eocene. Subsidence causes relative sea level rise, creating space for a sequence II delta. Delta progrades over 'platform' created by sequence I.

88

A.M. JOY

delta to prograde some 30 km basinwards of its Late Paleocene predecessor. Sand supply to the basin was dominated during sequence II time by this central area delta; the sand bypass zone was less well developed in this area than during sequence I time, and the sand-supply routes can generally be followed from the delta top to the base of the palaeoslope (Fig. 6).

Controls on Eocene sand deposition in the central North Sea Basin This study raises some interesting points about the importance of the various controls on Eocene sand supply and deposition in the central North Sea Basin. The main conclusion is that the pattern of sand supply was dominated by the locations of the deltas, and that the positions of the deltas were critically dependent upon accommodation space. During the late Paleocene, the position of the Moray Delta was controlled by the accommodation space provided by the Moray Firth Basin. By the start of sequence I time this accommodation space had been completely filled and fluvial sand supply was diverted around the northern and southern edges of the Moray Delta to areas where accommodation space was still available. By the start of sequence II time, a rise in relative sea level had created accommodation space over the Moray Delta, and this space was exploited as the main deltaic depocentre returned, broadly speaking, to its earlier location in the central part of the study area. The inferred rise in relative sea level may also have had the effect of reducing sand supply to the basin in the southern part (and to a certain extent in the northern part) of the study area. It seems likely that the rise in relative sea level referred to above was primarily tectonic rather than eustatic in origin. This is because: (1) it allowed the accumulation of a sequence II deltaic depocentre several hundred metres thick, and this is too great for a second-order eustatic fluctuation (Haq et al. 1987); and (2) according to the seismic data, the sequence II delta top was not eroded during the subsequent sea-level fall which would be expected in a eustatic cycle. Basin modelling studies have shown that the Eocene was a period of unusually rapid tectonic subsidence in the central North Sea Basin (Joy 1993). Palaeobathymetry also controlled the degree to which the deltas were able to prograde basinward. In the northern and the southern areas the deltas appear, from seismic-stratigraphic evidence (Milton et al. 1990; Joy 1993), to have been confined to the basin margins by deep water, which was in turn controlled by the structure of the Mesozoic rifts. The pattern of delta development in these areas was therefore dominated by aggrada-

tion. However, replacement of a significant proportion of the water column by sequence I sediment in the central area allowed the sequence II delta to subsequently prograde a significant distance (tens of kilometres) basinward of earlier deltas. On a smaller scale the distribution of gravity-flow sandstones was strongly controlled by fault-controlled palaeobathymetric features, such as the Crawford Ridge in the north of the study area and the L NW-SE structures on the Western Platform in the~ south. Since sand supply to deep-water depositional settings is sometimes used to infer relative sea-level change, attribution of a eustatic cause to stratigraphic phenomena in the central North Sea Basin might yield conflicting results. A study of southern Quadrant 21, for instance, might conclude that relative sea level was low during sequence I time, resulting in erosion in the marginal areas and sand supply to the basin, and that relatively high sea levels during sequence II time had resulted in the reduction of sand supply to the basin. By contrast, a study of southern Qua&ant 16 might reach the opposite conclusion with respect to relative sea level, since sequence II sandstones are more voluminous than sequence I sandstones. Needless to say, both interpretations of apparent sea level change cannot be attributed to eustasy. In fact it is difficult to discern any eustatic signature on this scale of investigation, since: (1) the pattern of sand supply through time and space was so varied; and (2) sand supply appears to have been so strongly controlled by delta location. Even during deposition of the mid-Eocene high-gamma zone, a widespread condensed sequence, which might be attributed to a eustatic highstand, sand supply continued uninterrupted in some parts of the South Viking Graben (Figs 2, 4 & 6).

Distribution o f hydrocarbon shows In the northern part of the study area shows and discoveries are abundant in sequence I sandstones while sequence II sandstones are almost invariably water wet. This indicates that the former are sealed by the thick upper sequence I mudstones. In the central area, by contrast, shows are almost exclusively confined to sequence II sandstones. Clearly, effective migration pathways must cross sequence I (Fig. 3), and it is not clear why no significant hydrocarbons have been discovered in this part of the section. It is hard to believe that this is exclusively due to infelicitous selection of well locations. It may be that the top seal to the sequence I sands has locally been breached by the numerous closely spaced intraformational normal faults which seismic data has revealed cutting the Eocene in this area (Higgs & McClay 1993; Joy 1993).

EOCENE SEDIMENTATION IN THE CENTRAL NORTH SEA BASIN

o

o

e-~

>..,

,..., ~

9

.o, ~

09

,.~

0"~ ~.~ ~ ~

m~ 9

,.~'~

9

=~

.H o

.= " ' ~

.,.~ ~

X4 __~

9

4 s~


"o

ga,

e,i

APECTODINIUM IN THE CENTRAL NORTH SEA

Selandian) are yet to be agreed and they are therefore not certain in well 22/10a-4. The well is located in a basinal setting (Fig. 1) and was recognized as providing a reference section for biostratigraphic, lithostratigraphic and sedimentological studies. The stratigraphy of the section is summarized in Fig. 2. Details of samples from the Lista and Maureen Formations are given in Table 1.

Palynology Apectodinium spp. Species of the genus Apectodinium are typically pentagonal in shape, covered in fine, nonparatabular processes and have a single-paraplate, quadra, intercalary archaeopyle. The genus comprises a plexus of forms with speciation chiefly based on the degree of development of the lateral, apical and antapical horns. In practice, many intermediate forms are found. Apectodinium species including A. augustum, A. homomorphum (Deflandre & Cookson 1955) Lentin & Williams 1977, A. hyperacanthum, A. paniculatum (Costa & Downie 1976) Lentin & Williams 1977, A. parvum (Alberti 1961) Lentin & Williams 1977 and A. quinquelatum (Williams & Downie 1966) Costa & Downie 1979 were found to be common to abundant throughout the Forties Sandstone Member of the Sele Formation and below that in unit S lb and the upper part of unit S I a to a depth of 2623.48 m. The percentage of the dinoflagellate cyst assemblage made up by Apecwdinium spp. is illustrated in Fig. 2. In addition to the expected occurrences of Apectodinium spp. in the Sele Formation, the genus was found to be present at 2658.35 m in the Lista Formation. The species present were A. augustum, A. homomorphum, A. paniculatum and A. parvum in a low-diversity dinoflagellate cyst assemblage. In addition, a large number of specimens transitional between A. augustum and A. paniculatum were encountered. Apectodinium spp. were also found to be present in the Maureen Formation, in basal unit M2 at 2701.68 m and in unit M1 at 2707.10 m.

Apectodinium augustum, A. homomorphum, A. parvum and A. quinquelatum plus some intermediate forms were present, again in lowdiversity assemblages. These results were considered sufficiently unusual to warrant resampling of the Maureen Formation core in much more detail. Apectodinium events were again encountered at similar levels, allowing for differences in sample quality and spacing. In the second sample set Apectodinium spp. were found in unit M2 at 2697.02 m and in unit M1 at 2706.40 m.

] 17

On the basis of current North Sea usage the

Apectodinium spp. events at 2706.40m and 2707.10 m are considered to be of latest Danian age.

Other dinoflagellate cysts Floods of dinoflagellate cysts belonging to the genera Areoligera and Glaphyrocysta, including Areoligera gippingensis Jolley 1992, were encountered at 2644.82, 2652.53 and 2655.48 m (Lista Formation). However, allocation of these strata to the A. gippingensis acme of Harland et al. (1992) cannot be made with confidence due to the generally poor productivity of the samples in question. Other palynological markers include the range base of Isabelidinium ? viborgense HeilmannClausen 1985 at 2701.22 m in unit M2. Since this species was found to be present in the uppermost sample of the Maureen Formation core but not in the lowermost sample of the Lista Formation core its range top is assumed to lie in the uncored interval. The range top of Alisocysta reticulata Damassa 1979 occurs at 2704.80 m in unit M1, although the range top of common or abundant A. reticulata lies at 2711.13 m in unit M1. The uppermost occurrence of this species was chosen as a marker for the top of the early Paleocene (Danian) by Harland et al. (1992). Additionally, HeilmannClausen (1994) stated that the lowermost occurrence of A. reticulata is in the upper Danian and that the uppermost occurrence of the species may prove useful for correlation with other regions although its precise position relative to North Sea zones is still uncertain. Another significant component of the dinoflagellate cyst assemblages in unit M1 is a distinctive gonyaulacacean species, Spiniferites sp. A, referred to as Spiniferites n. sp. by Stewart (1987) and Spiniferites cf. supparus by Powell (1992). Its range top at 2701.45 m lies just above the top of unit M1 in 22/10a-4. A number of authors, including Powell (1992) and Harland et al. (1992), have indicated that the uppermost occurrence of this species could be used to mark the top of the Danian in the North Sea Basin. In general, dinoflagellate cyst assemblages from the Maureen Formation of 22/10a-4 were dominated by Spiniferites spp., Cleistosphaeridium spp., Hystrichosphaeridium spp. and Areoligera spp. Also common was Palaeoperidinium pyrophorum (Ehrenberg 1838) Sarjeant 1967, its range top in Well 22/10a-4 occurring at 2625.00 m (unit Sla). Thalassiphora delicata Williams & Downie 1966 was present to a depth of 2703.13 m. Confirmation that the records of Apectodinium spp. in the Maureen Formation are true stratigraphic records, and not the result of misplaced core, is provided by the occurrence of Alisocysta

118

J . E . THOMAS

Table 1. Details of samples from the Lista and Maureen Formations of Well 22/10a-4 Core depth (ft)

8630.00 8636.00 8640.10 8645.00 8650.00 8655.00 8660.40 8665.00 8670.00 8675.00 8680.30 8685.00 8690.00 8695.75 8699.40 8704.00 8713.80 8723.40 8803.50 8805.00* 8807.00* 8809.00 8815.00" 8816.50 8819.75 8826.25* 8830.00 8837.00* 8839.00* 8840.00* 8840.75* 8841.50 8846.25* 8847.00 8849.50* 8851.75" 8855.00* 8857.00* 8859.00 8862.50* 8866.00 8869.00* 8871.00 8872.50* 8876.00 8879.00* 8881.50 8886.00 8888.00* 8891.00 8896.00

Log depth (ft)

Log depth (m)

Lithostrat unit

Dinocyst diversity (in No. taxa)

Apectodinium (%)

Marine (%)

8650.00 8656.00 8660.10 8665.00 8670.00 8675.00 8680.40 8685.00 8690.00 8695.00 8700.30 8705.00 8710.00 8715.75 8719.40 8724.00 8733.80 8743.40 8823.50 8825.00 8827.00 8829.00 8835.00 8836.50 8839.75 8846.25 8850.00 8857.00 8859.00 8860.00 8860.75 8861.50 8866.25 8867.00 8869.50 8871.75 8875.00 8877.00 8879.00 8882.50 8886.00 8889.00 8891.00 8892.50 8896.00 8899.00 8901.50 8906.00 8908.00 8911.00 8916.00

2637.19 2639.02 2640.27 2641.76 2643.29 2644.82 2646.46 2647.86 2649.39 2650.91 2652.53 2653.96 2655.48 2657.23 2658.35 2659.76 2662.74 2665.67 2690.09 2690.54 2691.15 2691.76 2693.59 2694.05 2695.04 2697.02 2698.17 2700.30 2700.91 2701.22 2701.45 2701.68 2703.13 2703.35 2704.12 2704.80 2705.79 2706.40 2707.01 2708.07 2709.15 2710.06 2710.67 2711.13 2712.20 2713.11 2713.87 2715.24 2715.85 2716.77 2718.29

Lista Lista Lista Lista Lista Lista Lista Lista Llsta Lista Lista Lista Lista Lista Lista Lista Lista Lista M2 M2 M2 M2 M2 M2 M2 M2 M2 M2 M2 M2 M2 M2 M1 M1 M1 M1 M1 M1 M1 M1 MI MI M1 M1 M1 M1 M1 M1 MI M1 M1

24 11 2 Barren Barren 6 21 21 Barren Barren 6 19 8 Barren 9 19 9 2 Barren 30 37 29 34 29 1 10 Barren 37 37 24 38 21 34 25 34 4O 33 23 25 42 Barren 38 27 42 34 44 29 29 40 28 15

0 0 0

16.77 75.42

0 0 0

99.21 70.92 72.08

0 0 0

67.25 84.17 88.23

47.50 0 0 0

70.83 33.22 40.55 90.50

0 0 0 0 0 0 45.50

67.95 59.60 57.34 65.55 65.82

0 0 0 0 11.60 0 0 0 0 0 1.93 2.60 0

55.37 64.95 62.40 98.90 57.57 88.58 91.98 94.95 93.40 94.55 79.50 67.07 92.90

0 0 0 0 0 0 0 0 0 0

96.00 96.25 93.65 97.89 91.84 96.82 93.02 97.00 88.88 91.78

* Indicates resampling of the Maureen Formation core.

52.40

APECTODINIUM IN THE CENTRAL NORTH SEA

reticulata and Spiniferites sp. A in the same assemblage as Apectodinium augustum at 2706.40 rn. Additionally, Palaeoperidinium pyrophorum was found to be present in the same assemblages as A. augustum at 2701.68, 2706.40 and 2707.01 m. The absence of such diagnostic species in the other Apectodinium-bearing samples is ascribed to the low-diversity nature of the assemblages (see below). Additional factors The relative abundance of marine versus nonmarine palynomorphs is illustrated in Fig. 2. and detailed in Table 1. The diversity of the dinoflagellate cyst assemblage (in number of taxa) is also given in Table 1. Generally, the Lista and Maureen Formation Apectodinium spp. events correspond to increases in the relative abundance of terrestrially derived palynomorphs, mostly pollen, and a limited decline in the diversity of the dinoflagellate cyst assemblages accompanied by a reduction in numbers of skolochorate (spinebearing) cysts. Furthermore, the amounts of amorphous organic matter are greater in these samples. Taken together with the restriction of the Apectodinium events to more sandy lithologies, these factors could be interpreted as indicating that the events do not represent in situ assemblages, but basin-margin assemblages that were swept periodically into the basin centre. Indeed, the 'Hystrichosphaera Association' of Downie et al. (1971), consisting of skolochorate cysts, was taken to represent more open marine conditions whilst the 'Wetzeliella Association' was taken to represent an inner neritic environment. Brinkhuis et al. (1994) mentioned an influx of continentallyderived material accompanied by the dominance of marginal marine dinoflagellate cysts at the

119

Danian-Selandian Boundary. The occurrence of a similar association in Well 22/10a-4 points to a possible relationship between basinal occurrences of Apectodinium spp. and periods of lowered sea level in the North Sea Basin. Evidence that Apectodinium spp. occurred in marginal North Sea environments before the major flourishing of the genus in biozone NP9 is provided by the record of Apectodinium spp. in Powell et al. (1993) from the Reculver Silts (NP8) in Kent. Additional data from such marginal deposits is needed to fully establish the distribution and environmental preferences of the genus. Although Brinkhuis et al. (1994) mentioned only the possible influence of climatic warming on extending the geographical distribution of Apectodinium spp., it is evident that other factors, such as sea-level change and palaeogeographical configuration, may have played an equally important role. It is clear from this study that a pattern of facies control is superimposed on the temporal distribution of the genus and that caution should be exercised in using the Base Apectodinium Datum as a time-constant biostratigraphic marker, even within a single basin. For the same reasons, it is not yet possible to say whether the 'early' appearances of Apectodinium spp., now reported from the central North Sea, southern England, France and Tunisia, represent events with the potential for regional correlation or whether they are of local significance only. The author benefited greatly from discussions with R. Harland and R. W. O'B. Knox during the course of this work. The paper was substantially improved by the comments of the referees C. Heilmann-Clausen, D. W. Jolley and J. B. Riding. This paper is published with the approval of the Director, British Geological Survey (NERC).

References ALBERTI, G. 1961. Zur Kenntnis mesozoischer und altterti~irer Dinoflagellaten und Hystrichosphaerideen von Nord- und Mitteldeutschland sowie einigen anderen europaischen Gebieten. Palaeontographica A, 116, 1-58. BRINKHUIS,H., ROMEIN, A. J. T., SMIT, J. & ZACHARIASSE, J.-W. 1994. Danian-Selandian dinoflagellate cysts from lower latitudes with special reference to the E1 Kef section, NW Tunisia. GFF, 116, 46-48. COOKSON, I. C. & EISENACK, A. 1965. Microplankton from the Dartmoor Formation, SW Victoria. Proceedings of the Royal Society of Victoria, 79, 133-137. COSTA, L. I. & DOWNIE, C. 1976. The distribution of the dinoflagellate Wetzeliella in the Palaeogene of north-western Europe. Palaeontology, 19, 591-694. -& 1979. The Wetzeliellaceae; Palaeogene dinoflagellates. Proceedings of the Fourth Inter-

national Palynology Conference, Lucknow (19761977), 2, 34-46. & MANUM, S. B. 1988. The description of the interregional zonation of the Paleogene (D1-D15) and the Miocene (D16-D20). Geologisches Jahrbuch, A100, 321-330. DAMASSA, S. P. 1979. Danian dinoflagellates from the Franciscan Complex, northern California. Palynology, 3, 191-207. DEFLANDRE, G. & COOKSON, I. C. 1955. Fossil microplankton from Australian late Mesozoic and Tertiary sediments. Australian Journal of Marine and Freshwater Research, 6, 242-313. DOWNIE, C., HUSSAIN, M. A. & WILLIAMS, G. L. 1971. Dinoflagellate cyst and acritarch associations in the Paleogene of south-east England. Geoscience and Man, 3, 29-35. EHRENBERG, C. G. 1838. Ober das Massenverhfiltniss --

120

J.E. THOMAS

der jetzt lebenden Kiesel-Infusorien und fiber ein neues Infusorien-Conglomerat als Polirschiefer von Jastraba in Ungarn. Abhandlungen der Preussischen Akademie der Wissenschaften, 1836, 109-135. HARLAND, R. 1979. The Wetzeliella (Apectodinium) homomorphum plexus from the Palaeocene/earliest Eocene of northwest Europe. Proceedings of the Fourth International Palynology Conference, Lucknow (1976-1977), 2, 59-70. , HtNE, N. M. & WILKINSON, 1. P. 1992. Appendix. Paleogene biostratigraphic markers. In: KNOX, R. W. O'B. & HOLLOWAY, S. Paleogene of the Central and Northern North Sea. In: KNox, R. W. O'B. & CORDEY, W. G. (eds) Lithostratigraphic Nomenclature of the UK North Sea. British Geological Survey, Nottingham, A1-A5. HEILMANN-CLAUSEN,C. 1985. Dinoflagellate stratigraphy of the uppermost Danian to Ypresian in the Viborg 1 borehole, central Jylland, Denmark. Danmarks Geologiske UndersOgelse, A7, 1~59. 1994. Review of Paleocene dinoflagellates from the North Sea region. GFF, 116, 51-53. JAN DU CHI~NE, R., GORIN, G. & VAN STUIJVENBERG,J. 1975. Etude g6ologique et stratigraphique (palynologie et nannoflore calcaire) des Gr~s des Voirons (Paleog~ne de Haute-Savoie, France). Geologie Alpine, 51, 51-78. JOLLEY, D. W. 1992. A new species of the genus Arealigera Lejeune Carpentier from the Late Palaeocene of the eastern British Isles. Tertiary Research, 14, 25-32. KNOX, R. W. O'B. & HOLLOWAY, S. 1992. Paleogene of the Central and Northern North Sea. In: KNox, R. W. O'B. & CORDEY, W. G. (eds) Lithostratigraphic Nomenclature of the UK North Sea. British Geological Survey, Nottingham. LENTIN, J. K. & WILLIAMS, G. L. 1977. Fossil Dinoflagellates: Index to Genera and Species. 1977 Edition. Bedford Institute of Oceanography, Report Series, B1-R-77-8, 1-209. & 1981. Fossil Dinoflagellates: Index to

Genera and Species. 1981 Edition. Bedford Institute of Oceanography, Report Series, B1-R-81-12, 1-345. MARTINI, E. 1971. Standard Tertiary and Quaternary calcareous nannoplankton zonation. In: Proceedings of the H Planktonic Conference, Roma, 1970, Vol. 2, 739-785. POWELL, A. J. 1988. A modified dinoflagellate cyst biozonation for latest Palaeocene and earliest Eocene sediments from the central North Sea. Review of Palaeobotany and Palynology, 56, 327-344. 1992. Dinoflagellate cysts of the Tertiary system. In: POWELL, A. J. (ed.) A Stratigraphic Index of Dinoflagellate Cysts. BMS Publications Series, Chapman & Hall, London, 155-251. --, BRINKHUIS,H. & BUJAK, J. P. 1993. Dinoflagellate cyst sequence biostratigraphy of the Thanetian in the type region. In: Correlation of the Early Paleogene in Northwest Europe, Programme and Abstracts. Geological Society, London. SARJEANT, W. A. S. 1967. The genus Palaeoperidinium Deflandre (Dinophyceae). Grana Palynologica, 7, 243-258. SCHRODER, T. 1992. A palynological zonation for the Palaeocene of the North Sea Basin. Journal of Micropalaeontology, 11, 113-126. SmSSER, W. G., WARD, D. J. & LORD, A. R. 1987. Calcareous nannoplankton biozonation of the Thanetian Stage (Palaeocene) in the type area. Journal of Micropalaeontology, 6, 85-102. STEWART,I. J. 1987. A revised stratigraphic interpretation of the Early Palaeogene of the central North Sea. In: BROOKS, J. & GLENNIE, K. (eds) Petroleum Geology of North West Europe, Volume 1. Graham & Trotman, London, 557-576. WILLIAMS, G. L. • DOWNIE, C. 1966. Wetzeliella from the London Clay. In: DAVEY, R. J., DOWNIE, C., SARJEANT, W. A. S. & WILLIAMS, G. Studies on Mesozoic and Cainozoic dinoflagellate cysts. Bulletin of the British Museum of Natural History, Supplement 3,215-235. -

-

An integrated palynological-palynofacies approach to the zonation of the Paleogene in the Forties-Montrose Ridge area, central North Sea S U S A N E. W O O D & R I C H A R D V. T Y S O N

Newcastle Research Group, Fossil Fuels and Environmental Geochemistry (Postgraduate Institute), D r u m m o n d Building, University o f Newcastle, Newcastle upon Tyne NE1 7RU, UK Abstract: A palynological and palynofacies study of late Paleocene and early Eocene sediments

of the central North Sea has been undertaken using 467 core samples from 11 wells on the Forties-Montrose Ridge. Conventional dinoflagellate cyst (dinocyst) biostratigraphy gives poor precision within the narrow (c. 2.5 million years) time interval studied. Furthermore, dinocysts are often present in only very low numbers and are frequently poorly preserved. An alternative local zonation scheme is proposed on the basis of downwell variations in quantitative palynofacies and palynomorph occurrence data, and on angiosperm pollen ratios. Integration of these three sets of data indicates a potential sevenfold subdivision of the mid-Sele Formation ($2) to upper Lista Formation (L3) interval in this area, with up to five divisions within the Forties Sandstone Member. This composite zonation scheme appears to work reasonably well for the studied wells, and preliminary comparisons with independent biostratigraphic data from two of our wells indicate a partial correlation with the dinocyst zones.

The Paleocene Forties Sandstone Member (FSM) (sensu Knox & Holloway 1992, p. 53-55) is one of the major exploration targets in the North Sea, and has been one of the basin's most prolific oil reservoirs. As has been widely recognized for many years, this sandstone unit was deposited in a submarine fan setting, with the main reservoir units corresponding to lower to middle fan channel sands (e.g. Thomas et al. 1974; Carman & Young 1981; Kulpecz & Van Geuns 1990). Submarine fan sands represent particularly challenging reservoir targets because of the complexity of the vertical and lateral (radial and proximal-distal) facies changes that result from the interactions of palaeogeographic, tectonic, eustatic and sedimentological factors. Successful production strategies therefore require sophisticated depositional models, which in turn require detailed local correlation schemes utilizing as wide a range of stratigraphic tools as possible. The integration of seismic stratigraphy, biostratigraphy and geophysical log correlations has been especially valuable (e.g. Stewart 1987; Milton et al. 1990; Mudge & Copestake 1992). This integrated approach is now almost universal in North Sea Paleogene studies (e.g. Whyatt et al. 1992; Armentrout et al. 1993; D e n Hartog Jager et al. 1993; O'Connor & Walker 1993; Vining et al. 1993). In addition to the general difficulties associated with submarine fan sands, biostratigraphic corre-

lation in the FSM is complicated by a number of other factors. The dark-coloured shale facies (characteristic of the Sele Formation) was deposited under dysoxic-anoxic depositional conditions, resulting in a strongly facies-controlled assemblage of calcareous microfossils (nannoflora and foraminifera) that is both sparse and rather poorly preserved (Stewart 1987). Paleogene palynology also has its problems; the taxonomy of both the plankton (dinocysts) and sporomorphs (spores and pollen) of this interval are complex and incompletely described; most of the available zonation schemes produced are of a local nature, and the relative influence of palaeoecological and evolutionary factors is not yet entirely clear. Consequently, the petroleum industry makes pragmatic use of a whole range of micropalaeontological (M) and palynological (P) 'bioevents' to provide a biostratigraphic framework for the North Sea Paleogene (Stewart 1987, p. 562; Mudge & Copestake 1992, p. 55). A lithostratigraphicbiostratigraphic-seismic sequence correlation is given in Table 1. This framework features only two bioevents which relate specifically to the FSM, and these correspond only to its top and base (P3 and M5 of Mudge & Copestake 1992, p. 56, respectively); it cannot be used in intra-FSM correlations. The poor biostratigraphic subdivision of the FSM undoubtedly relates to the short time period

From Knox, R. W. O'B., Corfield, R. M. & Dunay, R. E. (eds), 1996, Correlation of the Earl), Paleogene in Northwest Europe, Geological Society Special Publication No. 101, pp. 121-128.

121

122

S.E. WOOD • R. V. TYSON

during which this 200 m thick unit was deposited. The estimated duration of only c. 0.8-1.0 million years is too short for major evolutionary changes in either plankton or terrestrial floras, and therefore a small number of conventional biozones is inevitable. Other approaches, such as ecostratigraphy event stratigraphy, and palynofacies are therefore essential if greater resolution is to be achieved. Non-palaeontological approaches have also been applied. The FSM (S 1) can sometimes be locally subdivided into two units based on gammaray log signatures (Knox & Holloway 1992; O'Connor & Walker 1993): a lower and often more argillaceous unit (Sla) is separated from an upper unit (Slb) by a gamma-ray spike, especially in more distal sections. These gamma spikes relate to thin mudstones which are thought to reflect basinwide condensation events (Milton et al. 1990; Mudge & Copestake 1992, p. 68), and thus to be approximately isochronous, although the potential for local condensation due to sediment bypass cannot be ruled out. The commercial importance and value of Paleogene palynological zonation schemes mean that most of the intensive studies that have been undertaken within the petroleum industry are as yet unpublished. Some aspects of the palynological zonation schemes of Simon-Robertson, Mobil, BGS-Unocal and Esso have recently been very

briefly mentioned in papers in the Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference (Armentrout et al. 1993, p. 53; Morton et al. 1993, p. 75-76; Vining et al. 1993, p. 20). However, the only detailed publications available are those of Powell (1988) and Schrtider (1992). The dinocyst zonation scheme of Powell (1988) shows that the 'Forties Formation' corresponds to the Apectodinium augustum biozone. Three dinocyst biozones (Aau, Ahy and Ama) are shown for the 'Forties interval' in a table by Powell (1992, p. 15), but no details are given. The Area (Alisocysta margarita) biozone is usually placed in the top of the Lista Formation (L3), below the FSM (e.g. Knox & Holloway 1992, p. 30), so only the first two of Powelrs (1992) biozones appear to apply to the FSM. The most detailed published palynological scheme, that of Schr6der (1992), provides a fourfold palynological subdivision of the 'Forties Formation' (PT 19.1-PT 19.4, where PT stands for palynological time). However, the M5 bioevent of Mudge & Copestake (1992), which defines the base of the FSM, is correlated with the middle part of zone PT 19.2 (Schr6der 1992, p. 124), so only the upper two and a h a l l out of the four PT zones, relate to the FSM. Schr6der's subdivision of the FSM is based on one dinocyst-related boundary which defines the base of zone PT 19.2 (the base

Table 1. Lithostratigraphic, seismic and biostratigraphic subdivision of the Paleocene-Early Eocene (Thanetian-

Ypresian pars) of the central North Sea Basin Knox & Holloway 1992 Lithostratigraphy Sand Shale

Stewart 1987 Code

Sequence

B2 & B 1

9

Seismic System tracts

Mudge & Copestake 1992

SchriSder 1992

Bioevent

PT zone

Basin oxygen

21

Anoxic

20

Anoxic

19.4 19.2-

Anoxic?

M7 Balder

LST M6

Sele

$3 & $2

8

HST P3

Forties

Sele

S1

7

TST 4-- LST M5

Lista

L3 (pars)

6

HST

Oxic M4 & P2

Balm'l

Lista

L3 & L2

5/4

TST

19.1 15

Oxic

15

Oxic

13

Oxic

P1 Andrew

Lista

L1

3

LST M3

Maureen

M2 & M1

2

HST 0.5 million years) separating the two units can be calculated from the magnetostratigraphicdata; C25n is missing from the stratigraphic record across almost all of southern England. However, the Upnor Formation in central London contains a record of C25n. A hiatus of c. 0.4 million years separates the late Paleocene Lambeth Group from the early Eocene Harwich Formation. The Lambeth Group, Harwich Formation and lower London Clay Formation were all deposited during C24r. The start of Chron C24n.3n is positioned at the base of Division B of the London Clay Formation.

The majority of the internationally recognized Paleogene stages were originally defined in outcrops in the North Sea Basin. Each of the stages was related to a lithostratigraphic unit; however, the base of many of these units corresponds to an unconformity. Stage boundaries have been extended across the globe by relating the base of each of the stratotypes to the standard biostratigraphic zonation framework. In recent times, the increased precision required by stratigraphers has led to demands for boundary stratotypes that more easily facilitate global correlations. Thus it is likely that many, if not all, of the Paleogene stage boundaries will eventually be 'relocated" in sections outside the North Sea Basin. For example, the early Paleocene Danian Stage, originally defined by Desor (1847) in an outcrop at Stevns Klint, Denmark, is today located in an outcrop at E1 Kef, Tunisia (Jenkins & Luterbacher 1992). Minimizing the age discrepancy between the level originally used to define the start of a stage and the point chosen in an alternative section to mark the start of the newly defined standard stage will ensure that the stage remains a valuable chronostratigraphic concept. The late Paleocene Thanetian Stage (defined in Kent, England) and the early Eocene Ypresian Stage (defined in Belgium) bracket the PaleoceneEocene boundary (Hardenbol & Berggren 1979; Jenkins & Luterbacher 1992). Unfortunately, critical portions of the uppermost Thanetian and lowermost Ypresian stratotypes lack key calcareous

microlossils and nannofossils, and linking these intervals to the global marine record is problematic (see discussion in Ali et al. 1993). Nannofossil zone NP9 is poorly represented and NP10 is not known. Dinoflagellate cysts have been used to provide a second-order correlation with upper NP9 and NP 10 in the stage stratotype areas. Magnetostratigraphic studies potentially provide useful datums for linking the Paleocene-Eocene boundary to the North Sea Basin sequences, in particular the levels corresponding to the start and end of Chron C24r. Thus, if the boundary is eventually located in a section outside NW Europe, palaeomagnetic investigations of the Thanetian and Ypresian stratotypes will provide valuable reference markers. Magnetostratigraphic studies of the Thanetian and Ypresian stratotypes also provide a chronostratigraphic framework from which it is possible to assess the timing and duration of the upper Paleocene and lower Eocene depositional sequences of the southern North Sea Basin. One of the principal research programmes of the Southampton University Palaeomagnetic Group is to determine the synchroneity of late Paleocene and early Eocene depositional sequences from a number of passive margins around the globe. The object of this programme is to provide a rigorous test of the eustatic sea-level model for a 12-15 million years window of geological time. The present study marks the completion of the investigations for the upper Paleocene of the southern

From Knox, R. W. O'B., Corfield, R. M. & Dunay, R. E. (eds), 1996, Correlationof the Early Paleogene in NorthwestEurope, Geological Society Special Publication No. 101, pp. 129-144.

129

130

J.R. ALl • D. W. JOLLEY

North Sea Basin. (Studies of the lower Eocene from the region have recently been presented by Ali et al. 1993.)

Depositional setting During the late Paleocene and early Eocene the North Sea extended across much of SE England and NW mainland Europe (see Ziegler 1982). As the sea level fluctuated, the shoreline periodically advanced and retreated across the region. The upper Paleocene and lowermost Eocene sediments preserved in southern England (Fig. 1) comprise outer neritic, inner neritic, marginal-marine and fluviatile sediments.

The upper Paleocene and lower Eocene of southern England Many of the lithostratigraphic units referred to in this paper were introduced last century (e.g. Prestwich 1850, 1852, 1854; Whitaker 1872). Cooper (1976), Curry et al. (1978) and King (1981) have revised and formalized some of these terms. A major review of the stratigraphic nomenclature of the upper Paleocene and lower Eocene of southern England (summarized in Fig. 2) has recently been carried out by Ellison et al. (1994). O r m e s b y Clay F o r m a t i o n

The oldest Tertiary sediments in southern England (Chron C26r) are assigned to the Ormesby Clay

Fig. 1. Outcrop and subcrop of the Paleogene deposits of SE England, northern France and Belgium, and the location of boreholes and outcrops referred to in this study.

THANETIAN/YPRESIAN MAGNETOSTRATIGRAPHY, SE ENGLAND

131

The formation is present in a triangular area between NE Kent, south Greater London and south Suffolk. The formation attains over 30 m thickness in north Kent, thinning to a few metres in south Suffolk, where it passes laterally into the Ormesby Clay Formation (middle and upper part). The Thanet Formation spans nannoplankton zones NP6-NP8 and magnetochrons C26n to C25r (Knox et al. 1994). Curry (1981) suggested that a significant part of the Thanet Formation was removed by erosion prior to deposition of the Woolwich & Reading Beds (Lambeth Group). L a m b e t h Group: U p n o r F o r m a t i o n

The Lambeth Group comprises three formations. The lowermost unit, the Upnor Formation, was previously referred to as the Woolwich Bottom Bed (Ellison et al. 1994). The Upnor Formation sediments were deposited in a shallow marine environment. The unit is typically 2-5 m thick unit and consists of fine- to medium-grained sands and silts with occasional pebbles and is rich in glauconite. It is present across the London Basin and the eastern part of the Hampshire Basin. Calcareous nannoplankton obtained from the Upnor Formation in both the Hampshire Basin (Siesser et al. 1988) and the London Basin (Ellison et al. 1996) indicate an NP9 age. Fig. 2. Summary of the upper Paleocene and lower Eocene stratigraphy of the southern England (based on Eilison et al. 1994 and Jolley 1996). Correlation with the Martini (1971) nannoplankton zonation-schemeis based on Aubry et al. (1986), Siesser et al. (1988), Knox et al. (1994) and Ellison et al. (1996).

Formation (Cox et al. 1985; Knox et al. 1990). The formation is restricted to eastern East Anglia where it is known only from boreholes. It consists of glauconitic mudstone. Sporadic, thin, ash beds occur in the lower part of the formation. The unit is over 25 m thick in east Norfolk, thinning to c. 10 m in Suffolk. A red-brown mudstone in the lower middle part of the formation in Norfolk constitutes a useful regional marker. It can be traced across Suffolk and southwards into Essex and Kent, where it is present as a sandstone just above the base of the Thanet Sand Formation (Knox et al. 1994). The upper part of the Ormesby Clay Formation is assigned to Chron C25r. Thanet S a n d F o r m a t i o n

The Thanet Sand Formation consists of bioturbated silts and fine sandstones which are rich in glauconite. The sediments were deposited in an inner-shelf to coastal setting (Ellison et al. 1994).

L a m b e t h Group: W o o l w i c h a n d Reading formations

The laterally equivalent Woolwich and Reading formations were deposited over a large part of southern England. The Woolwich Formation occurs in the eastern half of the London Basin and consists of grey and grey-brown interlaminated sands, silts and clays. The formation is rich in plant debris, and includes thin lignites. Shell beds, up to 10-30 cm, occur. The formation is typically 10-15 m thick and represents a variety of marginal marine, low to high energy environments, with occasional freshwater intercalations (Ellison et al. 1994). The Reading Formation (fluviatile facies) comprises red mottled clays with thin sandy units. In the western London Basin and the Hampshire Basin the formation is typically 25-40 m thick. The Woolwich and Reading Formations interdigitate in the London area. The Woolwich and Reading Formations are also represented in north France, where they are referred to as the Sparnacian. T h a m e s Group: H a r w i e h F o r m a t i o n

Ellison et al. (1994) include in the Harwich Formation all the sediments between the top of

132

J . R . ALI & D. W. JOLLEY

the Lambeth Group and the base of the London Clay Formation Walton Member. The dominant lithologies of the Harwich Formation are glauconitic sandy clays and glauconitic fine sands. Tephras, up to 4 cm thick, are common in the upper part of the formation in north Essex and Suffolk. Elsewhere, disseminated ash occurs sporadically throughout the formation. The Harwich Formation was deposited in a shallow-shelf environment. The formation is typically 15-20 m thick, although it attains over 40 m in eastern Norfolk. In parts of Greater London the Harwich Formation is absent. Correlation of the Harwich Formation tephras with those in Goban Spur DSDP sites provides a second order link to the early Eocene NP10 nannoplankton zone (Knox 1984). The unit is assigned to the uppermost part of the Apectodinium hyperacanthum dinoflagellate zone, which Powell (1992) considers to be of early Eocene age.

Thames Group: London Clay Formation (basal part) Ellison et al. (1994) positioned the base of the London Clay Formation at the base of the Walton Member. The transgression surface marking the base of this member has been identified over the whole of the London Basin and the eastern part of the Hampshire Basin (King 1981). The sediment consists of silty clays; no ash layers have been recorded in this unit. In the eastern London Basin the Walton Member is typically c. 15 m thick, thinning westwards and southwards to c. 5 m at Whitecliff Bay, Isle of Wight. The first appearance datum (FAD) of the dinoflagellate Wetzeliella astra is positioned at c. 2 m above the base of the Walton Member at the type locality (King 1981). The Walton Member is of early Eocene age (?NP10).

Palynomorph association sequences Interpretation of the palynological data used in correlation for this study followed the method outlined by Jolley (1992, 1996), involving the recognition of palynofloral associations. Associations, i.e. groups of assemblages with shared characteristics of composition, were defined by empirical inspection of the data, with the aid of cluster analysis and detrend correspondence analysis. This enabled identification of groups of successive assemblages which showed shared characteristics of composition. In some cases, lower density sampling resulted in assemblages from single samples being referred to as associations, where the preceding and succeeding assemblages showed a significant difference in composition. Associations were defined individually for each section, without reference to other

studied sections. Each association was subsequently assigned a name based on the specific epithets of two of its characteristic taxa. In correlating the individual sections, comparison of associations was applied only between nearest neighbours which provided the most complete record of the palynofloras. This method of comparison between adjacent sections was utilized to minimize the effects of any biofacies control on the palynofloral assemblages. Neighbouring sections contain palynofloras representative of similar biofacies, whereas those from a distant section may contain biofacies which would make reliable correlation difficult. From the comparisons of the associations it is possible to identify laterally continuous, stratigraphically comparable sequences of associations within the study area. These are termed 'association sequences', and are regarded as containing palynofloras deposited in sediments of equivalent age. Each association sequence identified here is numbered (oldest to youngest), the number being prefixed by the letter T (Thanetian) and Y (Ypresian). In all, nine T association sequences and seven subsequences were identified (Jolley 1992), together with nine Y association sequences and ten association subsequences (Jolley 1996).

Palaeomagnetic methods The samples used in this study were collected between 1987 and 1993. Most of the analysed specimens are either 8 or 14 cm 3 cubes, although the recent acquisition of a whole-core cryogenic magnetometer has allowed measurements to be carried out on core pieces of 150-300 cm 3. Samples were collected from each section with a stratigraphic spacing of 0.3-1.0 m. As most of the sections are from boreholes, the specimens could not be oriented with respect to north, and were 'way-up' oriented only. Approximately half of the specimens were measured using a Molspin spinner magnetometer. The alternating field (a.f.) method of demagnetization was applied to specimens at steps of 5 mT rising to peak values of 30-60 mT using a Molspin tumbling-specimen system. The remaining specimens were examined using a 2G Enterprises cryogenic magnetometer, which has an in-line threeaxes demagnetization system capable of generating fields of 60 roT. Both Zijderveld (1967) plots and equal-area stereographic plots were used to determine the stability of remanence of each specimen in order to define the characteristic direction. Examples of response to a.f. demagnetization are shown in Fig. 3. The data are categorized in the manner described by Ali et al. (1993). A stable end-point direction (SEPD) is defined when a high

13 3

THANETIAN/YPRESIAN MAGNETOSTRATIGRAPHY, SE ENGLAND

(a) BRADWELL: 217

j/jol~

0t 0

(b)

F

BD4

J l0

,

F

SAMPLENO Jo (mA/m) BD4 15.73

~

~

BRADWELL: 217

,

J

r

'

F

20 DcmagFicld(naT)

'

30

;

"

40

BD42

SAMPLENO Jo (naA/m) BD42 18.40

j/jol? ~ . . ~ . . . . X . _ . ~ _ X _ _ _ ~ X 0I 0 10 20 DernagField(mT) ,

,

~

J

,

X ~

30

X-----X J__~ 40

Fig. 3. Examples of responses of typical samples to alternating field demagnetization. (a) Stable end-point (SEP) behaviour; (b) directional trend but with no SEP. In each case the magnetic vector after each demagnetization step is plotted on a stereographic projection (declination is arbitrary because the specimens are from unoriented borehole cores). On the stereographic projections solid symbols represent positive inclinations (plotted in the lower hemisphere) and open symbols represent negative inclinations (plotted in the upper hemisphere). The values in the range 0-40 represent the applied field treatment (in mT). Also shown for each specimen is a plot of the normalized magnetic intensity (J/Jo) v. applied field (mT).

stability component of magnetization is isolated in a specimen. Specimens which do not achieve an SEPD, but for which a reliable polarity determination can be made based on the trajectory of the remanence vector towards a particular polarity state, are referred to as 'trending'. Specimens which exhibit wide fluctuations in both direction and/or intensity between steps are classified as 'erratic'. Such behaviour is most commonly observed in very weakly magnetized sediments where the magnetic signal may be of comparable magnitude to the noise level of the magnetometer. On the magnetostratigraphic log for each section SEPD, trending and erratic behaviour are indicated. As none of the borehole cores are oriented, the polarity assignment for each specimen is based on the dip of the characteristic remanence inclination angle (downward directed indicates normal polarity, whereas upward dipping indicates reverse polarity). Isothermal remanent magnetization (IRM)

analyses were performed on representative specimens in order to determine the principal remanence carriers in each of the boreholes sampled. A Molspin pulse magnetizer, with a peak direct field of 0.86 T was used to generate the IRM. The IRM was measured between each of the 14 progressive field increments using a Molspin spinner magnetometer. Two distinct types of behaviour were observed. In the first case (e.g. Fig. 4a & b) the magnetization saturates in an applied field of c. 0.3 T, suggesting that magnetite is present in the specimen. In the second case (e.g. Fig. 4c & d), the magnetization does not saturate at the peak field, indicating that the remanence is carried by iron oxide in a higher oxidation state. The 'IRM ratio', defined by Ali (1989) as the IRM at 0.3 T as a ratio of the IRM at the peak field, is a convenient method for quantifying the shape of the IRM curve. Magnetite-rich specimens have IRM ratios typically >0.9 whereas for hematite-bearing sediments the value is typically 0.6-0.9.

134

J.R. A L I & D . W. JOLLEY

4000

2000

0

,

0.0

0.2

i, 0.4

a.

b.

JU47

SZ32

(0.92)

(0.97)

i r i 0.6 0.8

0

0.0

i, 0.2

i i i, 0.4 0.6

C.

400

V 0

i 0.8

d.

(0.76)

V

I'

I'

I'

I'

I

0.0

0.2

0.4

0.6

0.8

APPLIED FIELD (T)

o

_ (0.70) .

l,

i,

i

0.0

0.2

0.4

, i 0.6

~ i 0.8

APPLIED FIELD (T)

Fig. 4. Examples of IRM acquisition curves for magnetite-bearing (a & b) and hematite-bearing (c & d) sediments. IRM is expressed in mA m2; the figure in parenthesis is the specimen's IRM ratio.

Palaeomagnetic results Ormesby, Norfolk

The stratigraphy of the Paleogene succession in the British Geological Survey (BGS) Ormesby borehole (TG 5148 1424) was reported on by Cox et al. (1985). In ascending stratigraphic order, the following units (names from Ellison et al. 1994 and Jolley 1995) were cored: Ormesby Clay Formation (27.45m), Hales Clay (Harwich Formation) (14.40m) and Wrabness Member (Harwich Formation) (27.80m). The palaeomagnetic interpretations published in Cox et al. (1985) were based on 94 8 cm 3 palaeomagnetic specimens collected from the Ormesby core. Following Cox et al. (1985), Townsend & Hailwood (1985) and Aubry et al. (1986), Knox et al. (1990) were able to provide a fuller interpretation of the Ormesby borehole magnetostratigraphic sequence. Their principal findings were: (1) the Ormesby Clay Formation spans Chrons C26r to C25r; (2) the unconformity separating the Ormesby Clay from the Hales Clay is marked by the absence of Chron C25n, indicating a stratigraphic break of > 0.5 million years; (3) the Hales Clay and Wrabness Member were deposited during the early part of Chron C24r. In this study, a further 19 specimens were collected from the Ormesby core in order to refine the magnetostratigraphy presented by Cox et al. (1985). Sixteen samples were collected from the Wrabness Member, one sample from the Hales Clay and two from the Ormesby Clay Formation. Twelve samples from the upper part of the Wrabness Member had initial natural remnant magnetization (NRM) intensities of < 0.6 m A m -1.

Samples from below c. 79 m below datum (m.b.d.) had values of 20-40 mA m -t. Ten specimens demagnetized to a stable end-point direction, with eight specimens exhibiting well-defined trend directions. The magnetostratigraphy of the Ormesby section is shown in Fig. 5 (modified from Cox et al. 1985, fig. 5). The revised polarity sequence includes a number of thin (c. 1 m) normal polarity intervals, based on one or two samples. These had previously been identified in a study by Johnston (1983), but were omitted from the Cox et al. (1985) synthesis. The normal polarity intervals identified in the Ormesby core have been relabelled OR-A to OR-H, in ascending stratigraphic order. However, while some of the thin normal polarity intervals reported by Johnston (1983) have been confirmed, several are somewhat thinner than previously thought. Johnston also identified a short reversed interval in the upper part of magnetozone OR-B (correlated with Chron C26n). This has now been confirmed by data from two samples. Hales, Norfolk

A multidisciplinary stratigraphic investigation of the BGS Hales borehole (TM 367 969) was carried out Knox et at. (1990). Here we present the details of the palaeomagnetic investigations. The Hales borehole recovered 45.77 m of Paleogene sediments. The following rock units (names from Ellison et al. 1994 and Jolley 1996) were present: Ormesby Clay Formation (25.59 m), Hales Clay (Harwich Formation) (15.62m), and Wrabness Member (Harwich Formation) (4.96 m). One hundred and eleven 8 cm 3 specimens were

135

THANETIAN/YPRESIAN MAGNETOSTRATIGRAPHY, SE ENGLAND FM. PAS POL. 70--

Y9

-

I

OR-H OR-G

Y8

co "-" 09 Y7c LM Y7b' Z

80--

133

,OR-F 9 OR-E

Y7a

90--

Y6b i Y6a

T

100 -

d

E

!y_~4

OR-D

O Y3c

G)

O

110--

._J _ I Y3b Y3a T9

~

,,= 120

OR-C

reverse polarity specimens within each formation record similar inclination values; (4) the Ormesby Clay Formation and the Hales Clay have quite different mean inclination angles (45.8 and 29.3 ~ respectively). The palaeogeographic maps of Smith et al. (1981) predict a southern UK late Paleocene palaeolatitude of 42 ~ N, which equates to a geomagnetic field inclination angle of 61 ~ followed by a steady northward drift to its present-day site (c. 51.5 ~ N). The mean inclination of the Hales Clay is approximately half that predicted by Smith et al. (1981). Assuming that the SEPD represent primary magnetizations (the Hales Clay and Ormesby Clay both include normal and reverse polarity specimens which suggests this is the case), and that geomagnetic secular variation has

T8

15 T7 i

f

OR-B >-

a-b T!

130 -

T3

T~I

UNIT PAS POL.

OR-A

I

-90

i

'

0

~

-r"

Y3a I

90

Fig. 5. Magnetostratigraphy of the upper Paleocene and lower Eocene at Ormesby, Norfolk. S, T and E denote stable end-point, trending or erratic directions, respectively. Data from Johnston (1983) are marked +. Palynomorph association sequences (PAS) from Jolley (1992, 1996). Normal polarity, black; reverse polarity, white. Normal polarity intervals are coded OR-AOR-H (ascending).

,.-.., vE 7I-O. UJ 0

HC1

Tx

OC4

T9 ~"

O

HL-E

HL-D

T7e

oo~ ~o~

~

~• .J 0

u~EO

us

~

0

"~

~

o~o~o

u9EO

EL I:L Z

G. Z

Z

13_ Z

Z

I

Z

~,~

""~ 0 ~

~

M

THANETIAN]YPRESIAN MAGNETOSTRATIGRAPHY, SE ENGLAND Chron C26r

The oldest upper Paleocene sediments in eastern England occur at subcrop in the Ormesby and Hales boreholes, Norfolk. There the lower part of the Ormesby Clay is correlated with Chron C26r. The base of the Ormesby Clay Formation at both Ormesby and Hales yields broadly similar ages; palynomoph association sequences T1 and T2, respectively (Jolley 1992 and unpublisfied data). The difference in thickness of C26r (cf. Figs 5 & 6) sediments indicates relative condensation at Hales with respect to the same interval at Ormesby. Assuming a fixed accumulation rate during deposition of the Ormesby Clay at Hales, we postulate that the base of the Ormesby Clay in Norfolk may be c. 0.65 million years older than the oldest part of the Thanet Sand Formation (deposited during Chron C26n). Chron C26n

In the Ormesby and Hales boreholes the start of Chron C26n is within palynomorpb association sequence T6 (Jolley 1992). In all sections between Halesworth (Suffolk) and north Kent, the basal Tertiary sediments carry a normal polarity magnetization. The majority of the sections that have been studied for palynomorphs indicate that the base of the Thanet Sand Formation (and the Ormesby Clay in Sizewell) is at a higher level within T6 (Jolley 1992). Nannoplankton studies of the Thanet Sand Formation at Bradwell (Knox et aL 1994) suggest an NP6 age for these levels, which is similar to that reported from the type section in Kent (Aubry et al. 1986; Siesser et al. 1988). The normal polarity magnetozone identified at the base of the Thanet Sand and in the lower half of the Ormesby Clay formation is correlated with Chron C26n. Chron C25r

The start of Chron C25r is recorded at a number of sections through the Ormesby Clay and Thanet Sand Formation. In the Thanet Sand Formation at Bradwell the reversal is positioned at a level which corresponds to the upper part of nannofossil zone NP6 (Knox et al. 1994). Recent palynological studies (Jolley 1992) indicate a mid T7 (T7c) age for the start of Chron C25r in all of the sections between Norfolk and Kent. The top of the Thanet Sand and Ormesby Clay formations are thought to have been eroded prior to deposition of younger sediments. In the Thanet Sand and Ormesby Clay Formation sections north of London the youngest preserved sediments are of T9 age. In London and

141

north Kent the top of the Thanet Sand Formation is assigned to association sequence T8, indicating a slightly deeper level of erosion. Chron C25n

Chron C25n is associated with the upper part of nannofossil zone NP8 and the lower part of NP9 (Berggren et al. 1985). The studies by Hamilton & Hojjatzadeh (1982), Aubry et al. (1986) and Siesser et al. (1988) indicate that the record of nannofossil zones NP8 and NP9 in southern England is rather poor. Aubry et al. (1986) deduced that Chron C25n was not preserved in southern England; the lack of such a record resulted from either non-deposition or erosion of material deposited during that interval. The nonsequence/erosional event was located at the unconformity separating the Thanet Sand Formation and the Upnor Formation. Biostratigraphic and palaeomagnetic studies of the upper Paleocene at Ormesby and Norfolk revealed a similar gap in the sedimentary record (Cox et al. 1985; Knox et al. 1990). (Although the paper of Cox et al. 1985 pre-dated that of Aubry et al. 1986, their interpretation relied heavily on the conclusions of the latter work.) Knox et al. (1990) recognized that the Hales Clay was somewhat younger than the Lambeth Group, implying that the Chron C25n unconformity separating the Ormesby Clay Formation and the overlying sediments (Harwich Formation) is greater in Norfolk than in the London Basin. Recent studies by Ellison et al. (1996) report a record of Chron C25n in a borehole section from central London; normal polarity sediments positively assigned to the lower part of nannoplankton zone NP9 were recovered. There the Upnor Formation comprises two distinct sediment packages separated by a pebble bed. The lowermost unit, which records the Chron C25n magnetization, extends across only a few square kilometres. It forms an isolated remnant preserved beneath sediments deposited during the main Upnor Formation transgression. In this paper we have reported an interval in the uppermost 4 m of the Thanet Sand Formation in the Sizewell borehole where samples carry either positive or shallow characteristic remanence inclinations. The 'anomalous' interval is considered to have resulted from the late Paleocene (C25n) weathering of the subaerially exposed Thanet Sand Formation prior to deposition of the Lambeth Group. However, although numerous cores have sampled this same stratigraphic interval, the C25n remagnetization event has not been seen outside of the Sizewell area.

142

J.R. ALl • D. W. JOLLEY

Chron C24r

The main body of the Upnor Formation is reversely magnetized. The formation includes the oldest Chron C24r sediments. The Woolwich, Reading and Harwich Formations, and lower London Clay Formation, were all deposited during the 2.56 million years of Chron C24r. Based on Berggren et al. (1985, 1995), these deposits straddle the Paleocene-Eocene boundary. Chron C24n.3n

The start of Chron C24n.3n coincides with the base of the London Clay Formation Division B at Alum Bay, Whitecliff Bay (Aubry et al. 1986) and Sheppey, Kent (Ali et al. 1993). At these sections the normal polarity magnetozone correlated with Chron C24n.3n terminates within Division B. At Varengeville, north France, a reverse polarity magnetization is associated with the highly glauconitic clay marking the base of Division B. However, Ali et al. (1993) considered this magnetization to be much delayed postdepositional remanence acquired during a later reverse polarity interval. In Belgium, a record of the middle and latter part of Chron C24n.3n is preserved in the type section of the Wardrecques Member (Ali et al. 1993). This normal polarity interval is from a level equivalent to Division B of the London Clay Formation (King 1990). It is not possible to locate the start of Chron C24n.3n in Belgium as intervals corresponding to the Division A-B junction (the level at which the start of Chron C24n.3n is positioned in southern England) are not exposed.

Chronostratigraphy of depositional units Integration of the palaeomagnetic and palynomorph data provides a chronostratigraphic framework for age calibration of the upper Paleocene and lowermost Eocene depositional units in southeast England (Fig. 12). The Ormesby Clay Formation was deposited as a series of south- and eastward onlapping claystone units which pass laterally into the Thanet Sand Formation, the basal Thanet Sands at Pegwell Bay being 0.65 million years younger than the basal units of the Ormesby Clay Formation in Norfolk. Within this interval two separate series of onlapping units are identified, one within the basal part of Chron C26n and the second initiated in the lowermost Chron C25r (Jolley 1992; Knox et al. 1994). The accumulation of sediment was terminated by a fall in relative sea level at c. 56.6 Ma, giving rise to a prominent sequence boundary, with lowstand deposits of equivalent age to the subsequent hiatus occurring in the North Sea

Fig. 12. Chronostratigraphy of the upper Paleocene and lowermost Eocene deposits of southern England.

Basin (Sele Formation unit Sla). A significant rise in relative sea level at c. 56.2 Ma initiated lower Upnor Formation glauconitic sand deposition (Chron C25n) across southeast England, of which only a single erosional remnant has been identified (i.e. central London). The hiatus between the termination of Ormesby Clay Formation sedimentation and this later transgression is thought to exceed 0.3 million years, a period marked by subaeriel weathering of the Thanet Sand Formation and erosion of its upper units. Deposition of the remainder of the Upnor Formation (Chron C24r) and the overlying Woolwich and Reading Formations was as a series of barrier-bar and floodplain units, at a time of increasing relative sea level. Within this barrier-bar depositional system, two significant flooding surfaces are evident, although the younger is of apparently lesser intensity. Termination of the final Lambeth Group flooding phase immediately prior to 55 Ma, by a significant fall in relative sea level, shifted sedimentation into the North Sea Basin, with no equivalents of the Sele Formation units S2a and S2b occuring in southeast England or East Anglia (55.0-54.5 Ma) (Jolley 1995). The uppermost occurrence of the

THANETIAN/YPRESIAN MAGNETOSTRATIGRAPHY, SE ENGLAND dinoflagellate cyst Apectodinium augustum around the Sele Formation unit Slb/S2a boundary has been used by Powell (1988, 1992) and Knox & Holloway (1992) to approximate to the P a l e o c e n e - E o c e n e boundary in the North Sea Basin. This datum is comparable to the upper limit of the Lambeth Group in southeast England. Sedimentation resumed in southeast England and East Anglia, at c. 54.5 Ma in north Norfolk, with basal sands of the Hales Clay, this basal sand facies onlapping onto the Lambeth Group, reaching Essex by 54.4 Ma. In all, seven units (parasequences) are determined in the Harwich Formation, occurring in two onlapping series, separated by a hiatus of c. 0.2 million years in the west London Basin and the Hampshire Basin. The highest of these parasequences provides evidence of late Harwich Formation prograding sedimentation prior to a major flooding surface at the base of the overlying Walton Member (London Clay Formation) which appears to have been deposited during considerably higher relative sea levels.

Conclusions The upper Paleocene deposits of SE England magnetobiostratigraphic important constraints on following.

and lowermost Eocene are now placed within a framework that provides our understanding of the

143

(1) How the original Thanetian and Ypresian stage stratotypes relate to the global marine record. (2) The nature and timing of the Thanetian and Ypresian depositional sequences of the southern North Sea Basin. During this interval, deposition of the shallow marine and continental sediments along the southwest margin of the North Sea was punctuated by two hiatuses, each > 0.4 million years, at c. 57.6 and 55.0 Ma. The high-resolution age calibration of the N W European Thanetian and Ypresian depositional sequences is a valuable contribution to our ongoing research aimed at providing a rigorous examination of the Exxon eustatic sea-level concept. Ultimately, the dataset we are constructing from a number of passive margins around the globe will be used to examine the nature and timing (particularly the synchroneity) of the model's third- and fourthorder cycles over a 12-15 million year window of the late Paleocene to early-mid Eocene. NERC is acknowledged for financial support to JRA during his PhD studies. We are grateful to various colleagues for helpful and stimulating discussions during the development of the research, in particular Ernie Hailwood, Robert Knox, Chris King, Richard Ellison, Nick Johnston, Tony Morigi and Norman Hamilton. Kate Davies helped with the drafting of figures. Robert Knox and Chris King provided constructive reviews of the manuscript. The British Geological Survey are thanked for providing access to the cores used in this study. This paper is a contribution to IGCP Project 308.

References ALI, J. R. 1989. Magnetostratigraphy of Early Palaeogene Sediments from N.W. Europe. PhD Thesis, University of Southampton. , HAILWOOD,E. A. & KING, C. 1996. The 'Oldhaven Magnetozone' in East Anglia: a revised interpretation. This volume. - - - , KING, C. & HAILWOOD, E. A. 1993. Magnetostratigraphic calibration of early Eocene depositional sequences in the southern North Sea Basin. In: HAILWOOD,E. A & KIDD, R. B (eds) High Resolution Stratigraphy Geological Society, London, Special Publication, 70, 99-125. AUBRY, M.-E, HAILWOOD,E. A. & TOWNSEND,H. A. 1986. Magnetic and calcareous-nannofossil stratigraphy of the lower Palaeogene formations of the Hampshire and London Basins. Journal of the Geological Society, London, 143, 729-735. BERGGREN, W. A., KENT, D. V & FLYNN, J. J. 1985. Palaeogene geochronology and chronostratigraphy. In: SNELLING, N. J. (ed.) Geochronology of the Geological Record. Memoir of the Geological Society London, 10, 141-95. , , SWISHER, C. C. III & AUBRY, M.-P. 1995. A revised Cenozoic geochronology' and chronostratigraphy. In: BERGGREN, W. A., KENT, D. V., AUI3RY, M.-E & HARDENBOL, J. (eds) Geochronology, Time Scales and Stratigraphic

Correlation: Framework for an Historical Geology. Society of Economic Paleontologists and Mineralogists, Special Volume, 54, Tulsa. CANDE, S. & KENT, D. V. 1995. Revised calibration of the geomagnetic time scale for the Late Cretaceous and Tertiary. Journal of Geophysical Research, 100, 6093-6095. COOPER, J. 1976. British Tertiary Stratigraphical and Rock Terms. Tertiary Research Special Paper, 1. Cox, F. C., HAILWOOD,E. A., HARLAND,R., HUGHES, M. J., JOHNSTON,N. & KNOX, R. W. O'B. 1985. Palaeocene sedimentation and stratigraphy in Norfolk, England. Newsletters on Stratigraphy, 14, 169-185. CURRY, D. 1981. Thanetian. In: POMEROL, C. (ed.) Stratotypes of Paleogene Stages. M6moire Hors S6rie du Bulletin d'Information des G6ologues du Bassin de Paris, 2, 255-265. - - - , ADAMS, C. G., BOULTER, M. C., DILLEY, E C., EAMES,E E, FUNNELL,B. M. & WELLS,M. K. 1978. A Correlation of Tertiary Rocks in the British Isles. Geological Society, London, Special Publication, 12. DESOR, E. 1847. Sur le terrain Danien, nouvel 6tage de la craie. Bulletin de la Socigt~ G~ologique de la France, 2, 179-182. ELLISON,R. A., ALl, J. R., HrNE, N. M. & JOLLEY,D. W.

144

J . R . ALl 8z D. W. JOLLEY

1996. Recognition of Chron C25n in the upper Paleocene Upnor Formation of the London Basin, UK. This volume. , JOLLEY,D. W., KING, C. & KNOX, R. W. O'B. 1994. A revision of the lithostratigraphical classification of the early Palaeogene strata in the London Basin and East Anglia. Proceedings of the Geologists' Association, 105, 187-197. HAILWOOD, E. A. 1977. Configuration of the geomagnetic field in early Tertiary times. Journal of the Geological Society, London, 133, 23-36. HAMILTON, G. B. & HOJJATZADEH, M. 1982. Cenozoic calcareous nannofossils - a reconnaissance. In: LORD, A. R. (ed.) A Stratigraphical Index of Calcareous Nannofossils. Ellis Horwood, Chichester, 136-167. HARDENBOL, J. & BERGGREN, W. A. 1979. A new Paleogene numerical time-scale. American

Association of Petroleum Geologists Studies in Geology, 6, 213-34. JENKINS, D. G. & LUTERBACHER, H. 1992. Paleogene stages and their boundaries (introductory remarks).

Neues Jahrbuch fiir Paliiontologie, Abhandlungen, 186,

1-5.

JOHNSTON, N. 1983. Magnetostratigraphic Stud)' of Paleogene Sediments from SE England. MSc Thesis, University of Southampton. JOLLEY, D. W. 1992. Palynofloral association sequence stratigraphy of the Palaeocene Thanet beds and equivalent sediments in eastern England. Review of Palaeobotany and Palynology, 74, 207-237. 1996. The earliest Eocene sediments of eastern England: an ultra high resolution palynological correlation. This volume. KING, C. 1981. The Stratigraphy of the London Clay and Associated Deposits. Tertiary Research Special Paper, 6. - 1990. Eocene stratigraphy of the Knokke borehole (Belgium). Toelichtingen, Verhandelingen Geologische en Mijnkaarten van Belgie, 29, 67-102. Knox, R. W. O'B. 1984. Nannoplankton zonation and the Palaeocene/Eocene boundary beds of NW Europe: an indirect correlation by means of volcanic ash layers. Journal of the Geological Society, London, 141,993-999. - - & HOLLOWAY,S. 1992. 1. Paleogene of the Central and Northern North Sea. In: KNOX, R. W. O'B. & CORDEY, W. G. (eds) Lithostratigraphic Nomenclature of the UK North Sea. British Geological Survey, Nottingham. - - , HJNE, N. M. &ALI, J. R. 1994. New information on the age and sequence stratigraphy of the type Thanetian of Southeast England. Newsletters on Stratigraphy, 30, 45-60. - - , MORIGI, A. N., ALl, J. R., HAILWOOD, E. A. &

HALLAM,J. R. 1990. Early Palaeogene stratigraphy of a cored borehole at Hales, Norfolk. Proceedings of the Geologists' Association, 101, 15-151. MARTINI, E. 1971. Standard Tertiary and Quaternary calacareous nannoplankton zonation. In: FARINACCI, A. (ed.) Proceedings of the II Planktonic Conference, Roma 1970. Edizioni Tecnoscienza, Rome, 739-785. POWELL, A. J. 1988. A modified dinoflagellate cyst biozonation for the latest Palaeocene and earliest Eocene sediments from the central North Sea. Review of Palaeobotany and Palynology, 56, 327-342. 1992. Dinoflagellate cysts of the Tertiary system. In: POWELL, A. J. (ed.) A Stratigraphic Index of Dinoflagellete Cysts. Chapman & Hall, London, 155-252. PRESTWICH,J. 1850. On the structure of the strata between the the London Clay and the Chalk in the London and Hampshire Tertiary Systems. Part I. Quarterly Journal of the Geological Society of London, 6, 252-281. - 1852. On the structure of the strata between the the London Clay and the Chalk in the London and Hampshire Tertiary Systems. Part III. The Thanet Sands. Quarterly Journal of the Geological Society of London, 8, 235-264. - 1854. On the structure of the strata between the the London Clay and the Chalk in the London and Hampshire Tertiary Systems. Part II. The Woolwich and Reading Series. Quarterly Journal of the Geological Society of London, 10, 75-170. SIESSER, W. G., WARD, D. J. ~r LORD, A. R. 1988. Calcareous nannoplankton biozonation of the Thanetian Stage (Palaeocene) in the type area. Journal of Micropalaeontology, 6, 85-102. SMITH, A. G., BRIDEN, J. C. 8z HURLEY, A. G. 1981. Phanerozoic Palaeocontinental World Maps. Cambridge Earth Science Series. TOWNSEND, H. A. & HAILWOOD, E. A. 1985. Magnetostratigraphic correlation of Palaeogene sediments in the Hampshire and London Basins, southern UK. Journal of the Geological Society, London, 142, 1-27.

WHITAKER, W. 1872. The Geology of the London Basin: Part 1. Memoir of the Geological Survey of the United Kingdom. ZIEGLER, R A. 1982. Geological Atlas of Western and Central Europe. Shell International Petroleum, Amsterdam. Z1JDERVELD, J. D. A. 1967. AC demagnetization of rocks: analysis of results. In: COLLINSON, O. W., CREER, K. M. & RUNCORN, S. K. (eds) Methods in Palaeomagnetism, Elsevier, New York, 254-286.

Upper Paleocene-Lower Eocene dinoflagellate cyst sequence biostratigraphy of southeast England A. J. POWELL t, H. BRINKHUIS 2 & J. R B U J A K 3

1Millennia Ltd, Unit 3, Weyside Park, Newman Lane, Alton, Hampshire GU34 2P J, UK 2Laboratory of Palaeobotany and Palynology, University of Utrecht, Heidelberglaan 2, 3584 CS Utrecht, The Netherlands 3The Lexis Group, Albion House, 9 Albion Avenue, Blackpool, Lancashire FY3 8NA, UK Abstract: Detailed study of the aquatic palynomorph assemblages, particularly dinoflagellate cysts, from the type-Thanetian and related sections in southeast England has enabled a detailed biostratigraphic and sequence biostratigraphic analysis to be carried out. The base of the Thanetian Stage at Pegwell Bay lies above the base of the Alisocysta margarita (Area) biozone; the overlying type-Thanetian is assigned to the Apectodinium hyperacanthum (Ahy) and Apectodinium augustum (Aau) biozones. The base of the Ypresian succession is drawn at the base of the Harwich Formation at Lower Upnor corresponding to the base of the Glaphyrocysta ordinata (Gor) biozone. The Wetzeliellaastra (Was) biozone is characteristic of the lower London Clay Formation at Herne Bay. Five Thanetian and three Ypresian sequences are identified through consideration of dinoflagellate cyst palaeoecology and sedimentological evidence. Deepening and shallowing trends enable transgressive and highstand systems tracts to be identified. The maximum flooding surfaces are identified and correlated into the North Sea Basin (Central Graben) as a series of eight primary condensed sections, two of which (characterized by the Areoligera and Apectodinium acmes) are of a second-order scale (the other six are third-order). Lowstand system tracts are represented onshore by a series of seven unconformities (type 1 sequence boundaries) with the amount of missing time below biozonal resolution. A single type 2 sequence boundary is also evident in the Ypresian succession at Wrabness. The unconformities may be correlated into more distal locations in the North Sea Basin where lowstand sandstone deposition is characteristic.

One of the objectives of the International Commission on Stratigraphy (of the International Union of Geological Sciences) is the development of a global standard stratigraphic scale. As part of this exercise, the remit of the International Subcommission on Paleogene Stratigraphy (ISPS) is to review and, if necessary, revise the definition of the ages/stages of the Paleocene Epoch/Series (see Schmitz 1994). The ISPS has decided that the Paleocene Series be divisible into three stages, namely Danian, Selandian and Thanetian (see Jenkins & Luterbacher 1992). The Paleocene Working Group of the ISPS has set itself the objective of assessing the relative merits of various global events which might be used to define the base of the Selandian and Thanetian stages, and to recommend the Global Stratotype Section and Point (GSSP) for each stage. In view of the fact that one of the requirements for a GSSP should be as complete a sedimentary succession as possible, it is important that the type-Thanetian should be reappraised from a sequence stratigraphic perspective. After all, from a classical perspective, the base of the Thanetian Stage can be no older than the

base of the type-Thanetian (i.e. the contact of the Thanet Sand Formation with the Chalk Group at Pegwell Bay, Kent, UK). It is primarily in this context that the present study has been undertaken.

Aims of the study The principal purpose of this study is to provide a detailed interpretation of the sequence stratigraphy of Upper Paleocene and Lower Eocene sediments exposed in southeast England on the basis of the aquatic palynological (primarily dinoflagellate cyst) stratigraphic record. The scale of inquiry is such that systems tracts may be deduced from the biostratigraphic data (see Powell 1992a), thus necessitating a closely-spaced suite of samples. By so doing, it is possible to test third-order cyclicity represented in the rock record. By considering the details of a well-exposed set of strata in a relatively proximal setting, it is possible to deduce the stratigraphic relations with those located more distally within the North Sea Basin. A number of predictions may be made concerning the proximal expression of primary condensed sections (Powell

From Knox, R. W. O'B., Corfield, R. M. & Dunay, R. E. (eds), 1996, Correlation of the Earl)' Paleogene in Northwest Europe, Geological Society Special Publication No. 101, pp. 145-183.

145

146

A.J. POWELL

1992a) expressed more fully in distal locations in the basin. In addition, unconformities apparent proximally (onshore) may be traced towards their correlative conformities in distal localities offshore as secondary condensed sections (Powell 1992a) and basin-floor fans.

Scope of the study The sections examined are exposed in Kent and Essex (Fig. 1). The stratigraphic position of the samples is shown in the stratigraphic summary logs (Figs 2-5). During April 1991, two of us (AJP and HB, with the guidance of Mr David J. Ward) collected 28 samples from Pegwell Bay (Fig. 2); two samples were obtained from the Upper Chalk, 18 from the Thanet Sand Formation (Base-Bed, Stourmouth Clays and Pegwell Marls) at the cliffend section, Pegwell Bay, and eight further samples from the adjacent hoverport car-park section (Pegwell Marls and Reculver Silts). At Herne Bay (Fig. 3), the eastern cliff section (four samples from the Reculver Silts), the foreshore section (12 samples from the Reculver Silts, through the Woolwich Bottom Bed of the Upnor Formation to

ETAL.

the Oldhaven Member of the Harwich Formation) were sampled (AJP, HB and DJW). A further four samples were also obtained from the western cliff section from the Oldhaven Member into the London Clay Formation. At the Lower Upnor sandpit (Fig. 4) the Reculver Silts of the Thanet Sand Formation (two samples), the Lambeth Group (the Upnor and Woolwich Formations - 12 sampies) and the Oldhaven Member of the Harwich Formation (five samples) were sampled (AJP and HB). At the cliff section of Wrabness, Essex (Fig. 5), 15 samples from the Wrabness Member of the Harwich Formation were obtained (AJP and HB). The average gap between samples were: at the cliff-end section (Pegwell Bay) 45 cm, at the hoverport car-park (Pegwell Bay) 43 cm, at the eastern cliff section (Herne Bay) 105 cm, at the foreshore section (Herne Bay) 135 cm, at the western cliff section (Herne Bay) 136 cm, at the sandpit (Lower Upnor) 127 cm, and at the cliff section (Wrabness) 102cm. The key references were: for Pegwell Bay Ward (1977), for Herne Bay Ward (1978), for Lower Upnor Kennedy & Sellwood (1970), and for Wrabness King (1981).

~_~---- ; ~ ~

~"

so .,-,,

_=,~

gambridg~ 52"

Wr~ LONDON

Harwich

.~ ---~~_r--...

9

52'

Herne Bay

) eOover 51:

51" Newhaven V'-

,. i

I

i

Fig. 1. Location map of the Pegwell Bay, Herne Bay, Lower Upnor and Wrabness sections.

PALEOCENE-EOCENE DINOFLAGELLATESEQUENCE BIOSTRATIGRAPHY The Laboratory of Palaeobotany and Palynology, University of Utrecht, undertook all the preparations using standard techniques. Heavy liquid (ZnC12) separation of the material was applied and residues were sieved using a 10 gm precision mesh sieve. After mixing to obtain homogeneity, four

Fig. 2. Stratigraphic summary log for Pegwell Bay.

147

slides were prepared using glycerine jelly as the mounting medium. The slides were studied both qualitatively and quantitatively. The quantitative analysis comprised ,a count of 200 aquatic palynomorphs where possible. For this purpose, six broad palynomorph

148

A . J . POWELL ET AL.

Fig. 3. Stratigraphic summary log for Herne Bay.

PALEOCENE--EOCENE DINOFLAGELLATESEQUENCE BIOSTRATIGRAPHY

149

Fig. 4. Stratigraphic summary log for Lower Upnor. categories were recognized: (1) dinoflagellate cysts (dinocysts); (2) miscellaneous algae (e.g. Pediastrum spp., Tasmanites spp.); (3) acanthomorph acritarchs; (4) foraminiferal test linings; (5) Leiosphaeridia spp.; (6) Paralecaniella spp. The remaining material was scanned qualitatively for additional dinocyst species. The results are displayed in Figs 6-13 and Tables 1-4. The dinoflagellate cyst taxonomy corresponds to that cited in Lentin & Williams (1993). All material

is filed in the collection of the Laboratory of Palaeobotany and Palynology, University of Utrecht, The Netherlands.

Background Lithostratigraphy The Kentish sections at Pegwell Bay and Herne Bay are internationally important because they

150

A . J . POWELL E T A L .

Fig. 5. Stratigraphicsummarylog for Wrabness.

comprise the type section of the Thanetian Stage, the youngest stage of the Paleocene Epoch (Curry 1981). By definition, the Thanetian Stage can be no older than the age at its lower contact with the Coniacian-Santonian Chalk. The section at Lower Upnor provides a lateral comparison of the Upnor Formation of the Lambeth Group, and the Oldhaven Member of the Harwich Formation (JoUey 1996) exposed at Herne Bay. The Wrabness section comprises the type section of the Wrabness Member of the Harwich Formation (Jolley 1996). It is beyond the scope of the present study to consider in detail the lithostratigraphy of the exposed sections; a detailed review of the Pegwell Bay and Herne Bay sections is given by Siesser et al. (1987). A recent revision of the lithostratigraphic nomenclature for the Upper Paleocene and

Lower Eocene of southeast England has been presented by Ellison et al. (1994) and this scheme has been adopted in the present study. M a g n e t o strati g raphy

Aubry et al. (1986) assigned the Pegwell Bay section to Chron 26n, and that at Herne Bay to Chron 25r. The Lambeth Group (Upnor and Woolwich Formations) and the Thames Group (Harwich and London Clay Formations) belong to Chron 24r. Aubry et al. (1986) and Knox (1990) points out that Chron 25n is absent and that an unconformity must therefore lie between the Thanet Sand Formation and the Lambeth Group. More recently, however, Ali (1994) and Ellison et al. (1994) have reported Chron 25n from the

151

PALEOCENE--EOCENE DINOFLAGELLATE SEQUENCE BIOSTRATIGRAPHY

Table 1. Aquatic palynomorph distribution (counts) at Pegwell Bay SamplesPB

03 04 0 5 0 6 0 7 0 8 0 9

10 11 12 13

14

15

16 17 18 19 20 21 22 23 24 25 26 27 28

Yaxa

A. senonensiss.l. D. denticulata

200

26

Foram test linings

4

P. magnificum C. medcalfii O. centrocarpum Leiosphaeridia spp. undiff. S. pseudofurcatus S. ramosus S. sepmtus

1

H. tubiferum A. alcicornu Cordosphaeridium spp. undiff. C. fibrospinosum A. margarita Alisocysta sp. 2 t C. inodes C. speciosum A. gippingensis Fibrocysta spp. undiff.

* 189 45 35 143 185 134 * 66 47

1

14 11 62 4

3

14 25

6

2

10

12

5

1 2 * *

1 5

* * 3 28 29

3 5

40

1 9

2

*

5

1 1

4

9 1

2

*

9

1

*

9 9

Pterospermella spp. undiff. T. delicata H. membraniphorum D. oebisfeldensis S. ancyrea G. divaricata A. ramulifera Tasmanites spp. undiff. P. crenulatum Apteodinium spp. undiff.

13 2 6 9 21 15

*

3

* * * 10 10 9 1

* 1 * 1 *

22 47 65

1 2

Acanthomorph acritarchs C. gracile Spinidinium? spp. undiff. Cerodinium spp. undiff. Hystrichokolpoma spp. undiff. L paradoxum Cribroperidinium spp. undiff. C. speciosum glabrum Gerdiocysta? sp. indet. G. intricata s.l. Subtilisphaera ? spp. undiff. t Source: Heilmann-Clausen 1985. * present outside of count.

2 13 14

2 26

1 14 * 2 10 10 7 3 2 5 8 7 4 10 5 2 1 3 5 5 3 5 25 56 44 48 49 21 32 3 9 10 10 8 3 1

2 1 *

6

*11

2 1 1

10 7

1

3

* 5

2

4

5 6

1 3 9 * 1 1 * * 19 11 6 10 9 11 11 17 20 60 23 42

*

4 2

1

*

1

1

*

1

3

1 1 7

* 3 3

5 *

9 6

12 7 7 24 10 10 15 11 5 12 37 10 11 40 27 5 1 1

19 38

3 *

1 20

lmpletosphaeridium spp. undiff. S. bentorii Rhombodinium sp. Lejeunecysta spp. undiff. Phelodinium ? sp. P. lidiae L. communis R. borussica Glaphyrocysta spp. undiff. M. pseudorecurvatum

2

*

2 4 1 11

*

1

1

2

1 21 22 23 24 25 26 27 28 1 2 2 8 6 12 8 2

1

21 28 2 3 2

3

3

1

1

1 4

7

5 2 2 1 I1 9

3 5

3

1

2 10 31 2

152

A.J.

POWELL ETAL.

%

%

\

\ A

A

i'

\ ! ~

I~ ~ ~ I~

~ l ~ l I ~ 9.., ~-

,....,

O i I

I

I

O

I r

,ta r

L~

.,,.~

~5 ,d

.a

PALEOCENE--EOCENE

DINOFLAGELLATE

SEQUENCE

153

BIOSTRATIGRAPHY

%

\

%

ZN

\

%

~

,4,,

%

\ %

===~====

======

s

fa0 o

.f..j.

~~i ~ ~

"

I

I

I

I

I

o .=,

o ,.Q

2

154

A.J.

P O W E L L ET AL.

%

%

\

\

\ \ !

/"

/~J

i

i

\

:! I J J

o

~

i

"~

~Z

I

I

I

I

I

F

O

.u b-, O

.a

PALEOCENE-EOCENE

DINOFLAGELLATE

SEQUENCE

BIOSTRATIGRAPHY

155

\ ~% .%

/\ % %

\

F~

\

/F, %

\ 'a

o o

E

o'~

T

T

]

.,..., "d

O

O

~,..., r

.a

156

A . J . POWELL E T A L .

\

%

\ %

\

\

.a

~, -~

"ua.I

:

NOIs

HDIM~[VHINO~VDINO:IHDIMqOOM

~ONdFI o

=:

g

.~!

o .,..~

g U

I

I

,,

{_ L

_

.,...

C2 ,m,M

PALEOCENE-EOCENE DINOFLAGELLATE SEQUENCE BIOSTRATIGRAPHY

%

157

\ A

%

\

%

\ J

\

% o

"us=I

NOI,LVI/TSO,..q HDIMqOOM

H:~IM~IVI-I

NOI,LVIAI~O.,-I ~ I O N d f ' l

o

o

E

0

, i

I

"

1

I t.,,t

9- '

I o . ....~

t,,..

o

~

t~~

~l

_L~_L

l

1

I

A. J. P O W E L L ET AL.

158

%

.\ \

%

\

\\ \ \ u~d X'r D N C K I N O q

NOI.I.VIAt~IOd HDI/VIHVI--I

~: ,...1

I

+

, -,...

PALEOCENE-EOCENEDINOFLAGELLATESEQUENCEBIOSTRATIGRAPHY

159

\ A

% ua

tj

\ \

t~ ~3

\ v--1

% .%

\ \

I

A - T T l I ~-v-~

'a

~u~alXV'IDNCKINO3

NOIZVI,~OJ

HDIM~IVH 2

! \..

e~o o 9 E

l

1

~

[..,

I

t

I

J

160

A . J . POWELL ET AL.

Table 2. Aquatic palynomorph distribution (counts) at Herne Bay SamplesHB Taxa A. gippingensis A. robustum Alisocysta sp. 2t Gerdiocysta ? sp. indet. M. pseudorecurvatum Thalassiphora spp. undiff. S. pseudofurcatus T. pelagica P. crenulatum S. septatus

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20

7 * 1 * 4 * * 1 * * 3 * 8 1 24 13

2

1 19 1 2 *

3

2

1 1

1 :#

*

3

1

6 4 21

1 2

3

*

1 1

5 11

3

9

6 18 13

23 5 6 1 * * 1 2 * * * * * * 1

Foram test linings O. centrocarpum Fibrocysta spp. undiff. G. divaricata s.1. G. ordinata Acanthomorph acritarchs C. gracile C. inodes Spinidinium? spp. undiff. C. speciosum

13 76 40 1 30 28 9 1 5 11 21 31 8 15 * 11 12 10 3 14 17 21 2 9 18 15 19 11 10 9 * 1 * * * 26 4 13 71 19 24 34 46 13 1 8 14 29 17 2 1 * * * 9 2 1 2 2 6 10 13 4 7 5 34 * 5 1 * 2 5 3 2 * 4 2 2 * 2 7 4 6 1 2 16 13 16 1 6 10 7 13 12 7 2 6 2 7 5 10 4 2 3 5 5 10 17 17 11 8 14 19 3 7 7 12 15 *

4 1 32 1 9

A. cornufruticosum S. ramosus A. ramulifera H. tubiferum A. alcicornu

1 58 31 43 23 25 47 42 31 61 56 54 27 39 39 42 5 13 19 13 16 12 8 5 3 6 11 20 18 29 10 6 * 1 1 3 2 5 1 4 1 * 3 1 1 3 6 3 1 1 2 1 1 7

2 12 1 37 21 10 1 11 3

*

9

9 39 15 23

*

9

*

1

6

1 1 4

A. senonensis s.l. G. pastielsii Operculodinium spp. undiff. C. ? minimum Polysphaeridium spp. undiff. R. borussica Hystrichokolpoma spp. undiff. L. communis M. fimbriatum C. fibrospinosum

*

7

1 6

*

*

3

5 iI

1

5

*

*

1

1

1

1 1

4

*

* 4

*

7

5

8

4

*

2

2 1 3 1 * 1 1

1 5 1 8 2

* present outside of count.

l o w e r m o s t U p n o r F o r m a t i o n ( L a m b e t h Group). Knox et al. (1994) presented a reappraisal of the m a g n e t o s t r a t i g r a p h y o f the t y p e - T h a n e t i a n succession.

Biostratigraphy

Apart from dinoflagellate cysts, the most intensively studied microfossils from the type-Thanetian have been the calcareous nannofossils and a full review is given b y Siesser et al. (1987). These authors have also carried out the most detailed analysis o f the stratigraphic distribution o f these microfossils from P e g w e l l and H e r n e Bays. Their findings suggest that the l o w e r m o s t 4.2 m of the

Thanet Sand Formation is assignable to biozones N P 6 - 7 , and the overlying part o f the formation belongs to biozone NP8. Knox (1990) believed an N P 7 biozonal allocation to be likely for the l o w e r Thanet Sand Formation. H o w e v e r , more recently, K n o x et al. (1994) assigned the l o w e r m o s t Thanet Sand Formation (Stourmouth Clays) to b i o z o n e N P 6 and the R e c u l v e r Silts to b i o z o n e N P 8 (the P e g w e l l Marls being unassignable). K n o x (1990) suggested that the U p n o r Formation o f the L a m b e t h G r o u p m a y be assigned to b i o z o n e N P 9 (with an u n c o n f o r m i t y at its base). A c c o r d i n g to K n o x (1990), the ' O l d h a v e n B e d s ' o f the H a r w i c h Formation, and the overlying L o n d o n Clay Formation, belong to b i o z o n e N P 1 0 (with an u n c o n f o r m i t y at the b a s e o f the ' O l d h a v e n B e d s ' ) .

161

PALEOCENE--EOCENE DINOFLAGELLATE SEQUENCE BIOSTRATIGRAPHY Table 2. Continued SamplesHB

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20

Taxa

A. homomorphum C. depressum P. lidiae Apteodinium spp. undiff. G. intricata Pterospermella spp. undiff. Leiosphaeridia spp. undiff. H. campanula Tectatodinium? sp. indet. C. medcalfii S. ancyrea L. cf. hyalina 14. membraniphorum Subtilisphaera? spp. undiff. D. simplex P. magnificum C. speciosum glabrum Homotryblium spp. undiff. Qyclopsiella spp. undiff. D. denticulata

2 1

9 1

5 *

*

1 4

*

1

6

2 2 2

19 5 2 2 7 3 2 5 1 1

2

1 2

7 15 20 15 7 2 1 10 2 1

7

8 18 4

9

9

1 14

1

4 16 36 28 * 2 13 *

3

1

6 20 1

1 *

44

3

3

2

6

*

*

3

1

3

*

2

2

*

1

1 1

4

6

4

8

2

1

3

3

3

4

1

P. golzowense L. hyalina Tasmanites spp. undiff. Phelodinium ? sp. indet. D. oebisfeldensis A. summissum T. delicata A. cf. quinquelatum A. cf. parvum A. parvum

8

3

1

1

4

1 1 7

1

2

3

1

1 46 *

1

1 * 2 * 1 1 1 4 1 5

1

4

2 1

2 9 1 1

1

1

24

1

4 8 7 42 12

Apectodinium spp. undiff. A. quinquelatum W. astra t Source: Heilmann-Clausen 1985. * present outside of count.

Previous dinoflagellate cyst studies A number of studies have been carried out on the dinoflagellate cyst assemblages from the typeThanetian and these have been reviewed by Powell (1992b). These previous studies had biostratigraphic or palaeoenvironmental objectives and were conducted on a relatively coarse scale. Thus, although broad correlations are possible throughout northwest Europe (and second-order palaeoe n v i r o n m e n t a l trends are apparent), detailed sequence biostratigraphic analysis has hitherto not been possible. To summarize, it is well k n o w n that Areoligera spp. and Alisocysta margarita are characteristic of the Thanet Sand Formation (Husain 1967; Downie

et aL 1971; Allen 1982; Jolley 1992). From a biostratigraphic perspective, both Costa & Downie (1976) and Costa et al. (1978) assigned the Thanet Sand F o r m a t i o n to the Deflandrea speciosa biozone and the Lambeth Group to the Wetzeliella (Apectodinium) hyperacantha biozone. Costa et al. (1978) allocated the L o n d o n Clay Formation of Kent to the Wetzeliella (W.) astra and W. (W.) meckelfeldensis biozones. Costa et al. (1978) were unable to allocate the 'Oldhaven Beds' of the Harwich Formation to any biozone due to the absence o f key species. K n o x et al. (1981) suggested that the 'Oldhaven Beds' lie within the A. hypercanthum biozone. H e i l m a n n - C l a u s e n (1985) was able to correlate the Thanet Sand Formation at Pegwell Bay (his biozone 4, and the

162

A.J. POWELL ET AL.

Table 3. Aquatic palynomorph distribution (counts) at Lower Upnor Samples LU

01 02 03 04 05 06 07 08 09

10

11 12 13 14

I5

16 17 18 19

Taxa

O. centrocarpum Leiosphaeridia spp. undiff. Acanthomorph acritarchs Pediastrum spp. undiff. H. abbreviatum A. hyperacanthum H. tenuispinosum A. quinquelatum A. parvum Apectodinium spp. undiff. A. homomorphum A. cornufruticosum S. ramosus L. communis M. fimbriatum Deflandreoids undiff. Batiacasphaera spp. undiff. G. divaricata Polysphaeridium spp. undiff. L. machaerophorum A. senonensis s.L G. ordinata G. intricata M. pseudorecurvatum P. indentata A. robustum Adnatosphaeridium spp. undiff. Cordosphaeridium spp. undiff. D. colligerum H. tubiferum C. inodes F. ferox H. plectilum Fibrocysta spp. undiff. Cribroperidinium spp. undiff. Apteodinium spp. undiff. S. membranaceus G. pastielsii T. delicata Pterospermella spp. undiff.

14 14 9 44 50 14 29 6 ! 11 7 1 1 8 41 50 7 33 15 5 53 40 18 12 9 3 8 46 11 4 46 7 15 1 5 1 7 * * 1 * 9 2 5 12 9 9 12 17 1 3 7 57 30 2 4 3 14 65 86 5 8 86 11 17 5 2 3 27 20 5 9 1 * 1 1 3 * 2 3 7 13 11 5 * 3 42 15 22 3 4 2 5 2 2

1 1 1

1 *

*

* *

5 4

2 1 48

1

3 9 1

3 1

1

7

1 11 14 6 4

3 13 1 3 3

22 17 11

1 :#

12 7

2 12 l 7 3

*

* present outside of count.

Ama biozone o f Powell 1992b) with the Holmehus Formation of the Viborg 1 borehole in Jutland, Denmark. Similarly, he correlated the formation at Herne Bay (his biozone 5, and the Ahy biozone of Powell 1992b) to the 'Grey Clay' lying between the Holmehus and Olst Formations in the same borehole, despite the apparent absence of Apectodinium spp. Powell (1992b) assigned the Thanet Sand Formation at Pegwell Bay to the Alisocysta margarita (Ama) biozone, and at Herne Bay to

the Apectodinium hypercanthum (Ahy) biozone. He placed the Lambeth Group in the Apectodinium augustum (Aau) biozone, and the 'Oldhaven Beds' of the Harwich Formation in the Glaphyrocysta ordinata (Got) biozone. The lower London Clay Formation was assigned to the Wetzeliella astra (Was) and W. mecketfeldensis (Wme) biozones. The only previous palynological studies of the Harwich Formation at Wrabness have been those

PALEOCENE-EOCENE DINOFLAGELLATE SEQUENCE BIOSTRATIGRAPHY of Jolley & Spinner (1989, 1991). They assigned the Harwich Formation to the Apectodinium hypercanthum biozone (sensu Costa & Downie 1976 and Costa et al. 1978) and restricted the succession above the Harwich Stone Band to the Deflandrea oebisfeldensis acme biozone.

Dinoflagellate cyst biostratigraphy of the studied sections Although the lowermost samples at Pegwell Bay and Lower Upnor are virtually barren of palynomorphs, most other samples contain reasonably well-preserved aquatic palynological associations suitable for quantitative analysis. The results show that dinoflagellate cysts dominate the associations in most cases. Other categories of aquatic palynomorphs are only common to abundant in particular intervals of the Lower Upnor section and in the upper part of the Wrabness section. Samples from the Oldhaven Member at Herne Bay contain virtually only Leiosphaeridia spp. and Pterospermella spp. In the dinoflagellate cyst assemblages usually only a few taxa or 'complexes' of morphologically similar taxa are quantitatively important. In most samples, Spiniferites ramosus, Achomosphaera ramulifera, Areoligera senonensis group,

Glaphyrocysta intricata, G. divaricata, Cerodinium speciosum or C. medcalfii are common to abundant, and range throughout the investigated interval. In addition, Apectodinium spp. (mainly A. homomorphum) and Deflandrea oebisfeldensis are common in particular intervals. Cordosphaeridium spp., usually represented by C. fibrospinosum, C. gracile or C. inodes, may be common in some samples, as are representatives of Operculodinium centrocarpum, Homotryblium spp., Impletosphaeridium spp. and Polysphaeridium spp. The biozonation scheme applied in the present study is that of Powell (1992b). This scheme is defined by the range bases of particular taxa; it has the additional advantage in that range tops may be extrapolated into the North Sea Basin. The major latest Paleocene and earliest Eocene palynological (dinoflagellate cysts) bioevents occurring in the North Sea Basin are shown in Fig. 14, while a comparison of published latest Paleocene and earliest Eocene North Sea palynological biozonations is presented in Fig. 15. The acronyms FO and LO stand for first occurrence and last occurrence respectively. The relative abundance categories used in the present study are: rare (< 6%), common (6-20%), abundant (20.5-60%), superabundant (> 60%) and present (outside of count).

163

Pegwell Bay dinoflagellate cyst biostratigraphy Barren interzone Age: Indeterminate. Samples: PB1 and PB2. Lithostratigraphy: Chalk Group. Comments: The two samples from the Chalk proved to be barren of palynomorphs and may not be assigned to any biozonei' The Chalk Group is Coniacian to Santonian in age (Siesser et al. 1987). Alisocysta margarita (Ama) interval biozone

Age: Thanetian (pars), late Paleocene (pars). Samples: PB3-PB28. Lithostratigraphy: Thanet Sand Formation (BaseBed, Stourmouth Clays, Pegwell Marls and Reculver Silts, pars). Comments: The succession in the Base-Bed of the Thanet Sand Formation contains neither the index taxon of the Ama biozone (Deflandrea denticulata) nor the nominate taxon (Alisocysta margarita). However, the superabundant occurrence and total dominance (100%) of members of the Areoligera senonensis group in sample PB4 is strong evidence for allocation of the Base-Bed to the Ama biozone. Furthermore, indicators of older biozones, notably Palaeoperidinium pyrophorum, are absent. The sample at PB6 from the base of the Stourmouth Clays contains a highly impoverished assemblage which includes representatives of the Areoligera senonensis group. Other samples from the lower Stourmouth Clays (PB7-PB10) are also impoverished in terms of both specimen and species numbers. However, the occurrence of Deflandrea denticulata at PB7 confirms allocation to the Area biozone. Phelodinium magnificum,

Cerodinium medcalfii, Spiniferites pseudofurcatus and S. septatus all have their FO at PB7. Alisocysta margarita, the nominate taxon for the Area biozone, has its FO at P B l l (upper Stourmouth Clays) together with those of Cordosphaeridium fibrospinosum, Alisocysta sp. 2 of Heilmann-Clausen (1985), Achomosphaera alcicornu, the Cerodinium speciosum and the Hystrichosphaeridium tubiferum groups. The assemblage at PB 12 is impoverished, but contains a reasonable recovery of both Areoligera senonensis and Cerodinium medcalfii. Moving into the Pegwell Marls (sample PB13), a number of FOs are apparent, including those of

Diphyes colligerum, Hystrichostrogylon coninckii (only present in this sample), H. membraniphorum, Thalassiphora delicata, Glaphyrocysta divaricata

164

A.J. POWELL ETAL.

Table 4. Aquatic palynomorph distribution (counts) at Wrabness SamplesWB

01

02

03

04

05

1 2 2

*

06

07

08

*

2 1 2 2 2 5

3 3

13 4

1 8 1 6 2 2 13 2 7 3

5 3 1 8 2 1 11 4 11 1

1 2 4 5 43 18 8 7 7

* 2 8 4 46 18 12 3 2

1 4 3

09

10

1 1

1 1 1 1 1

9

4

11

12

13

14

15

4 4 5

5 5 4 1

2 4 4

7 15 10 1

Taxa

G. exuberans O. complex P. subtile L. disjuncture D. cladiodes s. Morg. E bipolaris C. multispinosum S. cf. chlamydophora C. giuseppei S. septatus

* 1 2 3 * 1 8 1 2 5

* 2

2 4

4

3 3

2 3

5 2

11 2 14 5

5 5 5 8 2 * 11 3 12 4

Hystrichokolpoma sp. A. C. gracile L. wetzelii O. centrocarpum A. alcicornu A. multispinosum A. ramulifera S. cf. membranaceus S. monilis Senegalinium? sp. A.

2 6 * 18 1 1 8 4 9 4

T. cf. pelagica T. delicata G. ordinata A. senonensis s.l. S. ramosus H. tubiferum G. divaricata D. oebisfeldensis Acritarch sp. A Foram test linings

1 * 9 3 38 23 2 15 4 5

1 14 6 37 22 7 9 3 4

Michrystridium spp. undiff. Acritarch sp. B. Leiosphaeridia spp. undiff. P. indentata

13 2 4 5

3 3 3 6

12 2

6

2 2 3 5

8 4 4 11 1

6 3 4 12 1

11 4 10 3

9 4 12 5

2 1 11 4 39 16 5 4 7 2

1 13 3 46 17 9 4 3 2

1 8 2 48 16 4 2 6 4

8 5

4 2

3

4

6 2 3 2

and Deflandrea oebisfeldensis. The Areoligera senonensis group is superabundant at PB15. Phthanoperidinium crenulatum has its FO at PB 18. A generally similar assemblage is present at PB19 and it is at this sample position that the top of an acme (> 20%) of Areoligera may be placed (equivalent to bioevent P1 of Mudge & Copestake 1992; biochronoevent BC-7 of Armentrout et al. 1993; and bioevent Aga of M u d g e & Bujak 1996). The assemblage at PB20 is highly impoverished but includes Areoligera senonensis. A large number of taxa have their FOs at PB21 (uppermost Pegwell Marls), the most significant being Cerodinium depressum, Lejeunecysta communis, Rottnestia borussica, Palaeocystodinium lidiae, Melitasphaeridium p s e u d o recurvatum and the Glaphyrocysta pastielsii group. The assemblage at PB22 (lowermost Reculver

* 3 4

4 5 2 7 2 * 9

4 3 3 5

4 4 1 3

49 18 4 5 18

2 5 5 5

1

1

2 1

4 4

3 2 3 3

4 1

Silts) is i m p o v e r i s h e d in terms o f s p e c i m e n numbers, but nevertheless is generally similar in species composition to that at PB21. The assemblage at PB23 contains the FO of Cordosphaeridium gracile, while that at PB24 contains the LO of Cerodinium medcalfii and the FO of Spiniferites cornutus. A generally similar assemblage is present in PB25 but with the additional FOs of I m p a g i d i n i u m p a r a d o x u m , Gerdiocysta cassiculus and Cerodinium speciosum glabrum. The LO of Alisocysta margarita lies at PB26 (equivalent to bioevent P2 of M u d g e & Copestake 1992; biochronoevent BC-8 of Armentrout et al. 1993; and bioevent A l m of M u d g e & Bujak 1996), while an impoverished assemblage is present at PB27 broadly similar to that at PB26 in terms of species composition. The richer assemblage

165

PALEOCENE-EOCENE DINOFLAGELLATE SEQUENCE BIOSTRATIGRAPHY Table 4. Continued SamplesWB

01

02

03

1

1

04

05

06

* 2

1 4

2 3

2

* 1 2

3 4 3

07

08

09

10

11

12

13

14

15

5

2

4

3

7

7

14

1

2

Waxa

A. biformoides E cf. vectense D. phophoritica C. pannuceum M. pseudorecurvatum T. pellitum C. depressum D. colligerum P crenulatum R. borussica

2 4 2 2 9 2

D. pastielsii G. pastielsii H. membraniphorum Pteroapermella spp. undiff. O. ? severinii S. cornutus I. californiense S. pseudofurcatus A. hyperacanthum A. homomorphum

2 * 3 2

1 1

1

3 1 9 2

5 2

2

2

2 3

6 2

4

4 1 1

1 1

Lejeunecysta spp. undiff. S. densispinatum A. parvum P. golzowense P.. magnificum P. inversibuccinum P. minusculum 11. rigaudiae C. speciosum C. inodes

*

6 2 2 1

* 1

1

1 2

1

7 * * 2 *

*

2 3 1 9 9 3 1

Pediastrum spp. undiff. Cymatiosphaera spp. undiff. Tasmanites spp. undiff.

* 1 * 1 3 3

1 2

1 * 2 1 1

2 1

2

*

4

5

3

* present outside of count.

at

PB28

is

characterized

by

abundant

Glaphyrocysta spp.

Herne

Bay dinoflagellate

cyst

biostratigraphy

A p e c t o d i n i u m h y p e r a c a n t h u m (Ahy) interval

biozone Age: Thanetian (pars), late Paleocene (pars). Samples: HB 1-HB 14. Lithostratigraphy: Thanet Sand F o r m a t i o n (Reculver Silts, pars); Lambeth Group, U p n o r Formation (Woolwich Bottom Bed, pars). Comments: The FO of Apectodinium homomorphum at HB 1 is good evidence for allocation to the Ahy biozone (calibrated close to bioevent M5

of M u d g e & Copestake 1992; biochronoevent BC-9 of Armentrout et al. 1993; and bioevent IA of M u d g e & Bujak 1996); Apectodinium cornufruticosum also has its FO at HB1, as do Adnatosphaeridium robustum and Muratodinium fimbriatum. Other FOs of note include Palaeocystodinium lidiae at HB2, P. golzowense at HB9 and Lejeunecysta hyalina at HB10, all within the Reculver Silts of the Thanet Sand Formation. There are, in addition, a number of LOs within the succession. At HB10, the LOs of Gerdiocysta

cassiculus, Melitasphaeridium pseudorecurvatum, Adnatosphaeridium robustum, Alisocysta sp. 2 of Heilmann-Clausen (1985) and Areoligera gippingensis (equivalent to bioevent P2 of Mudge & C o p e s t a k e 1992; B i o c h r o n o e v e n t BC-8 of Armentrout et al. 1993; and bioevent Ag o f Mudge & Bujak (1996) are all present, while Spiniferites

A.J. POWELL ET AL.

166 CHRONO-,BIOZONES STRAT. (Powell '92)

w

~

9 Was

z w

Gor

,5 ~

MAJOR PALYNOLOGICAL (DINOFLAGELLATE CYST) BIOEVENTS ~

'9'

...i

9

uJ z W

0.9 whereas for hematite-bearing samples the value is typically 0.6-0.9 (see Fig. 5). IRM ratios and peak IRM values are listed in Table 1. IRM ratio data are also shown in Fig. 6. A composite magnetostratigraphic section (Fig. 5) has been constructed from polarity data from the four cores (Table 1), using the top of the Thanet Sand Formation as datum. It shows that the bulk of the Lambeth Group is of reverse polarity, but a normal polarity interval, coded CL-A, is identified in the lower part of the Upnor Formation in Jubilee borehole 404T. Two normal polarity levels identified in the Woolwich Formation and upper leaf of the Reading Formation in Borehole A4A are based on single specimens only, and the stratigraphic significance of these data is not yet fully understood. In order to assess the reliability of the magnetostratigraphic data, in particular the normal polarity magnetozone CL-A, the magnetic properties of each formation are considered. The five specimens from CL-A that were included in the IRM analyses have ratios between 0.85 and 0.95 (Fig. 6a & b). This suggests that a low-coercivity mineral, probably magnetite, is the dominant remanence carrier at these levels. The specimen (JU34) immediately above the top of CL-A has an IRM ratio of 0.96,

191

CENTRAL

24

20

LONDON

--

I

I

-

I

I

m

FM. POL.

t,.,I--

s

r I

-

10 A m 2, respectively. Above this level the values are much lower (0.75 and < 5 A m 2, respectively). The NRM intensity data suggest that the transition takes place at c. 2.3 m above the base of the section (Fig. 6). This behaviour is again probably related to changes in the magnetic mineralogy, with the magnetic properties of the lower part of the section being dominated by magnetite and those of the upper parts by hematite.

Harwich The Harwich Stone Band is exposed at low tides at Dovercote near Harwich. Townsend & Hailwood (1985) drilled seven specimens from two sites

'OLDHAVEN MAGNETOZONE' IN EAST ANGLIA

201

Table 1. NRM intensity and IRM acquisition data for the Harwich Formation Wrabness Member and the London Clay Formation Walton Member in East Anglia Locality Levington

Walton-on-the-Naze

Wrabness

Unit

Spec.

Pol

Ht (m)

NRM (mA m-1)

IRM ratio

Peak IRM

H H H H W W W W H H W W H H H H H H

LV11.2 LV7.1 LV5.2 LV2.2 WN16.2 WN 10.1 WN7.1 WN6.1 WN4.1 WNI. 1 WR2.4 WR2.2 WR1.2 WR1.3 WR1.8 WRl.14 WR1.19 WR1.23

N N N N R R R R R R R R N N N N R R

8.15 4.65 2.85 0.40 11.20 5.40 4.05 3.90 2.75 0.00 11.68 10.63 9.18 8.33 5.99 3.46 1.29 0.00

2.11 3.09 12.39 30.16 0.31 0.38 2.14 5.25 7.92 15.13 1.14 1.45 1.82 8.57 1.10 2.41 26.43 30.96

0.74 0.79 0.87 0.97 0.73 0.77 0.90 0.85 0.94 0.96 0.83 0.74 0.70 0.81 0.71 0.76 0.95 0.95

1427 1866 3512 12502 1311 1132 4658 3437 8446 1320 682 1201 1167 264 1214 1951 14749 20920

H, Wrabness Member; W, Walton Member; N, normal polarity; R, reverse polarity. in the stone band. NRM intensities averaged 300 mA m -I, and all specimens exhibited a reverse polarity characteristic magnetization (Fig. 6). Summary

Magnetostratigraphic data have now been obtained from the Wrabness Member at Levington, Waltonon-the-Naze, Wrabness and Harwich (Fig. 6). The two most distant localities are only 25 km apart (Levington and Walton-on-the-Naze) and there are only minor variations in the lithology between the sections. The Harwich Stone Band in the Wrabness Member and the glauconitic base of the overlying Walton Member are the most important lithostratigraphic markers. The core-drilled specimens taken from the Harwich Stone Band at Levington, Wrabness and Harwich all exhibit a reverse polarity characteristic magnetization. However, at Levington, all sediments sampled above and below the Harwich Stone Band display normal polarity. At Wrabness, a normal polarity zone commences at c. 3.0 m above the Harwich Stone Band and continues to the top of the Wrabness Member. Two reverse polarity sites were identified in the lower part of the Walton Member. At Walton-on-the-Naze, the Wrabness Member displays reverse polarity, as did the Walton Member. At Harwich, only the reversely magnetized Harwich Stone Band was sampled. At first sight, the normal polarity zones at Levington and Wrabness appear to be at least

partially correlative. However, for reasons discussed below, we believe that they probably owe their normal polarity to recent weathering, which oxidized original magnetite in these sediments to hematite and resulted in a magnetic overprint in the recent normal polarity geomagnetic field. Firstly, the few metres of sediments above and below the distinctive and readily correlatable Harwich Stone Band are of normal polarity at Levington, but reverse polarity at Wrabness. Both of these sections are in degraded river cliffs, where the rates of erosion are likely to be much lower than in the sea cliff at Walton-on-the-Naze (where the cliffs are receding at c. 1 m per year). Consequently, there is a higher probability of recent weathering processes affecting the magnetism of the Levington and Wrabness sections than that of the Walton-on-theNaze section. However, it is necessary to explain why the reverse polarity Walton M e m b e r at Wrabness appears to have escaped weathering re-magnetization processes, even though it is located near the top of the section, where such processes might be expected to be more intense than at depth. Results from the less weathered section at Walton-on-the-Naze provide a possible explanation. The Harwich Formation and Walton Member in this section are of reverse polarity. However, the NRM intensity and IRM data indicate that the magnetic properties of the two members are quite different; the Wrabness Member is dominated by magnetite and the Walton Member by hematite. Assuming a hematite-rich original composition for

202

J.R. ALI ET AL.

the Walton Member at Wrabness, the magnetic minerals of this unit would be more resistant to further oxidizing processes and more able to withstand the normal polarity overprint which appears to have affected the upper part of the Wrabness Member which is believed to have been originally magnetitie-rich. The argument is strengthened further by the fact that in the section at Wrabness the level of the downward change from normal to reverse polarity (c. 3.0 m above the top of the Harwich Stone Band) occurs very close to that of the downward transition from hematite-rich to magnetite-rich sediments (c. 2.3 m above the top of the Harwich Stone Band) based on the IRM analyses. The normal polarity magnetization observed at Levington is believed to be the result of weathering also. The top of the Levington section is more altered (the magnetic mineralogy being dominated by hematite), with the base being only partially oxidized to maghemite. The lithified Harwich Stone Band appears to have resisted weathering and retained its original reverse polarity magnetization.

Borehole data from the Harwich Formation, East Anglia Since the pioneering studies of Townsend & Hailwood (1985) and Aubry et al. (1986) on outcrop sections, a large volume of lower Tertiary magnetostratigraphic information has become available from the East Anglia region through study of cores obtained by the British Geological Survey (Cox et al. 1985; Knox et al. 1990; Ali & Jolley, 1996). A major advantage of studying drill cores is that the recovered sediments are usually in pristine condition, having been protected from surficial weathering. All published data (summarized in Ali & Jolley 1996) indicate that across southern England the Harwich Formation Wrabness Member carries a dominantly reverse polarity magnetization (acquired during Chron C24r).

Conclusions It is concluded that both the Harwich Formation Wrabness Member and the lower part of the Walton Member of the London Clay Formation in the Ipswich area of East Anglia are characterized by reverse magnetic polarity. Recent studies of the borehole sections through the Harwich Formation of East Anglia support this observation. The preliminary result of Townsend & Hailwood (1985), indicating a normal polarity for the upper part of the Wrabness Member at Wrabness, is now believed to reflect recent weathering processes on this part of the section. Since the Wrabness Member and Walton Member together span the Apectodinium hyperacanthum to Wetzeliella astra dinoflagellate zones they can be correlated with the middle part of Chron C24r. The results presented in this paper refer to the Harwich Formation of the East Anglia region. There remains uncertainty over the reliability of the 'Oldhaven magnetozone' in Herne Bay, Kent (and Harefield, Middlesex). The magnetozone at Herne Bay was defined in lithologies which are not ideal for palaeomagnetic investigations (cross-bedded sands) and only a.f. demagnetization was used on the samples from there. It would be beneficial for the section to be re-studied using thermal demagnetization and detailed magneto-mineralogical anlyses to explore the validity of the 'Oldhaven' normal polarity magnetozone in this region.

The Natural Environment Research Council is gratefully acknowledged for financial support to JRA during his PhD studies. We would like to thank Nick Johnston and Kevin Padley for their invaluable help in the laboratory and field, and Kate Davis for drafting the figures. We are grateful to various colleagues for helpful and stimulating discussions during the development of the research, particularly Robert Knox and Norman Hamilton. Robert Knox, Niels Abrahamsen and Paul Montgomery are thanked for their constructive reviews of the manuscript.

References ALI, J. R. & JOLLEY, D. W. 1996. Chronostratigraphic framework for the Thanetian and lower Ypresian deposits of Southern England. This volume. , K1NG, C. & HAILWOOD, E. A. 1993. Magnetostratigraphic calibration of early Eocene depositional sequences in the southern North Sea Basin. In: HAILWOOD,E. A & KIDD, R. B. (eds) High Resolution Stratigraphy. Geological Society, London, Special Publication, 70, 99-125. AUBRY, M.-E, HAILWOOD, E. A. & TOWNSEND, H. A. 1986. Magnetic and calcareous-nannofossil stratigraphy of the lower Palaeogene formations of the Hampshire and London Basins. Journal of the Geological Society, London, 143, 729-735.

BERGGREN, W. A., KENT, D. V., SWISHER, C. C. HI & AUBRY, M . - R 1995. A revised Cenozoic geochronology and chronostratigraphy. In:

BERGGREN, W. A., KENT, D. V., AUBRY,M.-E & HARDENBOL, J. (eds) Geochronology, Time Scales and Stratigraphic Correlation: Framework for an Historical Geology. Society of Economic Paleontologists and Mineralogists, Special Volume, 54, Tulsa. CANDE, S. C. t~ KENT, D. V. 1995. Revised calibration of the geomagnetic polarity time scale for the Late Cretaceous and Cenozoic. Journal of Geophysical Research, 100, 6093-6095. Cox, F. C., HAILWOOD,E. A., HARLAND,R., HUGHES,M.

'OLDHAVEN MAGNETOZONE' IN EAST ANGLIA J., JOHNSTON, N. & KNOX, R. W. O'B. 1985. Palaeocene sedimentation and stratigraphy in Norfolk, England. Newsletters on Stratigraphy, 14, 169-185. ELLISON, R. A., ALl, J. R., nINE, N. M. & JOLLEY,D. W. 1996. Recognition of Chron 25n in the upper Paleocene Upnor Formation of the London Basin, UK. This volume. , JOLLEY, D. W., KING, C. & KNOX, R. W. O'B. 1994. A revision of the lithostratigraphical classification of the early Palaeogene strata in the London Basin and East Anglia. Proceedings of the Geologists' Association, 105, 187-197. GEORGE,W. & VINCENT,S. 1976. Some river exposures of the London Clay in Suffolk and Essex. Tertiary Research, 1, 25-28. JOLLEY, D. W. 1996. The earliest Eocene sediments of eastern England: an ultra-high resolution palynological correlation. This volume. KING, C. 1981. The Stratigraphy of the London Clay and Associated Deposits. Tertiary Research Special Paper 6. KNOX, R. W. O'B. 1984. Nannoplankton zonation and the Palaeocene/Eocene boundary beds of NW Europe: an indirect correlation by means of volcanic ash layers. Journal of the Geological Society, London, 141, 993-999. - - , HINE, N. M. & Ali, J. R. 1994. New information on the age and sequence stratigraphy of the type Thanetian of Southeast England. Newsletters on Stratigraphy, 30, 45-60.

203

MORIGI, A. N., ALI, J. R., HAILWOOD, E. A., & HALLAM, J. R. 1990. Early Palaeogene stratigraphy of a cored borehole at Hales, Norfolk. Proceedings of the Geologists' Association, 101, 145-151. MARTINI, E. 1971. Standard Tertiary and Quaternary calcareous nannoplankton zonation. In: FARINACCI, A. (ed.) Proceedings of the lI Planktonic Conference, Roma 1970. Edizioni Tecnoscienza, Rome, 739-785. POWELL, A. J. 1992. Dinoflagellate cysts of the Tertiary system. In: POWELL, A. J. (ed.) A Stratigraphic Index of Dinoflagellete Cysts. Chapman & Hall, London, 155-252. RIDDY, P..J. & HAILWOOD,E. A. 1989. Stationary-sample A.E demagnetisation using a coupled cryogenic magnetometer/demagnetiser system. (Abstract). Geophysical Journal Royal Astronomical Society, 88, 597. SIESSER,W. G., WARD, D. J. & LORD, A. R. 1988. Calcareous nannoplankton biozonation of the Thanetian Stage (Palaeocene) in the type area. Journal of Micropalaeontology, 6, 85-102. TOWNSEND, n. A. & HAILWOOD, E. A. 1985. Magnetostratigraphic correlation of Palaeogene sediments in the Hampshire and London Basins, southern UK. Journal of the Geological Society, London, 142, --,

1-27.

ZIJDERVELD,J. D. A. 1967. AC demagnetization of rocks: analysis of results. In: COLLINSON,D. W., CREER, K. M. & RUNCORN, S. K. (eds) Methods in Palaeomagnetism. Elsevier, New York, 254-286.

Mammalian biostratigraphy across the Paleocene-Eocene boundary in the Paris, London and Belgian basins J. J. H O O K E R

Department of Palaeontology, Natural History Museum, Cromwell Road, London SW7 5BD, UK

Abstract: Problems of resolution and poor superpositional evidence in mammalian biostratigraphy through Paleocene-Eocene boundary strata in NW Europe are solved by applying parsimony analysis to taxa shared between localities. On this basis, five biozones are established in the area for the interval formerly delineated by mammalian biostratigraphers as MP7-MP9. Integration with other biostratigraphies (dinocyst, calcareous nannoplankton, charophytes) aids correlation between the London, Belgian and Paris Basins, and supports the earlier idea of diachronism of the 'argile ~t lignites' facies. The advent of 'Sparnacian' mammal faunas in Europe may coincide with a carbon isotope excursion recently recognized in the Paris Basin. This would support recent views on essential synchronism of the beginnings of both the North American Wasatchian and European 'Sparnacian' land mammal ages.

One of the most important events in mammalian history during the Cenozoic, and certainly the most important within the Northern Hemisphere Paleogene, was that which took place at or around the Paleocene-Eocene boundary. This event was a rapid faunal turnover with large numbers of extinctions in mammal groups that had been dominant in the Paleocene, accompanied by origins at ordinal and family level. The event is best represented and documented in western North America, where long continental sequences contain an essentially continuous record of mammalian fossils (Gingerich 1989; Gingerich et al. 1980; Rose 1981; Schankler 1980). In Asia, the event is best documented in Mongolia, where continental sequences have a more sporadic mammalian record (Dashzeveg 1988). Europe has the most disjointed mammalian record (Russell 1975; Russell et al. 1982a,b; Hooker 1991), but the event is striking and the area is classic for containing all the stratotypes of the globally recognized Paleocene and Eocene stages (Pomerol 1981). In Europe (as in North America), the main Paleocene groups to suffer decimation were the order Multituberculata, the Plesiadapiformes (primate relatives) and the archaic ungulates (paraphyletic order 'Condylarthra'). The incoming groups in both continents were the orders Perissodactyla, Artiodactyla, Primates (s.s.), probably Chiroptera (although not recorded in the very earliest post-event faunas) and the family Hyaenodontidae (order Creodonta). The suddenness and morphological distinctness of the appearances imply dispersal from elsewhere, but the source has not been identified, although 'the

south' is usually invoked, e.g. Africa, Central America (Gingerich 1976) or India (Krause & Maas 1990). Other incoming European groups, the orders Rodentia and Apatotheria, the marsupial family Didelphidae (Paleocene records no longer upheld: Gheerbrant 1991) and the pantodont genus Coryphodon, are thought to have their origin in North America, because of distinctly earlier appearances there (Gingerich 1989; Rose 1981). Interchange was probably via land bridges connecting Greenland to each continent (McKenna 1983). The new fauna in Europe is often termed the Hyracotherium-Coryphodon fauna, after the dominant elements in old collections, and is taken to characterize the 'Sparnacian Stage', but there are problems with this definition (see below).

Biostratigraphic problems In 1987, at the International Symposium on Mammalian Biostratigraphy and Palaeoecology of the European Paleogene, in Mainz, a mammalian biochronology was established for the Paleogene of Europe (Brunet et al. 1987). It consists of numbered units with the prefix MP. Workers are unanimous that MP6 is Paleocene and MP10 is Eocene on the criteria of any of the main organisms used to define the Paleocene-Eocene boundary (i.e. planktonic or benthic forams, calcareous nannoplankton, dinocysts, mammals). MP7-9 lie in a transition zone, with the major mammalian faunal turnover between MP6 and MP7. There is currently poor biostratigraphic resolution within the important MP7-MP9 interval mainly for two reasons. Firstly, although workers have normally

From Knox, R. W. O'B., Corfield, R. M. & Dunay, R. E. (eds), 1996, Correlation of the Early Paleogene in Northwest Europe, Geological Society Special Publication No. 101, pp. 205-218.

205

206

J.J. HOOKER

accepted that the fauna from the Paris Basin locality of Mutigny is older than that of nearby Avenay, the consensus opinion in Mainz was that the differences were minor and that the combined faunas should be designated MP8+MP9 (Godinot in Brunet et al. 1987). Secondly, the Paris Basin localities of Pourcy and Meudon were placed in MPT, but the former was already known to have yielded several MP8+MP9-defining taxa and the latter was later shown to yield such taxa too (Russell et al. 1988) (see Hooker 1991). Russell et al. (1982b, fig. 2) show the location of all the major mammal localities of late Paleocene and early Eocene age in the London, Belgian and Paris Basins. The Paris Basin has the largest cluster, but despite this the superpositional evidence for faunal succession here is extremely difficult to find. This is because of a combination of rapid lateral facies change and poor exposure. Nevertheless, certain superpositional evidence is well established. C o r y p h o d o n , whose earliest occurrence is in MP7, was recorded in shelly lignitic sands and clays stratigraphically well above the Conglom6rat de Cernay, yielding the main MP6 fauna, in the small outlier of Mont de Berru (Dep6ret 1906). The Sables ~t Unios et T6r6dines, which at several sites in the vicinity of Epernay yield MP 10 faunas, consistently overlie an 'argile ~t lignites' facies, which in its upper part yields the MP8+MP9 fauna of Mutigny (Feugueur 1963; Riveline 1984). A complication is that the MP8+MP9 Avenay fauna occurs in sands of Sables Unios et T6r6dines facies, immediately overlying the argile ~t lignites. Michaux (1964), however, considered that the lithofacies at Avenay was subtly

different from that of typical Sables ~t Unios et T6r6dines.

Biostratigraphic solutions Methods

To avoid circular reasoning in considering the age relationships of these MP7-MP9 faunas, I have subjected them to parsimony analysis, and subsequently examined the available evidence for superposition to assess its support or otherwise for the analysis. Alroy (1992) has introduced the use of parsimony into taxonomic distributional studies. His statistical method involves the distinction between overlapping (conjunct) ranges and nonoverlapping (disjunct) ranges, together with the creation of hypothetical distributional spaces to overcome the inaccuracies caused by absent records ('apparent disjunctions') due to taphonomic or collecting biases. I have instead used a program called Phylogenetic Analysis Using Parsimony (PAUP 3.0: Swofford 1990). This program is much employed in phylogenetic analysis, but has been adopted for ecological analysis too (Lambshead & Patterson 1986). It avoids the need for hypothetical distributional spaces by simply expressing 'apparent disjunctions' as homoplasies. In the data matrix (Table 1), in contrast to a phylogenetic analysis, the locality names take the place of taxa, and taxa (numbered) take the place of characters, as in an ecological analysis. Only taxa that occur in more than one and fewer than all the localities have been used, as it is the principle of shared taxa that is being applied in order to relate the localities. A

Table 1. Data matrix of taxa and localities 00000000011111111112222222222333333333344444444 12345678901234567890123456789012345678901234567 Sezanne-B. Conde-en-Brie Avenay Mutigny Pourcy Abbey Wood Soissons Meudon Dormaal Suffolk P.B. Erquelinnes Try Berru

00000000000000000000001011000011000001001111111 00000000000000000000011111000101001101111111111 00000000000000000000111111000101111111011111110 00010000000000101101110111001011110111111100000 00100000111011100001110110111111111111000000000 00000000110000100000010100111111000000000000000 000000010110000100?0000101000000000000000000000 00010001111000010011111110000000000000000000000 000001110011110Ill00000000000000000000000000000 00001111101111100000000000000000000000000000000 00001101011000000000000000000000000000000000000 11101111110000000000000000000000000000000000000 11110000000000000000000000000000000000000000000

Localities span MP6-MP9. Taxa are restricted to those which occur in more than one and fewer than all localities within the MP7-MP9 group. Numbers attached to taxa relate to those listed in Fig. 3.

MAMMALIAN BIOSTRATIGRAPHY AND PALEOCENE--EOCENE BOUNDARY

taxon occurring at only one locality would simulate an autapomorphy in phylogenetic analysis and thus would not aid the analysis, but misleadingly increase the consistency index. Thus, the relationships between localities are established on the basis of taxa shared amongst them, minimizing the number of 'apparent disjunctions' that need to be invoked (i.e. it identifies the most parsimonious pathway linking localities). The localities are grouped into a tree, which is subsequently rooted by selection of one or more localities known to be stratigraphically the oldest (i.e. by outgroup). In this analysis, the site of Berm is used as the outgroup. This is justified because Berru together with Cernay are MP6 sites within the Sables de Rilly of the Mont de Berru outlier, which have been demonstrated to be stratigraphically below an MP7-MP9 fauna (Dep6ret 1906). Use of the Dollo-up character type in PAUP 3.0 (Swofford 1990, pp. 9-12) is essential since it ensures that all homoplasy takes the form of reversals, preventing a taxon from originating more than once in parallel. Thus, a synapomorphy simulates an origination and a reversal simulates an extinction. More than one reversal of the same taxon on different branches indicates either a local extinction or a collection failure due to taphonomic or methodological bias, within the total range of that taxon (i.e. = 'apparent disjunction'). Choice of taxa or taxonomic rank depended largely on whether there had been a recent revision and to an extent on reliability of occurrence. For instance, tillodonts were recently revised by Baudry (1992), but the occurrence of each species is so sporadic that they have been lumped here as Esthonychidae. Carnivores have been omitted. MP7-MP9 multituberculates are only partially described and have thus not been included in the analysis. Lophiodon is dealt with at genus level, at which it is readily recognizable, but its species require extensive revision (Marandat 1987). The new Meudon fauna is undescribed and I here rely on the published list (Russell et al. 1988). Clearly, much taxonomic work remains to be done and future additions to faunal lists will improve resolution.

Results Analysis, by means of a branch-and-bound search, of 13 localities and 47 taxa results in three maximum parsimony trees, each with 125 steps. The consistency index excluding uninformative taxa is 0.371. The successive nesting of the crown localities Stzanne-Broyes, Condt-en-Brie, Avenay and Mutigny, respectively, and the pairing of Pourcy and Abbey Wood at the next lower node are constant in all. Dormaal branches off at a node

207

above the Suffolk Pebble Beds in one tree, but the two form sister localities in the other two. Soissons is the most unstable, being relatively poorly represented faunally. It is the sister locality to Meudon in two trees, but sister locality to Erquelinnes in the third. An Adams consensus of the three trees shows Soissons and Meudon on the one hand and Dormaal and the Suffolk Pebble Beds on the other as forming trichotomies with the respective crown groups (Fig. l a). Analysis of the same taxa, but omitting the Soissons locality, results in four maximum parsimony trees each with 116 steps. The consistency index excluding uninformative taxa is 0.400. The only differences between them are that the relationship of Dormaal and the Suffolk Pebble Beds varies as in the original analysis and that in two trees Pourcy branches off at a node higher than Abbey Wood, the two forming sister localities in the other two. The Adams consensus is shown in Fig. lb. Removal of Soissons has slightly destabilized the relationship between Abbey Wood and Pourcy. An Adams consensus of 16 trees including all those of 116, 117 and 118 steps is shown in Fig. lc. It shows the three following locality pairs, Pourcy and Abbey Wood, Dormaal and the Suffolk Pebble Beds, and Erquelinnes and Try, as forming trichotomies with the respective crown groups, indicating the relative weakness of the evidence for the hierarchy of these localities in the maximum parsimony trees. Table 2 is a chart of the mammal occurrences used in the analysis plus some MP10 ones linking St Agnan with the Sables ~t Units et Ttrtdines localities in the vicinity of Epernay. It can be used as a range chart provided that it is recognized that the order of the Suffolk Pebble Beds and Dormaal on the one hand and of Meudon and Soissons on the other is interchangeable. The chart shows that Coryphodon does not typify the entire MP7-MP9 span, but becomes extinct part way through the sequence. The genus Hyracotherium has in the past been used in a grade sense for almost any primitive horse-like perissodactyl. These are here segregated amongst the genera Cymbalophus, Pliolophus, Propachynolophus, 'Propachynolophus" and Hyracotherium s.s. (see Hooker 1994b). The implied time order of species of the plesiadapid Platychoerops coincides with the order of nodes in an independent cladistic character analysis recently conducted on this genus (Hooker 1994a), suggesting that it is as important biostratigraphically as its Paleocene precursor, Plesiadapis (Gingerich 1976).

Biozonation To attempt some objectivity in constructing a biozonation from these data, I have summed the last

208

J.J. HOOKER

a

SEZANNE-BROYES CONDE-EN-BRIE AVENAY MUTIGNY POURCY ABBEY WOOD SOISSONS MEUDON DORMAAL S U F F O L K P.B. ERQUELINNES TRY BERRU

I

m

b

C

I

I

SEZANNE-BROYES CONDE-EN-BRIE AVENAY MUTIGNY POURCY ABBEY WOOD MEUDON DORMAAL SUFFOLK P.B. ERQUELINNES TRY BERRU SEZANNE-BROYES CONDE-EN-BRIE AVENAY MUTIGNY POURCY ABBEY WOOD MEUDON DORMAAL S U F F O L K P.B. ERQUELINNES TRY BERRU

Fig. 1. Adams consensus trees derived from analysis of the data matrix in Table 1. (a) From three maximum parsimony trees obtained from the full data matrix; (b) from four maximum parsimony trees excluding the Soissons locality; (c) from 16 trees of 116, 117 and 118 steps excluding Soissons.

occurrences of one locality and first occurrences of the next to obtain a turnover figure between each successive pair (Fig. 2). There would be little change if the order of the Suffolk Pebble Beds and Dormaal were reversed or if Soissons were deleted. These peaks also coincide with the more robust nodes of the cladograms. I have chosen the peaks to determine the zone boundaries. The zones are concurrent range zones, named below and numbered for ease of reference as PE (for Paleocene-Eocene) I-V. They differ from the MP system in that they are biozones (sensu Hedberg 1976), not reference levels (sensu Thaler 1966), and they apply to N W Europe (onshore North Sea Basin) only, not the whole of Europe. Overlapping first appearance datum (FAD) and last

appearance datum (LAD) are indicated in the definitions.

P E I - Platychoerops georgei-Cymbalophus cuniculus C o n c u r r e n t R a n g e Z o n e

Definition: Total range of Platychoerops georgei, coincident with that of one or more of the following: Cymbalophus cuniculus, Atria cf. junnei and Microparamys nanus. The zone can also be recognized by concurrence of Teilhardina belgica, Cantius eppsi and Coryphodon (FAD) with Pleuraspidotherium aumonieri and Orthaspidotherium edwardsi (LAD), provided that the last two taxa are truly contemporaneous (see

Table 2. Occurrence chart for main localities ranging from MP6 to MP IO in the Paris, London and Belgian Basins

1. 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. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47.

Pleuraspidotherium aumonieri Orthaspidotherium edwardsi Plesiadapis remensis Berruvius lasseroni et cf. Cymbalophus cuniculus Platychoerops georgei Arfia cf. junnei Teilhardina belgica Cantius eppsi Coryphodon Paschatherium dolloi Microparamys nanus Microhyus musculus Landenodon woutersi Hyopsodus wardi Palaeonictis gigantea Peratherium constans Amphiperatherium brabantense Platychoerops russelli Hyracotherium aft. leporinum Apatemys sigogneaui et cf. Paramys ageiensis et cf. Peratherium matronense Neomatronella Amphiperatherium maximum Peradectes louisi Palaeonictis cf. occidentaIis Pliolophus vulpiceps Arcius fuscus Microparamys russelli s.s. Phenacodus lemoinei Esthonychidae

Bunophorus cappettai Placentidens lotus Microparamys chandoni Platychoerops daubrei Apatemys mutiniacus Diacodexis varleti et cf. Lophiaspis maurettei Peradectes mutigniensis 'Propachynolophus' maldani et aft. Cantius savagei Propachynolophus levei Arcius lapparenti Donrussellia gallica Amphiperatherium bourdellense Lophiodon Nannopithex zuccolae Ailuravus michauxi Buxolestes Propachynolophus gaudryi Cuisitherium lydekkeri Protodichobune oweni Platychoerops richardsonii

B

T

E

S

D

M

S

A

P

M

A

C

S

S

E

R

R

U

0

E

0

B

0

U

V

0

E

T

R

R

Y

Q

F

R

U

I

B

U

T

E

N

Z

A

A

X

X

X

X

X

X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X X

X

X X

G

X

X X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X X X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X

X X X

X X

X

X

X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X X

X

X X

X X

X

X

X

X

X

X

X

X

X

X

X X

X

X

x

X

X

X

X

X

X X

x

X

X

X

x

X

X

x

X

X

x

x

X

X

x

x

X

X

x

X

x

x x x x H

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

Numbered taxa are those used in the cladistic analysis (Table 1). Those unnumbered are additional MPI 0 taxa. H means unrecorded but occurs in higher strata. Palaeonictis gigantea and Coryphodon are added to the Soissons list as they are recorded from the Argile ~t Lignites du Soissonnais in the Soissons area, although not specifically from Soissons. Relevant taxonomic works are listed in Hooker (1991, p.77). More recent additions are: Baudry (1992), Gunnell & Gingerich (1991) and Hooker (1994a, b). Abbreviations: BER, Berru; ERQ, Erquelinnes; SUE Suffolk Pebble Beds localities (Kyson and Ferry Cliff); DOR, Dormaal; MEU, Meudon; SOI, Soissons; ABB, Abbey Wood; POU, Pourcy; MUT, Mutigny AVE, Avenay; CON, Condd-en-Brie; SEZ, Sdzanne-Broyes; STA, St Agnan; GRA, Grauves and other MPI 0 localities in the Epemay area in the Sables ~ Unios et Tdrddines.

210

J.J. HOOKER FA

GRAUVES etc

o

ST AG NAN

73 o3

SEZANNE-B.

PE III - Platychoerops daubrei-Cantius eppsi Concurrent Range Zone

LA

Definition: Coincident ranges of Palaeonictis cf. occidentalis and Pliolophus vulpiceps. Concurrence of Arcius fuscus, Phenacodus lemoinei, Microparamys russelli (s.s.), M. chancloni, Esthonychidae, Bunophorus cappettai, Diacodexis varleti, Placentidens lotus, Platychoerops daubrei and Apatemys mutiniacus (FAD) with Plesiadapis remensis, Cantius eppsi, Coryphodon, Paschatherium clolloi, Microhyus musculus and Landenodon woutersi (LAD). Reference localities: Pourcy (Marne), France

CONDE~n-BRIE AVENAY

4

3

MUTIGNY

4

6

POURCY

6

8

ABBEY WOOD

6

o

SOISSONS

1

3

MEUDON

~

o

DORMAAL

3

3

SUFFOLK P.B.

4

1

ERQU ELII~.'NES

1

0

TRY

6

2

BERRU

~

(

~

33

PTj MP6

40

_~0

'~0

fo

I.A+FA

(Falun de Pourcy, within the Argile h Lignites d'Epernay); Abbey Wood, England (Blackheath Beds); Harwich, England (Harwich Stone Band in Harwich Member, London Clay Formation).

Fig. 2. Graph of taxonomic turnover between sites as listed in Table 2. Turnover figures are obtained by summing last appearances (LA) at one site with first appearances (FA) at the next. That between Berru and Try includes the entire fauna of Berru; that between the other localities is derived from Table 2. seudoextinctions are not accounted for as it is the morphological change that is here considered important for biostratigraphic purposes.

below). FAD and acme of Paschatherium dolloi occurs in this zone. Reference localities: Try (Marne), France ('Conglom6rat h Coryphodon', at the base of the Marnes Blanches de Dormans); Erquelinnes, Belgium (Erquelinnes Sand Member, Landen Formation); Kyson/Ferry Cliff, England (Suffolk Pebble Beds); Dormaal, Belgium (Dormaal Sand Member, Landen Formation). (Note: only Platychoerops georgei occurs at all the localities.)

PE II -Platychoerops russeUi-Teilhardina

belgica Concurrent Range Zone Definition: Total range of Platychoerops russelli. Concurrence of Hyracotherium aft. leporinum, Apatemys sigogneaui, Paramys ageiensis, Peratherium matronense, Amphiperatherium maximum and Neomatronella (FAD) with Teilhardina belgica and Palaeonictis gigantea (LAD). Reference localities: Meudon (Paris), France (Conglom6rat de Meudon at the base of the Argile Plastique Bariol6e); Soissons (Aisne), France (Argile ~ Lignites du Soissonnais). Sinceny (Aisne) (Sables de Sinceny) may also belong to this zone although it contains none of the zonal indicators (see below).

PE IV - Cantius savagei-Arcius fuscus Concurrent Range Zone

Definition: Concurrence of 'Propachynolophus' aft. maldani, Lophiaspis maurettei, Peradectes mutigniensis and Cantius savagei (FAD) with Hyopsodus wardi, Peratherium constans, Amphiperatherium brabantense, Arcius fuscus and Hyracotherium aft. leporinum (LAD). Reference locality: Mutigny (Marne), France (near the top of the Argile h Lignites d'Epernay).

PE V -Donrussellia gallica-Apatemys sigogneaui Concurrent Range Zone

Definition: Total range of Propachynolophus level Concurrence of Arcius lapparenti, Donrussellia gallica and Amphiperatherium bourdellense (FAD) with Bunophorus cappettai, Apatemys sigogneaui, Placentidens lotus, *Peradectes mutigniensis, *Phenacodus lemoinei, *Lophiaspis maurettei, *Platychoerops daubrei, *Microparamys chandoni, *M. russelli (s.s.) and *Paramys ageiensis (LAD); and of Lophiodon (FAD) with the asterisked taxa only.

Reference localities: Avenay (Marne), France (Sables ~ Unios et T6r6dines); Cond6-en-Brie (Aisne), France (Sables de Cuise); S6zanne-Broyes (Marne) (Sables h Unios et T6r6dines).

Minor faunas Harwich.

The Harwich Stone Band within the Harwich Member of the London Clay (= Wrabness Member, Harwich Formation of Jolley 1996) has yielded Pliolophus vulpiceps (the holotype) which is restricted to PE III. Between Harwich and St Osyth, the holotype jaw of Coryphodon eocaenus

MAMMALIAN BIOSTRATIGRAPHY AND PALEOCENE--EOCENE BOUNDARY

was found by offshore dredging. Although no horizon is recorded for this specimen, dredging was in the last century used to mine the Harwich Stone Band for cement manufacture (Whitaker 1885, p. 17), so it is likely that the Coryphodon came from a nearby horizon within the Harwich Member.

Sinceny. Amongst the small fauna from the Sables de Sinceny, only Pliolophus aft. vulpiceps (see Hooker 1994b) and Dipsalidictis cf. transiens (see Gunnell & Gingerich 1991, p. 177) give a clue to the age. P aft. vulpiceps otherwise occurs only at Soissons, whilst D. cf. transiens otherwise occurs in Europe only at Meudon, both PE II (personal observation).

Evidence for superposition Paris Basin The Avenay (PE V) fauna occurs in Sables h Unios et T6rddines facies immediately overlying Argile Lignites facies (Guernet 1981); the Mutigny (PE IV) fauna occurs 3 m below the top of the Argile Lignites (Riveline 1984, p. 145-146) at a lateral distance of 1.5 km from the Avenay quarry. The Pourcy (PE III) fauna occurs in a sandy coquina (falun) within Argile ~ Lignites facies (Laurain & Barta 1985, p. 42-43) 13 km from the Mutigny locality and so is impossible to stratify with respect to the latter. The best way of demonstrating mammal succession seems to be by means of associated dinocyst and charophyte zonal taxa, whose succession is documented. Thus, Pourcy (PE III) has yielded two species of dinocyst (D. E. Russell, pers. comm.) that, according to Powell (1992), occur no earlier than zone D7B (= W5), the highest dinocyst zone recorded from the Argile a Lignites facies of the Montagne de Reims area (i.e. at Verzenay and Mailly; Gruas-Cavagnetto et al. 1980). On this basis, it is likely to be close in level to Mutigny. The suggested partial reworking of the Pourcy fauna, thus giving it an overall older aspect (Cavelier 1987, p. 263-264) is a possibility, but the rolled appearance of the isolated teeth is insufficient evidence on its own and should in any case affect only certain faunal elements. Mutigny (PE IV) has yielded the charophyte Peckichara piveteaui (see Riveline 1984). This has also been recorded in the upper part of the Argile Lignites of the Fosse-Parisis quarry at Mt Bernon, Epernay (Grambast 1977), where, in a borehole, a different species, P disermas, occurs throughout the underlying Marnes Blanches du Mt Bernon (Riveline 1984). P. disermas also occurs in the Cendrier and Argile Plastique Bariolde at Passy in

211

the Paris area (Riveline 1984) just above the horizon with the Meudon mammal fauna (PE II). At Soissons, the Argile ~ Lignites with the PE II mammal fauna near the top (apparently in the Sables h Paludines at the Grande Sdminaire pit; de Lapparent 1939) are overlain by the Sables de Sinceny (Faluns Sableux) with a diversity of dinocysts of the genus Apectodinium. In the overlying Falun Supdrieur (Argiles ~t Cyrbnes et Huitres), Apectodinium dominates the dinocyst assemblage (Bignot et al. 1981). The same stratigraphic distribution of dinocysts is present at Sinceny (Gruas-Cavagnetto 1968, p. 21-22), where two mammals uniquely shared with the PE II Soissons locality occur in the Sables de Sinceny. At neither locality is there any sign of Wetzeliella, which occurs at Mt Bernon near the base of the Argile ~ Lignites (namely, W. meckelfeldensis, a D6B indicator; Laurain et al. 1983); so these strata at Soissons and Sinceny must pre-date D6. The dinocysts therefore demonstrate that PE II is below PE III. Try (PE I) is the most difficult site to relate stratigraphically. Its fauna is presumed to have come from the vertebrate-rich 'conglomdrat ~t Coryphodon', although, apart from this taxon, the remaining elements were found in quarry spoil (Louis et al. 1983). Suggestions by these authors of mixed ages for the fauna were based on the association of three Cernaysian and two Sparnacian taxa. One of the former, Plesiadapis tricuspidens, has now been reidentified as Platychoerops georgei (Hooker 1994a) and typifies PE I. Bearing in mind the rarity of mammaliferous horizons and the consistency of preservation type in the assemblage, it seems equally likely that the fauna is homogeneous and that the two remaining MP6 representatives (Pleuraspidotherium aumonieri and Orthaspidotherium edwardsi), as well as a champsosaur (D. E. Russell, pers. comm.), are survivors from an earlier time. A final resolution to the problem can only come from recollecting in situ. Whichever the outcome, a PE I locality with MP6 survivors, or closely superposed MP6 and PE I faunas at the same site, Try has great biostratigraphic potential. The 'conglom6rat b, Coryphodon' is sandwiched between 15 m of Marnes Blanches de Dormans above and the marnes calcaires Paludina aspersa (a probable equivalent of the Calcaire de Rilly) below (Hdbert 1853; Feugueur 1963, p. 334). The Marnes Blanches are capped by Argile h Lignites with brackish molluscs (Hdbert 1853). The succession of thick white marls followed by lignitic shelly clays is similar to that documented at Mt Bernon (Laurain et al. 1983). The 'conglom6rat ~t Coryphodon' has been equated with the Conglom6rat de Meudon (Feugueur 1963) as it occurs at the base of a 'Sparnacian' succession.

, C~

Q

u~,..~.~ LAk x l 1 ~ '-series'-33 -39~ Grey Clay

fJll llrl

i

Holmehus Form. , .

Fig. 2. Stratigraphy and age of the Danish ash-bearing deposits and their North Sea equivalents (based on information from Knox 1984; Heilmann-Clausen 1985, 1988; Heilmann-Clausen et al. 1985; Knox & Morton 1988). Only NP Zones (Martini 1971) marked with an asterisk are identified as biozones in Denmark. Other NP Zones are suggested from dinoflagellate correlation between Denmark and other West European areas and on an ash layer correlation (see main text). Hatched area indicates the interval in which the Paleocene-Eocene boundary occurs. Chronology after Aubry et al. (1988) and Swisher & Knox (1993).

278

B. SCHMITZ ET AL.

the Fur Formation in northwestern Denmark and the Olst Formation in the southeast (HeilmannClausen et al. 1985; Nielsen & Heilmann-Clausen 1988). The former is c. 60 m thick and contains about 170 ash layers of predominantly basaltic composition (Pedersen et al. 1975; Pedersen & Surlyk 1983). The Olst Formation contains the same ash layers with a further six at its base. This formation varies in thickness, usually from 9-29 m (Heilmann-Clausen et al. 1985). In Denmark the ash layers of volcanic phases 2a and 2b have been numbered consecutively from -39 to +140 (BCggild 1918; Gry 1941). Biostratigraphy for the non-calcareous sediments underlying the RCsna~s Clay Formation relies on dinoflagellate and ash-layer correlations (Fig. 2). Relying on an ash layer correlation, the boundary between nannoplankton zones NP9 and NP10 is probably close to the base of the Fur Formation (Nielsen & Heilmann-Clausen 1988). An ash layer identical to the Danish l a y e r - 1 7 is identified in the lowermost part of Zone NP10 (c. 8 m above the base of the zone) in DSDP Hole 550 (Goban Spur, SW of Ireland) (Knox 1984; Aubry 1995; Berggren & Aubry 1996). The Fur Formation was probably deposited at a higher sedimentation rate, and the base of the formation is located 13 or 14 m below the ash layer -17. By means of 4~ dating an age of c. 54.5 Ma has been obtained on single feldspar crystals from layer -17 (Swisher & Knox 1993). A nannoplankton (NP) zonation for the noncalcareous Danish deposits is suggested by means of dinoflagellate correlation between Denmark and other West European areas (Heilmann-Clausen 1985) (Fig. 2). For example, the Danish dinoflagellate Zone 6 (the acme of Apectodinium spp.) is correlated with the Woolwich Beds in southern England. At one locality the basal Upnor Formation (formerly the Woolwich Bottom Bed) yields calcareous nannoplankton indicating Zone NP9 (Siesser et al. 1987). Dinoflagellate Zone 6 also correlates with the lower part of the Alpine 'Association ~t W. homomorpha' (Jan du Chine et aL 1975), of which the base is in the lower part of Zone NP9. The middle part of the 'zone W. hyperacantha' defined in the Pyrenees also correlates with dinoflagellate Zone 6 and is suggested to be restricted to Zone NP9 (Caro 1973; Caro et al. 1975). In conclusion, there is clear evidence that the base of Danish dinoflagellate Zone 6 occurs in Zone NP9 and is possibly confined to it. The RCsn~es Clay Formation consists mainly of very fine-grained, reddish or yellowish brown, calcareous clay (Heilmann-Clausen et al. 1985). It has previously been assigned to the early Eocene nannoplankton zones NPI1 and NP12 (Thiede

et al. 1980; Heilmann-Clausen et al. 1985). The

formation occurs throughout the distribution area of the Danish Eocene. A comparable fine-grained red-brown clay unit also occurs over extensive areas of the North Sea and in northern Germany. In northern Germany, south of a line connecting Bremen and Hamburg, the red-brown clay is laterally replaced by a more sandy, glauconitic facies, reflecting a position closer to land. The London Clay Formation in England and the leper Formation in Belgium represent more landproximal sediments that formed in the North Sea Basin at the same time as the RCsna~s Clay. These formations are more coarse-grained and about an order of magnitude thicker than the Rcsn~es Clay Formation. The thickness of the Rcsna~s Clay Formation in Denmark varies between 3.25 and 20 m. The formation has previously been divided into seven distinct beds (Heilmann-Clausen et al. 1985). The lowermost bed constitutes the Knudshoved Member, which is restricted to the western Limfjord area. The remaining six beds have been named consecutively, from the base upwards, R1 to R6. Some of the beds are quite thin but, despite this, show only minor lateral variation over wide areas. A detailed description of the subdivision of the RCsna~s Clay Formation is given in Heilmann-Clausen et al. (1985). The Rcsn~es Clay Formation grades upwards into the mainly grey-green, non-calcareous and extremely finegrained Lilleb~elt Clay Formation (Fig. 2).

The Paleocene-Eocene boundary in Denmark The placement of the Paleocene-Eocene (P-E) boundary in the sedimentary record of Denmark must await the completion of the ongoing international revision of the definition of this boundary. Since the introduction of the term Paleocene in the nineteenth century, the position of the P-E boundary has been controversial. Various stratigraphic levels have been suggested in the classical Paleogene sections of northern France, Belgium and southern England. These levels include the base of the Sparnacian in northern France, the base of the Ieper Clay in Belgium and the base of the London Clay/Oldhaven Beds in southern England. A biostratigraphic correlation of these boundary levels to deposits in Denmark was made by Heilmann-Clausen (1985) on the basis of dinoflagellates. Thus, the base of the Sparnacian coincides with the base of the Danish dinoflagellate zone 6, which is at the base of the Olst Formation (Fig. 2). The base of the Ieper Clay coincides with the base of the Wetzeliella astra zone (De Coninck

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK

1990), which in Denmark occurs at the base of the Knudshoved Member (Fig. 2). The base of the Oldhaven Beds is diachronous according to Knox et al. (1983), and seems to span strata ageequivalent to the Danish dinoflagellate zone 7. Hence the various suggestions for the P-E boundary in NW Europe may be traced to Denmark where they correlate with either the base or the top, or some intermediate level, of the r Formation. On the other hand, if the P-E boundary is placed at the level of the global benthonic foraminiferal extinction event and associated stable isotopic shifts (see Kennett & Stott 1991; Berggren & Aubry 1996), this would mean that the boundary in Denmark lies somewhere in the lower part of the Olst Formation or in the upper Grey Clay (Fig. 2). Here we will consider the P-E boundary as corresponding to the NP9-NPI0 boundary (see, for example, Berggren et al. 1985; Hardenbol 1994), which in Denmark lies close to the base of the Fur Formation.

The late Paleocene--early Eocene section at Alb~ek Hoved RCsnces Clay and Olst Formations For this study we have collected samples from the RCsna~s Clay Formation in ice-rafted clay mounds at Alba~k Hoved (55042 ' N, 9058' E) on eastern central Jutland. Here the RCsna~s Clay Formation is particularly expanded and well exposed (HeilmannClausen 1990). In the basal part of the mounds there are a few metres of the non-calcareous Holmehus Formation (Fig. 3). A hiatus, probably representing a substantial amount of time, separates the Holmehus Formation from the overlying ashrich Olst Formation. The Olst Formation is only 2 m thick at Alba~k Hoved and contains about 40 of the ash layers (+79 to + 118) that were deposited during volcanic subphase 2b (Fig. 2). The RCsna~s Clay Formation overlies the ~lst Formation disconformably at this site and attains a thickness of c. 20 m. The lowermost member of the RCsna~s Clay Formation, the Knudshoved Member (Fig. 2), is missing, indicating the presence of a hiatus. In NW Denmark the non-calcareous Knudshoved Member is only about 4 m thick (HeilmannClausen et al. 1985). The member is predominantly made up of silty clay, implying that it was deposited relatively rapidly. The biostratigraphic succession across the RCsna~s Clay Formation at Alba~k Hoved has recently been studied in detail (Heilmann-Clausen, pers. comm.). The most important results as regards nannoplankton zonation, foraminiferal distribution, and local Danish dinoflagellate zonation are summarized in later sections and in Fig. 3.

279

Both the detailed nannoplankton and the dinoflagellate (Fig. 3) stratigraphies indicate that the RCsn~es Clay Formation at Alba~k Hoved is relatively complete over the interval from lower NP11 to the base of NP13. This stratigraphic interval corresponds to a time period of somewhat more than 2 million years, using the timescale of Aubry et al. (1988). The 20 m thick section thus appears very condensed, having formed at sedimentation rates as low as c. 1 cm per thousand years.

The +11 to +16 m interval o f the RCsnces Clay Formation The interval from 11-16 m above the base of the RCsna~s Clay Formation was studied for highresolution isotopic distribution because some prominent lithological 'event-beds' occur in this interval (Fig. 4). The major part of the interval contains abundant, well-preserved foraminifera. In the lower 3 m of the interval there is highly calcareous, light reddish brown clay. Above this follows 2 m of calcareous, light greenish grey clay (Heilmann-Clausen etal. 1985). Heilmann-Clausen (1990) identified three distinct lithological beds in the interval, numbered in ascending order from 6 to 8. Layer 6 is a 10 cm thick ash layer (no. V16 in Fig. 3) below a white clay layer, c. 15 cm thick. The white clay occurs c. 10-20 cm above the ash and may reflect a separate event. We therefore identify the ash as no. 6a and the white layer as no. 6b. Layer 7 is a prominent 20 cm thick white clay layer. Layer 8b is a black clay layer, extremely rich in dinoflagellates and other organic materials. The layer probably formed in connection with a plankton bloom. It is an important marker bed, that occurs also at other sites in Denmark where the RCsna~s Clay has been found. The layer is similar in appearance to the Cretaceous-Tertiary boundary clay (the Fish Clay) at Stevns Klint (Schmitz 1988). In the interval just below layer 8b there are abundant pyrite spherules similar to those that occur in the Fish Clay (Schmitz et al. 1988). We analysed samples from layer 8b with the Iridium Coincidence Spectrometer at the Lawrence Berkeley Laboratory (see Schmitz et al. 1991) and found 0.4 ppb Ir. This is a prominent enrichment compared to average shale (< 0.05 ppb Ir), but orders of magnitude lower than the Ir content of the Fish Clay at Stevns Klint (c. l l 0 p p b Ir; Schmitz 1988). The small Ir enrichment in layer 8b probably represents volcanic Ir that has been remobilized from surrounding sediments. The adjacent sediments are also generally high in Ir (0.05-0.2 ppb), because they contain reworked volcanic Ir-rich ash from the earlier explosive basaltic volcanism (see Schmitz 1994). Layers 8a and 8c are less distinct

280

B. SCHMITZ ETAL. AG[ FORMATIONI NP I DINO

AS~ES I zo.E ZONE

~T

.

.

.

~13 C

.

I

.

.

.

~18 O

.

.

INP13!

+18 ,

+16

c~o~eo c.

__I

NP12

D

NP11

=

.

1,.-.

.

.

.

.

.

.

.

.

.

/J'"

12~ .- . . . .

.

.

.

--,.-o-

o~

.

L e n t i c u l i n a spp. :- O r i d o r s a l i s u m b o n a t u s = S u b b o t i n a spp.

NP8

.._i

.._.1 O 'l-

Fig. 3. Stratigraphy and foraminiferal stable-isotopic results for clay section at Albmk Hoved. Dinoflagellate zones refer to local Danish zonation (Heilmann-Clausen 1985, 1988); NP zonation refers to Martini (1971). Units A to G refer to planktonic/benthonic ratio variations described in Fig. 17. In the 2 m-thick Olst Formation 44 ash layers have been found, whereas the overlying RCsnaes Clay Formation contains about 20 ashes.

beds, and are characterized by streaks of black organic-rich clay. The sediments surrounding layers 8a-8c are light grey to whitish grey. The white colour is a weathering phenomenon, and occurs only at the surface of the outcrop. At a few centimetres depth the sediment is instead darker grey, indicating high organic matter content. The same applies for the white layers 6b and 7.

Isotopic studies: materials and methods Samples for stable isotopic studies were collected at 0.1-0.5 m intervals through the major part of the RCsnms Clay Formation at Albmk Hove& Many of the samples, however, could not be used because of insufficient calcite content. In addition, a highresolution isotopic profile was established for the interval from 11-16 m above the base of the formation. The major part of this interval was sampled at 5 cm resolution, and some parts at 1 0 c m resolution. The R0sn~es Clay at Albmk Hoved appears generally in a very fresh, unweathered state at some centimetres depth below the outcrop surface.

Erosion by sea waves constantly exposes new clay. Tree and plant roots, a feature characteristic of many weathered land sections, are generally absent. The isotopic analyses of the low-resolution profile spanning the major part of the R0snms Clay Formation were performed on picked foraminiferal samples and on bulk-rock samples. The former included monospecific assemblages of the benthonic foraminifera Oridorsalis umbonatus and of planktonic Subbotina spp., respectively (all or most of the Subbotina spp. belong to Subbotina ex gr. linaperta, i.e. Subbotina patagonica of Gradstein et al. 1992). In addition, we picked monogeneric samples of the benthonic foraminifera Lenticulina spp. We chose O. umbonatus because it has a very long range in the RCsnms Clay Formation. Lenticulina spp. also occurs throughout most of the section, and has the additional advantage of being thick-shelled, which facilitates examination for possible diagenetic alteration and infillings. For isotopic analysis, 0 . 3 - 0 . 4 m g foraminiferal calcite was used. We used about 30 individuals of O. umbonatus in the size range 125250 ~tm and 60-80 individuals of Subbotina spp. in

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK

281

powdered rock were separated for isotopic analyses. For the high-resolution profile in the +11 to +16m interval we analysed assemblages of the benthonic foraminifera Cibicidoides ungerianus and planktonic Subbotina spp. For each analysis, we used three to seven individuals of C. ungerianus, which corresponds to 0.05-0.1 mg calcite, and ten to 15 individuals (0.04-0.09 mg calcite) of Subbotina spp. (185-250 ~tm size). Isotopic analyses were also performed on assemblages of O. umbonatus (125-250 ktm) and Subbotina spp. (185-250mm) from the early Eocene part of DSDP Hole 550. For these analyses small samples (0.03-0.08 mg) were used. Foraminifera were examined in detail with the light microscope for calcite infillings and overgrowths. Only clean, well-preserved individuals were used for establishing the isotopic records presented here (see next section). The isotopic analyses of foraminifera from the profile spanning the major part of the ROsna~s Clay Formation were performed according to standard procedures using a VG Micromass 903E mass spectrometer at the Department of Marine Geology in Grteborg (see Schmitz et al. 1992). (From a few levels, smaller foraminiferal samples, 0.030.05 mg calcite, were analysed with a Finnigan Mat 251 at the Institute of Energy Technology, Kjeller, Norway; Table 1.) The foraminiferal samples of the high-resolution profile were analysed with a VG Prism Series II mass spectrometer installed at the Department of Marine Geology in Grteborg in November 1993. Bulk-rock samples from the Rcsmes Clay Formation and DSDP 550 foraminiferal samples were analysed with a VG Prism mass spectrometer at the Department of Earth Sciences in Oxford. Intercalibration between the laboratories was attained by analyses of NBS standards 19 and 20. All isotopic data are given as 6 per mil values normalized v. the PDB standard.

Preservation of foraminifera Fig. 4. Lithology of the interval from 11-16 m above the base of the ROsn~esClay Formation. For highresolution stable isotopic profiles, see Figs 9-12. The numbering of the event beds refer to that introduced by Heilmann-Clausen (1990).

the size range 185-250 ktm. For the analyses of Lenticulina spp., six to eight individuals (250500 ktm large) were used. For the bulk-rock isotopic analyses 3-4 g of dried sediment was ground in an agate mortar. After calcite determination by EDTA titration, aliquots of 0.3-0.5 mg of

The effects of diagenetic processes may obscure original isotopic signals in fossil foraminiferal tests. The most common problems are: (1) replacement of original test calcite; (2) isotopic exchange and re-equilibration between test and pore water; (3) secondary infillings and/or encrustations attached to original tests (Barrera et al. 1987; Barrera & Huber 1990). Prior to analysis, foraminifera from all sampled levels were examined under the binocular microscope for possible diagenetic alteration. In addition, tests from some levels were studied by scanning electron microscope (SEM) (Fig. 5). We examined the outer surface and, after breaking the tests, also

282

B.

SCHMITZ ET AL.

Table 1. Isotopic results (%0 v. PDB) for foraminiferal assemblages from ROsnces Clay Fm Metres above base

O. umbonatus ~513C

1.0 1.5 2.1 2.5 3.4 3.9 4.7 4.7 4.8 4.9 5.2 5.5 5.7 5.9 6.1 6.4 6.6 6.7 6.9 7.2 7.4 8.5 9.0 9.2 9.2 9.7 9.7 10.2 10.7 11.2

~18O

-2.70 -0.81

-2.03 -2.12

-0.73 -1.92 -1,8" -1.17

-1.78 -2.03 -3.1" -1.64

-1.13 -1.15 -1.6" -1.13 -1.22 -1.46 -1.24 -0.97 -0.65 -0.98 -0.72

-1.74 -1.50 -1.8" -1.61 -1.49 -1.74 -1.61 -1.34 -1.08 -1.37 -1.42

-0.84

-1.34

-0.60 -0.89

-1.29 -1.43

11.7

-0.7*

-1.7*

12.2 12.7 13.2 13.4 13.5 13.5 13.7 14.2 14.7 14.8 14.9 15.2 15.7 15.7 16.1 17.6

-0.67 -0.91 ~).56

-1.49 -1.57 -1.26

-1.2" -0.59 -0.9* 0.0" -0.9*

-1.34 -1.9" -1.7" -2.4*

-0.54

-1.44

Lenticulina spp. ~13C

~18O

-1.60 -1.50 -1.89 -0.63 -2.45 -2.91 -1.09 -0.67

-1.70 -1.65 -1.68 -1.57 -2.15 -2.55 -2.08 -1.94

-1.75 -1.8"

-2.21 -2.0*

-1.84

-2.11

-1.62

-1.83

-0.63 -1.61 -1.20 -0.79 -1.28 -1.23 -1.27 -1.72 -2.02 - 1.73 -1.17 -1.40 -0.74 -2.65 -2.65 -3.05 -0.91

-1.46 -1.95 -1.81 -1.84 -1.88 -1.98 -1.93 -2.25 -2.39 -2.26 -2.03 -2.20 -1.80 -2.73 -2.76 -2.83 -1.85

-0.60 -1.04 -1.87 -1.1" -1.22 -0.97 -2.30

-1.89 -2.19 -2.66 -1.8" -1.81 -1.87 -2.72

Subbotina spp. ~13C

~180

-0.64 +0.47

-4.12 -2.46

+0.82 +0.90

-2.17 -2.32

+1.19 +0.79 +1.24 +0.36 +0.29 +0.59 +0.48 +0.52 +0.76 +1.80 +0.41 +0.59 +0.85

-3.96 -3.63 -3.25 9-2.01 -3.64 -2.03 -2.13 -1.96 -2.65 -3.95 -1.93 -2.53 -2.92

+0.61

-2.07

+0.62 -0.35 +0.04 +0.10 +0.49 -0.04 +0.58 -0.33 -0.25

-1.78 -2.45 -2.41 -2.32 -2.25 -2.67 -2.12 -2.62 -2.87

+0.70 +1.12 +1.15 +0.46 +0.20 +1.11 +0.91

-2.57 -3.04 -2.40 -2.62 -2.81 -3.22 -3.21

+0.10 -0.19

-3.11 -3.08

* = Sample analysed at Institute of Energy Technology, Kjeller, Norway

their interiors. Throughout the major part o f the R0sn~es Clay Formation, from 4 m a b o v e its base to its top, the tests appear excellently preserved, and infillings or encrustations are absent (Fig. 5b, c, e & f). The benthonic foraminifera typically s h o w an enamel-shining surface. Pores and other minute

c h a m b e r surface features appear as new. Original growth-related textural features are present, such as the prismatic structure of Lenticulina walls (Fig. 5c), and growth-lamellae parallel to test surface in O. umbonatus (Fig. 5a & b). In fact, the foraminifera o f the R0sna~s Clay Formation in

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK

283

E

Q ~

;,,..= ~ +

"~+~ ~..=

9~

ej

~eq ID

~

e~

'-"

r.~ , - ~

--t

284

B. SCHMITZ ET AL.

general represent some of the best preserved early Paleogene specimens that can be found. Within the basal 4 m of the formation the tests also appear well-preserved, but in some specimens there are secondary, euhedral calcite crystals or encrustations on the interior chamber walls (Fig. 5d). In some cases thin layers of compressed coccoliths occur on the inner walls (Fig. 5a). The interior walls of tests devoid of overgrowths differ from those with secondary calcite, by being covered by a thin brown organic coating. This coating is easily detached from the wall, and most likely represents the remains of the soft tissue of the foraminifera All picked Lenticulina tests were individually broken so that their interior chamber walls were exposed. The test fragments were treated with ultrasound for 10-20 s, and thereafter examined in detail under the binocular microscope. Only fragments that were definitely free from secondary calcite were used for isotopic analyses. Many specimens of O. umbonatus, C. ungerianus and Subbotina from different levels were broken and examined for occurrence of secondary calcite. It was soon realized that secondary calcite is only present at levels where the Lenticulina specimens have infillings. Secondary calcite infillings were only observed at some levels in the lower 4 m, although we extensively searched for them throughout the entire formation. Isotopic results for O. umbonatus or Subbotina from levels where calcite infillings of Lenticulina tests were observed, have been disregarded.

Stable isotope records ~13C variations through the RCsnces Clay Formation

Both the planktonic and the benthonic foraminiferal 513C records through the RCsn~es Clay Formation show substantial, rapid and seemingly random fluctuations. The strongest variations are observed for Subbotina and Lenticulina, whereas Oridorsalis shows a more stable trend (Fig. 3; Table 1). The benthonics Lenticulina and Oridorsalis give relatively similar values throughout most of the section, but at some levels such as between +10 and +12 m relative to the base of the RCsna~s Clay Formation the two records diverge. At the maximum divergence in this interval, Lenticulina shows 1.3%o lower ~513C values than Oridorsalis. The latter shows a stable trend over the interval, whereas Lenticulina displays a gradual, negative 813C excursion. At 13.4m, where the distinct greyish-white layer 7 occurs (Fig. 4), there is also a pronounced divergence. Here Lenticulina shows

a dramatic negative ~13C excursion, reaching values as low as -3%o, whereas Oridorsalis only shows a small negative shift from values of c. -0.6 to -1.3%o. Most benthonic foraminifera do not secrete calcium carbonate in isotopic equilibrium with ambient pore or bottom waters (Woodruff et al. 1980; Graham et al. 1981; Vincent et al. 1981; Grossman 1987). So called 'vital effects' lead to isotopic fractionations, mostly resulting in depletions in the heavy isotopes. Incorporation of metabolic carbon-oxygen compounds in the calcium carbonate is the most likely explanation to these 'vital effects' (Grossman 1987). Different species show more or less strong fractionation. For carbon isotopes it has been shown that both Oridorsalis and Lenticulina strongly discriminate 13C relative to 12C when precipitating their calcite. Both typically show 13C depletions of 0.5-1.5%~ compared to ambient water (Grossman 1987). In general, the extent of isotopic fractionation in a benthonic foraminiferal species is constant, leading to parallelism in multiple monospecific isotopic curves through time. Microhabitat environmental differences, however, may distort this parallelism (Grossman 1987). For example, below the sediment surface, bacterial oxidation of 13C-depleted organic matter may induce lower pore-water 513C values than at the seafloor. Infaunal species may therefore register more negative 513C values than epifaunal species (McCorkle et al. 1990). Recent findings show that O. umbonatus is a transitional infaunal taxon living in the sediment from 0 to c. 4 cm below the surface (Rathburn & Corliss 1994). Lenticulina show similar depth habitats, being either epifaunal or shallow infaunal, thriving at 1-2 cm depth in the sediment (Corliss & Chen 1988; Corliss 1991; Corliss & Fois 1991). Thomas & Shackleton (1996) found that Oridorsalis generally shows lower 513C values than typically epifaunal species, supporting a habitat at some depth. The great similarity in 813C values for Lenticulina and Oridorsalis at most levels in the ROsna~s Clay Formation, indicate that they fractionate carbon isotopes in a similar manner. It also indicates generally the same depth habitat in the sediment, or possibly different depth habitats, but absence of a vertical ~13C gradient in the uppermost sediment. The divergences between the two records at some levels reflect changes in depth habitat or 513C gradient. These changes may be primarily related to variations in the flux rate of organic matter to the seafloor. Subbotinid foraminifera lived in the deep part of the planktonic realm according to isotope depthranking approaches (Boersma & Premoli Silva 1983; Shackleton et al. 1985; Corfield & Cartlidge 1991). Subbotina ~513C values fluctuate consider-

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK ably between extreme values o f - 1 and 1.8%o (Fig. 3). The foraminifera specimens giving extreme values are excellently preserved, supporting that short-term isotopic variations occurred, possibly related to an unstable upper water mass. The subbotinids in the RCsn~es Clay Formation give about one to two per mil more positive 813C values than Oridorsalis, reflecting a well-developed, vertical negative 813C gradient between intermediate depths and the seafloor. Benthonic foraminifera data in this study indicate water depths of c. 600-1000 m during deposition of the RCsna~s Clay Formation. The generally great difference in 813C between Subbotina and Oridorsalis is consistent with such water depths. In shallow water (< 200 m) environments, the deep-dwelling planktonic foraminifera typically register similar 813C values as the bottom-dwelling foraminifera (Schmitz et al. 1992). At many places in the present-day ocean there is a vertical negative 813C gradient over the upper 500-1000 m in the water column (e.g. Kroopnick 1980, 1985). The vertical 813C profile in the North Sea during the early Eocene may have been similar to, for example, the present-day type carbon-isotopic profile at GEOSECS stations 212 and 213 at mid-

latitudes in the North Pacific (Kroopnick 1985; McNichol & Druffel 1992). At these stations, there is a gradual one to two %o negative 813C shift over the interval from where the subbotinids dwelled down to 600-1000 m depth. Decreasing 813C values with depth are related to the sinking and decomposition of ]3C-depleted organic matter (813C c. -25%o). During the decomposition of the organic material oxygen is utilized, thus the 813C depth profile in general parallels the oxygen curve (Kroopnick 1974, 1980, 1985). The water mass pH is inversely related to the amount of CO 2 and carbonic acid, which primarily derive from the decomposed organic detritus. Therefore, the 813C depth profile also gives an indication of how corrosive bottom waters were. In the North Sea in the early Eocene oxygen content and pH apparently decreased with depth, whereas CO 2 and carbonic acid increased. The oxygen minimum level probably lay just above the seafloor. This indicates that replenishment of deep water was generally slow, or that deep waters entering the North Sea were 'aged', with already low oxygen content and being corrosive. The Subbotina-Oridorsalis 813C difference (A~13C) fluctuates throughout the section (Fig. 6).

A 8~3C

v~ iii z

:~

w

z

I

NP12

I

A 8~80

~

S

y14--

Q r"r v.

NP11

-2

" ~

03 I

O uJ "J ._i

--~ NP81 -r" LLI _ _ ~ ~; O :l-

0

1

2

285

3

0

-1

Fig. 6. Subbotina-Oridorsalis A8]3C and A8180 variations through the RCsn~esClay Formation.

286

13. SCHMITZ ET AL.

The Ag13C values are generally higher in the interval from +4 m up to c. +9 m, than in the interval from +9 to +15.8 m. In the lower interval most of the A~13C values lie in the range 1.6-2.4%o, compared with 0.8-1.4%o in the upper interval. The higher A813C values in the lower interval reflect higher 813C values in Subbotina and lower in Oridorsalis compared to the upper interval. The higher ~513C values for Subbotina probably reflect that addition of decomposing organic material sinking from the surface was slower than when the upper interval formed. Thus, in the earlier interval, biological productivity in the upper water mass was low compared with later interval. During the earlier interval bottom waters were also lower in ~513C because of strong density stratification of the water mass. The renewal of bottom waters was slow and 13C-depleted carbon diffusing out from the sediment and from sinking plankton accumulated in the bottom water. The presence of corrosive, low-pH bottom waters in the North Sea during this period is also supported by the low planktonic/benthonic foraminifera ratios and low CaCO 3 content observed in the lower interval (see later sections). The smaller A~513C gradients in the upper interval can be explained by increasing

~3 (%)

addition of decomposing organic matter at middepth related to higher surface water productivity. Bottom water (Oridorsalis) ~13C values increase because of invigorated bottom water circulation. The reduction in Subbotina 513C values and the inferred increase in surface water productivity in the upper interval coincides with the major divergence between the Lenticulina and Oridorsalis 513C records, which may reflect a change in response to increased influx of organic matter to the seafloor. Possibly Lenticulina moved deeper into the sediment, whereas Oridorsalis stayed at or near the sediment-water interface. Bulk sample ~513C and CaCOa results are presented in Fig. 7. With respect to 8r3C values, the RCsn~es Clay Formation can be divided in three pans. In the lower 7 m of the formation 813C values generally lie in the interval -1.0 to 0%0. From c. +7 to +10 m, 513C values are higher, in the range +0.4 to +2. In the upper third of the section values are again lower, fluctuating in the range -0.2 to +0.6. Analyses of different aliquots of ground bulk samples from levels with low calcite content give sometimes widely scattered results, indicating inhomogeneties. In particular, in the samples with low calcite content, the 813C signal of

I

813C (bulk)

$18 0 (bulk)

~ 0...4

.._t~

I

~,.eo

g tl,---,O

H M

"El e

e.e J.e

I

",'r--"

o.~-. N

_C Z-"

le

I

4i

I

25

50

75

-1

0

+T~

-2

j

-1

-2

-3

-4

-5

Fig. 7. CaCO3 content and stable isotopic composition of bulk samples through the RCsn~esClay Formation. (For additional information, see also Figs 3 and 17.)

287

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK

heterogeneously distributed, diagenetic and authigenic calcite may sometimes overprint the nannofossil-calcite isotopic signal. In general, bulk-sample calcite is primarily made up of coccoliths. Coccolithophores grow and calcify primarily in the upper 100-200 m of the oceans. In the present-day Pacific, around the equator and north of 30~ ~ N, they are most abundant within the upper 50-100 m of the water column (Okada & Honjo 1973). In the midlatitudes the standing crop is more or less evenly distributed over the upper 200 m. The bulk-sample values most likely give a rough estimate of the average 813C composition of the water mass over the entire depth habitat of the coccolithophores, and therefore gives an idea of the isotopic composition closer to the surface than where the deep-dwelling Subbotina lived. This is supported by the oxygen isotopic results, discussed in the next section. The bulk-sample 813C values, however, are relatively similar to the Subbotina values. Usually, in the water column, there is a strong vertical 8~3C gradient between the surface and the base of the euphotic zone (Kroopnick 1974, 1980, 1985). Biological productivity in surface waters leads to transport of light 12C with decaying organic tissue from the surface to the deep. Reduced salinities or any other property of surface waters in the North Sea, however, may have suppressed productivity near the surface, leading to a reduction of the 813C gradient within the uppermost water mass. Wellpreserved nannofossil calcite usually registers the higher 813C values in the upper water mass (Margolis et al. 1975), and there is no strong reason to believe that influence of 'vital effects' has severely distorted the record. Considering the spread in bulk sample 813C values at some levels it

seems likely, however, that diagenesis has obscured original features of the record to some extent. The generally low 813C values in the lower 6 m, for example, may reflect diagenesis. Over the lower 4 m we found diagenetic 12C-enriched calcite infillings in foraminifera tests. The CaCO 3 content in the RCsn~es Clay Formation correlates with planktonic/bentonic foraminifera ratios. Based on planktonic/benthonic ratios, the RCsn~es Clay Formation has been divided into seven units, A-G (Fig. 7 and later section). Units with low planktonic/benthonic foraminifera ratios, such as unit D, also show low calcite content. This indicates that both parameters are related to calcite dissolution rates in the water column. The reason for the large 813C fluctuations in the foraminiferal assemblages in the RCsn~es Clay Formation is not clear. Present-day northwestern European freshwater is enriched in light carbon, usually showing 813C values between-9 and-13%o (Mook 1968). In The Netherlands, at the southern shore of the present North Sea, estuarine waters with salinities of c. 31 ppt, show 813C values two to three per rail lower than fully marine waters off the coast. One possibility is that the fluctuations in 813C in the R~sn~es Clay Formation are partly related to variations in organic carbon influx from rivers and estuaries. There is, however, no clear positive correlation between 813C and 8180 in support of this view (Fig. 8).

~ 8 0 variations through the RCsnces Clay Formation The foraminiferal oxygen isotopic records also show considerable, rapid fluctuations in a

O. umbonatus

Subbotina spp. []

%.

B 9 [] mm nnn 9

[] []

[] []

-,z

[]

[]

[]

mm

9

[]

[]

[]

in[]

oo -3

[]

m []

-2

[]

[]

[]

I

-3 Fig. 8. 813C versus

I

-2 5180

I

I|

813C

I

"I

0

'I

-1

I

I

0 813C

correlation plots for O. umbonatus and Subbotina spp., respectively.

288

B. SCHMITZ ET AL.

seemingly random fashion (Fig. 3; Table 1). The isotopic results most likely reflect rapidly changing conditions in a marine-marginal, semi-enclosed basin. It is possible that minor salinity changes, and not temperature changes, are the prime cause of the isotopic variations. Throughout the section there appears to be a relatively constant isotopic offset between Lenticulina and Oridorsalis. Lenticulina generally gives c. 0.3-0.8%~ more negative values than Oridorsalis, although in a few cases the values coincide. The pattern is consistent with the data about foraminiferal 'isotopic behaviour' compiled by Grossman (1987), suggesting that Lenticulina may record lower 8180 values than Oridorsalis due to 'vital effects'. Detailed studies by Vincent et al. (1981) and Graham et al. (1981), show that Oridorsalis precipitates calcite close to oxygen isotopic equilibrium with ambient water. Belanger et al. (1981), however, present results indicating disequilibrium precipitation. Based on extensive studies of fossil Oridorsalis, Shackleton et al. (1984) suggest that this genus is close to equilibrium for 8180, but erratic for 813C. Both the Oridorsalis and the Lenticulina records are relatively stable over the interval from +8 to +13 m, but below and above this the values fluctuate a great deal. In the lower interval Oridorsalis values fluctuate between -1.8 and -3.1%o. The oxygen isotopic record for Subbotina spp. shows even greater fluctuations than the benthonic records. The Subbotina values range between -1.8 and -4. l%o, with the most negative values and the most dramatic fluctuations registered between +4 and +8 m. The larger isotopic variations for Subbotina most likely reflect more unstable conditions at mid-depth than at the seafloor. The fluctuations may, however, also to some extent reflect foraminiferal blooms at different seasons or variations in the depth of test precipitation (Bouvier-Soumagnac & Duplessy 1985; Corfield & Cartlidge 1991). Throughout most of the section there is a strong Subbotina-Oridorsalis 8180 difference (A8180; Fig. 6). The A8180 gradient fluctuates considerably without apparent systematic trend, and varies between -0.3 and -2.6%o, with most values in the interval -0.6 to -1.5%o. This mid-depth to bottom 8180 gradient indicates 2-6~ lower temperatures or 1-2 ppt more saline waters at the seafloor than at mid-depth. Alternatively, a mixed effect of decreasing temperature and increasing salinity with depth may account for the 8180 gradient. The oxygen isotope values in the RCsn~es Clay Formation are lower than most published 8180 values for contemporaneous fully marine environments. Oridorsalis 8180 values lie in the range

-1.1 to -3.1%~, whereas Subbotina varies between -1.8 and -4.1%~. If these data are interpreted in terms of palaeotemperatures, and assuming that fully marine conditions prevailed and that Eocene sea water had a 8180 composition of-1.28%o (icefree world; Shackleton & Kennett 1975), this would imply that bottom and subsurface waters had temperatures in the range 16-25~ and 19-30~ respectively. These calculated high temperatures may, however, be artefacts due to lower salinities. In the present North Atlantic a 1%~ reduction in salinity corresponds to a 0.6%~ decrease in 8180 (Craig & Gordon 1965), which is equivalent to a 2-3~ increase in temperature. The lowest Oridorsalis 8180 values, in the range -2 to -3.1, most likely reflect temporary salinity reductions rather than unusually warm water masses. Subbotina temperatures of 30~ appear unrealistically high, indicating that very negative 8180 values reflect a freshwater admixture. The majority of the bulk samples give 8180 values in the range -3 to -5%o, which is very low. In open marine environments, such as at DSDP Leg 74 sites in the southern Atlantic, early Paleogene bulk samples give 8180 values typically in the range -0.8 to 0.6%0 (Shackleton & Hall 1984). Bulk-sample 8180 values at Alb~ek Hoved are generally one to two per mil lower than those measured in Subbotina. Most likely the coccoliths, which dominate the bulk-rock calcite, preferentially formed somewhat higher up in the water column, thus registering lower salinities than the Subbotina tests. Coccolithophores have a tolerance for somewhat reduced salinities (Okada & Honjo 1973). Alternatively, the low values could be related to disequilibrium isotopic fractionation during coccolith formation. Dudley et al. (1986) have grown different species of coccolithophores in the laboratory and show that there are profound species effects in the uptake of oxygen isotopes. Some species precipitate calcite enriched 1%,~ in 180, whereas others are depleted as much as 2.5%o relative equilibrium. It is not likely, however, that the coccolithophore flora was dominated by 18tdiscriminating species to the extent that it can explain the strong 180 depletions throughout the R0sna~s Clay Formation. Low salinity in the upper 200 m of the water mass is also consistent with other findings in this study, such as Subbotina low 8180 values and the absence of surface-dwelling Morozovella foraminifera (see later section). Meteoric diagenesis may also lead to low 8180 values. Although diagenesis may explain some features of the bulk-sample 8180 record, it is not likely that diagenesis accounts for a several per mil negative shift in 8t80 for all the calcite throughout the section. Only in sections that have been strongly affected by diagenesis do such major

289

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK shifts occur. In sections with very negative diagenetic 8180 values it is impossible, or very difficult, to find the kind of well-preserved foraminifera that characterize the Rcsn~es Clay Formation. The fact that Buchardt (1978) measured 8180 values as low as -4.7%0 in well preserved mollusc shells formed at 30-150 m depth in the Eocene North Sea, gives further support for reduced salinities.

High-resolution 613C and 6180 profiles in +11 to +16 m interval In the +11 to +16 m interval both the oxygen and carbon isotopes show prominent negative excursions at the three major lithological event levels (Layers 6b, 7 and 8b; Figs 4, 9 and 10; Table 2). The 813C negative shifts are on the order of 1.5 to 2%0 for both Subbotina and C. ungerianus, which implies that the entire water column experienced

m

8180 ---..-a.--

.

~.

.

813C Subbotina

16-

C. ungerianus . .

i.

o

8c

8b 15

14[

~ 8a

13] ..........................................

iiiiiiiiiiiiiii

m

spp.

Subbotina

16

......

7

"

6u 6a

12

spp. I

11 15-

!

-1

-3

-4

Fig. 10. High-resolution 8180 profiles for Subbotina spp. and C. ungerianus through interval from 11-16 m above the base of the R0snees Clay Formation. For lithology and event-bed denotation, see Fig. 4.

14-

13-

12-

> 11-2

-2

-I

0

I

Fig. 9. High-resolution 813C profiles for Subbotina spp. and C. ungerianus through interval from 11-16 m above the base of the R0sn~es Clay Formation. For lithology and event-bed denotation, see Fig. 4.

dramatic changes in its carbon isotopic composition. The negative ~13C shifts are almost as large as the famous negative carbon isotopic (2 to 2.6%o) excursion at the P-E boundary (Kennet & Stott 1991). The Subbotina-C. ungerianus A813C gradients show no clear relation to the lithological event levels (Fig. 11). Instead, there is a gradual change, with A513C gradients generally increasing upward in the section. This change is also observed in the Subbotina-Oridorsalis A813C curve (see Fig. 6). As will be discussed later, the gradual increase in A813C is probably related to increasingly restricted water exchange between the North Sea and the open ocean, which is also reflected in the lithological change over the interval, from highly calcareous, reddish brown to calcareous, greenish light-grey clay (Fig. 4). The oxygen isotopic excursions at the lithological event beds are not as large as the carbon

290

B. SCHMITZ ET AL.

A~I3C

8c 8b 8a

7

6b 6a

r

0

'T

I"

!

1

2

Fig. 11. Subbotina-C. ungerianus high-resolution ASI3C profile through interval from 11-16 m above the base of the R0snres Clay Formation. For lithology and event-bed denotation, see Fig. 4.

this is not seen. Nor is there any clear positive correlation between 813C and 6180 (Fig. 13). The lithological event beds and associated isotopic shifts appear to reflect short-term (103-105 year) events when conditions in the North Sea were quite unusual. The chemistry of the entire water mass changed, which could only have happened over a time period longer than the residence time of the water mass in the North Sea Basin. If unrestricted water exchange existed with the Norwegian-Greenland Sea to the north (and perhaps with the Arctic Ocean) a much larger water mass was involved. Thus the isotopic shifts may reflect large-scale regional palaeoceanographic events. In the newly rifted area between Greenland and Norway, there may have been small semienclosed basins with unusual water-mass chemistry. The isotopic shifts in the upper RCsnres Clay Formation may reflect that connections evolved between any of these basins and the North Sea, leading to mixing of two different water masses. Alternatively, the excursions may reflect intermittent water exchange with a semi-isolated Arctic Ocean, with unusual water-mass chemistry. The three lithological event beds are all characterized by a dark grey or black colour in unweathered condition. We think that the dark colour reflects high organic matter content (and/or occurrence of monosulfides, related to reduced oxygen concentrations). In Layer 8b the high organic-matter content is most likely related to a distinct, prominent plankton bloom, whereas, for layers 6b and 7 increasing surface-water biological productivity in general can account for the reducing conditions in the sediments.

Comparisons with DSDP Hole 550 isotopic shifts. There seems to be short lags in the 8180 shifts with respect to the 813C shifts (Figs 9, 10), and the peak negative ~5180 values occur typically 5-10 cm above the lithological event beds The largest ~5180 excursion occurs at the black Layer 8b, where Subbotina 8180 values change from -2.5 to -4%o. (We have no data for benthonic foraminifera at this level, because of insufficient recovery.) At Layers 6b and 7 the shifts are smaller, in the range 0.5 to 1%,~.The 8s80 shifts are of the same size in the benthonics and the planktonics, showing that deep and bottom waters were influenced to the same extent. This is seen also in the A8180 values that do not increase in connection with the isotopic excursions (Fig. 12). A short-term episode of freshwater influx, from for example, river run-off, would have resulted in a larger 8180 change for Subbotina than for the benthOnic foraminifera, leading to an increase in A8180, but

The stable isotopic records for the early Eocene part of DSDP Hole 550 in the North Atlantic are presented in Fig. 14, and comparisons with the North Sea are presented in Figs 15-16. Hole 550 is located in the Goban Spur area southwest of Ireland (Fig.l). The present-day water depth is 4400 m. Palaeodepth backtracking indicates a depth of 4000-4300 m in the early Eocene (see Miller et al. 1985). Nannofossil data indicate a relatively continuous record in the lower Eocene at Hole 550 (Aubry 1995). For stratigraphic correlations we have used the NP Zone boundaries (this study, and Aubry 1995) as reference levels, and depth-age interpolations in the intervals between. The isotopic records for benthonic foraminifera at the two sites (Fig. 15), show that bottom-water chemistry was more stable at the abyssal DSDP Site 550, than in the semi-enclosed North Sea. This is natural considering the much larger volume and longer mixing time of the North Atlantic deep-

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK

29 1

Table 2. Isotopic results (%~ v. PDB) for high-resolution profiles 10.90-15.85 m above the base of RCsnces Clay Fm Metres above base

10.90 11.00 11.10 11.20 11.30 11.40 11.50 11.60 11.70 11.80 11.90 12.00 12.05 12.10 12.15 12.20 12.25 12.30 12.35 12.40 12.45 12.50 12.55 12.65 12.75 12.85 12.95 13.05 13.15 13.25 13.35

Subbotina spp.

C. ungerianus

~13C

~180

~13C

-0.285 -0.363 -0.093 +0.123 +0.348 +0.497 +0.266 -0.149 +0.123 +0.131 +0.581 +0.098 +0.160

-2.590 -2.612 -2.618 -2.534 -2.416 -2.370 -1.987 -2.278 -2.329 -2.434 -2.019 -2.376 -2.657

+0.621 +0.623 +0.726 +0.715 +0.462 +0.504 -0.118 -0.922 -0.359 -0.048 +0.136 +0.388 +1.019 +0.846 +0.541 +0.718 -0.274

-2.100 -2.202 -2.073 -2.233 -2.363 -2.417 -2.622 -2.825 -3.183 -2.912 -2.618 -2.192 -1.872 -2.224 -2.527 -2.425 -2.974

-0.135 --0.171 -0.492 -0.373 -0.091 -0.048 +0.231 -0.227 -0.448 -0.108 +0.129 -0.341 -0.357 +0.226 +0.048 +0.268 +0.103 +0.107 +0.051 +0.200 -0.148 -1.483 -0.943 -0.873 -0.449 -0.285 +0.293 +0.401 -0.048 +0.018 -1.286

81SO -2.127 -2.186 -2.042 -1.417 -2.261 -2.230 -2.160 -2.425 -2.522 -2.558 -2.192 -2.200 -2.541 -2.102 -2.242 -2.194 -2.216 -2.319 -2.128 -2.157 -2.398 -1.937 -2.864 -2.744 -2.465 -2.395 -1.955 -1.877 -2.473 -2.178 -2.708

water reservoir compared with North Sea bottom waters. During Biochron NP12 Oridorsalis show similar 813C values in the North Sea and in the northern Atlantic, however, in the N P I 1 interval the North Sea values are generally lower and more variable. Because of a core loss at DSDP Hole 550 comparisons are difficult across the upper NP12 interval. The benthonic 8180 values in the North Sea are throughout 0.6 to 2.5%0 lighter than in the northern Atlantic, with the greatest 6180 difference in the middle and upper N P l l interval, and a convergence near the NP 11-NP 12 boundary. The generally more negative 6180 values of bottom waters in the North Sea may reflect a combined salinity/temperature effect. The Subbotina isotopic records also show m u c h more stable trends at D S D P Hole 550 compared with the North Sea (Fig. 16). Unfortunately, Subbotina specimens are poorly preserved in the uppermost NP12 part (above the core loss) at DSDP Hole 550, which excludes detailed com-

Subbotina spp. Metres above base 13.45 13.55 13.65 13.75 13.85 13.95 14.05 14.15 14.25 14.35 14.45 14.55 14.65 14.70 14.75 14.85 14.90 15.00 15.05 15.10 15.15 15.20 15.25 15.30 15.35 15.40 15.45 15.55 15.65 15.85

813C

C. ungerianus

8180

~13C

~180

+0.065 +1.110 + 1.047 +1.361 +1.172 +0.590 +0.991 +1.441 +0.941 +1.329 +1.213 +1.204 +0.477 +1.078

-2.806 -2.473 -2.766 -2.636 -2.735 -2.045 -3.085 -3.441 -2.429 -3.046 -2.963 -2.495 -2.617 -2.939

-1.272 -0.275 -0.163 +0.610 +0.690 +0.760 +0.268

-2.817 -2.470 -2.069 -1.820 -1.960 -2.114 -2.225

+0.116 +0.214 +0.572 +1.157 +0.774 +0.291 -0.013 -0.168 -0.049 +0.337 +0.165 +0.305 +0.668 +1.133 +0.656

-2.981 -3.099 -2.832 -3.039 -2.361 -2.468 -2.820 -3.577 -4.100 -3.665 -3.505 -3.608 -3.198 -3.042 -3.195

+0.236 +0.403 +0.390 +0.083 -0.380 +0.621 -0.511 -0.616 -0.322 +0.104 +0.205 -0.706 -1.063

-2.271 -2.160 -2.170 -1.714 -1.311 -1.922 -2.441 -2.578 -2.734 -2.458 -2.485 -1.699 -1.891

parisons in this interval. In the lower NP12 part Subbotina 8180 values appear to be generally 1 to 2%0 lower in the North Sea than at DSDP Hole 550, whereas in the upper N P l l part the North Sea values are often as m u c h as 3 to 3.5 lower. The Subbotina records are of particular interest, because they probably reflect the isotopic composition of the water mass at similar depths. A s s u m i n g that water mass temperatures at Subbotina depths were the same at the two sites, the difference in 8180 between the two sites can be attributed to salinity differences alone. Palaeosalinity estimates using this approach indicate generally 2 to 3 ppt, and intermittently 5 to 6 ppt lower salinity at Subbotina depths in the North Sea than in the North Atlantic. The bulk sample 8180 values are 1 to 2 ppt lower than Subbotina values. Thus North Sea upper water mass (i.e. coccolithophore-depths) m a y have had salinities approximately ranging between 26 and 30 ppt. Latitudinal temperature gradients were extremely

292

B. SCHMITZ ETAL.

m

A~I80

16-

.......................................................................

8c 8b

15, ................

-

- .................

a 8a

low in the early Eocene, thus there is no reason to believe that the Subbotina ~180 difference between the two sites is related to temperature differences, although Alb~ek Hoved lies 600 km further to the north than DSDP Hole 550. A possibly complicating circumstance in the salinity reasoning, however, may be that Subbotina had adjusted to different depths in the North Sea and the North Atlantic (see Corfield & Cartlidge 1991), but there is no way to test whether this was the case.

Biostratigraphic results Foraminifera distribution

14-

13-" .......

....

6b 6a 12-

i

i -I

0

i

Fig. 12. Subbotina-C. ungerianus high-resolution ~]80 profile through interval from 11-16 m above the base of the R0sn~es Clay Formation. For lithology and eventbed denotation, see Fig. 4.

The foraminifera in the Rc~sna~s Clay Formation can be assigned to four broad groups: noncalcareous agglutinating, calcareous agglutinating, calcareous benthonic and planktonic foraminifera. There are marked vertical changes in the relative proportions of these groups, which can be used to subdivide the R0sn~es Clay Formation at Alb~ek Hoved into seven successive biostratigraphic units, here lettered A to G in ascending order (Fig. 17). These units can be identified in all sections of the ROsn+es Clay Formation throughout Denmark, and in subsurface sections in similar facies in the southern North Sea. The vertical assemblage changes are identified by plotting two percentages for each sample analysed: (1) percentage of planktonic foraminifera in the total foraminifera assemblage (P% of King 1989); and (2) percentage of benthonic foraminifera which are non-calcareous agglutinants (NCA% of King 1989). Non-calcareous foraminifera form 100% of the assemblage in the lowest fossiliferous interval

C. ungerianus

Subbotina spp. []

-2 [] []

~m

[]

9

I

In

m.~ []

[] ~ . r % , r 9

[]

9 oo

[]

-2

-3

&

[]

n

9 []B

D me I

[]

9 []

[]

m

m

[]

m

[] B

I

[]

-4

[]

[]

i|

I

-2

I

0

~513C

I

I

-i

!

[]

I

o

I

!

~13C

Fig. 13. 6J3C versus 5]80 correlation plots for C. ungerianus and Subbotina spp., respectively.

293

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK

I DEPTH mbsf i

NP ZONE

320-

NP 13

8 180

8 130

I

2

330 NP 12 340 -

350 1

NP 11

360 -2

-1

0

+1

+1 =

I

I

'

0

-1

-2

Subbotina spp. umbonatus

= O.

Fig. 14. Stable isotopic results for DSDP Hole 550.

NP Zone

(~ 1 8 0

8 13C /"

k\

l

12

q,,;P

Q:

\

I

I

!

I

-3

-2

-1

0

-

+1

,I

--9

b

I

I

0

-1

'

I

I

t

I

-2

-3

-4

-5

= DSDP 550 =NorthSea

Fig. 15. Comparison of benthonic foraminiferal isotopic results at DSDP Hole 550 and Alb~ek Hoved. Basal five data points for the North Sea are measured on Lenticulina spp. All other analyses on O. umbonatus.

294

B. S C H M I T Z E T A L .

8 130

8180

,/ k

/ 4 //

-3

-2

-1

0

+1 - ~ - =

+2

i +1

0

-1

-2

-3

.4

-5

DSDP 550 = North Sea

Fig. 16. Comparison of Subbotina isotopic results at DSDP Hole 550 and Albaek Hoved. (unit A). The low-diversity assemblage is characteristic of the 'Rhabdammina-biofacies" (King 1989), including Ammodiscus cretaceus, Glomospira charoides and Glomospirella sp. The 'Rhabdammina-biofacies' is characteristic of restricted (usually poorly oxygenated) carbonatepoor bathyal and abyssal environments in extratropical areas (see Charnock & Jones 1990). The assemblages in unit A resemble Late Cretaceous (Maastrichtian) assemblages (Kuhnt et al. 1989) interpreted as 'middle slope' (middle bathyal, 500-1500 m water depth). Their occurrence in brown clay with low organic carbon content at Alb~ek Hoved indicates adaption to an oxygenated environment. The absence of calcareous foraminifera in this unit is therefore probably due to deposition below the CCD. This unit can probably be correlated with Division A of the London Clay Formation in southern England (King 1981 ). In units B to G, NCA% is very low (generally < 1%), and assemblages are composed predominantly of calcareous benthonic and planktonic foraminifera. Foraminiferal abundance is moderately high (typically > 10 specimens/g of sediment), and the benthonic associations are of moderately high diversity (20-40 taxa/sample). The benthonic foraminiferal associations are relatively similar in composition throughout

units B to G. They are characterized by the consistent occurrence of Gaudryina hiltermanni, Anomalinoides aft. capitatus, Angulogerina abbreviata, Cibicidoides eocaenus, C. aft. ungerianus and Oridorsalis umbonatus. In addition to these, there are some common taxa that are relatively short-ranging and can be used in biostratigraphic correlation (see Fig. 17). Based on the data summarized in Morkhoven et al. (1986, fig. 6), the occurrence of Anomalinoides (aft.) capitatus, Gaudryina hiltermanni, Turrilina brevispira, Vaginulinopsis decorata and Valvulina haeringensis, which have their upper depth limit near the neritic/bathyal boundary (c. 200 m), indicates bathyal depths for the entire interval. Nuttalides truempyi (common in parts of units D and E) and Hanzawaia ammophila (present in unit G) have their upper depth limit within the upper bathyal zone (c. 500 m), while Aragonia aragonensis, which is frequent in a number of samples from unit B to unit G, has an upper depth limit in the lower bathyal zone (c. 1000m) according to the same source. Assessment of the overall assemblages seems to favour a middle bathyal environment (600-1000 m depth) for most of the interval. More precise data on absolute and relative depths of deposition are under study. The changes in the relative abundance of plank-

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK

295

Fig. 17. Distribution of some key foraminifera across the RCsn~es Clay Formation. Percentages of planktonic (P) and non-calcareous agglutinated (NCA) foraminifera. Palaeounits A to G represent a subdivision of the R0sn~es Clay Formation, based on variations in the relative proportions of these groups.

tonic foraminifera, used to define units C to G, are believed to reflect changes in water-depth and circulation patterns in the North Sea Basin, as well as extent of water exchange with the world ocean. The latter may be related both to sea level variations in the northeastern Atlantic region (King

1989), as well as to the tectonic history of gateways to the open ocean. The semi-enclosed geography of the basin in the early Cenozoic (Ziegler 1988, 1990; Fig. 1), led to restricted circulation in deep water environments with only intermittent exchange of water between the Basin and the North

296

B. SCHMITZ ET AL.

Atlantic Ocean. In late Paleocene to earliest Eocene times, during deposition of the Sele Formation, low sea levels, perhaps associated with tectonic uplift, led to the almost complete isolation of the North Sea Basin (Knox et al. 1981). The resulting anoxic seafloor environments led to the almost complete extinction of marine benthic faunas in all areas below wave-base. Sea levels began to rise in the earliest Eocene, with each significant rise in sealevel leading to inflow of oceanic waters and an associated influx of benthonic and planktonic foraminifera. Thus foraminifera evolving elsewhere were able to enter the North Sea Basin intermittently. Most biostratigraphic events (first occurrences) identified within the basin reflect episodes of sea level rise, or regional tectonic events leading to connections with the world ocean. Therefore many of the events may be significantly later than apparently correlative events in the open oceanic environments. Unit B is characterized by an influx of calcareous benthonic foraminifera, with very low proportions (< 10%) of planktonics. Non-calcareous agglutinated foraminifera continue from unit A, but comprise only c. 10% of the assemblage. The total faunal turnover at the base of this unit reflects a lowering of the CCD. It is believed on the basis of regional evidence to represent a sea level rise, and can be correlated with the base of Division B in the London Clay Formation. It marks the first major early Eocene influx of planktonic and benthonic taxa from open oceanic environments into the North Sea Basin. Unit B is very thin, indicating either a high degree of condensation, or the presence of a stratigraphic break between this unit and unit C. The base of unit C is marked by a major influx of planktonics, which comprise over 80% of the foraminiferal assemblage. The planktonics are dominated by Subbotina ex gr. linaperta and allied taxa. This influx can be identified at the base of Division C in England, and probably represents the highest sea-levels attained during the early Eocene. Subbotina-dominated planktonic assemblages of this age occur throughout the North Sea Basin (King 1989). The base of unit D is marked by an abrupt decrease in the proportion of planktonics, which comprise less than 10% of the assemblage through most of the unit. A correlative event can be identified throughout the North Sea Basin, reflecting possibly a major fall in sea level or a rise in CCD. Nuttalides truempyi appears abundantly within the upper part of this unit, reflecting a water depth in the southern North Sea greater than 500 m. There is a brief increase in NCA% at the top of unit D, which may prove to be of regional significance. The base of unit E is defined at an increase in P

to between 10-50%. The base of unit F is defined at a further rise in P to between 50-80%. At or just below this event, the first occurrence of the planktonic foraminifera Pseudohastigerina wilcoxensis and Planorotalites spp. are recorded. This event marks a major rise in sea level, which is well-documented from contemporaneous neritic environments, and correlates with the marine flooding events at the base of Division E in southern England (King 1981). The first occurrences of the benthic taxa Bulimina sp. A (King 1989) and Cancris sp. A (King 1989), within unit F, are coincident biostratigraphic events which can be recognized throughout the North Sea Basin (King 1989). The base of unit G is defined at a decrease in P to less than 50%. In the lowermost part of the succeeding Lilleb~elt Clay Formation, P falls to less than 10%, reflecting a fall in sea level marked at the basin margin by the abrupt replacement of open marine by marginal-marine sediments at this time (King 1981). This sea-level fall terminated deposition of the London Clay Formation in southern England.

Calcareous nannoplankton distribution The calcareous nannoflora of the RCsna~s Clay Formation at Alb~ek Hoved is highly diversified. Eighty-five species have been recognized, of which approximately 25% are useful for the biostratigraphic subdivision of the formation and for correlation within the Ypresian of the North Sea Basin (Fig. 18). The underlying Olst Formation and the overlying Lilleba~lt Clay Formation are decalcified, except for the lowermost 1 m of the latter, which presents poorly preserved, low to moderately diversified assemblages assignable to Zone NP13. In the present paper King subdivides the RCsn~es Clay Formation into seven biostratigraphic units, labeled A to G in ascending order, on the basis of vertical changes in the relative proportions of the four main groups of foraminifera (the noncalcareous agglutinating, the calcareous agglutinating, and the calcareous benthonic and planktonic forms). These units, except G, can also be differentiated through their calcareous nannoflora. The boundaries between units A to F correspond to substantial vertical changes in the nannofossil abundance or to special events in the vertical distribution of biostratigraphically important taxa (first or last occurrences, recurrences, base or top of peak abundances) (Fig. 18). They coincide with the boundaries between Steurbaut's nannoplanktonunits I to VI, defined in the proximal shallow-water facies of the southern North Sea Basin (Steurbaut

297

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK

'

~

! J

" ~ ' numbe~ ot specime= ",,..,~,~ - . . . . . . . , ~ ! ~. per mm "~ .t~L,r~u L,I~I~I~U[ m

E b

I [

2'5

'

~

7~s l& UNITS

'

UNITS ]

=

]

CALCAREOUS NANNOFOSSILS

I (1) Steurbaut, 1988 & 1991

,.a~

--1

I~u ,.,1

I i

--20

IX

~.

~ --18

.......(.IL _

VIIlb

[~

=

~

........................................

i .-~ [ l .

60

sa

--is

Z0

~t

]

~

~'i ~

~ ~=

12

44

-lo!

F

'VI

E

V

36 ~

4 D

~

C

~\ \ \ \ ~ ' ~

~

I -Ill ,

;;;iiiii "

~..1~1

1 IIIa2

,-,-., ~o

'

IVmb

--e

= -

.~

,

~

*

_It

I-

--14

--

~

VIIIa

Illal ~

[~

| "

9 ~ !.,~,",~ 9

--

0

1

~

,

...~2 ;.'.'." . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 'J.L'.'~

l

i

Fig. 18. Qualitative and quantitative distribution of calcareous nannofossils across the Rosn~es Clay and basal Lilleb~elt Clay Formations at Alb~ekHoved (sample numberings after Heilmann-Clausen, pers. comm.; position in metres above the base of the Rcsmes Clay Formation).

1988, 1991; Steurbaut & King 1994). This indicates that these changes in nannoflora have a basin-wide biostratigraphic significance and reflect variations in water exchange with the world ocean. The lowermost calcite-containing sample studied

(A1-2; standard numbering according to HeilmannClausen, pers. comm.) occurs 0.33 m above the base of the RCsnaes Clay Formation. This sample, which probably marks the base of unit A, is characterized by a rich nannoflora, with rather

298

B. SCHMITZ ET AL.

common Discoaster diastypus, D. binodosus, D. kuepperi and Tribrachiatus orthostylus. The presence of these species, together with the absence of Tribrachiatus nunni, T. contortus and Rhabdosphaera sola, indicate the lower part of NP Zone 11 (nanno-unit I of Steurbaut 1991). How far down in N P l l is difficult to say, as in the underlying sediments in the North Sea Basin there is a rather large interval (about 1 to 2 million years) without any calcareous nannofossils (NP10 has never been recorded). With reference to the Goban Spur region (off Ireland) and Spain, however, it appears that the lowermost R~snaes Clay Formation does not correlate with the basal part of NP11, but formed somewhat later (D. kuepperi does not occur in the basal part of NP 11). Unit A has a maximum thickness of 20 cm at Alba~k Hoved. The composition of its nannofossil assemblage (only sample A1-2 was recovered) suggests correlation with the upper part of nannounit I, which, in Belgium, is over 15 m thick. The base of unit A is believed to correspond to the lower boundary of the third order depositional sequence Y-C of Steurbaut, which has been identified in the upper part of the Orchies Clay Member in Belgium and in the upper part of Division A in England (Steurbaut, pers. comm.). Unit B lies approximately between 0.5 m and 0.8 m above the base of the R~sna~s Clay Formation. Only one sample has been recovered (A1-3). Its nannoflora is marked by the incoming of rare Rhabdosphaera sola, which was used to define the base of Steurbaut's nanno-unit II. The nannofossil assemblage of unit B is less diversified and less rich than that of the underlying unit A, although there are many similarities. This suggests that only minor fluctuations must have occurred in the chemistry and temperature of the North Sea surface waters at the start of the formation of unit B. The base of unit C (AI-4) is marked by a sharp rise in nannofossil abundance and by the incoming of the various species, among which the marker Chiphragmalithus calathus, which have not been recorded from the underlying strata. This major influx of nannofossils is known to occur at the base of nanno-unit IIIal. It has been identified within the Roubaix Clay and Mons-en-P6v~le Sand Members in Belgium, just above a major omission surface, representing the main flooding surface of the fourth order depositional sequence (sub-sequence) Y-D1 (Steurbaut & King 1994). Throughout unit C the nannofossil associations are moderately well-preserved, quantitatively rich (over 2000 specimens/mm2 glass-slide; see Steurbaut & King 1994, for details on the counting techniques) and of high diversity (35 to 40 taxa/ sample).

The nannofossil associations in unit D present a rather high, but variable degree of decalcification. The base of unit D is characterized by an abrupt decrease in nannofossil abundance (over 90%), by the end of the acme of Rhabdosphaera sola and by the disappearance of Nannifula dupuisii. These events, which can be identified throughout the North Sea Basin, mark the boundary between nanno-units IIIal and IIIa2 (see Fig. 18). The reason for this drastic fall in nannofossil abundance (?productivity) is not clear, although there is strong evidence for calcite dissolution. In Belgium this event occurs at a medium-grained glauconite horizon, which separates underlying calcareous, nummulitic silty fine sands from overlying less fossiliferous silty clays. Unit D can be subdivided in four sub-units on the basis of qualitative and quantitative differences in the nannoplankton associations. These of unit D1 are poorly preserved and of low diversity. They belong to nanno-unit IIIa2 (absence of common R. sola; presence of Neochiastozygus rosenkrantzii). Sub-unit D2 is marked by a 100% calcite dissolution. Sub-unit D3 contains partially dissolved, low-diversity assemblages (20 taxa/ sample). The first occurrence of Micrantholithus mirabilis lies within this interval. In the less decalcified assemblages from Belgium this event, which occurs at the base of the third order depositional sequence Y-E (Steurbaut, pers. comm.), is slightly posterior to the first consistent occurrence of Discoaster lodoensis, and, thus, falls within Zone NP12. Assemblages from D4 are less dissolved, and characterized by frequent M. mirabilis, Imperiaster obscurus and Pontosphaera sp. The first D. lodoensis are recorded at the base of this interval at Alb~ek Hoved. The overall aspect of the nanno-assemblages of units D3 and D4 refers to periods with more restricted conditions and lower salinities, occurring during periods of relatively low sea-level. The base of unit E is marked by an abrupt increase in nannofossil abundance, by the first occurrence (also acme) of Chiphragmalithus barbatus and by a first marked rise in abundance of D. lodoensis. In Belgium, these events are known to occur at the base of the fourth order depositional sequence Y-E2. The nannofossil associations of unit E show considerable calcite dissolution, except for the base and the top. The base of unit F is characterized by a major influx of nannofossils and by the last occurrence of Toweius pertusus, defining the base of nanno-unit VI. This coincides with the start of the peak abundance of D. lodoensis and the entry of very rare, atypical Rhabdosphaera crebra. These events, which mark a major rise in sea level, have been recorded at the base of the Aalbeke Clay Member

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK in Belgium and at the base of Division E in southern England. The quantitatively and qualitatively richest associations are recorded at 13.75 m above the base of the RCsn~es Clay Formation (Sample A1-49: c. 9000 specimens per m m 2 glass slide; 41 taxa). This level is marked by the start of the peak abundance of R. crebra, which defines the base of Steurbaut's nanno-unit VII, by the common occurrence of Helicosphaera seminulum, and by the incoming of the genus Scyphosphaera. The first occurrence of the benthonic foraminifera Bulimina sp. A (sensu King 1989) and Cancris sp. A (sensu King 1989) seem to be coincident with these events, which all can be recognized throughout the North Sea Basin. Higher up in the R~sn~es Clay Formation, between 14 m and 18 m above the base, the assemblages are moderately diversified (20 to 30 taxa/ sample), although partially dissolved at some levels. The main events within this interval are the first consistent occurrence of Discoaster cruciformis at 15.55 m and the first occurrence of Nannoturba robusta at 17.60 m, which respectively define the base of nanno-units Villa and VIIIb. In Belgium, nanno-units VII and VIII correlate with the shallow marine Egem Sand Member (see Steurbaut & Nolf 1986; and Steurbaut 1995), which has been formed during a sea-level lowstand. Just above volcanic ash-bed V18, at c. 2 m below the top of the RCsna~s Clay Formation, nannofossil abundance and diversity increase again (> 35 taxa).

299

This might reflect a next sea-level rise, after a substantial period of low sea levels. This rise is slightly posterior to the first occurrence of N. robusta and slightly prior to the last occurrence of Tribrachiatus orthostylus, and consequently falls within the extreme top of NP12. In Belgium it corresponds to the transgressive event at the base of the Hyon Sand Formation (Steurbaut & King 1994). Within this interval with progressively increasing nannofossil abundance and diversity, the highest values are recorded at the top of the ROsna~s Clay Formation (A1-67). In the lower part of the succeeding Lilleb~elt Clay Formation the nannofossil abundance falls to less than 10%, reflecting the general fall in sea level.

A semi-enclosed North Sea in the early Eocene The results in this study indicate that in the early Eocene North Sea the major characteristics of water mass chemistry, sedimentation and biology, and changes in these conditions, were determined by the extent of water exchange with the open ocean. Water exchange was controlled by regional sea level and tectonic changes affecting the width and depth of important gateways. We discern principally three different conditions (strongly, moderately and somewhat restricted water exchange) that existed alternatingly in the North Sea (Table 3). These conditions can be related to the division of

Table 3. The North Sea in the early Eocene - three principal conditions

Condition I: Strongly restricted water exchange with open ocean (very low regional sea level) * Strong calcite dissolution; calcite is absent or very rare in sediment. * Non-calcareous agglutinated foraminifera dominate. * Grey or greenish grey clays dominate; dark brown clays occur; low oxygen content in sediment.

Condition II: Moderately restricted water exchange with open ocean (low regional sea level) * Calcite dissolution; low to intermediate calcite content in sediment. * Impoverished planktonic foraminiferal assemblages; low planktonic/benthonic foraminifera ratios. * Surface salinities 4-8 ppt lower than in the open ocean; fluctuating and repeatedly very negative Subbotina 5180 values (-2 to -4%0, related to reduced salinities also at mid-depths. * High Subbotina-benthonic A513C gradients; low surface-water biological productivity and reduced supply of organic 12C at mid-depths; temporarily strong density stratification of water mass, bottom waters corrosive because of slow renewal rates. *Variable lithology; different varieties of reddish brown clays and greenish or whitish grey clays; generally low to intermediate oxygen content in sediment. Condition III: Somewhat restricted water exchange with open ocean (high regional sea level) * Calcite-rich sediments dominate. * Rich planktonic foraminiferal assemblages; high planktonic/benthonic foraminifera ratios. * Surface salinities a few per rail lower than in the open ocean; Subbotina 8180 values (c. -2 to -3%0 indicate generally higher salinities and more stable conditions at mid-depths compared with condition II. * Low Subbotina-benthonic AS13C gradients, reflecting high surface-water productivity as well as invigorated bottomwater circulation; less strong water mass density stratification; less corrosive bottom-water. * Homogeneous reddish brown clay or marl; relatively high oxygen content in bottom-water.

300

B. SCHMITZ ET AL.

the RCsn~es Clay Formation in the units A to G, which mainly reflects the variations in planktonic/ benthonic foraminifera ratios (Fig. 17). The same variations are observed at many sites in the southern North Sea (King 1989). According to King (1989) the variations in planktonic/ benthonic ratios can be explained by migration of planktonics via the Faeroe Trough or the English Channel into the North Sea in connection with high sea levels. The results in this study, however, show that instead the principal governing factor may have been variation in calcite dissolution rates. This is supported by the strong positive correlation between bulk-sample CaCO 3 content and planktonic/benthonic ratios in the Alb~ek Hoved section (Figs 7, 17). In the intervals with low planktonic/benthonic ratios nannofossils also show stronger influence of dissolution. Thus the planktonic/benthonic ratios may mainly reflect the position of the lysocline, which in the North Sea was related to the rate of water exchange with the open ocean. With slow rates of water mass replenishment, there was a CO 2 build-up in the North Sea, inducing low pH and leading to calcite dissolution and reduced planktonic/benthonic ratios. If increasing calcite dissolution is related to a reduction in water exchange with the open ocean this would be reflected in lower salinities and more negative 8180 values. Indeed, the calcite-poor unit D corresponds to the interval with fluctuating and unusually negative foraminiferal 5180 values (Figs 3 and 6). The isotopic comparisons with DSDP Hole 550 also indicate that unit D formed while the North Sea was more isolated (Figs 15 and 16). In the major part of unit F, where planktonic/ benthonic ratios and calcite content generally are high, 5180 values are more positive and the isotopic pattern is more stable than in unit D. However, in the upper part of unit F calcite concentrations become lower (Fig. 7). (This decrease coincides with a lithological change from reddish brown to whitish grey clays.) In this interval 5180 values are again more negative and show fluctuations similar to those in unit D. Most likely, the upper part of unit F represents the beginning of a return to conditions similar to those in unit D. Eventually this gradual change led to the deposition of the non-calcareous Lillebzelt Clay, reflecting a return to a strongly isolated North Sea. Unit E is intermediate and transitional between units D and F in 8180 behaviour, calcite content as well as planktonic/benthonic ratios. In the dissolution unit D the S u b b o t i n a Oridorsalis A813C values are generally about a factor 1.5 to 3 higher than in the more calcite-rich parts of unit F (Fig. 6). In the upper, less calcareous part of unit F, the A813C gradients gradually

decrease towards values similar to those in unit D. Unit E shows an intermediate range of A813C values. The higher vertical 813C gradients in unit D and upper unit F are related both to higher Subbotina 813C values and lower Oridorsalis 813C values. An increase in 813C for surface-dwelling foraminifera is generally interpreted in terms of an increase in productivity. With increasing productivity the light carbon isotope, 12C, is preferentially removed from the water mass into organic tissue of plankton. With increasing productivity in the euphotic zone, the downward flux of decomposing 12C-rich organic matter increased at mid-depth where Subbotina lived. Thus we interpret the higher Subbotina 813C values as reflecting reduced surface-water productivity. The overall sedimentological and chemical data indicate that high Subbotina 13C values coincide with increased isolation of the North Sea and possibly low sea levels. Basin compartmentalization and freshwater admixture induced a more rigid density-stratification. This inhibited surfacewater productivity because of a decrease in upwelling of nutrients. Slow replenishment of the bottom water led to accumulation of CO 2 and carbonic acid from organic matter, and intensification of the oxygen minimum zone near the sea floor. This also explains why benthonic 813C values are low when Subbotina 813C is high. The lowermost part of the RCsn~es Clay Formation in Denmark is non-calcareous. This applies for the Knudshoved Member (missing at Alba~k Hoved) and the 0.3 m interval below unit A at Alb~ek Hoved. In the dark brown clay of unit A at Alba~k Hoved only non-calcareous agglutinated foraminifera are present. Gradstein & Berggren (1981) show that water depth is not the principal factor determining the occurrence of agglutinated foraminifera. More crucial is instead the deposition of 'fine-grained organic-rich carbonate-poor clastics under somewhat restricted bottom water circulation in compartmental basins'. The agglutinated foraminifera of unit A may be relics from the latest Paleocene and earliest Eocene (NP10), when the North Sea was almost completely isolated. At this time organic-rich, noncalcareous sediments with agglutinated foraminifera were ubiquitous in the North Sea. Bottom water 'aging' and CO 2 build-up was even more pronounced than during deposition of calcitepoor unit D in the early Eocene. In Table 3 the three different principal conditions of the North Sea in the early Eocene are summarized. Condition I reflects strongly restricted water exchange with the open ocean. It is represented by the sediment of the lowermost RCsna~s Clay Formation and the Lilleb~elt Clay Formation that began to form at the end of the early Eocene

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK (Fig. 2). Condition II, with moderately restricted water exchange, typically prevailed in connection with deposition of unit D. In the upper part of unit F there is a gradual return to these conditions (Figs 6 and 11). Condition III, with more open water exchange, correlates with the major part of unit E Unit E represents a transitional state between conditions II and III. Unit C, below the low-calcareous unit D, is rich in calcite and shows high planktonic/benthonic foraminifera ratios. Possibly this unit reflects a period of more open water exchange with the world ocean before a return to more isolated conditions. Unfortunately, diagenetic calcite inflllings in the foraminifera prevented detailed reconstruction of water mass conditions at this time. Analyses of well preserved foraminifera from unit C may add important information on the relation between isolation of the North Sea and water-mass chemistry.

Palaeogeography of the North Sea and adjacent areas in the early Eocene The palaeogeographical reconstructions indicate that in the late Paleocene-early Eocene the North Sea and its northward extension, the embryonic Norwegian-Greenland Sea, were delimited from the North Atlantic to the west by the British Isles and plateau basalts between Scotland and Greenland (Fig. 1; Ziegler 1988; King 1989). Based on ridge subsidence calculations it has been argued that water exchange between the North Sea and the Atlantic across the Greenland-Scotland volcanic ridge did not begin until the middle Eocene (Thiede & Eldholm 1983). This is consistent with land mammal distributions (McKenna 1983) indicating that a land bridge existed between Scotland and Greenland until the middle Eocene. On the other hand, Berggren & Schnitker (1983) found that early Eocene plankton faunas and floras in the Norwegian-Greenland Sea and the North Sea were similar to those at the Rockall Plateau in the northeastern Atlantic. They suggest that already at the beginning of the Eocene there was substantial water exchange through the Faeroe Channel, which overlies continental crust and may be a graben structure. Hulsbos et al. (1989) noted that early Eocene calcareous plankton in the Norwegian Sea, the North Sea and the North Atlantic belong to the same biogeographical province. Because of poverty of fauna and flora in the Norwegian Sea, however, they argued that migration did not occur across the Greenland-Scotland Ridge, but through the epicontinental seas of NW Europe. Major water exchange at this time may have occurred along the line of the present English Channel (King 1989,

301

1993). Via the Norwegian-Greenland Sea to the north there may also have been a shallow water connection with the Arctic Ocean (Marincovich et al. 1990). This is supported by the similarity of post-Danian to early Eocene faunas of molluscs and ostracods in the Arctic Ocean with coeval North Sea Basin faunas. Thick sediment accumulations in the Lofoten Basin, west of northernmost Norway, indicate sediment contribution from the Barents Shelf, and an open connection also with the Barents Sea (Thiede et al. 1986). There may also have been a connection with the Tethys across North Germany and Poland into southern Russia (King 1989). The North Sea Basin and its northern extensions were surrounded by vast continental drainage areas (Fig. 1). Rivers from Greenland, Fennoscandia, Britain and central Europe all transported fresh water into this relatively small semi-enclosed sea. Despite this, the fossil microfauna and -flora throughout the R0sna~s Clay Formation reflects more or less typical marine conditions. The isotopic data also indicate almost fully marine conditions. The average salinity in the euphotic zone of the southern North Sea may have been around 2630 ppt through most of the early Eocene. These salinities are still high compared to recent seas like the Baltic Sea, Hudson Bay or Black Sea, that are similarly surrounded by vast land areas. Accordingly, a substantial ocean-water influx must have balanced the riverine input in the North Sea in the early Eocene. Probably saline waters entered continuously across epicontinental shallow water (< 200 m) connections, such as the vast Barents Shelf and the English Channel. However, deepwater connections, where the rich cosmopolitan, bathyal, benthonic foraminiferal fauna could enter, must have existed, at least intermittently. Oxygen isotopic data for the North Sea in the late Maastrichtian and early Danian (Schmitz et al. 1992), the early Selandian (Mattsson & Schmitz 1994) and the late Eocene/early Oligocene (Burman & Schmitz, unpublished data) indicate fully marine conditions (salinity c. 34 ppt) at these times. From the mollusc isotopic data of Buchardt (1978) it appears that, at least in the southern North Sea, reduced surface salinities existed during most of the Eocene. Most likely, the North Sea had unusually low salinities in the latest Paleocene, but the absence of biogenic calcite in sediments from this time makes this difficult to test. The reduced salinities measured at Alba~k Hoved may be representative of the entire North Sea, and possibly also the Norwegian-Greenland Sea (and perhaps the Arctic Ocean). Early Eocene microplankton assemblages at Alb~ek Hoved are similar to faunas and floras of other areas of the North Sea (Gradstein et al. 1992), indicating representative conditions at Alba~k Hoved. Based on studies of agglutinated

302

B. SCHMITZ E T AL.

foraminifera in ODP Hole 643 Kaminski et al. (1990) suggest a few per mil reduced salinity for surface waters in the Norwegian-Greenland Sea in the early Eocene. The RCsn~es Clay Formation at Alb~ek Hoved formed at considerable distance from any shore, which explains why it is very fine grained, resembling pelagic red clays. The closest shore was probably situated 300-400 km to the south in northern Germany. The benthonic foraminifera fauna as well as the stable isotope depth gradients indicate substantial water depths, possibly corresponding to a middle bathyal environment, which means depths in the range 600-1000m. The typically reddish brown colour of the sediment and neglible amounts of organic carbon (HeilmannClausen et al. 1985), indicate mostly oxygenated bottom conditions, which is consistent with voluminous inflow of ocean water. The decrease in benthonic 813C values in connection with episodes of calcite dissolution are indicative of higher CO 2 pressure and lower oxygen content in the bottom water at these times. Low oxygen content would have promoted formation of anoxic laminated clays, rather than red clays. The formation of red clays under such conditions, however, could have taken place if sedimentation rates were low. Such a general situation is indicated both by grain-size and the completeness of the biostratigraphic record. Moreover, although red-brown clays dominate, there is a large variation in the lithology across the RCsn~es Clay Formation. Some intervals indeed consist of greyish or laminated sediments, indicative of suboxic conditions. Pyrite occurs in many samples throughout the section. Probably the seafloor redox conditions balanced on the boundary between suboxic and oxic. Even a minor change in sedimentation rates or organic matter influx would have shifted sedimentation conditions in a new direction. The water flow across different sea ways connecting the North Sea with the open ocean may have varied with time, depending on palaeogeographical configurations and regional isostasy. In particular the isotopic profiles for Oridorsalis, but also for Subbotina, show a trend towards convergence with North Atlantic data in latest Biochron NP11 and early NP12. This indicates that the Faeroe and the English Channel connections may have been deepened or widened at this time, which may be related to subsidence of the British Isles and the Greenland-Scotland Ridge, or to a general sea-level rise in the northeastern Atlantic. In the middle and late Biochron NPll, during formation of unit D, water exchange with the North Atlantic may have been more restricted and other water passages may have been of more importance. The stable isotope composition at mid-depth as

well as in deep waters in the North Sea fluctuated significantly and depletions occurred in the heavy isotopes (13C and 180). The instability of the isotopic trends may reflect greater local or regional influence on the water-mass chemistry. Only little is known about the northward and the eastward connections of the North Sea. These connections may have been of greater relative importance for North Sea water chemistry at times when water exchange with the North Atlantic resumed. The sometimes very low 813C values compared with the North Atlantic, could reflect influx of CO 2- and nutrient-rich waters into the North Sea. Waters intermittently entering the North Sea from north via the Norwegian-Greenland Sea could also have been low in 813C, because of admixed magmatic CO 2 (813C c. -5%0) or organic carbon (813C c. -25%0) from river influx. There is no correlation between occurrences of volcanic ash in the RCsna~s Clay Formation and negative 813C, arguing against significant admixture of magmatic CO 2 with the water. Gas emissions, however, may have occurred in connection with episodes of flood basalt formation, which may not have given any clear imprints in the sedimentary record. The more negative 813C and 8180 values during middle and late Biochron N P l l compared with later could reflect a higher influence of water deriving from the Arctic Ocean or any newly evolved semienclosed basin to the north. Such basins and the semi-enclosed Arctic Ocean probably had water mass chemistries quite different from the world ocean (see Marincovich et al. 1990). The prominent negative isotopic excursions in late NP12 may be related to water exchange with such a semienclosed basins. A substantial influx of water from the eastern Tethys does not appear likely considering, for example, that keeled planktonic foraminifera, such as morozovellids, are absent from the RCsmes Clay Formation. These foraminifera occur at quite high latitudes on the eastern North Atlantic margin during the Eocene (King 1993). If voluminous eastern Tethys water masses entered the North Sea across Poland and Germany, one would expect to find in the North Sea abundant morozovellids, a characteristic element of Tethys waters. However, one may wonder why the morozovellids did not invade the North Sea from the North Atlantic across the English Channel. One possibility is that reduced surface salinities prevented the surfacedwelling morozovellids from invading.

Summary During the early Eocene, the North Sea was semienclosed with restricted contact with the open ocean. The benthonic foraminiferal fauna indicate

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK water depths of 6 0 0 - 1 0 0 0 m where present D e n m a r k lies. Throughout the early Eocene, conditions in the southern North Sea were essentially marine, but salinities were somewhat reduced. Average salinities in the upper 100 m of the water column may have fluctuated between 26 and 30 ppt. Water exchange with the open ocean determined the water mass properties in the North Sea, inducing principally three different types of sedimentation: non-calcareous, low-calcareous and calcareous. During periods of restricted water exchange, density stratification of low- and highsalinity waters was more rigid, inhibiting water circulation and effective replenishment of bottom waters. Low sea levels and associated basin compartmentalization added to this effect. Slow bottom-water renewal led to CO 2 build-up, more strongly developed vertical ~13C gradients between mid-depth and the seafloor and a rise in the lysocline. Biological productivity decreased due to reduced nutrient upwelling. Non-calcareous sediments occur in the lowermost part of the RCsna~s Clay Formation, reflecting restricted water exchange with the world ocean. This was followed in early Biochron NP11 by more open water exchange. Calcite-rich sediments with marine nannoplankton and Subbotina-dominated planktonic foraminiferal faunas began to form. In late N P l l , water exchange was somewhat restricted and calcite dissolution increased. For this interval, Subbotina 5180 trends are very unstable, with repeated positive excursions because of fresh water influence. Calcite-rich sedimentation began

303

again at the N P l l - N P 1 2 boundary and prevailed until late NP12. At this time Subbotina 5180 values are more positive with a more stable trend, indicating less freshwater admixture at mid-depth. At the end of the early Eocene there was a gradual return towards a more isolated North Sea. Three prominent lithological event beds occur in the upper NP12 part of the Rcsna~s Clay Formation. Associated with these beds are major negative shifts (1 to 2%,~) in carbon and oxygen isotopes. The chemistry of the entire water column changed, implying that important regional palaeoceanographic events are registered. It is possible that water masses entering the North Sea from the north (embryonic NorwegianGreenland Sea or Arctic Ocean) explain some of the differences in water mass chemistry between the North Sea and the Atlantic Ocean. The shortterm isotopic events in late N P I 2 may be related to rapid water exchange between the North Sea and other semi-enclosed basins with unusual water chemistry to the north. Financial support for this study was obtained from the Swedish Natural Science Research Council, The Bank of Sweden Tercentenary Foundation, and the Erna and Victor Hasselblad Foundation. Samples from DSDP Hole 550 were provided by the Deep Sea Drilling Project. We thank O. Gustafsson for isotopic analyses and E Asaro for iridium analyses. Laboratory assistance was provided by T. Alavi, T. Andinsson and J. Burman, and M. Eliasson did the artwork. Reviews by M.-P. Aubry, W. B. Berggren and E. Thomas greatly improved this paper. The paper is a contribution to IGCP Project 308.

References AUBRY, M.-E 1995. From chronology to stratigraphy: assessing the continuity of the stratigraphic record. In: BERGGREN,W. A., KENT, D. V., AUBRY,M.-E & HARDENBOL, J. (eds) Geochronology, Time Scales and Stratigraphic Correlations: Framework for an Historical Geology. Society for Economic Paleontologists and Mineralogists, Special Publication, 54. ~, BERGGREN, W. A., KENT, D. V., FLYNN, J. J., KLITGORD, K. D., OBRADOVICH,J. D. & PROTHERO, D. R. 1988. Paleogene geochronology: An integrated approach. Paleoceanography, 3, 707742. BARRERA,E. & HUBER,B. T. 1990. Evolution of Antarctic waters during the Maestrichtian: Foraminifer oxygen and carbon isotope ratios, Leg. 113. Proceedings of the Ocean Drilling Program, Scientific Results, 113, 813-827. - - , - - , SAVIN,S. M. & WEBB, P.-N. 1987. Antarctic marine temperatures: Late Campanian through early Paleocene. Paleoceanography, 2, 21-47. BELANGER,P. E., CURRY,W. B. & MATTHEWS,R. K. 1981. Core-top evaluation of benthic foraminiferal isotopic ratios for paleoceanographic interpretations,

Palaeogeography, Palaeoclimatology, Palaeoecology, 33, 205-220. BERCGREN, W. A., & AUBRY, M.-P. 1996. A late Paleocene-early Eocene NW European and North Sea magnetobiochronological correlation network. This volume. -& SCHNITKER,D. 1983. Cenozoic marine environments in the North Atlantic and NorwegianGreenland Sea. In: BOTT, M. H., SAXOV, S., TALWANI, M. & THIEDE, J. (eds) Structure and Development of the Greenland-Scotland RidgeNew Methods and Concepts. Plenum, New York, 495-548. - - - , KENT, D. V. & P-LYNN, J. J. 1985. Jurassic to Paleogene: Part 2. Paleogene geochronology and chronostratigraphy. In: SNELL~G, N. J. (ed.) The Chronology of the Geological Record. Geological Society, London, Memoir, 10, 141-195. BOERSMA, A. & PREMOLI SILVA, I. 1983. Paleocene planktonic foraminiferal biogeography and the paleoceanography of the Atlantic Ocean. Micropaleontology, 29, 355-381. BOGGILD,O. B. 1918. Den vulkanske Aske i Moleret samt en Oversigt over Danmarks ~eldreTerti~erbj~ergarter.

304

B. SCHMITZ ET AL.

Danmarks Geologiske UndersCgelse, Raekke 2, 33. BOUVIER-SOUMAGNAC, Y. & DUPLESSY, J.-C. 1985. Carbon and oxygen isotopic composition of planktonic foraminifera from laboratory culture, plankton tows and recent sediment: Implications for the reconstruction of paleoclimatic conditions and of the global carbon cycle. Journal of Foraminiferal Research, 15, 302-320. BUCHARDT, B. 1978. Oxygen isotope palaeotemperatures from the Tertiary period in the North Sea area. Nature, 275, 121-123. CARO, Y. 1973. Contribution h la connaisance des dinoflagellrs du Palrocene-Eocene des Pyrenfes espagnoles. Revista Espanola de Micropaleontologia, 5, 329-372. --, LUTERBACHER, H., PERCH-NIELSEN, K., PREMOLISILVA, I., RIEDEL, W. R. & SANFILIPPO, A., 1975. Zonations h l'aide de microfossiles prlagiques du Palrocbne suprrieur et de l'l~ocbne infrrieur. Bulletin de la Socidt~ G~ologique de France, S~rie 7, 17, 125-147. CHARNOCK, M. A. & JONES, R. W. 1990. Agglutinated foraminifera from the Palaeogene of the North Sea. In: HEMLEBEN, C., KAMINSKI, M. A., KUHNt, W. & ScoTT, D. B. (eds) Paleoecology, Biostratigraphy, Paleoceanography and Taxonomy of Agglutinated Foraminifera. NATO Advanced Study Institutes, Series C, 327, 139-244. CORFIELD, R. M. & CARTLtDGE, J. E. 1991. Isotopic evidence for the depth stratification of fossil and recent Globigerinina: A review. Historical Biology, 5, 37-63. CORLISS, B. H. 1991. Morphology and microhabitat preferences of benthic foraminifera from the northwest Atlantic Ocean. Marine Micropaleontology, 17, 195-236. & CHEN, C. 1988. Morphotype patterns of Norwegian Sea deep-sea benthic foraminifera and ecological implications. Geology, 16, 71 6-719. & FoIs, E. 1991. Morphotype analysis of deep-sea benthic foraminifera from the northwest Gulf of Mexico. Palaios, 5, 589-605. CRAIG, H. & GORDON, L. I. 1965. Deuterium and oxygen 18 variations in the ocean and the marine atmosphere. In: TONGIORGI,E. (ed.) Stable Isotopes in Oceanographic Studies and Pateotemperatures. Consiglio Nazionale delle Ricerche Laboratorio di Geologia Nucleare, Pisa, Italy, 1-122. DE CONrNCK, J. 1990. Ypresian organic-walled phytoplankton in the Belgian Basin and adjacent areas. Bulletin de la Soci~t~ Belge de G~ologie, 97, 287-319. DUDLEY, W. C., B1ACKWELDER, P., BRAND, L. & DUPLESSY, J.-C. 1986. Stable isotopic composition of coccoliths. Marine Micropaleontology, 10, 1-8. ELDHOLM, O. 1990. Paleogene North Atlantic magmatictectonic events: Environmental implications. Memorie della Societ~ Geologica Italiana, 44, 13-28. & THOMAS, E. 1993. Environmental impact of volcanic margin formation. Earth and Planetary Science Letters, 117, 319-329. GRADSTEIN,E M. & BERGGREN,W. A., 1981. Flysch-type agglutinated foraminifera and the Maestrichtian to -

-

Paleogene history of the Labrador and North Seas. Marine Micropaleontology, 6, 211-268. ~ . , KRISTIANSEN,I. L., LOEMO, L. & KAMINSKI,M. A. 1992. Cenozoic foraminiferal and dinoflagellate cyst biostratigraphy of the central North Sea. Micropaleontology, 38, 101-137. GRAHAM, D. W., CORLISS, B. C., BENDER, M. L. & KEIGWlN, L. D. JR. 1981. Carbon and oxygen isotopic disequilibria of recent deep-sea benthic foraminifera. Marine Micropaleontology, 6, 483-497. GROSSMAN, E. L. 1987. Stable isotopes in modern benthic foraminifera: A study of vital effect. Journal of Foraminiferal Research, 17, 48-61. GRY, H. 1941. De istektoniske Forhold i Moleromraadet. Meddelelser fra Dansk Geologisk Forening, 9, 586-627. HARDENBOL, J. 1994. Sequence stratigraphic calibration of Paleocene and Lower Eocene continental margin deposits in NW Europe and the US Gulf Coast with the oceanic chronostratigraphic record. GFF, 116, 49-51. HEILMANN-CLAUSEN,C. 1985. Dinoflagellate stratigraphy of the uppermost Danian to Ypresian in the Viborg 1 borehole, central Jylland, Denmark. Danmarks Geologiske UndersCgelse, Series A7. 1988. The Danish Subbasin: Paleogene dinoflagellates. Geologisches Jahrbuch Reihe A, 100, 339-343. -1990. Tertia~ret ved Alb~ek Hoved: Et fedt profil. Varv, 1990 (4), 99-118. --, NIELSEN, O. B. & GERSNER, E 1985. Lithostratigraphy and depositional environments in the Upper Paleocene and Eocene of Denmark. Bulletin of the Geological Society of Denmark, 33, 287-323. HULSBOS, R. E., KROON, D., JANSEN, H. S. M. & vAN HINTE, J.E. 1989. Lower Eocene benthic foraminifera and paleoenvironment of the outer V0ring Plateau, Norwegian Sea (DSDP Site 338). Micropaleontology, 35, 256-273. JAN DU CHI~NE, R., GORIN, G. & VAN STUIJVENBERG,J. 1975. Etude g6ologique et stratigraphique (palynologie et nannoflore calcaire) des Gr~s des Voirons (Pal6og6ne de Haut-Savoie, France). G~ologie Alpine, 51, 51-78. KAMINSKI, M. A., GRADSTEIN, E, GOLL, R. M. & GREIG, D. 1990. Biostratigraphy and paleoecology of deep-water agglutinated foraminifera at ODP Site 643, Norwegian-Greenland Sea. In: HEMLEBEN,C., KAMINSKI, M. A., KUHNT,W. • SCOTT, D. B. (eds) Paleoecology, Biostratigraphy, Paleoceanography and Taxonomy of Agglutinated Foraminifera. NATO Advanced Study Institutes Series C, 327, 345-386. KENNETT, J. P. & STOry, L. D. 1991. Abrupt deep-sea warming, palaeoceanographic changes and benthic extinctions at the end of the Palaeocene. Nature, 353, 225-229. KING, C. 1981. The stratigraphy of the London Clay and associated deposits. Tertiary Research, Special Paper, 6, Backhuys, Rotterdam. 1989. Cenozoic of the North Sea. In: JENKINS,D. G. & MURRAY, J. W. (eds) Stratigraphical Atlas of Fossil Foraminifera. Ellis Horwood Limited, Chichester, 418-489. -

-

STABLE ISOTOPE AND BIOTIC EVOLUTION IN NORTH SEA, DENMARK 1993. The channel connection: Tropical foraminiferids in the early Eocene of the southern North Sea Basin. In: IGCP 308 Meeting: Correlation of the Early Paleogene in Northwest Europe, London, 1-2 December 1993, Programme and Abstracts. KNox, R. W. O'B. 1984. Nannoplankton zonation and the Palaeocene/Eocene boundary beds of NW Europe: an indirect correlation by means of volcanic ash layers. Journal of the Geological Society, London, 141,993-999. -& MORTON,A. C. 1988. The record of early Tertiary N Atlantic volcanism in sediments of the North Sea Basin. In: MORTON, A. C. & PARSON, L. M. (eds) Early Tertiary Volcanism and the Opening of the NE Atlantic. Geological Society, London, Special Publication, 39, 407-419. , HARLAND,R. & ICdNG,C. 1983. Dinoflagellate cyst analysis of the basal London Clay of southern England. Newsletters on Stratigraphy, 12, 71-74. , MORTON, A. C. & HARLAYD, R. 1981. Stratigraphical relationships of Palaeocene sands in the UK Sector of the Central North Sea. ln: ILLING, L. V. & HOBSON, G. D. (eds) Petroleum Geology of the Continental Shelf of North-West Europe, Heyden & Sons, London, 267-281. KROOPNICK, E 1974. The dissolved O2-CO2-13C system in the eastern equatorial Pacific. Deep-Sea Research, 21, 211-227. - 1980. The distribution of 13C in the Atlantic Ocean. Earth and Planetary Science Letters, 49, 469-484. - 1985. The distribution of 13C of ~CO 2 in the world oceans. Deep-Sea Research, 32, 57-84. KUHNT, W., MORLOTTI, E., WINKLER, W. & KAMINSKI, M. A. 1989. Alpine late Mesozoic and early Cenozoic deep-water sequences and their agglutinated foraminifera. Excursion Guidebook, Ttibingen, Germany. MCCORKLE, D. C., KEIGWIN, L. D., CORLISS, B. H. & EMERSON, S. R. 1990. The influence of microhabitats on the carbon isotopic composition of deep sea benthic foraminifera. Paleoceanography, 5, 161-185 MCKENNA, M. C. 1983. Cenozoic paleogeography of North Atlantic land bridges. In: BOTT, M. H., SAXOV, S., TALWANI, M. & THIEDE, J. (eds) Structure and Development of the GreenlandScotland Ridge - New Methods and Concepts. Plenum, New York, 351-399. McNICHOL, A. E & DRUFFEL, E. R. M. 1992. Variability of the 813C of dissolved inorganic carbon at a site in the north Pacific Ocean. Geochimica et Cosmochimica Acta, 56, 3589-3592. MARGOLIS, S. V., KROOPNICK, P. M., GOODNEY, D. E. DUDLEY, W. C. & MAHONEY, M. E. 1975. Oxygen and carbon isotopes from calcareous nannofossils as paleoceanographic indicators. Science, 189, 555-557. MARINCOVICH,L. JR., BROUWERS,E. M., HOPKINS, D. M., & MCKENNA, M. C. 1990. Late Mesozoic and Cenozoic paleogeographic and paleoclimatic history of the Arctic Ocean Basin, based on shallow-water marine faunas and terrestrial vertebrates. In: GRANTZ, A., JOHNSON, L. & SWEENEY,J. E (eds) The Geology of North America,

305

Vol. L. The Arctic Ocean region. Geological Society of America, Boulder, Colorado, 403-426. MARTINI, E. 1971. Standard Tertiary and Quaternary calcareous nannoplankton zonation. In: FARINACCI, A. (ed.) Proceedings of the 2nd Planktonic Conference, Rome 1970. Editizione Technoscienza, Rome, 739-785. MATTSSON, E. I. & SCHMITZ, B. 1994. Stable isotopic study of the basal type Selandian (Viborg borehole 5): preliminary results. GFF, 116, 58. MILLER, K. G., CURRY,W. B. & OSTERMANN,D. R. 1985. Late Paleogene (Eocene to Oligocene) benthic foraminiferal oceanography of the Goban Spur region, Deep Sea Drilling Project Leg 80. Initial Reports of the Deep Sea Drilling Project, 80, 505-538. MOOK, W. G. 1968. Geochemistry of the stable carbon and oxygen isotopes of natural waters in the Netherlands. PhD Thesis, Rijksuniversiteit te Groningen. MORKHOVEN, E P. C. M. VAN, BERGGREN, W. A. & EDWARDS, A. S. 1986. Cenozoic Cosmopolitan Deep-Water Benthic Foraminifera, Bulletin des Centres de Recherches Exploration-Production Elf Aquitaine, Memoir, 11, Pau, France. MORTON, A. C. & KNOX, R. W. O'B. 1990. Geochemistry of late Palaeocene and early Eocene tephras from the North Sea Basin. Journal of the Geological Society, London, 147, 425-437. NIELSEN, O . B . & HE1LMANN-CLAUSEN, C. 1988. Palaeogene volcanism: The sedimentary record in Denmark. In: MORTON,A. C. & PARSON,L. M. (eds) Early Tertiary Volcanism and the Opening of the NE Atlantic. Geological Society, London, Special Publication, 39, 395-405. OKADA, n. & HONJO, S. 1973. The distribution of oceanic coccolithophorids in the Pacific. Deep-Sea Research, 20, 355-374. OWEN, R. M. & PEA, D. K. 1985. Sea-floor hydrothermal activity links climate to tectonics: The Eocene carbon dioxide greenhouse. Science, 227, 166169. PEDERSEN, A. K., ENGELL,J. & RONSBO,J. G. 1975. Early Tertiary volcanism in the Skagerrak: New chemical evidence from ash layers in the mo-clay of northern Denmark. Lithos, 8, 255-268. - & SURLYK, E 1983. The Fur Formation, a late Paleocene ash-bearing diatomite from northern Denmark. Bulletin of the Geological Society of Denmark, 32, 43--65. RATHBURN,A. E. & CORLISS, B. H. 1994. The ecology of living (stained) deep-sea benthic foraminifera from the Sulu Sea. Paleoceanography, 9, 87-150. REA, D. K., ZACHOS, J. C., OWEN, R. M. & GINGERICH, E D. 1990. Global change at the Paleocene-Eocene boundary: Climatic and evolutionary consequences of tectonic events. Palaeogeography, PaiReDclimatology, Palaeoecology, 79, 117-128. ROBERTS, D. G., MORTON, A. C. & BACKMAN, J. 1984. Late Paleocene-Eocene volcanic events in the northern North Atlantic Ocean. Initial Reports of the Deep Sea Drilling Project, 81, 913-923. SCHMrrZ, B. 1988. Origin of microlayering in worldwide distributed It-rich marine Cretaceous-Tertiary boundary clays. Geology, 18, 87-94.

306

B. SCHMITZ ETAL.

1994. Iridium enrichments in volcanic ash layers from the r Formation (Lower Eocene) in Denmark. GFF, 116, 62. --, ANDERSSON, P. & DAHL, J. 1988. Iridium, sulfur isotopes and rare earth elements in the CretaceousTertiary boundary clay at Stevns Klint, Denmark. Geochimica et Cosmochimica Acta, 52, 229-236. --, ASARO, E, MICHEL, H. V., TH1ERSTEIN, H. R. & HUBER, B. T. 1991. Element stratigraphy across the Cretaceous/Tertiary boundary in Hole 738C. Proceedings of the Ocean Drilling Program, Scientific Results, 119, 719-730. --, KELLER, G. & STENVALL,O. 1992. Stable isotope and foraminiferal changes across the CretaceousTertiary boundary at Stevns Klint, Denmark: Arguments for long-term oceanic instability before and after bolide-impact event. Palaeogeography, Palaeoclimatology, Palaeoecology, 9 8 , 233-260. SHACKLETON,N. J. & HALL, M. A. 1984. Carbon isotope data from Leg 74 sediments. Initial Reports of the Deep Sea Drilling Project, 74, 613-619. -& KENNETT, J. P. 1975. Paleotemperature history of the Cenozoic and the initiation of Antarctic glaciation: Oxygen and carbon isotope analyses in DSDP sites 277, 279, and 281. Initial Reports of the Deep Sea Drilling Project, 29, 743-755. , CORFIELD, R. M. & HALL, M. A. 1985. Stable isotope data and the ontogeny of Paleocene planktonic foraminifera. Journal of Foraminiferal Research, 15, 321-336. --, HALL, M. A. & BOERSMA, A. 1984. Oxygen and carbon isotope data from Leg 74 foraminifers. Initial Reports of the Deep Sea Drilling Project, 74, 599-612. SIESSER, W. G., WARD, D. J. & LORD, A. R. 1987. Calcareous nannoplankton biozonation of the Thanetian Stage (Paleocene) in the type area. Journal of Micropaleontology, 6, 85-102. STEURBAUT, E. 1988. New Early and Middle Eocene calcareous nannoplankton events and correlations in middle to high latitudes of the northern hemisphere. Newsletters on Stratigraphy, 18, 99-115. 1991. Ypresian calcareous nannoplankton biostratigraphy and palaeogeography of the Belgian Basin. In: DuPuIS, C., DE CONINCK, J. & STEURSAUT, E. (eds) The Ypresian stratotype. Bulletin de la Socidt~ Beige de Gdologie, 97 (1988), 251-285. -1995. Calcareous nannoplankton, a major key for developing high-resolution biochronologies and for unravelling Earth's Tertiary geological history. Nouvelles de la Science et des Technologies, 13 (2-3), 4p. Brussels.

& KING, C. 1994. Integrated stratigraphy of the Mont-Panisel borehole section (151E340), Ypresian (Early Eocene) of the Mons Basin, SW Belgium. Bulletin de la Socidtg Belge de Gdologie, 102 (1-2) (1993), 175-202. - & NOLF, D. 1986. Revision of Ypresian stratigraphy of Belgium and Northern France. Mededelingen van de Werkgroep voor Tertiaire en Kwartaire Geologie, 23, 115-172. SWISHER,C. C. & KNOX, R. W. O'B. 1993. Single-crystal laser-fusion Ar4~ 39 dating of Early Eocene tephra layers from the North Sea Basin: Calibration points for the Paleogene time-scale. Correlation of the Early Paleogene in Northwest Europe, Programme and Abstracts. Geological Society, London, 1-2 December 1993. THIEDE, J. & ELDHOLM, O. 1983. Speculations about the paleodepth of the Greenland-Scotland Ridge during late Mesozoic and Cenozoic times. In: BOTT, M. H., SAXOV, S., TALWANI, M. & THIEDE, J. (eds) Structure and Development of the GreenlandScotland Ridge - New Methods and Concepts. Plenum, New York, 445-456. - - . , DIESEN, G. W., KNUDSEN,B.-E. & SNARE, T. 1986. Patterns of Cenozoic sedimentation in the Norwegian-Greenland Sea. Marine Geology, 69, 323-352. , N I E L S E N , O . B . & PERCH-NIELSEN, K. 1980. Lithofacies, mineralogy and biostratigraphy of Eocene sediments in northern Denmark (Deep Test Viborg 1). Neues Jahrbuch fiir Geologie und Paliiontologie Abhandlungen, 160, 149-172. THOMAS, E. & SHACKLETON,N, J. 1996. The PaleoceneEocene benthic foraminiferal extinction and stable isotope anomalies. This volume. VINCENT, E., KILLINGLEy,J. S. & BERGER, W. H. 1981. Stable isotope composition of benthic foraminifera from the equatorial Pacific. Nature, 289, 639-643. WHITE, R. S. 1988. A hot-spot model for early Tertiary volcanism in the N Atlantic. In: MORTON, A. C. & PARSON, L. M. (eds) Early Tertiary Volcanism and the Opening of the NE Atlantic. Geological Society, London, Special Publication, 39, 395-405. WOODRUFF, E, SAVIN, S. M. & DOUGLAS, R. G. 1980. Biological fractionation of oxygen and carbon isotopes by recent benthic foraminifera. Marine Micropaleontology, 5, 3-11. ZIEGLER, P. A. 1988. Evolution of the Arctic-North Atlantic and the Western Tethys. American Association of Petroleum Geologists, Memoir, 43. -1990. Geological Atlas of Western and Central Europe (second edition). Shell Internationale Maatschappij B.V., Den Haag, Netherlands. - -

A late Paleocene-early Eocene NW European and North Sea magnetobiochronological correlation network W. A. B E R G G R E N

1 & M.-E

AUBRY 2

1 Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA 2 Institut des Sciences de l'Evolution, Universit~ Montpellier II, 34095 Montpellier Cedex 5, France Abstract: Published and unpublished data on bio-, chemo- and magnetostratigraphic events spanning the late Paleocene-early Eocene are reviewed and calibrated to a revised Geomagnetic Polarity Time Scale (GPTS), itself already in need of revision. This timescale serves as a template for placing the upper Paleocene-lower Eocene stratigraphic succession of NW Europe (Anglo-Paris-Belgian Basins) in a sequence stratigraphic framework. It is clear that an approach that integrates sequence- and magnetobioisotope stratigraphy provides a unifying correlation framework within which to delineate the disjunct depositional histories of marginal and deep marine basinal stratigraphies and allows the stratigrapher to move from rock to time. Significant conclusions of this integration include the following: 1. The major climatic warming, weakening of atmospheric circulation and faunal extinction events, which are seen in the deep sea stratigraphic record, are seen to be closely associated with the mid-part of the (redefined) planktonic foraminiferal Zone P5 and calcareous nannoplankton Zone NP9. 2. A relatively rapid (c. 1000 years) decrease of c. 3 to 4 per mil in the ~13C of marine carbonates has now been recognized in high southern and northern latitude sites and at several scattered, intermediate locations. A similar excursion has been observed in mammalian teeth and soil carbonates near the base of the Wasatchian North American Land Mammal Age in the Big Horn Basin a few metres above the Argile plastique bariolre ('type' Sparnacian) in the Pads Basin. This 813C excursion is seen to be a truly global event which occurs in both marine (c. midZone NP9) and terrestrial stratigraphies at a level estimated at c. 55.5 Ma. 3. An evaluation of the calcareous plankton biostratigraphy and stable isotope records at several deep sea sites suggests the presence of multiple unconformities in the Paleocene/Eocene boundary interval. There would appear to be no unequivocally demonstrated continuous stratigraphic section through the c. 2.55 million year interval of Chron C24r in the deep sea record or from outcrops on land, and a composite record is required to construct the sequence of events that occurred during this interval. 4. The Paleocene/Eocene boundary (which awaits the determination of a Global Stratigraphic Section and Point [GSSP]) is bracketed by the base of the leper Clay Formation (= Ypresian Stage), estimated here at > 54.6 Ma, and the P5/P6a zonal boundary (LAD Morozovella velascoensis) at c. 54.7 Ma (above), and by the ~13C spike (and associated events) in midZone NP9 at c. 55.5 Ma (below). The 'boundary interval' encompasses the NP9/10 zonal boundary at 55 Ma and the base of the Harwich Formation (base of the London Clay Formation Oldhaven Beds of previous authors) at c. 54.8 Ma. 5. Calcareous nannoplankton studies of recently recollected Thanetian localities in the eastern part of the Paris Basin have revealed that the Thanetian Sables de Ch~lons-sur-Vesles are older (Zones NP6 and NP7) than the Sables de Bracheux s.l. (= Sables du Tillet; Zone NP8). This leads to a revised framework of correlation between the upper Paleocene formations of northwest Europe. 6. The base of the Sparnacian base, taken as the Paleocene/Eocene boundary by most vertebrate palaeontologists, is correlative with a level within Zone NP9, c. 0.5 million year older than the NP9/10 zonal boundary (used by some marine micropalaeontologists to delineate the Paleocene/Eocene boundary) and > one million years older than the base of the Ieper Clay (base Ypresian) in Belgium.

The marine stratigraphic succession d e v e l o p e d in the sedimentary basins on the passive margins o f the Baltic Plain surrounding the North Sea (AngloBelgian-Paris-Danish-North G e r m a n Basins) has

c o m e to s e r v e as the standard for m o n d i a l P a l e o g e n e cbronostratigraphy. I G C P Project 308 ( P a l e o c e n e / E o c e n e B o u n d a r y Events in Space and Time) is currently examining representative marine

From Knox, R. W. O'B., Corfield, R. M. & Dunay, R. E. (eds), 1996, Correlation of the Early Paleogene in Northwest Europe, Geological Society Special Publication No. 101, pp. 309-352.

309

310

W.A. BERGGREN & M. E AUBRY

and terrestrial stratigraphic sequences spanning this boundary in an attempt to provide criteria suitable for locating and positioning a global boundary stratotype section and point (GSSP). Within the framework of this project we present a comparative anatomy and correlation of the stratigraphic sequences in this region spanning an approximately six million year interval from early Thanetian to early Ypresian. In the first part of this paper we review the calcareous plankton and benthic foraminiferal biostratigraphy, as well as stable isotope and magnetostratigraphic data in deep sections spanning the Paleocene-Eocene boundary interval in order to erect a global magnetobiostratigraphic correlation network within which to place the standard NW European stratigraphic succession. We have examined samples from some deep sea drilling holes (in particular 549, 550 and 690) and have evaluated the published calcareous microfossil stratigraphies for others. The main lithostratigraphic units are then placed in a geochronological framework based on a newly constructed Geomagnetic Polarity Time Scale (GPTS) (Cande & Kent 1992, 1995) which is derived from an analysis of (primarily) South Atlantic seafloor anomaly profiles in which anomaly spacings are constrained by finite rotation poles and averages of stacked profiles. The revised magnetochronology was generated by using a spline function to fit a set of nine calibration points (including anchoring calibrations of 55.0 Ma on the supposed NP9/NP10 zonal boundary at/near the Paleocene/Eocene boundary and 66.0 Ma (Cande & Kent 1992; 65 Ma in Cande & Kent 1995) at the K/P boundary plus the zero age ridge axis to the composite polarity sequence. This magnetochronology has served as the template for a revised magnetobiochronology of the standard calcareous biostratigraphy of the Paleocene/Eocene boundary interval, based primarily on data from DSDP Hole 550 in the NE Atlantic, and for the Paleocene based primarily on a magnetobiostratigraphic record recently obtained at DSDP Hole 384 (NWAtlantic). Cross-correlation of the recently 4~ dated -17 and +19 ashes (of Denmark and North Sea area) to Chron C24r in Hole 550 and to the base of the Wrabness Member (ex Harwich Member) and Hales Clay of the Harwich Formation (ex London Clay division A1 - see Ellison et al. 1994; Jolley 1996), respectively, of the London-Hampshire Basin provides a direct tie between the deep sea and marginal marine record at the Paleocene/Eocene boundary. We then place the stratigraphic succession of NW Europe and the North Sea in a revised sequence stratigraphic framework and correlate this record to the newly revised geochronology. This

makes it possible to estimate the duration of the hiatuses associated with the major/minor unconformities separating the various lithic units, to situate the associated terrestrial mammalian levels within a chronostratigraphic framework and to make reasonable estimates of the timing/duration of major prochoresis/faunal turnover events.

Paleocene/Eocene b o u n d a r y sections In this section we present a review of data on bio-, magneto- and isotope stratigraphy and climatic events spanning the Paleocene/Eocene boundary observed in a number of deep sea drilling sites. Chronology is that of Berggren et al. (1985) with exceptions noted below. In the second part of the paper dealing with geochronology of the Paleocene/Eocene boundary and with NW European chronostratigraphy we switch to a revised chronology based on the newly developed Geomagnetic Polarity Time Scale (GPTS) by Cande & Kent (1992, 1995). With regard to the discussion of planktonic foraminiferal biostratigraphy below it is important to point out that a basic modification to planktonic foraminiferal Zones P5 and P6 (spanning the Paleocene/Eocene boundary) has been made in the revised Paleogene integrated magnetobiochronology of Berggren et al. (1995). This has been necessitated by the recognition that Zones P5 and P6a as defined in Berggren & Miller (1988) are essentially equivalent owing to the concurrent range of Morozovella subbotinae and M. velascoensis following the LAD of Globanomalina pseudomenardii, nominate taxon of the underlying Zone P4. The newly defined/emended zonation is presented here so that the reader may make the transition from the zonal definition of Berggren & Miller (1988) and that currently in use in the review presented below (Fig. 1). The magnetochronology used in the section on zonal definitions (below) is that of Berggren et aI. (1995) and it is taken directly from that paper. Zonal definitions from Berggren et al. (1995) Zone P5. Morozovella velascoensis. Interval Zone (Bolli 1957; P5 and P6a of Berggren & Miller 1988) Definition: biostratigraphic interval between the LAD of Globanomalina pseudomenardii and the LAD of Morozovella velasacoensis. Magnetostratigraphic calibration: Chron C25n(y)Chron C24r (midpart). Estimated age: 55.9-54.7 Ma. Remarks: Zone P5 with a different denotation

LATE PALEOCENE--EARLYEOCENE MAGNETOBIOCHRONOLOGYOF NW EUROPE

Berggren & Miller (1988)

311

Berggren et al. (1995)

_.

Zones

Criteria

Z one

P7

P7

i.!

_2_

P6c

6b

8 P6b

P6a -v- -

T :,~

P6a r~

P5

P5

r~ ,.M

D

P4

!

P4

Fig. 1. Comparison of biostratigraphic criteria used in defining planktonic foraminiferalZones P5-P7 in Berggren & Miller (1988) and revised zonation used in this paper and Berggren et al. (1995).

(partial range of the nominate taxon between the LAD of Gl. pseudomenardii and the FAD of Morozovella subbotinae) was defined (Berggren & Miller 1988) before definitive evidence of the juxtaposition/overlap in range of Gl. pseudomenardii and M. subbotinae, nominate forms for Zone 1'4 (top) and P6 (base), respectively, became available (cf. Blow 1979, p. 265-267, although we would disagree with Blow on the upper limit of pseudomenardii in his Zone P7 = P6b of Berggren & Miller 1988). However, in some instances, the FAD of M. subbotinae has been observed to be delayed because of the widespread occurrence of (a) distinct dissolution event(s) that span(s) the lower part of magnetozone 24.3r and Zone P6a. Thus the chronology of a zone based on the FAD of M. subbotinae is obviously very approximate. For this reason we have decided to revert to the previous, relatively unequivocal usage of Bolli (1957) in which two sequential LADs are used to define a distinct interval which would appear to span the Paleocene/Eocene boundary as currently recognized by at least some (bio)stratigraphers. As a result of this modification to Zone P5, Zone P6 of Bergren & Miller (1988) is also modified. Zone P5 (as revised) essentially contains the concurrent range of M. subbotinae (FAD) and M. velascoensis

(LAD), but it is defined as an interval zone because the definition of the top of P4 is the LAD of Gl. pseudomenardii. Characteristic features of this subzone include the relatively closely spaced appearances of Morozovella marginodentata, M. formosa gracilis, Igorina broedermanni, Acarinina wilcoxensis, Turborotalia pseudoimitata, and the relatively common occurrence of strongly muricate 'large acarininids' (soldadoensis, coalingensis-triplex group). The Paleocene/Eocene boundary is usually correlated with the P5/P6 (= P6a/P6b of Berggren & Miller 1988) boundary by planktonic foraminiferal specialists and is estimated here at 54.8 Ma (cf. discussion in Berggren et al. 1995, on radioisotopic calibrations for the Paleogene chronology adopted here). Calcareous nannoplankton specialists usually consider the NP9/10 zonal boundary as definitive for this boundary. The base of the Harwich Formation (Oldhaven Beds = Hales Clay; = base Eocene = Thanetian/Ypresian boundary in some usage) has been cross correlated to the -17 ash in DSDP Hole 550 and dated in NW Europe at 54.5 Ma (see discussion in Berggren et al. 1995). However, we choose to use an age of 54.8 Ma for the base of the Harwich Formation

312

w.A. BERGGREN • M. P. AUBRY

based on sedimentation rates in DSDP Hole 550, rather than on the -17 ash date. This is discussed in greater detail below. The base of the leper Clay (= base Ypresian s.s.) in Belgium is located one fourth order cycle higher in the stratigraphic record with an estimated age of c. 54.6 Ma here (see discussion below). The problems associated with the identification and delineation of events suitable for the determination of an appropriate Paleocene/ Eocene boundary GSSP are being currently examined by IGCP 308 (Paleocene/Eocene Boundary Events in Time and Space) and are discussed in greater detail below and elswhere (Aubry et al. 1996).

manner that Zone P6 (as redefined in Berggren

et al. 1995) corresponds to Subzones P6b and P6c of Berggren & Miller (1988, p. 371).

Subzone P6a. Morozovella velascoensis-Morozovella formosa formosa and/or M. lensiformis. Interval Subzone (P6b of Berggren & Miller 1988; emended in Berggren et aL 1995) Definition: biostratigraphic interval between the LAD of Morozovella velascoensis and the FAD and/or of Morozovella formosa formosa M. lensifonnis Magnetostratigraphic calibration: mid to late part of Chron C24r.

Age estimate: 54.7-54 Ma; earliest Eocene (earliest Zone P6. Morozovella subbotinae. Partial Range Zone (redefined in Berggren et al. 1995) Definition: biostratigraphic interval characterized by the partial range of the nominate taxon between LAD of Morozovella vetascoensis and FAD of

Morozovella aragonensis Magnetostratigraphic calibration: Chron C24r (midpart)-Subchron C23n.2r (earliest part).

Estimated age: 54.7-52.3 Ma. Remarks: Berggren & Miller (1988, p. 370) defined the Morozovella subbotinae Partial Range Zone (P6) as the partial range of the nominate taxon between the FAD of the nominate taxon and that of M. aragonensis. Investigations on a number of deep sea sites and outcrop sections have now shown that the FAD of M. subbotinae essentially coincides with the LAD of Gl. pseudomenardii and that the supposed stratigraphic gap between the LAD of Gl. pseudomemardii and the FAD of M. subbotinae is illusory, although delayed entry of the latter taxon is often caused by strong dissolution in sections within the upper part of Zone NP9. Thus P5 and P6a, as defined by Berggren & Miller (1988, p. 370) are essentially equivalent. As emended here, Zone P6 coincides essentially with the M. subbotinae Zone of Premoli Silva & Bolli (1973) and Luterbacher & Premoli Silva in Caro et al. (1975) except for the (apparently brief) temporal interval between the LAD of M. velascoensis and the LAD of the small, enigmatic taxon M. edgari (Primoli Silva & Bolli 1973) (? = M. finchi Blow 1979) which was shown by Toumarkine & Luterbacher (1985, fig. 5, p. 100) to occur only slighly below the simultaneous FADs of M. formosa formosa and M. lensiformis. (cf. Blow 1979, figs 48 and 50 in which the LAD of M. finchi is shown to occur at essentially the same level). The temporal span between the LAD of M. velascoensis and the FAD of M. formosa formosa is estimated here at c. 0.8 million years In order to maintain numerical and biostratigraphic continuity with the zonation of Berggren & Miller (1988), we redefine Zone P6 in such a

Ypresian).

Remarks: In open ocean stratigraphic successions the sequence of FAD of M. subbotinae, LAD of M. velascoensis acuta, FAD of M. formosa formosa/M, lensiformis and FAD of M. aragonensis serve as a means of providing a discrete biostratigraphic subdivision of Zone P6 (Berggren & Miller 1988) and that subdivision is followed here with minor modification. This subzone has been redefined as an interval subzone to avoid conceptual confusion/overlap with the use of M. subbotinae as nominate form of both Zone P6 and Subzone P6a in its original definition (see Berggren & Miller 1988, p. 370, 371) and for Zones P6 and Subzone P6a (emended in Berggren et al. 1995). At the same time we remove Pseudohastigerina wilcoxensis as one of the nominate forms of Subzone P6a (P6b in Berggren & Miller 1988); this form appears under ideal conditions at the P5/6 zonal boundary in fully tropical assemblages but has a demonstrably delayed entry in mid-high latitude regions within the P6b-7 (or correlative) biostratigraphic interval. The LAD of Subbotina velascoensis occurs within this subzone. Additional comments on this subzone (as P6b) are to be found in Berggren & Miller (1988, p. 371).

Subzone P6b. Morozovella formosa formosa/ M. lensiformis-Morozovella aragonensis. Interval Subzone (P6b, emended in Berggren et al. 1995 = P6c of Berggren & Miller 1988, p. 371; P8a of Blow 1979). Definition: biostratigraphic interval between the essentially simultaneous FADs of M. formosa and/ or M. lensiformis and the FAD of Morozovella aragonensis. Magnetostratigraphic calibration: Chron C24r (late) to Chron C23r.

Estimated age: 54-52.1 Ma; early Eocene (early Ypresian).

Remarks: This is a biostratigraphically distinct interval (see also Primoli Silva & Bolli 1973; Blow

LATE PALEOCENE-EARLY EOCENE MAGNETOBIOCHRONOLOGY OF NW EUROPE 1979) characterized by the essentially simultaneous FADs of the n o m i n a t e taxa and the L A D s o f

M. subbotinae, M. marginodentata and M. aequa.

Review of data from DSDP-ODP sites In the discussion below, we utilize the revised chron nomenclature of Cande & Kent (1992, 1995) as well as that of Berggren et al. (1985) where appropriate. Correlation of nomenclature and estimated ages of the Chron C24n to C25n are shown below: Berggren et al. 1985 Nomenclature

Estimated age (Ma)

Subchron C24 (younger) (--- Subchron C24A) Subchron C24 (older) (= Subchron C24B) Chron C25n

55.14-55.37 55.66-56.14 58.64-59.24

Cande & Kent (1995) Nomenclature

Estimated age (Ma)

Subchron C24n. In Subchron C24n.2n Subchron C24n.3n Chron C25n

52.364--52.663 52.757-52.801 52.903-53.347 55.904-56.391

Site 5 4 9 ( G o b a n Spur, Irish continental margin, N E Atlantic) (Fig. 2) Biostratigraphic data (Snyder & Waters 1985 and our own re-examination of material from this site) indicate that Zones P5-P7 (or their correlative biostratigraphies) are represented at this site. Globanomalina peudomenardii is present in the stratigraphic interval encompassed by 549/20/1-5 (c. 370376 m), absent in 549/19/2 (c. 361+ m) as well as in the overlying stratigraphic interval between cores 19/1 and 16/6 (c. 360-340 m) which is characterized by strong dissolution (Snyder & Waters 1985, 448, 449, fig. 6); a lone occurrence is noted at 16/5:57-60 cm (c. 338 m) located in the middle part of a reversed interval (= Chron C24r), in the upper part of Zone NP9 (Mtiller 1985; Aubry et al. 1996) and c. 3 m below the FAD of Morozovella subbotinae, M. marginodentata and Acarinina wilcoxensis at c. 335 m (associated with an unconformity; see below). Chron C25n is situated between c. 349.5-354 m (within the dissolution interval and is associated with the NPS/NP9 boundary). The LAD of pseudomenardii may lie within the dissolution interval and near the Chron C25n interval; we have not found Gl. pseudomenardii despite a diligent search of samples above the dissolution interval in cores 16 and 17. Thus the P4/5 zonal boundary cannot be determined precisely at this site.

313

The FAD of Morozovella lensiformis denotes the Subzone P6a/P6b boundary (Berggren et al. 1995). In Hole 549 the first occurrence of this taxon (Snyder & Waters 1985) is within a c. 8 m interval between samples in core 15R/1 and 15R/6 (c. 332-340 m). There is an unconformity at c. 335 m separating upper Zone NP9 from upper Zone NP10 and it is most likely that M. lens(formis (and Subone P6b) extends down to the level of the unconformity. This is consistent with the LAD of M. aequa between core 15R/6 AND 14R/6 (c. 331 m: Snyder & Waters 1985) which occurs in nearby Hole 550 in Subzone P6b). It is also consistent with correlations of the 13C record in Holes 549 and 550 (Aubry et al. 1996; Stott et al. 1996) which make sense over the stratigraphic interval of c. 335-332 m in Hole 549 and c. 380-370 m in Hole 550, within upper Zone NP10. The FAD of Morozovella aragonensis occurs in Core 14R between c. 313 and 320 m and delineates Zone P6/P7 boundary. At sites 549 and 550, the P6a/b zonal boundary occurs in the upper part of Chron C24r and within Zone NPI 1. The FAD of Pseudohastigerina wilcoxensis occurs in Core 14R/1 (c. 313 m) in the mid-part of Subzone P6b, c. 20-21 m above that of M. subbotinae and Ac. wilcoxensis in Core 16. Elsewhere the FAD of Ps. wilcoxensis has been shown to occur within the uppermost part of Zone NP9 (e.g. Site 605), or NP10 (e.g. Sites 401, 690). It is difficult to determine whether this difference is due to diachrony or taxonomic discrimination of the planispiral pseudohastigerinid morphology. The equivocal nature of the palaeomagnetic records at these sites adds to the problem as well. The calcareous nannofossil succession in the NP9NP10 zonal interval is discussed in detail in Aubry et al. (1996). These authors show the presence of two successive unconformities at and close to the NP9/NP10 zonal boundary as indicated by the close stratigraphic succession of the LAD of Fasciculithus tympaniformis (at 335.50 m), the FAD of Tribrachiatus bramlettei (at 335.30 m) and that of T. contortus (Morphotype B, see Aubry et al. 1996) (at 335.16 m). The presence of unconformities near the Paleocene/Eocene boundary in Hole 549 is supported by the absence in Hole 549 of the series of more than 55 bentonitic ashes (distal equivalents of North Sea-Danish ash series) which occur in the lower to mid-part of Zone NP10 in Hole 550 (from c. 386-403 m) and by studies on the benthic foraminiferal fauna of Hole 549 by Reynolds (1992, MSc Thesis, Univerisity of Maine, Orono) which suggest that c. 12.6 m of sediments equivalent to the lower part of Zone NP10 are missing at c. 335 m. Principal component analysis of the benthic foraminiferal fauna (Reynolds 1992, MSc Thesis, University of Maine, Orono) reveals a definite Paleocene fauna (PC2), a definite Eocene fauna (PC1), and two short-lived transitional faunas. Allowing for a hiatus of c. 0.8 million years at c. 335.40 m (near the NP9/10 boundary), the rate of faunal turnover was shown to be essentially synchronous in the high latitude North Atlantic and Antarctic (Thomas 1990). We would interpret the data of Reynolds to indicate that the LAD of the Stensoina beccariformis fauna occurs between 336 m and 332 m and is associated with the hiatuses at c. 335.40 m and 335.22 m; sporadic specimens of St. beccariformis above this level are interpreted as having been reworked.

314

w . A . BERGGREN & M.P. AUBRY

Fig. 2. Stratigraphic framework of events spanning the Paleocene/Eocene boundary: DSDP Hole 549.

LATE PALEOCENE--EARLY EOCENE MAGNETOBIOCHRONOLOGY OF NW EUROPE The 813C spike has been identified at 338 m in Hole 549 (Sinha & Stott 1993a, b; Stott etal. 1996), only 2.6 m below the unconformity (c. 335.40 m) which separates (upper) Zone NP9 from Zone NP10 (see Aubry et al. 1996).

Site 5 5 0 ( P o r c u p i n e A b y s s a l Plain, S W o f s e a w a r d edge o f G o b a n Spur, N E Atlantic) (Fig. 3) Subchrons C24n.ln, C24n.3n and ?Chron C25n (partim) bracket a relatively long (62.85m) Chron C24r (Townsend 1985) making this a supposedly ideal site for detailed, integrated magnetobiostratigraphy (Table 1). Calcareous nannofossil (MUller 1985) and planktonic foraminiferal (Snyder & Waters 1985) data indicate the presence of Zones NP9-NP12 and P4 (equivalent), P5, P6a, P6b and P7, respectively (Tables 3-5). We have examined material from Hole 550 kindly provided by S. Snyder as well as material studied for stable isotopes by A. Sinha and L. Stott and our results are incorporated here in the text as well as in tables below. The FAD of D. lodoensis (346.73m) ( = N I l / 1 2 = CP9/10 zonal boundary) occurs in Subchron C24n.l-2r, consistent with earlier records. The stratigraphic range of Tribrachiatus contortus (which defines Subzone CP9a of Okada & Bukry (1980) in the upper part of Zone NP10 of Martini (1971) was shown by Mailer (1985) to span a 40.62 m-thick interval of Chron C24r, between 406.35 m and 365.72 m. Reexamination of the NP9-NP10 zonal interval (Aubry et al. 1996) has shown that the range of T. contortus is restricted to the upper part of Zone NP10 between 381.9m and 378.4m (Morphotype A) and between 372.65 m and 375.72 m (Morphotype B).

315

The NP9/NP10 zonal boundary occurs between 4 0 8 . 0 2 m and 407.75 m based on the FAD of Tribrachiatus bramlettei at this latter level (Aubry et al. 1996). The LAD of Fasciculithus tympaniformis at 408.02 m immediately precedes the FAD of T. bramlettei. Integration of calcareous nannofossil and carbon isotope stratigraphies (Aubry et aL 1996) shows that the NP9/NP10 zonal contact is unconformable, and that the unconformity corresponds to the lithological contact at 408 m. The sequential FADs of M. subbotinae (407 m in Snyder & Waters 1985; 409 m, this paper; but see below), Ac. wilcoxensis (404.5 m), M. formosa gracilis, Igorina broedermanni (395 m) and LAD ofM. acuta (396 m), and FAD of M. lensiformis (377 m) and M. aragonensis (341 m) support recognition of Zones P5, P6a, P6b and P7 respectively. The P5/P6a and P6a/b zonal boundaries lie within the mid- and upper part of chronozone C24r, respectively. However the FAD of M. subbotinae is certainly delayed owing to the presence of an intense dissolution zone which spans the interval of cores 36/2 to 34/4 (early Citron C24r interval). The P6a/b boundary (using a midpoint of 377 m for the FAD ofM. lensiformis) lies in the upper part of chronozone C24r. The LAD of the Stensioina beccariiformis benthic foraminiferal fauna occurs between samples 35/1: 62-65 cm and 34/4: 62-6 cm (413.62-408.65 m; personal observation, WAB) within a dissolution facies (characterized by abundant radiolarians) which spans the interval from Core 36/2 to Core 34/5 (c. 425-410 m), in lower Chron C24r, within Zone NP9 and (apparently) mid-Zone P5. Snyder & Waters (1985, fig. 6) place the FAD of M. subbotinae in sample 34/2:62-65 cm (405.65 m), c. 3 m above the top of the dissolution facies. Our study (WAB) of samples from cores 33-35 indicates that specimens transitional between M. aequa and

Table 1, Estimated (minimum) thickness of Chron C24r in some representative deep sea sections Hole

Depth (m)

Thickness (m)

577

79.95-85.45

5.50

690B

185.42-135 (uncons

50.42

549

317.5-350

37.5

550

359.65-422.50

62.85

605

557-598-7

>41.3

Remarks

Stratigraphic section possibly missing at/between core breaks of core 9 and 10 (= 82.8 m) and perhaps in lower part of core 9 (absence/nominal representation of Zone CP9a (c. 81 m). Probably minimum estimate. 'Normal events' between 154-144 m interpreted as C24R based on biostratigraphy. Unconformity between NP11/12 at about 135 m in reversed interval interpreted as C23R/C24R boundary. Estimate is considered a minimum. >56 bentonitic ashes in NP10 at Site 550 not found in NP10 at 549. NP10 unusually short (c. 4 m) compared to NP11 above (c. 30 m) suggesting lower NP10 missing at Site 549. Estimate considered a minimum. Unconformity at 408 m between NP9 and NP10. Strong dissolution facies from 422-410 m (_=_Zone NP9). Estimate considered a minimum. No data above c. 557 m within C24R, NP10 and CP9a. Estimate is minimum value.

Data source: DSDP/ODP Initial/Science Reports

316

W.A. BERGGREN & M . E AUBRY

Fig. 3. Stratigraphic relationship of events spanning the Paleocene/Eocene boundary: DSDP Hole 550.

M. subbotinae and referable to M. subbotinae occur in the lowest sample examined (34/4:62-65 cm = 408.65), c. 1.5 m above the dissolution facies. This sample belongs to Zone P5 and the LAD of the St. beccariiformis fauna would be consistent with (most) occurrences elswhere, including several outcrop sections discussed below. The benthic foraminiferal event in Hole 550 consists in the replacement of a relatively small, but typical, Paleocene

fauna characterized by i.al., Stensioina beccariiformis, Cibicidoides velascoensis, Nuttallinella florealis (up to sample 3 5 / 1 : 6 2 - 6 5 c m = 4 1 3 . 6 1 m), by an influx of minute abyssaminids, pleurostomellids, pulleniids and Cibicidoides eocaenus. (sample 34/4:62-65 m = 408.65 m). Nuttallides truempyi is a characteristic component of both the late Paleocene and early Eocene faunas. Dissolution is so intense in the intervening 5 m

LATE PALEOCENE--EARLY EOCENE MAGNETOBIOCHRONOLOGY OF NW EUROPE that all calcareous benthic and planktonic foraminifera have been dissolved. Thus if our interpretation is correct the FAD of M. subbotinae is a minimum estimate and the LAD of the St. beccariiformis fauna is a maximum estimate. The true FAD of M. subbotinae may lie lower within the dissolution facies interval; indeed it probably does inasmuch as its FAD has been recorded in Chron C25n in Hole 577A by Corfield 1987 and by Liu & Olsson (pers. comm. 1992); see also Miller et al. 1987; p. 744, fig. 2, who record its FAD in mid-Chron C24r (based on lower sampling density and/or perhaps slightly different taxonomic concept). The true LAD of the St. beccariiformis fauna may lie within the interval of intense dissolution represented by samples 34/6: 62-65cm and 3 4 / 5 : 6 2 - 6 5 cm (411.62-410.12m) in which no benthic fauna was observed. A series of c. 40-50 volcanic ash beds span the interval of Cores 27-32, but are primarily developed in Cores 32-33 (c. 384.85-404.6 m; Fig. 3). These ashes are distal representatives of phase 2a and 2b ashes known from the North Sea and NW Europe, particularly Denmark. The distinctive -17 and +19 ash beds occur within the range of Tribrachiatus bramlettei, bracket the P5/P6a zonal boundary and lie approximately one-third up within Chron C24r. The -17 ash occurs in Zone P5, at the same level (400 m) as the FAD of M. marginodentata, 6 m above the FAD of M. subbotinae 18.1 m below the FAD of Morphotype A of Tribrachiatus contortus, 27.35 m below that of Morphotype B of this species, and c. 22.5 m above the Chron C24r/?C25n boundary. The +19 ash occurs at c. 393 m in Zone P6a. It is located c. 2 m above the FADs of M. formosa gracilis and Ac. broedermanni, 4 m above the LAD of M. acuta (397 m), 11.1 m below the FAD of Morphotype A of T. contortus, 20.35 m below that of Morphotype B of this species, and at c. 29.5 m above the Chron C24r/?C25n boundary. Recent 4~ dates of 54.5 and 54.0 Ma have been obtained on the -17 and +19 ash Beds, respectively (see discussion in Berggren et al. 1995; and below in this paper). Owing to the intense dissolution in the lower part of Chron C24r in Hole 550 it has proven impossible to obtain a complete ~13C record; however a portion of the major excursion has been recognized at c. 409 m (Sinha & Stott 1993a, b; Stott et al. 1996), c. 1 m below the unconformity delineated in Aubry et al. (1996) in what we would interpret as the mid-part of the truncated Zone NP9. Finally, the inability to derive a precise chronology in Hole 550 is due to the fact that the normal event identified at c. 425-435 m may not be Chron C25n. At Hole 549 a rhyolitic ash layer occurs at 352.66 m in the upper part of Chron C25n and basal Zone NP9 (Knox 1985). This ash does not occur at the corresponding level in Hole 550 at the base of NP9. It is uncertain whether the magnetic reversal between 419.51-425.5 m (Snyder & Waters 1985; Townsend 1985) corresponds to the Chron C25n/C24r boundary as indicated by Snyder & Waters 1985). The lowest occurrence of D. multiradiatus, normally associated with a level within Chron C25n, is at 424.85 m (Mtiller 1985) in an (unsampled) interval of unknown polarity. In addition the sharp lithological boundary at 426.5 m between Units 2b and 3a (Graciansky et al. 1985) occurs within Zone NP5, not at

317

the NP5/NP9 zonal boundary. Thus the upper 50 cm of the normal polarity interval which extends from 425.5434.04 m may not represent Chron C25n. We suspect that the upper surface of the unconformity between Zones NP5 and NP9 in Hole 550 is younger than the Chron C25n/C24r reversal and that Chron C25n is not represented in Hole 550. If this is true, we have lost an important calibration point in estimating the chronological position of the -17 and +19 ashes and the 813C spike in this hole (see further discussion below and in Aubry et al. 1996).

Site 401 ( M e r i a d z e k Terrace, B a y o f Biscay, N E A t l a n t i c ) (Fig. 4) Relatively incomplete recovery, sparse sampling and lack of stratigraphic range data (Krasheninnikov 1979) preclude accurate delineation of biostratigraphic datum events and zonal boundaries. Neither Tribrachiatus bramlettei nor T. contortus are reported from Hole 401. Zone NP10 in this hole is delineated by Mtiller (1979, table 7) based on the LAD of Fasciculithus tympaniformis (between 204.90 m and 204.15 m) and the FAD of T. orthostylus (between 193.75 m and 193.08 m). The association of Morozovella velascoensis, M. subbotinae and Ps. wilcoxensis in Core 13 CC (196 m) is used to differentiate Zones P6a and P6b which are equivalent to P5 and P6a here, and occur within Zone NP10. The FAD of M. formosa (core 13,191 m) delineates the P6a/P6b zonal boundary within Zone N P l l (cf. Aubry et al. 1988; Berggren & Miller 1988). The FAD of M. aragonensis in the lower part of core 12 delineates the Zone P6b/P7 boundary, approximately equivalent to the NPll/NP12 zonal boundary (cf. Aubry et al. 1988; Berggren & Miller 1988). The LAD of the St. beccariiformis fauna occurs between 202 m and 202.60 m (Schnitker 1979; Pak & Miller 1992) and is thus located in the lower part of Zone NP10 as delineated at its base by the LAD of E tympaniformis. (We recognize, however, that this constitutes only an approximation of this base which would perhaps more appropriately be taken at the FAD of D. diastypus at c. 198.5 m.) If the calcareous nannofossil biozonal interpretation is correct, we infer the presence of an unconformity in core 14, section 3, between 202.60 m and 202.15 m from the close association of the benthic foraminiferal extinction and the NP9/NP10 boundary. Sporadic bentonitic ashes (equivalent to those in Hole 550) occur at c. 195-196 m within Zone NP10 near the level of the LAD of M. velascoensis and FAD of M. subbotinae and above the LAD of the St. beccariiformis benthic foraminiferal association, generally consistent with observations elsewhere.

Site 6 0 5 ( N e w J e r s e y C o n t i n e n t a l Rise, NW Atlantic) Rather sparse sampling (Saint-Marc 1987) precludes precise determination of planktonic foraminiferal datum

318

W.A. BERGGREN & M . E AUBRY

Fig. 4. Stratigraphic relationship of events spanning the Paleocene/Eocene boundary: DSDP Hole 401.

levels/zonal boundaries, whereas relatively close sampling allows more accurate delineation of the calcareous nannofossil zonal boundaries (Applegate & Wise 1987; Lang & Wise 1987). The NP9/NP10 zonal boundary, as recognized by the FAD of T. bramlettei, lies between 558 and 559.5 m (Applegate & Wise 1987, table 1). Zone NP9 is c. 43 mthick whereas Zone NP10 is less than 9 m-thick. The FAD of Discoaster diastypus is immediately above the boundary.

The LAD of GL pseudomenardii is located within core 46 in a reversed interval identified as Chron C24r, c. 28 m above a normal event identified as Chron C25n, and within the middle of Zone CP8 (= Zone NP9). The virtually simultaneous LAD of M. velascoensis and FAD of M. subbotinae (c. 577 m) and subsequent (c. 562 m) FADs of Ps. wilcoxensis and M. formosa gracilis, the LAD of M. acuta in the upper part of Zone CP8 (= Zone NP9) and the FADs of M. marginodentata and Ac. wilcoxensis (577 m) support distinction of Zones

LATE PALEOCENE--EARLY EOCENE MAGNETOBIOCHRONOLOGY OF NW EUROPE P5 and P6a at c. 577 m within a reversed polarity interval assigned to Chron C24r. The extreme thinness of Zone P5 suggests an unconformity at c. 577 m. There are no planktonic foraminiferal data above core 43 CC (c. 557 m) owing to poor preservation. The LAD of the St. beccariiformis fauna (core 44, c. 565 m) ocurs within the upper part of Zone P6a and lies within the upper part of Zone NP9 in Chron C24r, consistent with observations elsewhere.

Hole 690B (SW Flank of Mause Rise, S o u t h A t l a n t i c ) (Fig. 5) The absence of tropical (low latitude) markers precludes direct correlation with standard planktonic foraminiferal biostratigraphic zonal scheme(s); instead a high latitude zonation was developed by Stott & Kennett (1990). In contrast, the standard calcareous nannofossil zonation of Martini (1971) is applicable to this site/region (see Aubry et al. 1996). The FAD of Globanomalina australiformis (in core 19H, at 170.64 m in Stott & Kennett 1990, p. 556 but at 171.38 m, in Stott & Kennett 1990, p. 555) is closely associated with the LAD of St. beccariiformis fauna at 170 m (Thomas 1990) and two short normal polarity 'events' within the mid-part of Chron C24r. By correlation with the sequence of events at other sites this level should be stratigraphically equivalent to a level within Zone P5 (see discussion of Hole 550 above). The FAD of Ac. wilcoxensis berggreni (core 18H, c. 159 m) and the sporadic occurrence of Ps. wilcoxensis in core 16 supports assignment of the 'normal-reversed' polarity sequence of cores 16H and 17H to Chron C24r (and not to Subchrons C24An and C24Bn (Berggren et al. 1985; = Subchrons C24n.1n to 3n; Cande & Kent 1992). The interpretation of the magnetic polarity stratigraphy for the upper Paleocene-lower middle Eocene section in Hole 690B is discussed in Aubry et al. (1996). Because the magnetic polarity stratigraphy in Hole 690B is so important to regional and global correlations across the Paleocene/Eocene boundary, we comment further on the integration of magneto- and biostratigraphy in this interval. A detailed magnetic polarity stratigraphy was developed in the lower Paleogene by Spiess (1990). A series of short normal events was recorded in this expanded stratigraphic section over the interval of cores 16H-17H (top) and correlated with Subchrons C24An and C24Bn of Berggren et al. (1995) (Spiess 1990, fig. 9, pp. 288, 289 and fig. 10, p. 292; using the new terminology of Cande & Kent (1992), this would correspond to Subchron C24n.ln and Subchron C24n.3n, respectively). A hiatus (of c. 1.5 million years duration) was shown to be present within core 15H eliminating the younger part of Chron C23n to the uppermost part of Chron C22r (Spiess 1990, p. 290) based on magnetobiostratigraphic age-depth relationships and sedimentation rates. The Chron C23r/C24n boundary was shown to lie at the normal event in the basal part of core 15H. By contrast Stott & Kennett (1990) and Pospichal & Wise (1990, fig. 4, p. 632) identified the normal polarity events in mid-

319

core 15H as Chron C22n entirely (we would identify these two normal events as Chrons C22n and C23n, respectively, and place an unconformity between the two, see Aubry et al. 1996), and another unconformity at c. 137.8 m between Zones NP10 and NPll. Kennett & Stott (1991, p. 555) indicated that the magnetic polarity stratigraphy of Hole 690B between cores 17H and 25H matches clearly the standard magnetostratigraphic scale of BKF85. Reference to their fig. 2 (1990, p. 552), however, shows that Chron C24r is drawn from mid-17H to 21H (top), whereas Spiess (1990, fig. 9, p. 289 and appendix B, p. 312-314) shows that the long reversed interval of Chron C24r extends from mid-17H to base 21H (not to 23H (Stott & Kennett 1990, p. 555) or top 21H (Stott & Kennett 1990, fig. 2)). On the assumption that the base of the series of normal events at c. 154.2 m and the top of the normal event at 185.2 m are Subchron C24n.3n and Chron C25n, respectively, Stott & Kennett (1990:555) estimated the position of the Paleocene/ Eocene boundary at c. 173.2 m (Berggren et al. 1985 = 57.8 Ma) or at 165 m (Aubry et al. 1990 ; 57.0 Ma) using the correct depth for the Chron C24r/C25n of 185.48 m (see Stott etal. 1990, p. 851, table 1; Spiess 1990, p. 279, table 4). The former level is near the FAD of G1. australiformis (171.38 m) and the LAD of the St. beccariiformis benthic faunal assemblage at c. 170 m (Thomas 1990). The latter lies approximately half way between the FAD of Gl. australiformis (171.38 m) and the FAD of Ac. wilcoxensis berggreni at 159.45 m. In contrast, Pospichal & Wise (1990) drew the Paleocene/Eocene boundary at c. 150m (core 17H/2: 28-30 cm), at the level of the CP8/CP9 zonal boundary that they delineate based on the FAD of T. bramlettei, not on that of Discoaster diastypus (the marker for the zonal boundary) at 148.89 m. The 150 m level (= base NP10) lies within the series of short 'normal' events correlated by Spiess (1990, fig. 10, p. 292) with Subchrons C24An and C24Bn (Berggren et al. 1985). Pospichal & Wise (1990) report the LAD of T. contortus (= CP9a/b boundary = NP10/NPll boundary) at 137.4 m (but see Aubry et al. 1996), in the basal part of the reversed interval identified as Chron C23r by Spiess (1990). However, the association of Zone NP10 with normal polarity events identified as Chron C24n is completely anomalous inasmuch as Zones NP10 and CP9a are known to lie within Chron C24r and the NP10/NPll = CP9a/b boundary (= LAD of T. contortus) to occur a short distance below the Chron C24n/C24r boundary (see Berggren et al. 1985 and discussion above and below). The identity/reality of the series of normal 'events' in cores 16H and 17H remains to be resolved. Pending resolution of this discrepancy we consider the magnetic polarity signatures in this part of the stratigraphic record as questionable (possibly a result of normal overprinting) and not suitable for magnetobiostratigraphic correlation and/or calibration. A marked 813C excursion has been identified over an apparently brief (c. 0.02 million years interval) centred at 170 m (essentially contemporaneous with the FAD of Globanomalina australiformis and the LAD of the St. beccariformis benthic foraminiferal association: Thomas 1990; Thomas et al. 1990; Thomas & Shackleton 1996) with an age estimated of 57.3 Ma (Berggren et al. 1985; Stott et al. 1990; Kennett & Stott 1991). This level

320

w . A . BERGGREN (~ M. P. AUBRY

Fig. 5. Stratigraphic relationship of events spanning the Paleocene/Eocene boundary: ODP Hole 690B.

was considered just to predate the Paleocene/Eocene boundary in Hole 690B, based on a magnetobiostratigraphically interpolated age (from the Subchron C24n.3n/C24r boundary) of 57.0 Ma for the 166 m level (Kennett & Stott 1991). However it will be seen that the 166-171 m interval in Hole 690B (spanning the benthic foraminiferal extinction event, the carbon isotope

excursion and the Paleocene/Eocene boundary as estimated by Berggren et al. (1985) at 57.0 Ma) lies well down within Zone NP9, c. 20 m below the NP9/NP10 boundary. Elsewhere Zone NP9 lies within the mid-to lower part of chronozone C24r. The PaleoceneflEocene boundary (revision by Aubry et al. 1988: defined as the base of the London Clay Formation, not approximated by

LATE PALEOCENE-EARLY EOCENE MAGNETOBIOCHRONOLOGY OF NW EUROPE the NP9/NP10 zonal boundary) with an estimated age of 57.0 Ma (Berggren et al. 1985) was suggested to lie within mid-upper Zone NP10, approximately correlative with the CP8/CP9 zonal boundary (= FAD of D. diastypus), which was believed earlier (Berggren et al. 1985) to correspond to a level in the mid to upper part of chronozone C24r. The reviews presented above (Holes 549, 550) are seen to compound the complexity of the issue because of the presence of an unconformity at both locations, rendering it difficult to determine the precise position of the ~13C excursion, the NP9/NP10 zonal boundary and the -17 ash within Chron C24r, and hence to establish a precise chronology for Chron C24n as well as for the Paleocene itself (see discussion below and in Aubry et al. 1996). The benthic foraminiferal extinction and carbon isotope events in Hole 690B are clearly older than 57.3 Ma in the magnetobiochronology/correlationof Aubry et al. (1988). The estimated age of the 170 m and 166 m levels in Hole 690B are c. 57.6 and c. 57.4 Ma, respectively, if one uses only the duration/length of Chron C25n (assuming that the entire normal interval between 185.48 m and 195.94 m represents Chron C25n) and extrapolates upward (using Berggren et al. 1985). The latter estimate is still well within the interval of Zone NP10 in the chronology of Aubry et al. (1988) indicating further problems with the calcareous nannoplankton biostratigraphy and/or the magnetostratigraphy of Hole 690B. The estimate of c. 57.4 Ma for the benthic foraminiferal extinction event is within previous estimates of the chronology of the lower part of Zone NP10, but this event has been shown to be linked with mid-Zone NP9 here in Hole 690B.

Site 577 (Shatsky Rise, NW Pacific) (Fig. 6 ) There are a relatively large number of datum events in both the planktonic foraminifera (Corfield 1987; Liu & Olsson, pers. comm. 1993) and calcareous nannoplankton (Monechi 1985) and a good magnetostratigraphy (Bleil 1985) spanning the interval of c. 75-85 m in Hole 577 and 70-80 m in Hole 577A, which allows the development of an integrated magnetobiostratigraphy (Monechi et aL 1985). However, it should be borne in mind that the c. 2.5 million year interval of Chron C24r is only c. 5 m thick at Site 577 precluding a clear separation of biostratigraphic events spanning the Paleocene/Eocene boundary. Unless otherwise specified the following data refer to Hole 577. Data on the relationship between the LAD of Globanomalina pseudomenardii and FAD of Morozovella subbotinae, nominate taxa for the Zone P4/5 and P5/6 boundaries, respectively, in the zonal scheme of Berggren & Miller (1988), are somewhat equivocal. Liu & Olsson (pers. comm. 1993 and verified by WAB by examination of their samples) indicate the stratigraphic overlap of these two taxa within Chron C25n, whereas Corfield (1987) indicated their overlap at the top of Chron C25n and basal Chron C24r. These observations have led to the revision by Berggren et al. (1995; see also above) to the Zone P4-6 part of the zonal scheme of Berggren & Miller (1988). Miller et al. (1987, p. 744, fig. 2) record the FAD

321

of M. subbotinae in mid-Chron C24r and the LAD of Gl. pseudomenardii in Chron C25n (the latter being consistent with both Corfield (1987) and Liu & Olsson (pers. comm. 1993)), indicating the presence of a 3--4 mthick Zone P5. However, sample spacing in these studies was relatively large; this was improved upon in the study by Pak & Miller (1992). The overlap of M. subbotinae and M. velascoensis (Zone P6a in Berggren & Miller 1988; Zone P5 of this work) over the interval of c. 86-82 m (upper core 10) extends to slightly above the midpoint of the reversed interval identified as Chron C24r, in close proximity to the LAD of Fasciculithus spp. (81 m), considered generally to occur at, or near, the top of Zone NP9 (= CP8). The LAD of M. velascoensis at c. 82-83 m is estimated to be at c. 57.8Ma (Berggren et al. 1985) by Miller et al. (1987, p. 745, table la) and 56.63 Ma by Corfield (1987, p. 100, table 6.2), apparently by interpolation between magnetostratigraphic datum levels (Bleil 1985). The benthic foraminiferal extinction occurs between 82 and 83 m. However, we believe that there is truncation at the level of the extinction inasmuch as a number of events occur near this level and there is a suggestion that Subzone CP9a is missing in this sequence (Monechi 1985; Monechi et al. 1985, fig. 2, who place an unconformity at the LAD of F. tympaniformis). However, Pak & Miller (1992) emphasize that closer sampling indicates that the LAD of M. velascoensis is within 577/10/1 and not between cores 9 and 10. The FAD of M. formosa (base of Zone P6b, this work) at c. 81 m (lower part of core 9) occurs just below a normal event identified as Subchron C24n.3n, consistent with earlier records (Berggren et al. 1985; Berggren & Miller 1988). The juxtaposition of the nominate criteria for the top of Zone P5 (LAD M. velascoensis: c. 82-83 m) and the base of Zone P6b (FAD M. formosa) in the upper half of Chron C24r reflects truncation by an unconformity. This is well supported by calcareous nannofossil stratigraphy. The NP10/NP11 zonal boundary is almost coincident with the Subchron C24n.3n/Chron C24r reversal (compare with Aubry et al. 1988). This, and the fact that the FAD of Discoaster kuepperi is at the same level as the FAD of Tribrachiatus orthostylus (see Monechi 1985, table 5), indicate that only the upper part of Zone N P l l is present. The mean age of the LAD of M. velascoensis has been estimated at 56.79 Ma (in the chronology of Berggren et al. 1985) based on data from Hole 577 and Site 527 (Corfield 1987), i.e. in the upper half of chronozone C24r, although this younger age estimate relative to Berggren et al. (1985) may reflect (in part) on the presence of unconformities at these sites. At Site 605 (see above) the LAD of M. velascoensis is well down within chronozone C24r, c. 23 m above the top of chronozone C25n. It may prove necessary to adjust the age of the chronozone P5/P6a boundary to younger within Chron C24r, but by what amount remains unclear. At Site 550 (above) we have used the LAD of M. acuta (just below the midpoint of chronozone C24r) as a proxy for the P5/P6 zonal boundary. The FAD of M. aragonensis (Zone P7) at c. 79 m (core 9) is associated with a normal event identified as Subchron C24n.3n, somewhat older than its suggested correlation with Subchron C24n.l-2n based on data from

322

W.A. BERGGREN & M . E AUBRY

rz Ig3

O

r 9

o

ID

g, ex0

~ ,....~

9O ,...i

z r .,,.~ ,.t=

,d

LATE PALEOCENE--EARLY EOCENE MAGNETOBIOCHRONOLOGY OF NW EUROPE the Gubbio section(s) (Berggren et al. 1985) and basal Subchron C23n.2r at Sites 550 and 549 (see above). At Site 527 the FAD of this taxon is shown to correlate with Subchron C23n.2r (Corfield 1987), but identification of the magnetostratigraphy at that site is not unequivocal. The correlation of the FAD of M. aragonensis (Zone P6/P7) with Subchron C24n.3n, if correct, would suggest that Zone P6b is relatively short (c. 0.3 million years). The FAD of D. multiradiatus at c. 86 m is correlative with the upper part of chronozone C25n, consistent with some records elsewhere, although there are records of a correlation with the lower part of chronozone C25n (Site 605; see also Berggren et aL 1985). At Hole 690B this datum event is correlated to a level in uppermost chronozone C25n, virtually at the subchronozone C24n.3r/ chronozone C25n boundary. The lone, rare occurrence of T. contortus in sample 577/9/5:117 cm (c. 81 m, in the reversed section just below Subchron C24n.3n and only slightly above the FAD of M. formosa (= Zone P6b) precludes precise recognition of the T. contortus Subzone (CP9a) of the D. diastypus (CP9) Zone (Monechi 1985, p. 307; Monechi et al. 1985, p. 794). The latter authors point out, however, that the LAD of Fasciculithus spp. occurs slightly below (c. 82 m) the lone occurrence of T. contortus and within the upper part (but not top) of Zone CP8 (= NP9), consistent with records in the South Atlantic (Shackleton et al. 1984). The relatively consistent LAD of Fasciculithus spp. within the lower part of chronozone C24r suggests that Subzone CP9a spans the middle to upper part of chronozone C24r, inasmuch as the LAD of Tr. contortus (nominate form for the CP9a/b boundary) is consistently reported from the upper part of chronozone C24r (Berggren et al. 1985; Monechi et al. 1985; see also above). Indeed Monechi et al. (1985, p. 793; table 3) estimated the duration of the T. contortus (CP9a) Subzone at 0.3 million years (56.3-56.0 Ma) at Site 577, although it is difficult to understand how this estimate was made inasmuch as the taxon is recorded only from a single sample at Hole 577 and does not occur in Hole 577A (Monechi 1985). Also, these authors indicate their belief that Subzone CP9a is absent at Site 577 (Monechi et al. 1985, p.789, fig. 2). The FAD of D. lodoensis is located at a level (Subchron C23n.2r) just above a normal event in both Hole 577 (74.6 m) and Hole 577A (74.43 m) identified as Subchron C24n. 1-2n, slightly younger than its reported association with the top of subchronozone C24n. 1-2n in the Gubbio section(s) (Berggren et al. 1985) but consistent with other reports from subchronozone C24n.2r (Berggren et al. 1985) and Hole 550 (above). A major (c. 3) 6~3C shift (lowering) is seen to essentially span the Chron C25n to C24n interval with a midpoint between 83-81 m in the middle of Chron C24r (within Zones NP9 and P5 as currently recognized) at c. 57.8 million years (Shackleton et al. 1985; Miller et al. 1987; chronology of Berggren et al. 1985) and assuming that the sequence is essentially continuous in Hole 577 (see remarks above). A sequential series of benthic foraminiferal extinctions occur over the interval of 89-83 m, but the major St. beccariiformis faunal extinction event is located between 83 m and 82 m apparently, in close proximity to

323

the carbon isotope shift and the Zone P5/P6a boundary (Miller et al. 1987, p. 749, 750; fig. 3; Pak & Miller 1992).

Sites 698, 700 (NE Georgia Rise) and 702 (Islas Orcadas Rise) (Figs 7, 8) Incomplete recovery precludes precise delineation of biostratigraphic subdivisions at these sites in the Atlantic region of the Southern Ocean (Crux 1991; Nocchi et al. 1991; Katz & Miller 1991). The interval between 235.7 m and 198 m in Hole 702 is characterized by the occurrence of Tribrachiatus orthostylus without Discoaster lodoensis Crux 1991). As this latter species occurs at c. 191.7 m (at a level which yields D. lodoensis and D. sublodoensis (Cieselski, Kristoffersen et al. 1988), thus assignable to Subzone NP14a), its absence between 235.7 m and 198 m indicates that the interval belongs to Zones NP10 and NP11 and does not include Zone NP12. Furthermore, since T. contortus and T. bramlettei occur at other high latitude sites (e.g. ODP Site 690), their absence in this hole can be taken to indicate Zone N P t l . This interpretation is supported by the similar correlations between planktonic foraminifera and calcareous nannofossil zones seen in Holes 549 and 550. The situation is similar at Sites 698 and 700. The FAD of Gl. australiformis at c. 70.5 m in Hole 698B, 220 m in Hole 700B and 245 m in Hole 702B allows approximate correlation of these levels with the 170m level in Hole 690B which occurs near the St. beccariiformis benthic foraminiferal extinction event and the carbon isotope shift, and in mid-Chron C24r. In Holes 698A and 702B the FAD of Gl. australiformis (Nocchi et al. 1991, p. 247, fig. 9) coincides closely with the LAD of F. tympaniformis (used as proxy for a level equivalent to the Zones NP9/10 zonal boundary) and the FAD of Tr. orthostylus (Crux 199l:169) which supports our interpretation of an unconformity at this level. The St. beccariiformis faunal extinction event is approximately correlative with the FAD of Gl. australiformis, the LAD of Fasciculithus spp. and lies within a major carbon isotopic excursion (as seen elsewhere) which would be equivalent to a level within chronozone C24r (see remarks above). A significant minimum peak in the carbon isotope record at c. 195 m (core 21) in Hole 700B within Zone NP 11 (Zone NP10-NP12 undifferentiated in Crux 1991) may be correlative with one of the two peaks associated with Chron C24n in Hole 577 and the composite Bottacione-Contessa Highway sections (Corfield et al. 1991).

The benthic foraminiferal extinction event in some outcrop sections We have examined the sequence of biostratigraphic events which occur in the calcareous nannoplankton (MPA) and planktonic and benthic foraminifera (WAB) in four outcrop sections from Spain, Israel and the North Caucasus (Georgia) in order to assess the relationship of the extinction of the St. beccariiformis fauna to calcareous

613CpDB o I

•

1 I

BFE

I N truompyj

Fig. 7. Stratigraphic relationship of events spanning the Paleocene/Eocene boundary at ODP Hole 698A.

2 J

325

LATE PALEOCENE-EARLY EOCENE MAGNETOBIOCHRONOLOGY OF NW EUROPE

,

_

,.~ r

~"

I~ o o

1

,:::::,

2 ,

ti m = .~

3

6

i,

~13CpDB

} I 0! 1

t 2.

1 1

2 I

3 ,1 J

?

~205.8 ~ i ~

!

t t:: I

/

r

E

/ 2 2 , 1 8 - -

B

~ / j

r

i

/

r /

//

234.3 ~ C4"

1

,~= ~...,.v-.~ ~ ~ 9 ~,

2430,

BFE

J

~ ~

/

/ ! P6a "

253.3-

---j /

7

'

/

Cibicidoides

spp.

-~. ~

9

N_ truempy~

TT

,/ /

// //

/

/ )" / /

272.3

z

o

Fig. 8. Stratigraphic relationship of events spanning the Paleocene/Eocene boundary; ODP Hole 702B.

326

w . A . BERGGREN ~; M. P. AUBRY

planktonic biostratigraphy. Our results are reviewed below.

Zumaya section (NW Spain; samples kindly furnished by E. Molina and correspond to those used in the study by Canudo & Molina 1992). In this deepwater flysch facies the LAD of the St. beccariiformis fauna occurs (sample 18) within Zone NP9, essentially equivalent to the top of the concurrent range of M. subbotinae and M. veIascoensis-acuta (= Zone P5/6 boundary) and within a dissolution facies (see also Canudo & Molina 1992 where this event is said to occur in sample 18, c. 26 m above the base of the local base line of this section). Calcareous microfossils are generally strongly recrystallized and poorly preserved. Caravaca section (SE Spain; samples kindly provided by E. Molina; another set of samples has been donated by Shell International Oil Company). In this more carbonate rich section the LAD of the St. beccariiformis fauna apparently occurs near the base of a dissolution interval (within which planktonic foraminifera are dissolved and calcareous nannoplankton are too heavily recrystallized to allow definitive identification), within the concurrent range of M. subbotinae and M. velascoensis-acuta (= Zone P5) and within Zone NP9. Zin Valley section (Negev, Israel; samples kindly provided by C. Benjamini). In this section the LAD of the St. beccariiformis fauna occurs in a dissolution interval characterized by a flood of radiolarians (extensively developed in the Tethyan seaway and extending from the Mediterranean to Caucasus region), at the top of the concurrent range of M. subbotinae and M. velascoensisacuta (= Zones P5/6 boundary) and (apparently) upper Zone NP9. Khieu River section (NW Caucasus, Georgia; samples kindly provided by N. Muzylov). The LAD of the St. beccariiformis fauna occurs in sample 321, within Zone P5 (c. 10 m above the FAD of M. subbotinae), within a dissolution facies spanning most of Zone NP9 characterized by silicified radiolarians and a sporadic agglutinated benthic foraminiferal fauna, c. 20 m below the FAD of Ps. wilcoxensis (= Zone P6a) and c. 28 m below the FAD of M. lensiformis (= Zone P6b) in basal Zone NP11. The results from a study of these four widely separated sections spanning the upper Paleocene-lowermost Eocene are consistent and indicate a close association between the LAD of the St. beccariiformis fauna and a level within Zones P5 and NP9.

Discussion of DSDP/ODP and onshore data There is a generally consistent correlation of the St. beccariiformis benthic foraminiferal faunal association extinction event to a level within Zone P5 and with a level within Zone NP9, the mid-point of the carbon excursion event and a level within the lower third of chronozone C24r. In

Hole 549 the St. beccariiformis faunal association extinction essentially coincides with the FAD of M. subbotinae and is closely bracketed by the LADs of F. tympaniformis (below) and the FAD of T. bramlettei (above), at an unconformity between Zones NP9 and NP10 and P5 and P6. In Hole 550 this event occurs in a dissolution facies just below an unconformity separating Zone NP9 and NP10 and within Zone P5. In Hole 690B this event coincides with the 513C spike and the FAD of Gl. australiformis and occurs within mid-Zone NP9 and Chron C24r. In Hole 577 these events are separated by a few metres and a stratigraphic gap occurs at the NP9/10 zonal boundary. Correlation of these events to southern high latitudes is afforded by the FAD of Gl. australiformis which occurs in Zone NP9 in Hole 690B at the level of the St. beccariiformis assemblage extinction and the carbon isotope shift, whereas in Hole 702B it appears to coincide with the Zones N P 9 / 1 0 boundary (as delineated by the L A D of E tympaniformis) and the FAD of T. orthostylus (the juxtaposition of which suggests an unconformity). The North Atlantic ash series occurs within the lower half of Zone CP9a and NP10. It spans the upper Zone P5 to the lower half of Zone P6a in Hole 550. The - 1 7 and +19 ashes, recently dated at 54.5 and 54.0 million years, respectively, lie within the lower part of Zone NP10 and CP9a, a short distance above the L A D of F. tympaniformis and the FAD of T. bramlettei, and bracket the Zone P5/P6a zonal boundary a short distance above the (delayed) entry of M. subbotinae and the FAD of Ac. wilcoxensis. The two ashes also bracket the FADs of M. formosa gracilis and Ig. broedermanni (below) and the L A D o f M. acutavelascoensis (above). The position of these events within Chron C24r is the remaining fly in the ointment, their determination being hampered by the presence of unconformities at Sites 549 and 550 and the lack of a reliable magnetostratigraphy at Hole 690B. The magnetobiostratigraphic correlation of the NP9/NP10 zonal boundary remains problematic but crucial to the questions surrounding the placem e n t of the P a l e o c e n e / E o c e n e boundary. The problem is exacerbated by, and probably owing to factors that include: (a) d i s c o n f o r m i t i e s / paraconformities within chronozone C24r juxtaposing various biostratigraphic datums; (b) inconsistent boundary determination owing to difficulty of determining the L A D of F. tympaniformis, uncertainty of the relationship of the base of Zone NP9a (= FAD of Discoaster diastypus) and the base of Zone NP 10 ( - F A D of T. bramlettei). An interpretation of the composite records of the magnetobiostratigraphic correlations in Holes 549, 550 and 690B (Aubry et al. 1996) suggests that the

LATE PALEOCENE-EARLY EOCENE MAGNETOBIOCHRONOLOGY OF NW EUROPE

various events observed near the Paleocene/Eocene boundary (i.e. carbon isotope minimum spike, oxygen isotope warming spike, Stensioina beccariiformis benthic foraminiferal association extinction, reduction in grain-size of wind blown dust) are closely aligned and occur within Zone P5 and within mid Zones CP8 and NP9. Romein (1979, fig. 49) demonstrated an evolutionary sequence linking T. bramlettei-contortusorthostylus in the Nahal Avdat section of Israel where these forms replace each other over a 4 m interval (fig. 13, p. 31; fig. 14, p. 34, 35). A similar sequence was observed in DSDP Hole 550 in the upper part of Zone NP10 (see above). However, two morphotypes of T. contortus (or two different species of Tribrachiatus) with disjunct ranges occur in this 62 m-thick section. The evolutionary sequence concerns only Morphotype B. The short range of Morphotype A is entirely comprised within that of T. bramlenei and no evolutionary link between the two taxa is apparent. These and related taxonomic problems as well as magnetostratigraphy and magnetochronology of the Chron C25n-C24r interval are crucial in deriving an integrated magnetobiochronology for Biochrons NP9 and NP10. If one uses the FAD of D. multiradiatus (sample 102) and the FAD of D. lodoensis (sample 123) in the Nahal Avdat section, equates them with the mid-point of chronozone C25n and the chronozone C23r/C24n boundary (Berggren et al. 1985), respectively, assumes continuity of the section (Romein 1979, p. 31, 34, 35) and calculates age~lepth relationships for the datum levels concerned, one finds the following: (a) the FADs of T. bramlettei, T. contortus and T. orthostylus fall at 57.8, 57.2 (56.8 Ma in Berggren et al. 1985) and 57.0 Ma (56.3 in Berggren et al. 1985), respectively; (b) these values imply a duration for lower NP10 (prior to the entry of T contortus) of c. 0.6 million years and a duration of c. 0.2 million years for the T. contortus (CP9a) Subzone; (c) the LAD of most Fasciculithus species was shown to occur in sample 106, 3 m below the FAD of T. bramlettei (Romein 1990, figs 13, 14) (although rare specimens of F. tympaniformis were recorded throughout the T. contortus Subzone), consistent with data reported in Berggren et al. 1985 and here. In view of the records compiled here, it would appear that the use of the LAD of F. tympaniformis to approximate the NP9/NP10 zonal boundary is not satisfactory, and that subzone CP9a is indeed not the precise correlative of Zone NP10 at its lower limit as suggested by some workers (e.g. Okada & Bukry 1980; Monechi 1985). In a companion paper to this paper (Aubry et al. 1996) we have estimated the duration of Zone NP10 to be 1.45 million years, and that of Zone NP9 to be > 0.904 million years

327

(within the constraints of the CPTS of Cande & Kent 1995). In our earlier work Zone NP10 was estimated to have had a duration of c. 1.3 million years (Berggren et al. 1985: Aubry et al. 1988) based on the proxy use of the LAD of Fasciculithus spp. for the NP9/10 zonal boundary in the absence of a magnetostratigraphic calibration for the nominate boundary taxon, T. bramlettei. The data reviewed here (and above) suggest that there is a general succession of events from LAD of Fasciculithus spp., FAD of T. bramlettei, FAD of D. diastypus, FAD of T. contortus and FAD of T. orthostylus. This sequence is seen to occur in Holes 550 and 577. However, there is a stratigraphic gap at 408 m in Hole 550, which encompasses the upper part of Zone NP9 and the lowermost part of Zone NP10. Similarly, there is a gap at c. 82.5 m in Hole 577, which encompasses upper Zone NP9 and a large part of Zone NP10. It would appear that the FAD and LAD of M. subbotinae and G1. pseudomenardii, respectively, are juxtaposed in chronozone C25n(_, and Y) that the delayed entry o f M. subbotinae, caused by the widespread dissolution event in early Chron C24r is the reason for the erratic/anomalous records (i.e. stratigraphic overlap or separation) of the relationship between these two taxa. An assessment of available deep sea drilling sites which span the Paleocene/Eocene boundary stratigraphic interval suggests that there is no complete/ continuous (i.e. ideal) section present in which one can discern the complete succession of either bioor magnetostratigraphic events from Chrons C25n to C23n.

The Paleocene/Eocene boundary: identity and geochronology N W E u rope

The marine stratigraphic succession developed in the sedimentary basins on the passive margins of the Baltic Plain surrounding the North Sea (the Anglo-Belgian-Paris-Danish-North German) has come to serve as the standard for mondial Paleogene chronostratigraphy (Berggren 1971; Pomerol 1973; Curry 1965; Curry et al. 1978). Unfortunately exposures in this region are often poor and stratigraphic sequences are limited in both vertical and lateral extent and riddled with unconformities. Repeated transgressions and regressions have left a characteristic imprint of rapid lateral and vertical facies changes, and late Neogene glacial tectonism has further added a final coup de grace in some areas (e.g. Denmark) in the form of physical dislocation of important parts of the stratigraphic record. Nevertheless a network of biostratigraphic

328

W.A. BERGGREN & M.R AUBRY

and magnetostratigraphic studies (including Aubry 1983, 1985, 1986; Aubry et al. 1986; Townsend & Hailwood 1985; Steurbaut & Nolf 1986; Knox et al. 1990, 1994; Ali et al. 1992, 1996), coupled with petrological studies on volcanoclastic deposits intercalated in the uppermost Paleocene-lowermost Eocene sedimentary sequence of northern Europe and the North Sea Basin (Morton et al. 1983; Knox & Morton 1988; Morton & Knox 1990; Knox 1990) is now providing a framework for an integrated Paleogene chronostratigraphy and an estimate of the position and numerical chronology of the Paleocene/Eocene boundary relative to the geomagnetic polarity time scale (GPTS). In the remaining part of the paper the chronology of the late Paleocene-early Eocene is based on the revised magnetochronological scale developed by Cande & Kent (1992, 1995) and not on that of Berggren et al. 1985, unless otherwise indicated. In epicontinental northwest Europe, the Paleocene Series is represented by the Danian, Selandian and Thanetian Stages. The base of the Thanetian (Pegwell Marls and stratigraphic equivalents in a borehole in south Essex: Knox et aL 1994) is within the Zone NP6-NP7 interval and C26n, and the succeeding Thanet Sands are referable to Zone NP8 (see Aubry 1983, 1985, 1986). The overlying Woolwich-Reading Beds belong to Zone NP9 (Siesser 1987). The lowest calcareous fossiliferous horizons in marine sediments representative of the Ypresian Stage have been assigned to Zone N P l l in the London Clay Formation in England (Aubry 1983, 1985, 1986), the Formation de Varengeville at Varengeville, Normandy, France (see Aubry 1983, 1985, 1986) and Belgium (Steurbaut & Nolf 1986; Steurbaut 1988). Zone NP 10 has not been recognized in NW Europe owing to the absence of calcareous facies. However, dinoflagellate biostratigraphy has enabled a correlation between calcareous and noncalcareous facies, and this has been discussed in detail elsewhere (Berggren et aL 1985). Various lithostratigraphic schemes have been proposed to describe the Paleocene series in the North Sea. Five lithostratigraphic units (A1 to E3) were described by Knox et al. (1981) to further subdivide the Ekosfisk Formation, the Montrose Group and Moray Group (both groups being equivalent to the Rogaland Group), which Deegan & Scull (1977) had differentiated. These lie between the top of the Cretaceous chalk and Balder Formation. These, in turn, have been correlated with the shallow water deposits of England (e.g. Cox et al. 1985). The position and correlation of the North Atlantic volcanoclastic tuff series with regard to standard biostratigraphy and magnetostratigraphy is critical to an evaluation of the position and age

of the Paleocene/Eocene boundary. The record of at least two phases of early Paleogene explosive volcanism in the NE Atlantic is found in the form of volcanoclastic deposits intercalated in marginal marine sediments of NW Europe (Knox & Morton 1983, 1988). Distal, attenuated representatives of the younger (Phase 2) ash series have been recorded over 1000 km to the southeast in deep sea sediments of the Goban Spur (Hole 550) and Bay of Biscay (Hole 403) (Knox 1985) in the NE Atlantic. The source of the NW European and eastern Atlantic pyroclastic deposits is thought to have been the so-called Hebridean-Thulean volcanic province of Scotland, the Faeroes and East Greenland (Morton & Parsons 1988, table 7). The pyroclastic deposits of the first (older) Phase 1 were restricted to volcanic activity in the British and Faeroe-Greenland province (Knox & Morton 1988); those of the younger Phase II have been linked with the Faeroe-Greenland igneous province, more specifically the proto-GreenlandScotland Ridge (Morton & Knox 1990; Miller & Tucholke 1983; Berggren & Schnitker 1983). Phase I tuffs occur in stratigraphic units C 1 and C2 and the lower part of Unit C3 (Lista Clay Formation) in Norfolk (Cox et al. 1985), correlated in turn with the upper part of Chron C26n. Calcareous nannoplankton biostratigraphy (Siesser 1987) of outcrops in SE England enabled Knox & Morton (1988) to suggest extension of Unit C3 into Zone NP8, indicating that the Forties Sand (= Unit D) is no older than Zone NP8. Phase I ashes have also been recorded from the lower Thanet Beds at Pegwell Bay which also belong to Chron C26n (Townsend & Hailwood 1985; Aubry et al. 1986) which, in turn, appears to be correlative with the younger half of Chron 26N (Berggren et al. 1985; Berggren et al. 1985, 1995). These ashes have also been correlated with the upper part of the A l i s o c y s t a m a r g a r i t a Zone by Knox & Morton (1988) (see also Powell 1988). It would appear that the Phase I ashes (at least those recorded from the lowermost Thanet Sands at Pegwell Bay) are to be more appropriately correlated with the lower part of the A. m a r g a r i t a Zone if the correlation of that zone with Zone NP7 is correct (Powell 1988). Phase 2 pyroclastic deposits have been divided into 4 subphases (Knox & Morton 1988) and occur in the Sele Formation (stratigraphic unit E of Knox et al. 1981) and Balder Formation of the North Sea and at equivalent stratigraphic levels in onshore and offshore boreholes, as well as in NE Atlantic drillsites (e.g. Sites 403, 550). In Denmark, a remarkable series of almost 200 ashes corresponding to subphases 2a and 2b occur in a finegrained diatomaceous unit known as the Mo Clay or Fur Formation (BCggild 1918; Andersen 1937). These ash beds are divided into a lower (negatively

LATE PALEOCENE--EARLYEOCENE MAGNETOBIOCHRONOLOGY OF NW EUROPE numbered series (-1 to -39) of mixed basalticrhyolitic composition (corresponding to phase 2a) and an upper (positively numbered) series (+1 to +140) of predominantly basaltic composition (corresponding to phase 2b). The negative series has been correlated with the following silicoflagellate zones (in ascending order): the Naviculopsis constricta Zone (ashes -39 to -35), the N. danica Zone (ashes -4 to -21A), the Dictyocha elongata Zone (ashes -19 to -17) and the Corbisema naviculoidea (partim) Zone (ashes -17 to 0), whereas the positive series has been correlated with the C. naviculoidea Zone (partim) (ashes +1 to +19) and the N. aspera Zone (ashes +20 to +130) (Perch-Nielsen 1976). The lower part of the negative ash series has been correlated with the upper part of the Apectodinium hyperacanthum (dinoflagellate) Zone. The upper part of the negative series, beginning with the -19 ash layer, and the positive ash series have been equated with the (?) local Acme Zone of Deflandrea oebisfeldensis in the upper part of the A. hyperacanthum Zone (Hansen 1979; Heilmann-Clausen 1982; Nielsen & Heilmann-Clausen 1988). The lowermost negative ashes have been (tenuously) correlated with the Sables d'Erquelinnes (Belgium) assigned to Zone NP9 (Heilmann-Clausen 1985). Subphase 2a ashes occur extensively in the North Sea Basin (Sele Formation and in onshore beds stratigraphically equivalent to the WoolwichReading Beds (Cox et al. 1985) as well as at deep sea drilling sites from Goban Spur (550) and the Bay of Biscay (403). In Hole 550 Knox (1984) has shown that a series of over 40 ashes (correlated with the Phase 2 ashes of the North Sea area) are restricted to Zone NP10. The -17 and +19 ashes are seen to lie within the lower third of Zone NP10 and the lower half of Chron C24r (see discussion above). Elsewhere the -17 ash has been shown to lie within the D. oebisfeldensis Acme Zone, equivalent with a level in the upper part of the Sele Formation, within the Hales Clay unit of the Harwich Formation and (probably) with a level close to (but above) the Woolwich-Reading/ Oldhaven Beds boundary (Knox 1990). There is probably a significant unconformity between the Woolwich/Reading Beds and the base of the Harwich Formation, however (Knox 1994). Recent magnetostratigraphic studies (All et al. 1992, 1996) have shown that the lower part of the Harwich Formation (Ellison et al. 1994, including the Oldhaven Beds and Division A1 of King 1981) and the lower part of the London Clay Formation (Divisions 2 and 3 (partim) of King 1981) can be assigned to Chron C24r. Until recently there was no documented record of Chron C25n in SE England (Townsend & Hailwood 1985; Aubry et al. 1986), its absence apparently owing to the uncon-

329

formity/hiatus between the Woolwich Shell Beds and the Oldhaven Beds (Harwich Formation). However, a recent study has revealed the presence of an areally restricted c. 2 m thick normal polarity interval assignable to Chron C25n at the base of the Upnor Formation (Bottom Bed) of the London Basin (Ellison et al. 1996; Ali & Jolley 1996). The presence at Sheppey of a (lower) normal polarity interval in Divisions B1-2 and an (upper) normal polarity event in Divisions C1-2 of the London Clay Formation (King 1981) assigned to Zones NPll and NPll/NP12, respectively, and in stratigraphically equivalent levels in Belgium, provides (familiar) constraints for determining the position of the Paleocene/Eocene boundary. The two normal events are assignable to Subchrons C24An and C24Bn (=Subchron C24n.ln (and probably C24n.2n) and Subchron C24n.3n, respectively; Ali et al. 1992). The lowest part of the Harwich sequence has reversed polarity (= Chron C24r; Ali et al. 1992). Knox (1990) has reported volcanic ash layers in sandy mudstones (Hales Clay of Knox et al. 1990) from Norfolk and Suffolk. A comparable mudstone with ash layers, including the -17 ash, was said to occur along the western margin of the southern North Sea. The similarity in stratigraphic position of the two mudstone sequences relative to the ash horizons led Knox (1990) to suggest correlation of the base of the 'London Clay' (now base Harwich Formation) to a level correlative with Zone NP10. Although the -17 ash has not been positively identified in outcrop sections of southern England, basaltic ashes equivalent to those of the North Sea and NE Atlantic (where they occur in Zone NP 10) have been identified in the Hales Clay (King's Division A1) of East Anglia. This latter is generally considered stratigraphically equivalent or only marginally younger than the Oldhaven Beds of the London-Hampshire Basins (King 1981; Knox 1990). Finally, Ali et al. (1992; fig. 19) correlated the base of the 'London Clay' (now Thames Group) (Division A1) transgression with the rapid rise in sea level in the middle part of cycle T.2.4 of Haq et al. (1987) and showed that it was correlative with the early part of Chron C24r (Ali et al. 1992; fig. 17). The position and age of the Paleocene/Eocene boundary in terms of the Geomagnetic Polarity Time Scale (GPTS) continues to elude stratigraphers. A significant component to the solution of this problem lies in the development of a correlation network which integrates data on the distinctive -17 (54.5 Ma) and +19 (54.0 Ma) ashes within the calcareous plankton biostratigraphic framework of DSDP Hole 550 (Fig. 9) and the combined ~13C record in Holes 549 and 550, the global deep sea stratigraphic record and placement

330

w . A . BERGGREN (~ M. P. AUBRY

~I

P7 340-.

o~

+ .-.....--...

.~

c"B'~Z.

"

!'~

-o

,~

o,,I

o

EL

lL

',r-

" I

.-...

-~-

350-~

EL

"5 t-. Q. t,_ 0

I:~ "~-

(b

0

t_ 0 -It-

9

,

"5 tv u)

370.~ r~

E 9

:

(b

: m"

380"

O~1

lL

13_

:

i Q}

::

~

9

~ -vvv.

i"

"

--

.,-vvvv

:

~

Lo 13..

. -

i

I

I

,-v-~ ,v.~ w~.,v,

I i

: :

L,vv.~ ,.,.,v~ ,.,.: .,..~-~

: :

o o o ~@o o ~ '~ o ,' o d ~ .o O'L - o '~ ~ S ~ " ^ o = o o .0 o o" 'oOoOoOo o OoO o oo%,^~ Oo( o PoO^o O o O otDOO'3:"~'C, O o O o ~ o" o o , o o o o~176 q|u"'o 6 o ~o~o 1 7 6 o o " oo ~, o'~----:,;~,~ OoO . ooo , o oo, o ooo~o ,.o o c ~o1 7 o~O r _ o o o ~o1 7 6o [ ~"-. ~ G ' ~ o ~o o o~176 o ~ 1 7 6 o o~176 o=r176176176176162176176176 ~oo~ 'o ~o ~)o?o ~ ~ 1 7 6 ~ . ~ o ~ ~o ~ 1 7 6 ~ . ~ . , ~ o ~ ~ ~ ~ ~~176 ~ ~oo ~ "~0~ o~ -

lL

".,.M

or

o~

! r i l l

-

420----

Positive ash

I~

NP5

series Negative ash series 0 r 1 6 2 +19 ash

o0o

-17 ash

i::::= ~

~

n

,-.

!

~-

:

LC

"'~(I)--

a:

L----

(P4)430"~.c2so

~

00

! ~ '1 !

44O

Z

Z

CP4

+

LAD 7 I n c l u d i n g ~" bracketing

} FAD . ) l i m i t s D

Zonal b o u n d a r y uncertainty

'~ Ckl

Benthic foraminifer extinction event

i (5

Fig. 9. Magnetobiostratigraphic correlations spanning the Paleocene/Eocene boundary in DSDP Hole 550.

LATE PALEOCENE-EARLYEOCENE MAGNETOBIOCHRONOLOGYOF NW EUROPE of the (standard) stratigraphic record in NW Europe in a sequence stratigraphic framework. Recent studies by J. R. Ali, E. A. Hailwood, J. Hardenbol, C. King, R. W. O'B. Knox, us and others have shown the following. 1. The base of the Harwich Formation (Oldhaven Beds = Hales Clay) is within early Biochron NP10 and probably within early to mid Chron C24r. 2. The North Sea Phase 2 ash series occurs primarily within Zone NP10, but may extend downwards into Zone NP9. The -17 ash is within the Deflandrea oebisfeldensis Acme Zone which is biostratigraphically equivalent to lower Zone NPI0. 3. In DSDP Hole 550 the -17 ash (54.5 Ma) lies at c. 400 m, about a third of the way up in Chron C24r; the +19 ash (54.0Ma) lies at c. 393 m slightly below the mid-point of Chron C24r (Fig. 8). 4. In DSDP Hole 550 the -17 and +19 ashes are bracketed by the (delayed) entry of Morozovella subbotinae and the FAD of Acarinina wilcoxensis (below) and the FAD of M. lensiformis (above). The FADs of Morozovella formosa gracilis and Igorina broedermanni and the LAD of Morozovella acuta (above) occur between the two ashes, i.e. they lie within (revised) basal Zone P6a and the P5/P6 zonal boundary, respectively. The -17 ash lies 8 m above the FAD of T. bramlettei, c. 27.8 m below its LAD and c. 27.35 m below the FAD of T. contortus Morphotype B. The +19 ash lies 15 m above the FAD of T. bramlettei, c. 20.8 m below its LAD, and c. 20.35 m below the FAD of T. contortus Morphotype B (Fig. 9). 5. The composite 8~3C records of Holes 549, 550 and 690B suggest that the prominent spike associated with the major turnover in deep water benthic foraminifera at other deep sea sites occurs in the early part of Chron C24r, within Zone NP9 and P5. There are problems with this straightforward interpretation, however, as underlined below and delineated in Fig. 10 (see also Aubry et al. 1996 for a more thorough discussion of this problem). These problems are associated with dissolution and an unconformity/hiatus at c. 408 m in Hole 550. We discuss the problem of dissolution first. Dissolution is moderate (upper half of chronozone C24r) to strong (lower half of chronozone C24r) in Hole 550. By carefully examining the c. 63 m-thick chronozone C24r stratigraphic interval in the course of a stable isotope study, Ashish Sinha (personal communication 1993) has estimated that c. 4.7 m (12.5%) of the stratigraphic interval between the -17 ash and the base of Subchron C24n.3n is a dissolution facies. Corresponding estimates for the interval from

331

the -17 ash to the top of C25n are 12-14m (60%). Sediment acumulation rates reflect this phenomenon: above the +19 ash rates are estimated at c. 5.1 cm/103 years, whereas below this level (in the interval of strong dissolution) they are c. 14-16 cm/103 years. However, these rates are quite unrealistic and alternative estimates are discussed in greater detail below. In an attempt to restore the c. 63 m stratigraphic interval of Chron C24r to its 'true' thickness, so as to determine the position of the -17 and +19 ash beds, that of the relevant biostratigraphic datum levels, and thus establish the appropriate ages for these datum events and, eventually, the age of the Paleocene/ Eocene boundary, we have added c. 5 m to the upper part of the Chron C24r interval and 13.5 m to its lower part. This results in a total thickness of c. 81 m for Chron C24r and a net change of c. 18.5 m (Fig.10). If we now assign the -17 and +19 ashes to their 'correct' (i.e. restored) level, we see that they lie at 398 m and 401 m, respectively (c. 5 m lower than their former (unreconstructed) level. Within the context of the revised thickness of Chron C24r, these now lie at c. 44% and 53%, respectively, of the way up in chronozone C24r, i.e. slightly higher than in the unrestored section (see Fig. 10). The net change in the position of the +19 ash is seen to be insignificant, whereas that of the -17 ash is relatively more significant (a 10% change upwards in chronozone C24r owing to the fact that the lower part of chronozone C24r has been expanded by c. 60% (Fig. 10). The second part of the conundrum is the fact that Chron C24r is not completely represented at Hole 550. We have already discussed (above) the fact that the normal/reversed magnetic polarity boundary at c. 4 2 2 m may not be the Chron C24r/C25n boundary. However, even if it is not, the biostratigraphic data reviewed above indicate quite clearly that the reversed polarity interval just above c. 422 m corresponds to the earliest part of Chron C24r. In any event, Chron C25n is present in Hole 549 and the thickness of the stratigraphic interval between the Chron C25n/C24r ~md the NP9/NP10 contact is essentially the same as that of the interval between the unconformity and the questionable Chron C24r/C25n boundary in Hole 550. Hole 549 could just as easily be substituted for the lower (pre-unconformity) interval. Of greater importance is the presence of an unconformity at c. 408 m heralded by the close juxtaposition of the FAD of Tribrachiatus bramlettei, the benthic foraminifieral extinction and the carbon isotope excursion - events which are separated by a discrete stratigraphic interval in continuous sections as in Hole 690B (Aubry et al. 1996). To determine the relative position of the

332

w . A . BERGGREN & M. P. AUBRY % C24r ISed- Rate Dissol. T in m I m/million]Interval Strat. I years (cumul.] Interv.

ZONES

NP

P

i

,rm r.)

RESTORED STRATI( IRAPHIC POSITION OI DATUM LEVEl

.,..,_

350 - -

3.~r NP11

;24n.2-3

P6b

TTT

--366~ _

--374--

_

P6a

_

0 0 >0r

400 - - o o _

~ NP10 ooooo

)oc

51.8

62.85

-4.7

.J_

P6a

O 0 +19 Ash, 54 Ma

ooooo

P6b

P5

NP5

T

Benthic Ext.

NP9

0 0 398

ooo-...~.. 0 0 0 0 0

0 0 -17 Ash, 545 Ma

ooooo

~408:~

450

NP11

_

i

NPIO

ooo

405

P5 ~

(422)

P5

(P4)

C25n-(441) C27n

O

Fig. 10. Revision of datum levels and restored thickness: DSDP Hole 550.

513C spike and the -17 and +19 ashes in Chron C24r in order to obtain a better estimate of the chronology of this latter and determination of the age of the 513C spike and the biochronology of the calcareous plankton events in Chron C24r, it is necessary to have a complete record of it in a deep sea core or outcrop (which we do not believe has been demonstrated in any published record to date) or to reconstruct a composite record. This has been attempted below and in Aubry et al. 1996). The problem with deriving an internally consistent chronology within Chron C24r and C24n is exacerbated by the fact that there is an unconformity at 408 m in Hole 550, at a level which separates mid-Zone NP9 from lower Zone NP10. A further, and more immediate, problem is deriving an internally consistent magnetobiochronology of calcareous plankton datum events in the Paleocene-Eocene boundary interval spanned by Chron C24r. Figure 11 illustrates (a part of) the problem and its construction and implications are discussed below.

Line A is drawn through the age v. depth calibration points in Hole 550 of magnetochron boundaries from Chron C25n/C24r to Subchron C23n.2n/C23n.lr using the ages estimates in the GPTS (Cande & Kent 1995; Tables 2 and 3). A relatively straight line over the c. 5 million years interval is seen with an inflection in the Chron C24n interval suggesting decreasing rates of sedimentation in the younger c. 2 million years part of the record. The fact that the NP9/NP 10 boundary at 408 m does not fall on the line reflects the fact that the NP9/NP10 (chrono)zonal boundary in the GPTS (CK92, 94) is located two-thirds of the way down in Chron C24r, thus differing slightly in position compared to Hole 550. Line B is constructed using the age v. depth calibration points on the -17 (400 m) and +19 (393 m) ashes and extrapolating down to the Chron C25n/C24r boundary at c. 423 m. The position of the -17 ash as dated by Obradovich (in Berggren et al. 1992) at 55.0 Ma is seen to lie on Line A at c. 400 m. The fact that the -17 and +19 ashes do not fall on Line A reflects

Table 2. Revised normal polariO, intervals (Chron C25n to C23n.2n; Cande & Kent 1995)

Chron/Subchron

Age (Ma)

Depth (m) in Hole 550

C25n

55.904-56.391

c. 423 (top)-?

C24n.3n C24n.2n C24n. in C23n.2n

52.903-53.347 52.757-52.801 52.364-52.663 51.047-51.743

C24n.3n(o)-NP9/10 boundary at 408 m

53.347-55.00

359.65-349.62 342.60-344.81 325.9-328.74/337.33 (or 333.03) 359.65-408

Remarks C25n truncated/concatenated with ?C27n in mid part. 55.0 Ma is calibration point at 408 m used by Cande & Kent (1995)

LATE PALEOCENE--EARLY EOCENE MAGNETOBIOCHRONOLOGY OF NW EUROPE

333

Table 3. Estimated sedimentation rates in DSDP Hole 550 Chron/Subchron

Time span (million years)

Cm/1000 years

C25n(y)-C23n.2n(y) C25n(y)-C23n.2n(o) C25n(y)--C24n. ln(y) C25n(y)-C24n.3n(y) C25n(y)--C24n.3n(y) C25n.3n(o)-408 m C24n.3n(o)-C24n. ln(y) C24n.3n C24n.3n(o)-C24n. ln(y) C27n.3n(y)-C23n.2n(y) C24n.3n(o)-C23n.2n(y) C24n.3n(o)-C23n.2n(o) Top ash series to 408 m

4.857 4.161 3.540 3.001 2.557 1.653 0.983 0.444 0.539 1.856 2.30 1.604 0.540

1.99 2.16 2.28 2.44 2.45 2.92 1.78 2.25 1.40 1.27 1.467 1.65 4.28

Top ash series to base ash series + 19 ash to C24n.3n(o)

0.467

4.28

0.653

5.107

-17 ash to C24n(o)

0.847

4.76

-17 to +19 ash

0.500

1.4

-17 to +19 ash

0.16

4.28

the imprecision in placing the calibration point (55.0 million years) in the chronology of Chron C24 owing to the fact that it is based on an estimated sedimentation rate, imprecision in the dating, and to a lesser extent, may be a function of the 'wiggle' inherent in a cubic spline which has been fit to calibration points spaced c. 10 million years apart. The unrealistic aspect of Line B is seen in the steep increase in sedimentation rate above the +19 ash required to satisfy the age of the Chron C24r/C24n boundary (Cande & Kent 1995). In the discussion below, we present a line of argumentation directed at resolving, at least for the present, the dilemma posed by the present GPTS (Cande & Kent 1995) in constructing a coherent and consistent chronology of events in the younger half of Chron C24r and its application to events associated with the Paleocene-Eocene boundary interval. Radioisotopic ages on the - 1 7 (54.5 Ma) and +19 (54.0 Ma) ashes in N W Europe (Swisher & Knox 1991) have played a key role in the development of a revised GPTS (Cande & Kent 1992, 1995). These ashes are situated 7 m apart, at 400 m and 393 m, respectively, in DSDP Hole 550, and a sedimentation rate (1.4 cm/1000 years) was used to estimate

Remarks Based on derived age of 54.46 Ma for top ash series using age model in item 8 (above) and 55.0 Ma calibration at 408 m level. Using assumptions above. Using age of 54.0 Ma on +19 ash and GPTS of Cande & Kent (1995) for C25/C24n boundary. Using age of 54.5 Ma for -17 ash and GPTS of Cande & Kent (1995) for C25r/C24n.3n(o) boundary. Using ages of 54.5 and 54.0 Ma on -17 and +19 ashes, respectively. Using assumptions for top ash series to 408 m.

the age of the (supposed) calcareous nannoplankton NP9/NP10 zonal boundary (based on the FAD of Tribrachatus contortus and L A D of Fasciculithus tympaniformis in Mtiller 1985) at c. 407 m. Our studies have shown (1) that: T. contortus is not present at 407 m, nor indeed, in the stratigraphic section i m m e d i a t e l y above. The T. contortus complex, in fact, does not appear until c. 380 m; (2) On the other hand, Tribrachiatus bramlettei, nominate taxon for the base of Zone NPI0, does appear at c. 408 m; and (3) that there is an unconformity at 408 m which separates lower Zone NP10 from mid-Zone NP9. The appearance of T. bramlettei in Hole 550 is thus a delayed entry and not a true FAD. The amount of stratigraphic section missing (and the c o r r e s p o n d i n g time interval not represented) at Hole 550 is virtually impossible to determine owing to the fact that we are unable to derive an internally consistent c h r o n o l o g y for C h r o n C24r. This is due, in turn, to problems with the manner in which the recently revised GPTS (Cande & Kent 1992, 1995) was constructed and with the calibration point used in Chron C24r. The age estimate of 55 Ma on the 408 m level in Hole 550 was used as one of nine calibration points to which a spline function was fit by Cande

334

w . A . BERGGREN & M. P. AUBRY

& Kent (1992, 1995) in the construction of the GPTS. The position of this calibration point within Chron C24r is obviously a critical element in determining the Chron C24r chronology and its constituent biostratigraphic datums. Its position was placed by Cande & Kent (1992, 1995) at Chron C24r.(0.66) based on Berggren et al. (1985). Our studies on Hole 550 have shown that the calibration point (55.0 Ma) is situated at an unconformity, and that significant dissolution has diluted the stratigraphic section in Chron C24r, particularly the lower part. If we attempt a restoration of the section to its original thickness (Fig. 11), the position of the unconformity lies somewhat higher in the expanded (but still incomplete) Chron C24r. Generally speaking the position of the (true) NP9/NP10 zonal boundary could lie somewhat higher in Chron C24r if sedimentation rates (which we cannot calculate because of the presence of an unconformity at c. 408 m) remained constant over the early part of Chron C24r (this point is discussed at greater length in Aubry et al. 1996). Table 2 lists the ages of boundaries between Chron C25n and Subchron C23n.2n from Cande & Kent (1995). Sedimentation rates in Hole 550 are then estimated (Table 3) for the stratigraphic sequence spanning all or parts of the corresponding nearly 5 million year interval. Observations on this approach are listed below. 1. At first glance sedimentation rates over the nearly 5 million year interval would appear to have varied from c. 1.4 cm/103 years to over 5 cm/ 103 years. These rates, however, are based on a variety of assumptions which we shall examine further below. 2. A rate of 2.44 cm/103 years is estimated for Chron C24r using Chron C25n(y) to Subchron C24n.3n(y). The problem is that this estimate runs across the unconformity at 408 m and does not account for the time not represented at this level. However, it is consistent with the GPTS (Cande

& Kent 1995) which was derived by using the calibration of 55.0 Ma at c. 408 m in deriving an age for the Chron C24r/C25n boundary in Cande & Kent (1995). 3. Sedimentation rates of c. 1.5 cm/103 years to 2.25 cm/103 years occurred over the Chron C24n to C23n interval. The fact that sedimentation rates for Subchron C24n.3n (2.25cm/103 years) and Subchron C24n.3n(y) to Subchron C23n.2n(y) (1.27 cm/103 years) were reduced by almost a factor of two attests to the gradual reduction in sedimentation rates over the later part of Chron C24r to C23n interval. 4. Sedimentation rates fluctuated significantly, if not radically, in Chron C24r if one uses both the polarity estimates at the lower and upper boundaries of the Chron and integrates the radioisotopic ages on the -17 and +19 ashes at face value. (a) A sedimentation rate of 1.4 cm/103 years is derived for the 7 m (and 0.5 million years) interval between the two ashes and estimated ages for the top (384.85 m) and base (404.60 m) of the 19.75 m thick ash series are 53.41 Ma and 54.82 Ma, respectively, for a span of 1.41 million years. (b) If we calculate a rate of sedimentation from the +19 (54.0 million years) ash to the Chron C24r/C24n boundary (to satisfy the constraints of the GPTS, Cande & Kent 1995), a rate of 5.1 cm/103 years for the upper c. 33 m of Chron C24r results. If the 54.5 Ma age on the -17 ash (400 m) and the 53.347 Ma age estimate on the Chron C24r/C24n boundary at 359.65 m (GPTS; Cande & Kent 1995) is used to calculate a sedimentation rate for the upper 40 m of Chron C24r, a rate of c. 3.5 cm/103 years is derived. These rates are counterintuitive to evidence from the lithostratigraphic record of the cores in the Chron C24rC23n interval of Hole 550. The sediment is a calcareous nannofossil ooze with significant dissolution in the interval between c. 400 to 425 m and

.

300 NPl1112 (b '~

325 A

~ 375

I~

-17 Ash P5/6

C24n'3n(o) NPlOI11 /

C24n.3n(y) J

~

l

024n.1 n(y) / ~

P6/7

C23n.2n(o ) ~

C23n.2n(y) ]

1 -

-

(+1g/Ash)

-

"NP9/10" (JD~)) l ~ ~ NP10/11 C25ny) ~ ~ ~ "~. P6aro 400-" 1 ~ ~ B ~ unconformity at ~408m ..,.,,T ~ ~ ~ ~ - - - "NP9/10" -17 Ash (CCS) +19 Ash (CCS) 42556.0.'g '.'a '.; '.'6 '.; '.: '.'3 '; '', 55.0 '9 ' 'a' '7' ', ''s' '4 ''3' ;' '~'5/.'0 '9' =s' ',' '6' 's' .'4' .'3' '2' ', 5~.0 '9' '8' '7' '6' '5 ' .', ' .'3 ' .'2 ' ', '5~.0.'9 ''8 ' .'7 ' .'6 ' .'5 ' : '.'3 ' '2 ' .', '5~.0

Time in million years

Fig. 11. Age versus depth framework of magnetostratigraphic boundaries and -17 and +19 ash beds in DSDP Hole 550. See text for discussion.

LATE PALEOCENE--EARLYEOCENE MAGNETOBIOCHRONOLOGYOF NW EUROPE only minor dissolution above. Sedimentation rates would be expected to have increased (not decreased) over the 20 m interval characterized by the input of volcanogenic detrital material to the normal pelagic carbonates. Sedimentation rates of c. 3.5 to > 5 cm/103 years in a normal pelagic carbonate (later part of Chron C24r in Hole 550) are unrealistic unless a significant increase in productivity can be demonstrated. No such evidence is seen in the carbonate record (see de Graciansky, Poag et al. 1985, DSDP Leg 80, Hole 550, p. 300-303 (Barrel Sheets) and p. 335-338 (core photographs)). Such high rates are contradicted by the significantly lower (c. 2-1.5 cm/103 years) rates of sedimentation in the pelagic carbonates of Chron C24n-C23n interval above and which would appear to be extrapolatable downwards into the carbonate section in the younger part of Chron C24r. These rates also assume constant/similar rates of sedimentation in both the volcanogenic part of the section (c. 385-405 m) and the non volcanogenic, carbonate part (c. 385 to 360 m, and above). We can only conclude that the use of a constant sedimentation rate over the entire Chron C24r interval is an appropriate working assumption only if the radioisotopic ages on the ashes are not used in deriving calibration points for the calculations. However, this assumption falters also on other grounds, which are discussed below (see point 5). (c) These observations suggest to us that the use of the radioisotopic ages on the -17 and +19 ashes to derive an estimated age for the 408 m level in Hole 550 and its use, in turn, as a calibration point in the GPTS (Cande & Kent 1995) is fraught with danger. While the isotopic ages on the ashes, at least on the -17 ash, may be analytically precise, we would query their geological accuracy. The 55.0 Ma age on the -17 ash by Obradovich (in Berggren et al. 1992) v. the 54.5 Ma age on the same ash by Swisher & Knox (1991) reinforces this uncertainty. Further to the point, Obradovich (pers. comm. 1994) informs us that he has recently dated the +19 ash at 54.5 ___0.5 Ma, the large error (comparable to that in the 54.0___0.5 Ma age of Swisher & Knox) being due to the heterogeneous nature of the ash with visible reworked detrital feldspars common. He cautions against using the +19 ash for dating purpose and indicates that the -17 ash remains the most homogeneous and 'cleanest' ash for dating the ash series. (d) There is no solution to this conundrum at the present time because the 55 Ma age at c. 408 m is an extrapolated age estimate at an unconformity whose position in Chron C24r remains elusive. The next generation of GPTS will require the integration of additional radioisotopic ages on well calibrated levels within a continuous/

335

complete Chron C24r and preferably within uniform lithic (carbonate) facies. 5. An alternative approach to estimating the sedimentation rate(s) within the Chron C24n to Chron C24r in Hole 550 involves the following three steps. (a) We have seen above (Table 3) that sedimentation rates decreased from c. 2.5 cm/103 years in Chron C24r (integrated over its extent) to c. 1.5 cm/103 years in the younger part of Chron C24n. An estimate of c. 2.25 cm/103 years for Subchron C24n.3n is seen to have decreased to c. 1.4 cm/103 years for the Subchron C24n.3n(y) to C24n.ln(y) interval and an integrated value of c. 1.78 cm/103 years is estimated for the Chron C24n interval. While an unconformity (and brief hiatus) cannot be excluded between subchronozone C24n.3n and C24n.ln (in the absence of definitive evidence for the presence of subchronozone C24n.2n), it is more likely that the section is continuous in this interval and that the absence of the relatively short (0.054 million years) Subchron C24n.2n is due to the moderate sedimentation rates. We proceed by using the 1.78cm/103 years estimated for the Chron C24n interval to estimate datum events in the Chron C24n to Subchron C23n.2r interval. (b) We use a sedimentation rate estimate of 2.25 cm/103 years (in Subchron C24n.3n; Table 3) for datum events in the carbonate section between the Chron C24r/C24n boundary (359.65 m) and the top of the ash series (384.85 m). We justify this rate by noting a sedimentation rate change in the Chron C24n-C23n interval (point 5a, above). The rate for Chron C24n is likely to be more appropriate for purposes of downward extrapolation in the pelagic carbonates of Chron C24r than in the interval of Chron C24n and C23n in view of the upward diminishing sedimentation rates (Table 3). Inasmuch as the ash series was almost certainly deposited at a higher rate of sedimentation than the pelagic carbonates above, we believe it justified to use this extrapolation down to the top of the ash series. (c) Downward extrapolation of the 2.25 cm/ 103 years sedimentation rate in the carbonate section below c. 360 m yields an estimated age of 54.46 Ma for the top of the ash series (384.85 m). If we then use the calibration of 55.0 million years for the 408 m level, we derive an estimated sedimentation rate of c. 4.28 crn/103 years) for the ash series (and c. 3.5 m of carbonates below the base of the ash series between 404.60 and 408 m). This rate is used to calculate the ages of datum events within the ash series. As a matter of interest using this rate results in an estimated duration of 0. 467 million years for the 19.75 m thick ash series, age estimates of c. 54.65Ma and c. 54.81Ma for the +19

336

w . A . BERGGREN & M. E AUBRY

and -17 ash, respectively, and an interval of c. 0.160 million years between the two. Needless to say these estimates are predicated on the closed loop of having used the 55.0 Ma calibration at 408 m to anchor the lower point of this interpolation. Finally, the estimated duration of the ash series in Hole 550 (0.467 million years) should be compared with the estimate of c. 1.4 million years (based on the dates themselves; point 4 above). The latter estimate is over half the duration of Chron C24r (Cande & Kent 1995) and is clearly unrealistic considering the discussion above. In this connection, a recent study by Fenner (1994) on the diatom stratigraphy of the ash series in the Fur-Olst formations in the Harre borehole in the southern Limfjord area of northwest Denmark are particularly germaine to our study of Hole 550. Fenner (1994) cross-correlated the range of stratigraphically important diatom species in the Danish diatomaceous ash series with their range in an upper P a l e o c e n e - l o w e r Eocene stratigraphic section in ODP Hole 752A in the northern Indian Ocean that contains both siliceous and calcareous microfossils. Conclusions pertinent to this study include the following. (i) The -30 to +30 ash series of the Fur Formation (and perhaps the entire Fur and Olst formations) are correlative with the upper Disr diastypus Subzone (CP9a) and/or upper Zone NP10. (ii) A time span of c. 0.5 million years is estimated for the deposition of the ash series of the Fur Formation (presumably by reference to Berggren et al. 1985). It will be readily seen that from our studies of Hole 550 (Fig. 9) that Fenner's results (1994) are in

close agreement with those in this paper. We would note, however, that the FAD of Dicoaster diastypus postdates the FAD of Tribrachiatus bramlettei (rather than being simultaneous) so that the two zones are not exactly equivalent. We cannot be certain that the entire Danish ash series is represented in Hole 550, but the c. 20 m thick series between c. 385 m and 405 m belongs to lower to mid Zone NP10/CP9a. The ash series in Hole 550 is terminated (c. 385 m) c. 12 m below the FAD of T. contortus (Morphotype B; c. 373 m) with an estimated age difference of c. 0.5 million years, supporting an early-mid (rather than late) Biochron CP9a/NP10 age for that portion of the ashes present in Hole 550. Previous estimates on the duration of the ash series which have varied from c. 3 million years (Perch-Nielsen 1976), c. 1 million years (Heilman-Clausen 1982, by correlation of the -19 to +140 ashes to the Deflandrea oebisfeldensis (dinoflagellate) Acme Zone of the Apectodinium hyperacanthum Zone), to c. 60 000 years (Bonde 1974) who interpreted the ash layers as laminated (seasonal) varves) are seen to have significantly over- and underestimated the probable duration (c. 0.5 million years) of the ash series. 6. Depth data on calcareous nannofossil and planktonic foraminiferal datum events in Hole 550 are tabulated in Tables 4 and 5, respectively. Age estimates of these datum events are tabulated in Tables 6 and 7, respectively, using the tripartite sedimentation rates discussed above (Point 5). Where then does the Paleocene/Eocene boundary lie in the context of the GPTS, and what is its

Table 4. Calcareous nannoplankton datum events in DSDP Hole 550 spanning Paleocene-Eocene Zones NP9-NP12

Datum event

Sample interval Core/Section (m)

FAD Discoaster lodoensis LAD Tribrachiatus contortus (morphotype B) FAD Tribrachiatus orthostylus LAD Tribrachiatus bramlettei LAD Discoaster multiradiatus FAD Tribrachiatus contortus (morphotype B) LAD Tribrachiatus controtus (morphotype A) FAD Tribrachiatus contortus (morphotype A)

Depth (m)

Midpoint value) (m

27/cc-2811:46--47 29/cc-30/1:20-23

346.50-346.46 365.50-365.72

346.23 346.61

30ll: 62-64 30/1:109-111 30/5:22-24 30/5:66-68 30/4:61-63 30/4:99-101 30/5:113-115 30/6:7-9 31/3:6-8 31/3:39-41 31/5:88-90 32/1:17-19

366.13-366.60

366.36

371.73-372.17

371.95

372.13-372.49

372.31

372.64-373.08

372.86

378.07-378.40

378.23

382.89-384.68

383.28

Remarks

NP 11/12 = CP 9/10 NP 10/11 = CP 9a/b boundary

Data from Aubry in Aubry et al. (1996) and from Mtiller 1985, table 10, p. 592-593, for the interval above 365.50 m.

LATE PALEOCENE--EARLYEOCENE MAGNETOBIOCHRONOLOGY OF NW EUROPE

337

Table 5. Planktonic foraminiferal datum levels in DSDP Hole 550 spanning Paleocene-Eocene Zones P5-P7 Datum event

Sample interval Core/Section (m)

Depth (m)

FAD Morozovella aragonensis

27/3: 41-4527/4:42-45

340.42-341.50

340.96

LAD Morozovella lensiformis

28/1: 50-5327/5:42-46 28/1: 50-5327/5:42-45 30/1: 49-5329/6:51-54 30/3: 49-5330/1:49-53 "31/2: 65-6731/2:105-107 33/2:59-61 33/2:59-61 33/1 : 59-6133/2:59-61 *33/2: 51-5333/2:70-72

347.0-343.42

345.21

347.0-343.42

345.21

366.03-363.94

364.99

368.99-366.03

367.51

375.17-375.55

375.36

394.51-396.09

395.05

394.51-396.09

395.05

396.02-396.21

396.11

399.52-400.03

399.78

403.72--405.22

404.47

405.65-408.62

407.14

409.02-409.20

409.11

LAD Morozovella marginodentata LAD Subbotina velascoensis

LAD Morozovella aequa FAD Morozovella lensiformis FAD Morozovellaformosa gracilis FAD Igorina broedermanni

LAD Morozovella acuta FAD Morozovella marginodentata FAD Acarinina wilcoxensis

FAD Morozovella subbotinae

*33/4: 101-10333/5:2-5 34/1 : 62-6534/2:62-65 a. 34/2:62-6534/4:62-65 b.* 34/4: 101-10334/5:119-121

Midpoint value) (m

Remarks

Sample 27/4:42-45 cm = barren zone; taxon not present in 27/5:42-45 cm (1.5 m below); P 6/7 zonal boundary.

P6a/b zonal boundary.

_=_P5/6 zonal boundary if LAD M. acuta is taken as proxy for LAD M. velascoensis.

Entry probably delayed entry here due to dissolution.

* This paper. Data from Snyder & Waters 1985, fig. 6: 448, 449, except as noted. Chronology based on Berggren et al. (1995). Depth (in m) of datum levels given in parenthesis.

age? There are essentially four Paleocene/Eocene boundary levels currently recognized by marine (bio)stratigraphers (with a fifth, the 5J3C spike, looming in the wings): (a) at the base of the Harwich Formation (Oldhaven Beds/Hales Clay), SE England; (b) at the base of the Ieper Clay Formation (base of the type Ypresian Stage), Belgium; (c) at the P5/P6a zonal boundary (planktonic foraminifera); (d) at the NP9/NPI0 zonal boundary (calcareous nannoplankton) (Fig. 12). North American vertebrate palaeontologists customarily link their Paleocene/Eocene boundary with the Tiffanian/Wasatchian N A L M A boundary, whereas their continental European counterparts place it at the base of the Sparnacian 'Stage'. The base of the Wasatchian contains a 513C spike in mammalian tooth and soil carbonates thought to correlate with a similar 8J3C spike in soil carbonates in the lower Sparnacian of the Paris

Basin and with the carbon excursion in marine sediments. It would appear that the Paleocene/Eocene boundary (as currently recognized in N W Europe at the base of the Harwich Formation) is correlative with a level in the lower part of Zone NP10 and upper part of Zone P5, somewhat below the midpoint of Chron C24r, near (but stratigraphically below) the -17 ash layer, dated at 54.5 Ma (but estimated here at c. 54.81 Ma). This differs significantly from the correlation proposed in Aubry et al. (1988, p. 734-735) in which the Paleocene/Eocene boundary = base Oldhaven Beds was estimated to lie in the upper part of Zone NP10. This difference reflects considerable improvement in the upper P a l e o c e n e - l o w e r Eocene lithobiostratigraphic correlations over the last few years. The base of the Ypresian Stage (as stratotypified in Belgium) commences one fourth-order cycle higher than the

338

W.A. BERGGREN & M.E AUBRY

POLARITY

HISTORY 51

CALCAREOUS LITHOSTRATIPLANKTON GRAPHIC

ZONES P

NR

P7

c~ E z

,,

L~176

P/E BOUNDARY

UNITS

CRITERIA/LEVELS

~

NALMA

Basin

n2

~ o

$2-

~

.1

n2 53

54

0

P6b

n3

,~.

~_

~

13_

~

~l~f

z

--"~'1 ~-

~

'

I

~

'

~

"~ (~

,._ P6a

~ ~

Oldhaven ~ ' .

/-(1) Base Ypresian (~-54.6 Ma) ~-(2) P5/6 Boundary (=54.7 Ma) - - (3) Base Harwich Frn. (=-17Ash,~54.8 Ma) - - ( 4 ) NP9/10 Boundary (_--55.0Ma)

(D

P5

< z

a:
55 Ma). An age estimate of 54.7 Ma for the P5/P6 zonal boundary, 54.8 Ma for the base of the Harwich Formation and of 55 Ma for the NP9/10 boundary based on the data reviewed here, may serve as a framework for an ultimate determination of the position and age of the Paleocene/ Eocene boundary. We may include the 813C spike as an appropriate criterion for denoting/correlating the position of the Paleocene/Eocene boundary. We are unable to provide a satisfactory age estimate for this event because it lies at a level which is within the early part of Chron C24r and which is distorted in the current GPTS (Cande & Kent 1992, 1995). However, a provisional age of 55.5 Ma is estimated for the 613C minimum and the benthic foraminiferal extinction event (BFE) based on the relative chronology established for Hole 690B (Aubry et al. 1996). This value is internally consistent with the GPTS (Cande & Kent 1995). Correlation with the multiple oceanographic events discussed above (carbon isotope minimum,

LATE PALEOCENE--EARLY EOCENE MAGNETOBIOCHRONOLOGYOF NW EUROPE oxygen isotope warming event, benthic foraminiferal extinction event, size reduction in wind blown aerosols) appears to have been resolved (Aubry et al. 1996). These events seem to line up approximately with a level in mid Zone NP9 and with a level within Zone P5 at several sites. Pervasive unconformities which juxtapose the NP9-NP10 and P5-P6a-b biostratigraphic events are believed to be responsible for the apparent diachroneity of some of these events.

Sparnacian problem and marineterrestrial correlation The ensuing discussion of marine-terrestrial correlations in NW Europe (Fig. 13) is placed within the framework of the recently revised Cenozoic magnetochronology of Cande & Kent (1992, 1995). The basic assumptions/methodology used in constructing this framework and the suggested correlation network between marine and continental stratigraphies that we propose are explained below. The basic magnetobiochronological framework is based on the revised magnetochronology of Cande & Kent (1992, 1995). This revision has resulted in the magnetochronological boundary estimates shown on p. 313. This revised magnetochronology is based on fitting a spline curve to a series of nine age-calibration points in the Cenozoic, including an age estimate for the NP9/NP10 zonal boundary of 55.0 Ma, based, in turn, on extrapolation from 4~ dates of c. 54.0 and 54.5 Ma on the +19 ash and -17 ash, respectively (Swisher & Knox 1991). It will be readily seen that the revised age estimates differ from those estimated in Berggren et al. (1985) by c. 2.5-3 million years. The standard calcareous microfossil stratigraphy is placed in a magnetochronological framework by using the tripartite sedimentation rate calculations in Hole 550 discussed above (Tables 4-7; Figs 3, 9). Dinoflagellate biostratigraphy is added based on previous (Berggren et al. 1985) and recent (De Coninck 1990; Ali et al. 1992) correlations to magnetostratigraphy. At this point it may be useful to review the basic methodology which has been used in developing a geochronological framework for the events associated with Chron C24 and the construction of Fig. 13: (a) The dates of 54.0 and 54.5 Ma on the +19 and -17 ashes in NW Europe (and occurring at 393 m and 400 m, respectively, in DSDP Hole 550) were used by Swisher & Knox (1991) to estimate the age (55 Ma) of the NP9/NP10 chronozonal boundary (which lies approximately the same distance (7-8 m) below the -17 ash as the distance

339

between the two ashes. This boundary is considered here to be located at an unconformity in Hole 550 but we see no way around the dilemma of estimating its true age and the value of 55.0 Ma is retained here for the present. (b) The age estimate of 55 Ma on the NP9/NP10 chronozonal boundary was used by Cande & Kent (1992, 1995) as a calibration point in their GPTS and the position estimated to lie approximately 0.33 of the way up in Chron C24r following the position estimated for this datum level in Berggren et al. (1985). This level is retained here, although we believe that the NP9/NP10 chronozonal boundary is actually located somewhat younger in Chron C24r based on the arguments presented in Aubry et al. (1996). (c) The NP9/10 zonal boundary is located at c. 408 m in Hole 550 which is c. 23% of the way up in (incomplete) Chron C24r (Figs 3, 9). Restoration of Chron C24r to its approximately 'true' thickness (taking into consideration the strong dissolution affecting the lower part of Chron C24r in particular) results in the relocation of the NP9/NP10 zonal boundary to c. 421 m and a total restored thickness of c. 81 m for chronozone C24r. This has the effect of situating the NP9/NP 10 zonal boundary c. 25% of the way up in chronozone C24r. Given the uncertainties in restoring the c. 62 m thick chronozone C24r to its true thickness, we see that the estimate made here in Hole 550 is consistent with placing the NP9/10 zonal boundary 0.33 of the way up in chronozone C24r (Berggren et al. 1985). (d) The age estimates on the -17 ash (54.8 Ma) and the +19 ash (54.65 Ma) based on a calculated rate of sedimentation (rather than their respective radioisotopic values of 54.5 and 54.0 Ma; Swisher & Knox 1991; see discussion above), are used to constrain the stratigraphic position of the base of the Harwich Formation (Oldhaven Beds = Hales Clay) and the base of the Wrabness Member, respectively. (e) An interval low in chronozone C24r, and spanning the chronozone C24r/C25n boundary, is used to constrain the base (Upnor Formation; formerly the Bottom Bed) of the Lambeth Group (formerly the Woolwich & Reading Beds), consistent with Ali (1994). A major feature of the regional stratigraphic correlation framework presented here is the location of the main lithostratigraphic units of the upper Paleocene-lower Eocene succession in NW European basins in a sequence stratigraphic framework (Fig. 13). This framework has been established by the collaborative efforts of members of IGCP 308 (Paleocene/Eocene Boundary Events in Space and Time) Working Group Members during the course of field excursions to the Hampshire-

340

W . A . BERGGREN & M . P . AUBRY

i i

~'i

NORTH SEA BASIN

MAGNETIC B I O S T R A T I G R A P H Y LONDONPOLARITY HAMPSHIRE HISTORY PLANKT. CALCAREOUSDINOFL. BASINS FORAM,NANNOPLANKTON

"

W6

P7

PARIS

BASIN

BELGIAN

Argile de

~~ E

-- ~b]es"E -

~

@

- - - -

52"

~

C23

z~

~1

w~

-~

Sab,ed'Ai,y

~,c2

_

~ B1

10

1~

E

8

o

Laon ~ =

W3

o=

C2

de , ~,

13 --

O

~

~ :~

~E ,~

- I _ J

....

SEISMICSTRAT. LFrHO6TRA.DINOFL. ASH SER. TIGRAPHYBIOSTR.PHI,SE.'

E

Laon

D~

BASIN

~

C1

~I

B1

.~ _c

~


62 mthick upper Paleocene to lower Eocene (Chron C24r) section from this site is becoming an important reference section. Its carbon isotopic record together with that from adjacent Site 549 (Stott et al. 1996) may serve as a reference for chemostratigraphic correlations and, ultimately, for chemochronology. Diachrony has no relevance with regard to deriving a chronology from a continuous section. On the other hand, a chronology derived from a section which contains unrecognized stratigraphic gaps will undoubtedly be inaccurate. It is thus critical to examine the upper Paleocene to lower Eocene stratigraphic successions in deep sea sites where the carbon isotope excursion occurs in order to evaluate their temporal continuity and establish accurately the stratigraphic position of the excursion.

Upper Paleocene-lower Eocene elements of correlation Marine correlations around the Paleocene/Eocene boundary are among the most difficult in the Paleogene. One reason is that the boundary falls within a long magnetic reversal, Chron C24r, estimated to be 2.5 million years long in the magnetochronology of Berggren et al. (1985) and 2.73 ! and 2.557 million years long, respectively, in the revised magnetochronology of Cande & Kent (1992, 1995). Magnetobiostratigraphic correlations, which are of critical importance for other chronostratigraphic boundaries, are thus of limited usefulness for the characterization of the Paleocene/Eocene boundary. Another reason for the difficulty is that the calcareous microfossils that serve as zonal markers are often absent. In many sections, the boundary is associated with a dissolution event which affects more or less severely the planktonic foraminifera. As for the calcareous nannofossils, the absence of the markers is mostly attributed to ecological exclusion, in particular for the representatives of the critical genus Tribrachiatus.

Planktonic foraminiferal stratigraphy The stratigraphic interval representing the c. 2.5 million years span of Chron C24r is correlative with (redefined) (sub)tropical planktonic foraminiferal Zones P5 and P6 (partim) (Berggren, in Berggren et al. 1995). The P4/P5 zonal boundary (= highest occurrence (HO) of Globanomalina pseudomenardii) is associated with Chron C25n.., and the P5/P6 zonal boundary ( = H O tYo)f Morozovella velascoensis/acuta) with a level in mid-Chron C24r. Zone P6 is divided into two subzones; the P6a/b zonal boundary (= lowest occurrence (LO) of M. lensiformis/M, formosa formosa) is associated with the younger part of Chron C24r. The P6b/P7 zonal boundary (= LO of M. aragonensis) is associated with the basal part of Chron C23r. With regard to the zonal bitstratigraphy of Holes 549, 550 and 690 the problems are the following. (i) At Holes 549 and 550 dissolution has affected the preservation of planktonic foraminifera over critical parts of the stratigraphic record, particularly at Hole 550, resulting in delayed entries of certain taxa (M. subbotinae). In addition Gl. pseudomenardii is either absent (e.g. DSDP Site 550) or extremely rare (e.g. DSDP Site 549) and thus not applicable to local biostratigraphy. Morozovella acuta is not present at Site 549. (ii) at Hole 690 the taxa used in (sub)tropical biostratigraphies are not present, rendering inter-regional correlation difficult. The LO of

UPPER PALEOCENE-LOWER EOCENE STRATIGRAPHYAND CARBON ISOTOPE EXCURSION G1. australiformis, considered by Stott & Kennett (1990) to be correlative with the P6a/P6b subzonal boundary of Berggren & Miller (1988) (= P5/P6 zonal boundary of this paper) is seen to be correlative with a level within mid-Zone P5 since it is correlative with the 813C excursion peak which lies within mid Zone NP9. Calcareous nannofossil stratigraphy

Calcareous nannofossil Zones NP9 to N P l l correlate with Chron C24r interval. The base of Zone NP9, defined by the LO of Discoaster multiradiatus, is associated with the younger part of Chron C25n. The NP9/NP10 and NP10/NPI1 zonal boundaries fall within Chron C24r but the upper part of Zone NP11 correlates with Subchron C24n.3n, Subchron C24n.2r and the older part of Subchron C24n.l-2n (Berggren et al. 1985, 1995; Aubry et al. 1988). Although the base of the Eocene, as defined by the base of the Harwich Formation/Oldhaven Beds (or by the base of the Ieper Clay), is located within Zone NP10 (Knox 1984; Aubry et al. 1988), the NP9/NP10 zonal boundary is often used in marine sections to approximate the Paleocene/Eocene boundary. Zone NP10 is a hybrid between a concurrent range zone and an interval zone, defined by taxa that are part of the same lineage. The LO of Tribrachiatus bramlettei defines the base of the zone while the HO of T. contortus defines its top (Martini 1971). Romein (1979) has shown how T. bramlettei evolved into T. contortus which itself evolved into T. orthostylus in the later part of biochron NPI0. Tribrachiatus orthostylus is common in many lower Eocene (uppermost Zone NP10 to top of Zone NP12) sections but it is more abundant at mid and high latitudes. Tribrachiatus contortus seems to be less restricted in occurrence than T. bramlettei. In the absence of T. bramlettei, the HO of Fasciculithus tympaniformis (or Fasciculithus spp.) is used to approximate the NP9/NP10 zonal boundary. This practice may be hazardous. In several sections which yield the NP9/NP10 zonal boundary and where T. bramlettei occurs, there is a short overlap between the range of this species and that of F. tympaniformis. This is seen, for instance, in the Possagno section, Italy (ProtoDecima et al. 1975) and in the Nahal Avdat (Israel) and Caravaca (Spain) sections (Romein 1979). However, in other sections (such as DSDP Site 527: Backman 1986; DSDP Sites 549, 550 and ODP Site 690B: this paper) there is a sharp decrease in the abundance of E tympaniformis much below the top of Zone NP9, so that it is difficult to differentiate levels where the species is in situ or reworked. Although the three species of Tribrachiatus clearly constitute a lineage, their stratigraphic relationships remain some-

355

what unclear. In the Nahal Avdat section, the species first occur in a sequential fashion with substantial overlap between successive species so that T. contortus co-occurs either with T. bramlettei (in its lower range) or with T. orthostylus (in its upper range). On the other hand, in the Possagno section, the LOs of T. contortus and T. orthostylus are at the same level so that their lower ranges overlap with the upper range of T. bramlettei. Still in other sections (DSDP Sites 549, 550: de Graciansky et al. 1985a, b; DSDP Site 577: Heath et al. 1985) the HO of T. contortus is immediately below the LO of T. orthostylus (Mtiller 1985; Monechi 1985; Backman 1986). In ODP Hole 690B, the three species first occur in a sequential fashion. In this hole, however, the cooccurrence of T. contortus and T. orthostylus at 137.80 and 137.4 mbsf may not be reliable because of strong bioturbation (Barker et al. 1988) and because of proximity to an unconformity (see below). Diachrony is often held responsible for unexpected co-occurrence of taxa. The close juxtaposition of the HO of F. tympaniformis and the LO of T. contortus seen in some sections could be interpreted as reflecting diachrony of the former event, of the latter, or of both. However, in geographically distant sections such as the Possagno (Italy) and the Nahal Avdat (Israel) sections, similar relationships are seen between the LO of T. bramlettei, the HO of F. tympaniformis and the LO of T. contortus. At North Atlantic DSDP Sites 403 and 404 (Montadert & Roberts 1979), and 553 (Roberts et al. 1984) located further north than DSDP Site 550 and that did not penetrate into Zone NP9, T. bramlettei (without F. tympaniformis) occurs in sediments assigned to Zone NP10 (Mtiller 1979; Backman 1984). Thus we do not believe that latitudinal diachrony is responsible for different stratigraphic relationships between the three species, as for instance in DSDP Hole 550 (see below). The latest Paleocene-earliest Eocene was the warmest interval of the Cenozoic (Rea et al. 1990; Stott 1992), when tropical water masses extended to the high latitudes as shown by the biogeography of calcareous micro- and nannoplankton (Boersma et al. 1987; Aubry 1990). As minimal biogeographical differentiation occurred in the calcareous nannoplankton then, diachrony of calcareous nannofossil species in uppermost Paleocene and lowermost Eocene sediments is expected to have been minimal. It should be noted here that diatoms have a wide biogeographical distribution in upper Paleocene and lower Eocene sediments (Fenner 1994), further supporting our expectation. We thus accept the following relative chronology of events: HO of F. tympaniformis, LO of

356

M.-P. AUBRY ET AL.

T. bramlettei, LO of T. contortus, HO of T. bramlettei, LO of T. orthostylus, HO of T. contortus. We recognize, however, that, because of the lack of integrated magnetobiostratigraphic studies on sections representing continuous deposition during Chron C24r (if such sections exist), numerical age estimates of these events are premature at this time (see below). O D P Site 6 9 0 (65~

"S; 1 ~

A c. 60 m-thick upper Paleocene-lower Eocene section, extending from calcareous nannofossil Zone NP9 to Subzone NP14a was recovered from Hole 690B (Fig. 1) located on the southwestern flank of Maud Rise in 2914 m water depth (Barker et al. 1988). The section contains two lithostratigraphic units. Unit III between 128.1 and 137.8 mbsf, consists almost exclusively of calcareous biogenic particles, whereas Unit IV includes various amounts of terrigenous components. Different interpretations have been given of the upper Paleocene-lower Eocene section recovered from Hole 690B but all agree that the section yields all magnetic reversals between Chrons C22n and C25n, and in particular, that Chron C24n is well developed (compare Spiess 1990; Stott & Kennett 1990; Stott et al. 1990; Thomas et al. 1990). We disagree with these interpretations and, based on biostratigraphic evidence, we demonstrate that Chron C24n is not represented. We also suggest that two significant stratigraphic gaps occur, at c. 137.8 mbsf and at c. 139 mbsf. As pointed out by Stott & Kennett (1990), the low latitude planktonic foraminiferal zonations cannot be applied to the sediments recovered from Maud Rise because most low latitude index species are absent. Among them are the late Paleoceneearly Eocene large, keeled morozovellids. However, the Paleocene-lower Eocene calcareous nannofossil assemblages recovered from Hole 690B yield the marker species used in Martini's zonal scheme (1971), so that all boundaries from Zone NP14 to Zone NP9 can be delineated, allowing identification of the magnetozones recognized in the hole on the basis of direct biostratigraphic correlations and reference to the magnetobiochronological framework of Berggren et al. (1985) and Aubry et al. (1988). We have re-examined the calcareous nannofossil assemblages in Cores 23H-4 to 15H-1 (196.10 and 128.50 mbsf). We essentially agree with the biozonal subdivision of this interval by Pospichal & Wise (1990) although we disagree on minor points that are discussed below where appropriate. However, contrary to these authors, we stress the facts that (i) preservation decreases uphole and is

poor above Core 16H-4, 41-44 cm (142.70 mbsf) leading to uncertain specific determination; (ii) reworking is common, in particular of Tribrachiatus orthostylus and Fasciculithus tympaniformis, this latter species occurring throughout the section; (iii) discoasters are exceedingly rare at some levels, particularly in the lower and upper parts of the section. When combined, these three facts lead to difficulties in positioning zonal boundaries, and in delineating LOs and HOs of taxa, particularly of Discoaster species. We agree essentially with Spiess (1990) and Thomas et al. (1990) in identifying the normal polarity intervals that occur between 131.17 and 132.10 mbsf and between 132.85 and 133.10 mbsf (see Spiess 1990, appendix B) as Chons C22n and C23n, respectively. The former interval correlates with Subzone NP14a as indicated by the cooccurrence of Discoaster sublodoensis, D. lodoensis and D. kuepperi (see Aubry 1991). The latter interval correlates with Zone NP12 as indicated by the co-occurrence of D. lodoensis and Tribrachiatus orthostylus. Zone NP11 (= Subzone CP9b of Okada & Bukry (1980; subzone defined by Bukry (1975) on the same criteria as Martini's Zone N P l l ) is represented in Hole 690B as a thin interval between the HO of Tribrachiatus contortus (at 137.41 mbsf following Pospichal & Wise 1990) and the LO of D. lodoensis (between 134.51 and 134.41 mbsf, combining our data with those of Pospichal & Wise 1990). In Hole 690B, Zone NP11 corresponds to an interval with reversed polarity which may represent Subchron C24n.lr, Subchron C24n.2r, or Chron C24r (using the terminology of Cande & Kent (1992); see Berggren et al. (1985) and Aubry et al. 1988 for correlations). Zone NP10, defined as the biostratigraphic interval between the LO of Tribrachiatus bramlettei and the HO of T. contortus (Martini 1971) is well characterized in Hole 690B. We mostly agree with Pospichal & Wise (1990) as to the extension of Zone NP10 in Hole 690B, but we disagree with those authors as to the range of the three species in the genus Tribrachiatus and their abundance. According to Pospichal & Wise (1990, table 3), T. bramlettei is rare to common between Core 17H-2, 28-30 cm and 15H-7, 30-32 cm (149.29-137.41 mbsf), questionably present in Core 15H-6, 30-32 cm (135.91 mbsf), and rare in Core 15H-3, 30-32 cm (131.41 mbsf). In contrast, we observe that T. bramlettei is exceedingly rare, and we record rare to common Rhomboaster cuspis. The lowest occurrence of T. bramlettei is extremely difficult to delineate. Pospichal and Wise (1990, table 3) locate it between Core 17H-2, 28-30 cm and 17H-3, 28-30 cm (149.29 and 150.79 mbsf). Although we have observed specimens questionably assignable to T. bramlettei

UPPER PALEOCENE--LOWER EOCENE STRATIGRAPHY AND CARBON ISOTOPE EXCURSION

357

Fig. 1. Stratigraphy of the upper Paleocene-lower Eocene section recovered from Hole 690B. Note the relationships between the carbon isotope excursion, the deep sea benthic foraminifera extinction and the selected calcareous nannofossil datums. See text for further explanation. (1) Lithology from de Graciansky et al. (1988); (2) magnetostratigraphy from Spiess (1990), based on very high density sampling; (3) our preferred magnetostratigraphic interpretation; (4) calcareous nannofossil biozonal subdivisions based on data from Pospichal & Wise (1990) and personal observation (MPA); (5) planktonic foraminifera biozonal subdivisions (Stott & Kennett 1990); (6) selected calcareous nannofossil and planktonic foraminifera datums (from Pospichal & Wise (1990), Stott & Kennett (1990) and personal observation (MPA)).

358

M.-P. AUBRY ET AL.

as low as Core 17H-3, 43-46 cm (150.94 mbsf), we have recorded the lowest typical specimens in Core 17H-l, 86-90 cm (148.40 mbsf). Accordingly, we prefer to delineate the NP9/NP10 zonal boundary between Core 17H-l, 86-90 cm and Core 17H-2, 40--42 cm (148.40 and 149.4 mbsf), but recognize that this represents a negligible difference with Pospichal & Wise (1990). We note that only early morphotypes of T. bramlettei occur in Hole 690B, suggesting the lowermost part of Zone NP10. Pospichal & Wise (1990, table 3) also report that the upper range of T. bramlettei overlaps with the range of T. contortus between at least Core 16H-1, 28-30 cm and Core 15H-7, 30-32 cm (138.09 and 137.41 mbsf) and with that of T. orthostylus between Core 15H-CC and Core 15H-7, 30-32 cm (137.80 and 137.41 mbsf). We did not examine samples from Core 15H-7 and -CC to avoid mixing problems around 137.80 mbsf (Barker et al. 1988), but we did not observe a range overlap of T. bramlettei and T. contortus in Core 16H-1 and -2. We place the HO of T. bramlettei in Core 16H-2, 41--43 cm (139.70mbsf) and the LO of T. contortus in Core 16H-I, 41-43 cm (138.22 mbsf). In addition to the lack of overlap in the range of the two taxa, we note a striking change in the composition of the assemblages between these two levels, in particular a decreased abundance of Zygrhablithus kerabyi, and the absence of Discoaster multiradiatus above 139.70 mbsf. Although exceedingly rare and poorly preserved, only Morphotype B of T. contortus (see discussion below) has been found in Hole 690B (see also Pospichal & Wise 1990, pl. 6, fig. 7). The mud-bearing nannofossil oozes that we assign to Zone NP10 yield primarily a strong normal polarity (Spiess 1990, fig. 9 and appendix B). Spiess (1990) interprets the normal polarity intervals between 138 and 140 mbsf and 144 and 155mbsf as Subchrons C24n.ln, C24n.2n and C24n.3n, respectively, and he is followed in this interpretation by Stott & Kennett (1990), Stott et al. (1990) and Thomas et al. (1990). Contrary to Spiess's claim, calcareous nannofossil stratigraphy does not support this interpretation. Biochron NP11 is correlative with Chron C24n (partim) which, thus, is not represented in Hole 690B (see above). Although the magnetic record is straightforward and clean, the normal polarity intervals which occur between 137.60 and 154.50 mbsf can only be interpreted as normal overprinting within Chron C24r, without global significance. Chron C24r thus extends from at least 137.60 to 185.70 mbsf. It may also extend up to 131.41 mbsf. The Chron C24r/C25n reversal is clearly delineated between 185.25 and 185.70 mbsf and is associated with the LO of Discoaster multiradiatus (between 185.70 and 185.90mbsf). However, we

caution that the extreme scarcity of discoasters in Core 22H-1, 50-52 cm (185.70 mbsf) and below does not allow a firm delineation of the NP8/NP9 zonal boundary. We thus interpret the upper Paleocene-lower Eocene section recovered from Hole 690B to consist of a discontinuous lower Eocene interval (between 128.1 and c. 137.8mbsf; Subzone NP14a to Zone NP11) and an expanded uppermost Paleocene-lowermost Eocene interval (c. 137.8c. 185.80 mbsf; Zones NP10 and NP9). At least four major stratigraphic gaps occur in the lower Eocene interval. The youngest occurs between 131.41 and 132.91 mbsf. It is indicated by the superposition of Zone NP12 and Subzone NP14a. The hiatus is at least 1 million years. Detailed calcareous nannofossil stratigraphy would be necessary to determine precisely the position of the unconformity and to confirm that the normal polarity interval between 131.17 and 132.10 mbsf corresponds to Chron C22n (and does not result from the concatenation of Chrons C22n and C23n.ln). Thomas et al. (1990) suggest that the presence of a thin debris flow with an erosional contact at c. 133 mbsf, and Spiess (1990, fig. 10) indicates that the younger part of Chron C23n is not represented in Hole 690B. A second unconformity occurs at the N P l l / NP12 zonal contact and is indicated by the absence of normal polarity sediments representative of Chron C24n. This unconformity occurs between 134.41 and 134.51 mbsf. It likely corresponds to heavy bioturbation at c. 134.5 mbsf (Barker et al. 1988). The hiatus represents over 1 million years. The third unconformity corresponds to the lithological change that occurs at 137.8 mbsf, from exclusively calcareous oozes to clay-bearing nannofossil oozes. It is biostratigraphically characterized by the close juxtaposition of the HO of T. contortus and the LO of T. orthostylus. Whether there is a natural overlap in the range of the two species is unclear. According to Pospichal & Wise (1990, table 3) there is a 40 cm overlap in the range of the two taxa. However, strong bioturbation at the lithological boundary and above (Barker et al. 1988) may have caused mixing in Core 15H-7 and -CC. The fourth unconformity is located between 138.22 and 139.70 mbsf, and is inferred from the juxtaposition of Morphotype B of T. contortus (at 139.70mbsf) and early morphotypes of T. bramlettei (at 138.22mbsf). This is well supported also by the change in assemblages between these two levels (see above). This indicates that the lowermost part of Zone NP10 underlies the mid to upper part of the zone (see below). We interpret the lower part of the section (between c. 139 and 185.80 mbsf) as being essen-

UPPER PALEOCENE--LOWEREOCENE STRATIGRAPHYAND CARBON ISOTOPE EXCURSION tially continuous, although we have little means to verify this assumption. The calcareous nannofossil assemblages are of limited diversity, and poor preservation, scarcity (e.g. Discoaster diastypus), or absence (e.g. Cruciplacolithus eodelus) of stratigraphically significant species do not allow objective evaluation of the section. However, we note the sequential LOs of Zygrhablitus kerabyi (in Core 20H-2, 40-42 cm; 149.20 mbsf), Pontosphaera plana (in Core 17H-6, 4 3 - 4 5 c m ; 155.45 mbsf) and Rhomboaster cuspis (in Core 17H-4, 40-41 cm; 152.40 mbsf; although questionable (poorly preserved) specimens may occur as low as Core 17H-6, 87-90 cm; 165.90 mbsf). The best indication that the section is continuous across the NP9/NP10 zonal boundary is the occurrence of early morphotypes of T. bramlettei (see Pospichal & Wise 1990, pl.6, fig. 6) that are compact and suggestive of a direct evolution from R. cuspis. As reported by Pospichal & Wise (1990), the relationship between the LO of T. bramlettei and the HO of Fasciculithus tympaniformis cannot be firmly established in Hole 690B. The abundance of this latter species varies greatly from common (e.g. in Core 18H-6, 40-42 cm; 163.60 mbsf) to very rare (e.g. in Core 18H-3, 32-35cm; 160.55 mbsf; 17H-4, 41-44 cm; 152.40 mbsf) in the interval assigned to Zone NP9, and it was found at all levels examined above c. 149mbsf (NP9/NP10 zonal boundary). Using the stratigraphic framework established above, and considering that Zone NP9 apparently contains no significant stratigraphic gap(s), the stratigraphic position of the carbon isotope excursion and of the deep sea benthic foraminiferal extinction can be precisely determined. Zone NP9 is c. 36.8 m-thick in Hole 690B (c. 149185.80mbsf). The sharp benthic foraminiferal extinction occurs between 170.63 and 170.67 mbsf (Thomas & Shackleton 1996) and the minimum 813C values occur at 170.26 mbsf (Stott et al. 1990; Kennett & Stott 1991). Thus, the deep sea events occur in the lower part of Zone NP9 (42.2% at Hole 690B). We note in passing that the deep sea Paleocene/Eocene boundary events occur at a stratigraphic position that is well below the currently accepted position of the Paleocene/ Eocene boundary (Aubry et al. 1988), which implies that these events are older than 57.3 Ma (in the chronology of Berggren et al. 1985) as determined through extrapolation using the incorrectly identified Chron C24n/C24r magnetic reversal in Hole 690B (Kennett & Stott 1991).

D S D P Site 550 (48~

13~

DSDP Site 550 (Fig. 2) is located on the Porcupine Abyssal Plain 10 km southwest of the seaward edge

359

of the Goban Spur, NE Atlantic Ocean. Hole 550, drilled in 4432 m water depth, recovered a 112 mthick section of upper Paleocene-lower Eocene (calcareous nannofossil Zone NP9 to Subzone NPI4a) marly nannofossil chalk (lithological Subunit 2a, 310.34-408 mbsf) and siliceous marly nannofossil chalk and mudstone (Subunit 2b, 408426.5 mbsf) (de Graciansky et al. 1985b). The stratigraphic interpretation of the upper part of the section (above 356 mbsf) is discussed in Aubry (1995). It has no relevance to our discussion because in Hole 550, the upper Paleocene-lower Eocene (Zones NP9 to N P l l ) magnetobiostratigraphic correlations are essentially straightforward (Fig. 2). Subchron C24n.3n is well represented between 349.86 and 359.27 mbsf, and Subchron C24n.3r extends at least down to 419.51 mbsf (Townsend 1985; Snyder et al. 1985). It is uncertain whether the magnetic reversal between 419.51 and 425.5 mbsf (Townsend 1985) corresponds to the Chrons C24r/C25n boundary as indicated by Snyder et al. (1985). The lowest reported occurrence of D. multiradiatus is at 424.85 mbsf (MOiler 1985) in an (6.5 m-thick) unsampled interval of unknown polarity (Fig. 2). In addition, the sharp lithological contact at 426.5 mbsf between lithological Units 2b and 3a (de Graciansky et al. 1985b) occurs within Zone NP5 (see MOiler 1985), not at the NP5/NP9 contact. Thus, the upper c. 50 cm of the normal polarity interval that extends between 425.5 and 434.04 mbsf may or may not represent Chron C25n. In Hole 549 (Fig. 3), a rhyolitic ash layer occurs at 352.66 mbsf in the upper part of Chronozone C25n. This layer does not occur in Hole 550 (Knox 1985). Thus, we strongly suspect that the upper surface of the unconformity between Zones NP9 and NP5 in Hole 550 is younger than the Chron C24r/C25n reversal, and that Chron C25n is not represented in Hole 550. Because of strong dissolution, no biostratigraphically diagnostic planktonic foraminifera are preserved between 409 and 425 mbsf (Snyder & Waters 1985; Berggren & Aubry 1996). This interval is assigned to calcareous nannofossil Zone NP9 by MOiler (1985) who, in the absence of Tribrachiatus bramlettei, used the HO of Fasciculithus tympaniformis to approximate the NP9/NP10 zonal boundary at c. 408 mbsf. This author (1985, table 10) reported the LO of Tribrachiatus contortus low in Zone NP10 (at 405.6 mbsf), i.e. 2.4 m above the NP9/NPI0 zonal boundary, which implied that the section between 408 and 365.96 mbsf (at which level M~iller reports the HO of T. contortus) belongs to the mid and upper part of Zone NP10. We have re-examined the calcareous nannofossil assemblages in the interval between Core 34-7 13-15 cm and Core 30-1, 21-23 cm (412.63 and

360

M.-P. AUBRY E T AL.

Fig. 2. Stratigraphy of the upper Paleocene-lower Eocene section recovered from DSDP Hole 550. Note the relationships between the carbon isotope excursion, the deep sea benthic foraminifera extinction and the selected calcareous nannofossil and planktonic foraminiferal datums. See text for further explanation. (1) Lithology from de Graciansky et al. (1985); (2) magnetostratigraphy from Townsend (1985): tick marks indicate position of samples analysed; (3) magnetostratigraphic interpretation; (4) calcareous nannofossil biozonal subdivisions (from Mtiller 1985, and personal observations, MPA); (5) planktonic foraminifera biozonal subdivisions (see Berggren & Aubry 1990); (6) selected calcareous nannofossil and planktonic foraminifera datums, based on MUller (1985), Snyder & Waters (1985) and personal observations (MPA and WAB).

UPPER PALEOCENE--LOWER EOCENE STRATIGRAPHYAND CARBON ISOTOPE EXCURSION

361

Fig. 3. (a-i) Tribrachiatus contortus Morphotype A. Note the flat morphology of the specimens. This form is reminiscent of early morphotypes of Tribrachiatus orthostylus but with long bifurcations. (j, k) Tribrachiatus bramlettei. (l) Rhomboaster cuspis. Sample 550-31-5, 17-19 (381.18 mbsf).

365.72 mbsf), using all samples which served for stable isotope analysis (Stott et al. 1996, table 2). We agree with MUller's interpretation (1985) of the section below 408 mbsf but we disagree with her interpretation of the section above this level. We locate the NP9/NP10 zonal boundary at 408 mbsf between Core 34-4, 2-4 cm and Core 34-3, 125127 cm (408.02 and 407.75 mbsf) based on the LO of Tribrachiatus bramlettei at this latter level. The HO of E tympaniformis in the section immediately precedes the LO of T. bramlettei. Zone NP10 extends up to at least Core 30-1, 21-23 cm (365.72 mbsf). This is in agreement with Mtiller (1985), who placed the NP10/NP11 zonal boundary between 365.96 and 365.30 mbsf. Tribrachiatus bramlettei ranges from Core 34-3, 125-127 to Core 30-5, 66-68 (407.75-372.17 mbsf). It is scarce to common at most levels except between Core 34-3, 47-49 and Core 34-1, 146-148 (406.98-404.97 mbsf) where it is rare to exceedingly rare.

In Hole 550, Tribrachiatus contortus is represented by two morphotypes (which probably correspond to two distinct species: MPA, work in progress) with a disjunct range. Morphotype A (Fig. 3a-i) ranges between Core 31-5, 88-90 cm and Core 31-3, 39-41 (381.9-378.4 mbsf). Morphotype B (Fig. 4a-h) occurs between Core 30-5 113-115 and Core 30-1, 21-23 (372.65365.72 mbsf). Transitional forms between T. bramlettei and T. contortus Morphotype B occur in Core 30-6, 7-9 (373.09 mbsf), while transitional forms between the latter and T. orthostylus occur in Core 30-1 141-143 cm and Core 30-1 109-111 cm (366.91-366.6 mbsf). In contrast, intermediate forms between T. contortus Morphotype A and T. bramlettei have not been observed. Morphotype A of T. contortus is flatter than Morphotype B, which has strongly twisted tips. Both morphotypes have been distinguished in the course of the study of other sections in the Atlantic Ocean and the middle East (MPA, unpublished data). Previous

362

M.-P. AUBRY ET AL.

Fig. 4. (a-h) Tribrachiatus contortus Morphotype B. Note the twisted tips of the arms which lie in different planes. Sample 550-30-4, 138-140 cm (371.40 mbsf). (i-I) Tribrachiatus bramlettei (Figs 10-12: same specimen focused at different levels.) Sample 550-34-3, 125-127 cm (407.75 mbsf).

lack of distinction between these two forms results in difficulties in comparing the succession of biostratigraphic events in Hole 550 with that in other sections. We believe that the evolutionary sequence

hole, the range of T. contortus is entirely within that of T. bramlettei (Morton et al. 1983, table 1). In Hole 550, we note that the HO of Morphotype A is close to the LO of Morozovella

T. b r a m l e t t e i -

lensiformis.

T. contortus

-

T. orthostylus

described by Romein (1979) from the Nahal Avdat section over a narrow interval in the upper part of Zone NPI0, is that observed in Hole 550 between 373 and 366.6 mbsf. This evolutionary sequence involves Morphotype B of T. contortus. It is likely that Backman (1984, table 4) encountered Morphotype B in his study of the lower Eocene in Hole 553A (Roberts et al. 1984) where a short overlap between T. nunnii (vel T. bramlettei) and T. contortus occurs in Core 22-1. Similarly, the very short overlap between T. contortus and T. orthostylus shown by Backman (1986, fig. 1) in DSDP Hole 577 (Heath et aL 1985) suggests that only Morphotype B of T. contortus was encountered in the hole. Instead, it is likely that Morphotype A alone occurs in the lower Eocene in DSDP Hole 117 (Laughton et al. 1972). In this

Because of poor preservation, it is difficult to delineate precisely the benthic foraminiferal extinction in Hole 550. It occurs between 413.62 and 408.65 mbsf. The 513C minimum occurs between 407.5 and 409 mbsf. Assuming that there is stratigraphic continuity over the NP9-NP10 zonal interval in Hole 550, it implies that (i) the benthic foraminifera extinction occurs just below the NP9/NP10 zonal boundary and (ii) the carbon isotope excursion straddles it. We observe in Hole 550 that the LO of Rhomboaster cuspis (Core 34-4, 89-91 cm, at 408.90 mbsf), a species which first occurs within Zone NP9 is closely associated with the HO of Fasciculithus tympaniformis. In addition, this latter biostratigraphic event is close to the LO of Tribrachiatus bramlettei. Finally, we note that the

UPPER PALEOCENE--LOWEREOCENE STRATIGRAPHYAND CARBON ISOTOPE EXCURSION NP9/NP10 zonal boundary is associated with a sharp lithological change at 408 mbsf.

D S D P Site 549 (49~

13~

DSDP Site 549 is located on the Goban Spur, Irish Continental Margin, NE Atlantic, c. 50 km N-NE of Site 550, near the Pendragon escarpment. Hole 549, drilled in 2533 m water depth, recovered 76.5 m (353 to 277.5 mbsf) of upper Paleocenelower Eocene (calcareous nannofossil Zone NP9 to Subzone NP14a) brown and grey nannofossil chalks which, on the basis of colour and silica content, are subdivided into four lithological subunits. The boundaries occur at 322 mbsf between Subunits 3a and 3b, at 335 mbsf between Subunits 3b and 3c and at 350.5 mbsf between Subunits 3c and 3d. The boundary between Subunits 3b and 3c is the most distinct (de Graciansky et al. 1985a). The stratigraphic interpretation of the upper part of the lower Eocene interval (above 303 mbsf) is discussed in Aubry (1995). It is not relevant to the present discussion because the magnetobiostratigraphic correlations for the upper Paleocene-lower lower Eocene interval in this hole are straightforward. Subchron C24n.3n is well represented in the upper part of Zone NP11, and the normal polarity interval between ?351.53 and 353.81 mbsf, associated with the NP8/NP9 zonal boundary (Mtiller 1985), represents Chron C25n (Townsend 1985; Snyder et al. 1985). Dissolution in the NP9 zonal interval is less intense than in Hole 550 and it affects a thinner interval. Yet, the marker species among the planktonic foraminifera are absent below 323 mbsf, and Zones P6a, P5 and P4 are only approximated (Fig. 5). Mtiller (1985) reported calcareous nannofossil assemblages of high diversity in the upper Paleocene recovered from Hole 549. She delineated the NP9/NP10 zonal boundary between Core 16-3, 105-106cm and Core 16-3, 50-51 (335.55 and 335 mbsf) and reported the HO of Fasciculithus tympaniformis and the LO of Tribrachiatus contortus in the former and latter sample, respectively. We have re-examined the calcareous nannofossil assemblages in the interval from Core 17-7, 24-26 cm to Core 16-1, 22-24 cm (350.25 to 331.70mbsf), using all samples that served for stable isotope analysis (see Stott et al. 1996, table 3). Based on this, we delineate more precisely the NP9/NP10 zonal boundary between Core 16-3 100-102 cm and 16-3, 78-80 cm (335.50 and 335.30 mbsf). The former sample yields Rhomboaster cuspis, Fasciculithus tympaniformis (few) and no representatives of the genus Tribrachiatus. In contrast, the latter sample yields Tribrachiatus bramlettei, Rhomboaster cuspis, no

363

Fasciculithus tympaniformis and no T. contortus. Tribrachiatus bramlettei (few) was encountered only at this level. Tribrachiatus contortus (Morphotype B, except at 335.16mbsf where both morphotypes occur) is common from Core 16-3, 66-68 cm to Core 16-1, 22-24 cm (335.16331.70 mbsf), and according to Mtiller (1985) it has its HO in Core 15-CC (c. 331.40 mbsf). The lithological boundary between Subunits 3b and 3c, delineated on the basis of colour and composition (de Graciansky et al. 1985a) has been associated with the NP9/NP10 zonal boundary (Knox 1985). Our detailed biostratigraphic analysis does not support this association. If located at 335 mbsf, the contact between Subunits 3b and 3c falls within Zone NP10. We note, however, that this lithological contact is not sharp (de Graciansky et al. 1985a, barrel sheets, p. 172 and core photographs p. 172). There is a c. 40 cm interval with fine banding, colour changes, and a sharp colour contact is reported at c. 80 cm in Core 16-3. We suggest that this colour contact corresponds to the NP9/NP10 zonal boundary, represents an important stratigraphic gap, and may be a better choice for the lithological boundary between Subunits 3b and 3c. The colour changes between 334.90 and 335.30 mbsf likely reflect sedimentary disturbances, and we suggest that the interval between the HO of F. tympaniformis (335.50 mbsf) and the LO of T. contortus (335.16 mbsf) is stratigraphically disturbed and comprises several stratigraphic gaps. Stratigraphic discontinuity near the NP9/NP10 zonal boundary has been recognized by Knox (1985) who observed that the upper Paleocenelower Eocene bentonitic ashes which occur in the NP 10 zonal interval in Hole 550 are absent in Hole 549. The juxtaposed LOs of MorozoveUa subbotinae, M. marginodentata and Acarinina wilcoxensis at 335 mbsf (or between 335 and 336.60 mbsf; Snyder & Waters 1985), events which are spread over a 10 m interval of Zone NP10 in Hole 550, are a clear expression of a stratigraphic discontinuity. The NP9/NP10 zonal contact is unconformable, as marked by the juxtaposition of the HO of Fasciculithus tympaniformis, the LO of Discoaster diastypus and that of late morphotypes of Tribrachiatus bramlettei. In addition, we suggest that a stratigraphic gap separates samples 16-3, 78-80 cm (335.30 mbsf) and 16-3, 66-68 cm (335.16 mbsf), which both belong to Zone NP10. The lower level yields T. bramlettei alone, whereas the upper one yields T. contortus solely. In Hole 550, the ranges of the two species overlap (see above). In addition, the presence of Morphotype A of T. contortus in an assemblage largely dominated by Morphotype B of this species in Core 16-3, 66-68 cm (335.16 mbsf) is seen as further support of an unconformity immediately below this level.

364

M.-P. AUBRY E T AL.

Fig. 5. Stratigraphy of the upper Paleocene-lower Eocene section recovered from Hole 549. Note the relationships between the carbon isotope excursion, the deep sea benthic foraminifera extinction and the selected calcareous nannofossil datums. See text for further explanation. Note the two unconformities at c. 322.20 and 335.40 mbsf and the restricted occurrence of T. bramlettei. (1) Lithology from de Graciansky et al. (1985); (2) magnetostratigraphy from Townsend (1985): tick marks indicate position of samples analysed; (3) magnetostratigraphic interpretation; (4) calcareous nannofossil biozonal subdivisions (Mfiller 1985, and personal observation (MPA)); (5) planktonic foraminifera biozonal subdivisions (see Berggren & Aubry 1996); (6) selected calcareous nannofossil and planktonic foraminifera datums, based on Miiller (1985), Snyder & Waters (1985) and personal observations (MPA and WAB).

UPPER PALEOCENE-LOWER EOCENE STRATIGRAPHY AND CARBON ISOTOPE EXCURSION Despite the definitive stratigraphic discontinuities at c. 335.40mbsf and above c. 335.20 mbsf, the extinction event in the benthic foraminifera and the minimum ~513Cvalues in Hole 549 are clearly associated with the NP9/NP10 zonal boundary. Both occur 2.60 m below the boundary.

Correlation between Holes 690B, 550 and 549 Comparison between the upper Paleocene-lower Eocene sections recovered at ODP Site 690 and DSDP Sites 549 and 550 shows that the NP9-NP10 zonal interval consists of three distinct stratigraphic intervals. The first (lowest) extends from the base of Zone NP9 to the carbon isotope excursion/benthic foraminifera extinction (CIE/ BFE), and is similar in the three holes. The second, which extends from the CIE/BFE to the LO of Tribrachiatus bramlettei, and the third, which corresponds mainly to the range of T. bramlettei, occur in all three holes but vary considerably in thickness between holes. The similarity between the lower interval in the three holes is striking. In Holes 690B and 549, where Chron C25n is represented, its thickness is 15.5 and 14.5 m, respectively. In Hole 550 where the lower part of Zone NP9 may be truncated, the interval is estimated to be 11-15 m thick. The late Paleocene sedimentation rates at the three sites are unknown; intensive dissolution probably results in reduced thickness in Hole 550 and, to a lesser extent, in Hole 549. Yet, it is likely that the intervals comprised between the base of Zone NP9 and the CIE/BFE in the three holes are essentially correlative and represent the early part of Chron C24r and of Biochron NP9. The second interval, between the CIE/BFE and the LO of T. bramlettei is 21.26 m thick in Hole 690B, but 2.6 m-thick in Hole 549 and non-existent to negligible in Hole 550 (the BFE in this hole is poorly delineated because of intense dissolution; see above). Based on the LO of Tribrachiatus bramlettei in the three holes, the levels where the NP9/NP10 zonal boundary is delineated in Holes 550 (408 mbsf) and 549 (c. 335.40 mbsf) should be essentially correlative with level 149 mbsf in Hole 690B. Yet, this is not true in Hole 549 where an unconformity occurs at 335.40 mbsf (the NP9/NP10 zonal boundary, Fig. 5). The absence of the lower part of Zone NP10 in this hole has been demonstrated previously (Knox 1985; see above). We now conclude from the proximity of the CIE/BFE to the NP9/NP10 zonal boundary in Hole 549 that the stratigraphic gap at 335.40 mbsf also includes the upper part of Zone NP9 (Fig. 6). This is well supported by isotope stratigraphy (see

365

below). The situation is similar in Hole 550. Zone NP10 is more complete in this section than in Hole 549 (Knox 1985; see also above and below), but the proximity of the CIE/BFE to the LO of T. bramlettei in this hole indicates that the upper part of Zone NP9 is missing. The unconformity in Hole 550 is at 408 mbsf and it corresponds to the contact between lithological Subunits 2a and 2b. In Hole 549, a lithological boundary is also closely associated with the unconformity at 335.40 mbsf. However, whereas Subunits 2b and 3c, respectively, in Holes 550 and 549, represent the same stratigraphic interval, Subunit 2a in Hole 550 is older than Subunit 3b in Hole 549. The third interval, which essentially corresponds to the range of T. bramlettei in the three holes, is over 35 m thick in Hole 550, c. 10 m thick in Hole 690B and no more than 34 cm in Hole 549. The difference in thickness between Holes 550 and 549 reflects the larger stratigraphic gap associated with the NP9/NP10 zonal boundary in Hole 549. We have shown above that an unconformity occurs at c. 139mbsf in Hole 690B below the contact between lithological Units III and IVa. Figure 6 attempts to correlate the three holes. It is extremely difficult to correlate the NP10-NP11 zonal intervals of Holes 550 and 549 because of the inconsistencies between biostratigraphic ranges and between them and magnetostratigraphy (compare Figs 2 and 5). The 19 m-thick interval (385.5 to 404.5 mbsf) with bentonite layers in Hole 550 is not represented in Hole 549 (Knox 1985). Thus the stratigraphic gap in Hole 549 includes at least the 22 m of sediments that lie immediately above the unconformity at 408 mbsf in Hole 550. Different early Eocene sedimentation rates at Sites 550 and 549 likely account for the difference in thickness of the upper part of Zone NP10 and Zone N P l l in the two holes. Knox (1985) reports the presence of disseminated ash particles in Core 15-5 and 16-1 similar to those associated with the bentonites in Hole 550, and scattered specimens of Stensioina beccariiformis occur in this interval (Reynolds 1992: unpublished M.Sc. thesis, Univ. Maine, Orono). This indicates reworking in Hole 549. Most of the c. 31 m-thick section in Hole 690B from a level just above the carbon isotope excursion to the unconformity at c. 139 mbsf, is not represented in Hole 549 (gap (2) in Fig. 6). It is also likely that the c. 28 m-thick section between the level of the carbon isotope excursion and the HO of T. bramlettei in Hole 690B corresponds to a stratigraphic gap in Hole 550 (gap (3) in Fig. 6). As pointed out above, the specimens of T. bramlettei that occur in Hole 690B are very compact whereas those which occur immediately above the unconformity in Hole 550 have longer and more

366

M.-P. AUBRY ET AL.

Fig. 6. Correlation between the upper Paleocene and lower Eocene sections recovered from ODP Site 690 and DSDP Sites 549 and 550. See text for further explanation. D.m.; Discoaster multiradiatus; T.b.; Tribrachiatus bramlettei; T.c.; T. contortus; T.o.; T. orthostylus; Et.; Fasciculithus tympaniformis; M.s., Morozovella subbotinae; A.w., Acarinina wilcoxensis; M.m., Morozovella marginodentata; M.I., M. lensiformis; BFE, deep sea benthic foraminifera extinction; CIE, carbon isotope excursion.

slender arms (Fig. 4i-1). Unless it can be demonstrated that these are e n v i r o n m e n t a l l y related differences between southern high and northern mid-latitude ecomorphotypes, it is reasonable to interpret t h e m as c o r r e s p o n d i n g to different

evolutionary stages, with earliest evolutionary morphotypes in Hole 690B and more evolved morphotypes in Hole 550. This is in agreement with our interpretation of stratigraphic continuity across the N P 9 - N P 1 0 zonal boundary in Hole

UPPER PALEOCENE--LOWER EOCENE STRATIGRAPHY AND CARBON ISOTOPE EXCURSION

690B (but we acknowledge a certain amount of circular reasoning) and the fact that the NP9/NP10 zonal contact is unconformable in Hole 550. Thus we consider that there is no overlap of the intervals assigned to Zone NP10 in Holes 550 and 690B. Finally, a part of the section present in Hole 550 between the LO of T. bramlettei and an undetermined level in Chron C24r (but within the range of T. contortus Morphotype B) is absent in Hole 690B where it corresponds to the unconformity at c. 139 mbsf (gap (4) in Fig. 6). We have established that Holes 690B, 550 and 549 yield discontinuous, but complementary upper Paleocene-lower Eocene (Chron C24r; Zones NP9-NP11) stratigraphic records. The section between the carbon isotope excursion in Hole 690B and the LO of T. bramlettei is almost non-existent in Hole 550 and very thin in Hole 549. On the contrary, the greater part of Zone NP10 is absent in Hole 690B. A numerical chronology is necessary for a temporal interpretation of the sections, in particular to estimate the duration of the hiatuses associated with the unconformities in each. Because the Hole 550 section was considered continuous across the NP9/NP10 boundary, the age estimate of this latter was used to constrain magnetochronology between it and the Cretaceous/Paleogene boundary in the GPTS of Cande & Kent (1992, 1995). The -17 and +19 ash layers, isotopically dated in Danish sections (Swisher & Knox 1991), occur in Hole 550. Extrapolation through these allowed determination of the age of the NP9/NP10 zonal boundary (and, incorrectly, the FAD of T. contortus), which, in turn served as a tie point in the magnetochronology of Cande & Kent (1992, 1995). Following our biostratigraphic revision of Hole 550, we point out that the biostratigraphic position of the bentonites layers (North Sea tephra phases 2a and 2b, Knox & Morton 1988; see Berggren & Aubry 1996) in this hole is in the lower to mid part of Zone NP10, below (c. 3 m) the LO of T. contortus Morphotype A and much below (13 m) that of T. contortus Morphotype B (and not within the lower range of T. contortus following Mtiller 1985). As we have now established the presence of a stratigraphic gap at the NP9/NP10 zonal boundary in Hole 550, we recognize that the age of the NP9/NP 10 zonal boundary as derived from this site is erroneous. Thus we have at present no reliable chronological framework for the earliest Eocene and the Paleocene (see below). The temporal interpretation attempted here (Fig. 7) is based on a conversion depth-time primarily based on the reconstructed record of magnetozone C24r at Site 550. The methodology followed is explained below. We emphasize that the temporal framework

367

used here is highly speculative because our reconstruction is based on the unverifiable assumption that the sedimentation rates were essentially the same at Sites 550 and 690. The objective of Fig. 7 is to show the relationships between the upper Paleocene-lower Eocene records in Hole 690B, 550 and 549, and the overlap between the hiatuses in the three sections.

Implications for isotope stratigraphy Detailed isotope stratigraphies based on the variations in isotopic composition of the tests of the deep-dwelling planktonic foraminiferal species Subbotina patagonica have been obtained for Holes 690B (Fig. 1; Stott & Kennett 1990, Stott et al. 1996), 550 and 549 (Figs 2, 5; Stott et al. 1996). In Hole 690B, the interval between c. 165 and 147.5 mbsf is characterized by relatively constant 813C values, which vary between 1.5 and 1.2%o. From 147.5 to 137.8 mbsf, a decrease occurs from 1.2%~ at 147 mbsf, to 0.6%~ at 138 mbsf. Despite the appearance, the low isotopic values (0.7 to 0.6%~ between 137.8 and c. 134 mbsf cannot be part of the decrease because of the unconformity at 137.8 mbsf (Fig. 1). In Hole 550, the interval between 406 and 380 mbsf also yields relatively constant values which vary between 1.5 and 1.0%o. A decrease occurs between 380 and 370 mbsf, from 1.6 to c. 0.6%0. The isotopic record in the interval between 165 and 137.8 mbsf in Hole 690B and between 406 and 370 mbsf in Hole 550 yields similar trends and values, which suggests that the two stratigraphic intervals are correlative (Fig. 8). This, however, is contradicted by their biostratigraphic age as discussed above. Relying upon biostratigraphy, the decrease occurs in the lowermost part of Zone NP10 in Hole 690B, whereas in Hole 550 it occurs in the upper part of this zone. Also, the interval characterized by isotopic values ranging between 1.5 and 1.0%c, which underlies the decrease in both holes, is assigned to the upperpart of Zone NP9 in Hole 690B, to lower Zone NP10 in Hole 550. There are two alternative interpretations to resolve the discrepancy between the isotopic and the biostratigraphic information. If the intervals between 165 and 137.8 mbsf in Hole 690B and between 406 and 375 mbsf in Hole 550 are correlative, which would account for the similarity in the isotope records, all the calcareous nannofossil stratigraphic events in the upper NP9 and NP10 zonal interval are diachronous (Fig. 8). In particular, this implies that the NP9/NP10 zonal boundary cannot be used for correlation to extreme

368

M.-P. AUBRY ET AL.

CHRONOLOGY MAGNETIC HISTORY

TEMPORAL INTERPRETATION

BIOCHRONOLOGY 1 I2 I

ODP HOLE 690B

DATUMS

DSDP 'HOLE 549

DSDP HOLE 550

1

to 13m

o

m

l

il

=_ BFE

~

J

CIE

--




~

o

~..,~

o'3 o-~

~~

,...~ t,.r

o.~ 9o ~ ~J .=.25

r,.)~

372

M.-P. AUBRYET AL.

Zone NP10 is represented in Hole 690B. In contrast, Hole 550 provides an excellent record of most of Zone NP10, and we believe that the section is continuous across the NP 10/NP11 zonal boundary. Yet, the lowermost part of Zone NP10 is not represented in this section, and only the lower part of Zone NP9 is present. Hole 549 plays a smaller role in the construction of a composite section of Chron C24r, but it provides a more satisfactory record of the Chron C25n/C24r boundary than the other two holes. Also, it should be borne in mind that our correlation of the 813C isotopic records on Subbotina patagonica in Holes 550 and 690B hinges on the integrated biostratigraphic and isotopic records in Hole 549. Based on both isotope and calcareous nannofossil stratigraphies, we do not believe that there is an overlap of Holes 550 and 690B in the lower part of Zone NP10. We have no means of assessing the importance of the gap until additional isotopic studies become available or until rates of evolution in Tribrachiatus bramletti become established. At this time, for the sake of simplicity, we assume that the gap is extremely small and less than 0.1 million years. Relative c h r o n o l o g y o f events in C h r o n C 2 4 r

To establish a relative chronology within Chron C24r, we now need to adjust the two sections 550 and 690B which serve for the composite reconstruction to a common vernier. This requires conversion of the stratigraphic information into a relative chronology (distance between events) taking into account differences in sedimentation rates. Relative c h r o n o l o g y in the N P 9 z o n a l interval

Relative chronology in this interval is primarily provided by Hole 690B. Zone NP9 is 36.90 m-thick

in this hole. Compared to other records of the correlation of the LO of Discoaster multiradiatus and Chron C25n, the latter LO may be delayed; we have indicated the rarity of discoasters below 184.70 mbsf. For the sake of simplicity, we ignore this uncertainty which appears to be very minor compared to those encountered further in this discussion. Reference to Barker et al. (1988, p. 209, fig. 29) indicates that the calcium carbonate content varies generally between 75% and 90% over the stratigraphic interval of 125-200m, suggesting that sedimentation rates remained relatively stable and uniform. However, Thomas & Shackleton (1996) propose a local increase in productivity around Site 690, immediately following the benthic foraminifera extinction (e.g. above c. 170 mbsf) in order to reconcile the increase of infaunal species indicative of higher productivity with the carbon isotopic data which show decreased surface to deep water gradients suggestive of lowered productivity. However, there is no supporting evidence of increased productivity and sedimentation rates at Hole 690B. Thus assuming that the sedimentation rates were constant, the CIE occurs at 42.11% and the BFE at 41.52% of the way up in Zone NP9. In addition, the FAD of G. australiformis is at 41.07%, the decrease in abundance of Fasciculithus tympaniformis (from few to very rare) is at 63.68% and the (tentative) FAD of Rhomboaster cuspis is at 81.62% of the way up (Table 1). R e l a t i v e c h r o n o l o g y in the N P I O z o n a l interval

Relative chronology in this interval is provided by Hole 550 which constitutes an excellent continuous sedimentary record of the lower Eocene from low Zone NP10 to Zone NP13 (above and Aubry 1995).

Table 1. Relative chronology of events in early Chron C24r (essentially correlative with Biochron NP9) as deduced from the stratigraphic succession of events in ODP Hole 690B Datum FAD T. bramlettei < E tympaniformis FAD R. cuspis Carbon excursion Benthic extinction FAD G. australiformis Chron C24r/C25n FAD D. multiradiatus Chron C25n/C25r

Samples 17-1, 86-90/17-2, 40-42 18-4, 39--42/18-4, 88-92 17-6, 43-45/17-6, 87-90 21-CC, 4--6/22-1, 49-51 22-1, 50-52/22-1, 68-72 23-3, 148-150/23-4, 48-50

FAD of R. cuspis is tentative (see text).

Depth interval (mbsf) 148.40-149.40 162.10-162.50 155.45-155.90 170.31-170.65 185.25-185.70 185.25-185.70 195.69-196.19

Mean depth (mbsf)

Chronology (%)

148.90 162.30 155.68 170.26 170.48 170.64 185.47 185.90 195.94

100 63.68 30.12 42.11 41.52 41.08 0 -

UPPER PALEOCENE--LOWEREOCENE STRATIGRAPHY AND CARBON ISOTOPE EXCURSION However, for the upper (late) part of Chron C24r in Hole 550, comprised between 408 m (the unconformable NP9/NP10 zonal contact) and 359.65 mbsf (the Subchron C24n.3n/Chron C24r boundary) the relative chronology of events cannot be established in the straightforward depth-relative age conversion that we have applied for the lower (early) part of Chron C24r. The NP10 zonal interval in Hole 550 includes a bentonite-rich interval between 384.85 and 404.60 mbsf. Sedimentation rates in this interval are expected to be higher than in the overlying calcareous nannofossil oozes. Also, the interval between 408 and 400 m has been strongly affected by dissolution with consequently lower (apparent) sedimentation rates. To explore the variations in sedimentation rates in Hole 550, we benefit from at least 5 tie points. The -17 and +19 ashes constitute two dated levels with respective ages of 54.5 and 54.0 Ma (Swisher & Knox 1991). The 408 mbsf level yields an estimated age of 55 Ma, derived through extrapolation of the sedimentation rates between the two ashes (Swisher & Knox 1991). The two other tie points are the magnetic reversals at 359.65 and c. 422 mbsf which bound Chron C24r, with respective ages of 53.347 and 55.904 Ma in the GPTS of Cande & Kent (1995). It should be borne in mind that the age of these reversals is dependent upon the age of 55 Ma for level 408 mbsf, so that if all correlations have been established correctly, internal consistency between the GPTS and the stratigraphic record in Hole 550 for Chron C24r should be maintained. Additional tie points are found at younger levels and correspond to the magnetic reversal boundaries of Chron C24n (Table 2). An average sedimentation rate of 2.92 cm/103 years is determined for the upper part of Chron C24r (between 408 mbsf with an age of 55 Ma and

373

359.65 m with an age of 53.347Ma). This compares with an average rate of sedimentation of 2.44 cm/103 years for the whole of Chron C24r (using the Chron C24r boundaries as tie points). This approach would be normally improper because of the unconformity at 408 mbsf, but it is acceptable here because this level was not seen as unconformable when it was selected to provide an age control in the GPTS of Cande & Kent (1992, 1995). A rate of 2.92 cm/103 years is, however, twice the rate of 1.4 cm/103 years calculated for the 7 m interval between the -17 ash (at 400 m) and the +19 ash (at 393 m) with ages of 54.5 and 54.0 Ma, respectively (Swisher & Knox 1991). This implies that the bentonite rich interval between c. 385 and 405 mbsf in Hole 550 was deposited at much lower rates than the overlying calcareous nannofossil oozes between c. 385 and 360 m. A sedimentation rate in excess of 5 cm/103 years is obtained for the oozes between c. 385 and 360 mbsf using the +19 ash and the Subchron C24n.3n/Chron C24r boundary as tie points. This is counter-intuitive in as much as the contribution of volcanogenic detrital material to an otherwise calcareous nannofossil ooze would be expected to increase normal carbonate sedimentation rates significantly. Indeed, we calculate sedimentation rates of 1.78 cm/103 years for the interval representing Chron C24n (Subchron C24n. In to Suchron C24n.3n: c. 342-360 mbsf). Although dissolution has occurred in the bentonite rich interval, it is insufficient to explain the low sedimentation rates that we calculate for it. One of us (AS) has estimated that c. 12.5% of the 40 m-thick stratigraphic interval between the -17 ash and the Subchron C24n.3n/Chron C24r boundary, and c. 60% of the 8 m interval between the -17 ash and the NP9/NP10 zonal boundary (408 mbsf) have been dissolved. This leads to a

Table 2. Position of magnetic reversals in the upper Paleocene-lower Eocene section (Chron C25n to Chron C24n) in DSDP Hole 550 Chron/Subchron

Time interval (million years)

Duration (million years)

Depth interval (mbsf)

C24n. in C24n. lr C24n.2n C24n.2r C24n.3n C24r C25n

52.364-52.663 52.663-52.757 52.757-52.801 52.801-52.903 52.903-53.347 53.347-55.904 55.904-56.331

0.299 0.094 0.044 0.102 0.444 2.557 0.487

342.06-344.81 344.81-349.62 344.81-349.62 344.81-349.62 349.62-359.65 359.65-c. 423 c. 423-?

The magnetochronology is that of Cande & Kent (1995). Note that Subchron C24n.ln to Subchron C24n.2n, Subchron C24n.2r, and Suchron C24n.3n in the terminology of Cande & Kent (1992, 1995) correspond to late Subchron 24n (=Subchron C24AN), Subchron C24Ar, and Suchron C24Bn, respectively, in the terminology of Berggren et al. (1985).

374

M.-P. AUBRY ET AL.

restored thickness of 57.80 m for the upper part (408-359.68 mbsf) of Chron C24r in Hole 550. Using these figures, the corrected sedimentation rate for the (now 7.84 m thick) interval between the two ashes is 1.56 cm/103 years (versus 1.4 cm/ 103years). This represents a minor correction which has little significance regarding the compared rates of deposition for the bentonite rich interval and the overlying carbonate oozes. On the other hand, the thickness correction between the -17 ash and the NP9/NP10 zonal boundary is important because it implies a greater age than 55 Ma for the 408 m level if the sedimentation rates for the +19 ash to -17 ash interval is used. Using a corrected rate of 1.56 cm/103 years, the age is 55.32 Ma. Using the non-corrected rate of 1.4 cm/103 years, the age is 55.41 Ma. Yet, we cannot use a revised age for level 408 mbsf because, as it was believed to correspond to the NP9/NPI0 chronozonal boundary, its age of 55 Ma was used as a tie point by Cande & Kent (1992, 1995) for deriving an early Paleogene magnetochronology. The age of the Subchron C24n.3n/Chron C24r boundary (and of all reversal boundaries between Chron C29 and Chron C21) is (are) fully dependent upon the 55 million years age of level 408 mbsf in Hole 550. Thus we cannot use new revised data (age and thickness) to estimate the average sedimentation rates for the upper part of Chron C24r. We recognize (i) that the section may have been more expanded than represented now, and (ii) that the NP9/NP10 chronozonal boundary may be 0.34).4 million years older than previously estimated. Yet, we cannot apply these new estimates because the age of 55 Ma for level 408 mbsf constitutes a fixed point in the GPTS of Cande & Kent (1992, 1995). This is a deadlock situation, which, as we will show below, is of greater gravity for establishing a numerical chronology for Chron C24r. Unable to resolve the problem until the position of the NP9/NP10 chronozonal boundary in Chron C24r is correctly established, we estimate the relative chronology of the events located above the bentonite bearing nannofossil oozes (above 384.85 mbsf) based on a sedimentation rate of 2.25 cm/103 years (the sedimentation rate for Subchron C24n.3n). We estimate the relative chronology of the events within the bentonite rich interval using a sedimentation rate of 4.28 cm/ 103 years (the rate determined by interpolating between the derived age of 54.46 Ma for the top of the ash-series using the 2.25 cm/103 years in the carbonate section above and 55.0 million years calibration at the 408 mbsf level; see Berggren & Aubry (1996) for additional discussion on these and related points). We emphasize that these two values are workable estimates in a 'desperate' attempt to derive a relative chronology for Chron C24r.

Comparison of the sedimentation rates calculated on the basis of magnetochronology for different segments of Chron C24r and the Subchron C24n.3n to Chron C23n interval indicates that there is a decrease in rates around or just above Subchron C24n.3n (Berggren & Aubry 1996). The rate for Subchron C24n.ln to Subchron C24n.3n is only 1.78 cm/103 years compared to the rate of 2.25 for Subchron C24n.3n that we have chosen to apply to the upper 25.20 m of Chron C24r. The rate of 4.28 cm/103 years compares poorly with the rate of 1.4 cm/103 years derived from the radioisotopic ages on the ash layers. We do not underestimate the fact that the age of 55 Ma for level 408 mbsf is tainted with uncertainty, particularly owing to the fact that intense dissolution occurred in the lower 8 m (400-408 mbsf) of the section. Studies of more complete sections in the near future are the only way we will establish a satisfactory chronology for the upper (younger) part of Chron C24r. If we accept the sedimentation rates of 2.25 and 4.28 crn/103 years for the upper part of Chron C24r, the relative chronology of the palaeontological and isotopic events are those given in Table 3.

Relative chronology for Chron C24r If the average rates of sedimentation were comparable in the NP10 zonal interval in Hole 550 and the NP9-NP10 zonal interval in Hole 690B, a direct conversion depth-relative time based on the total thickness of the NP9-NP10 zonal interval in the composite section of > 89 m (considering that there is probably no overlap of the NP10 zonal intervals in the two sections) would be possible. If that were correct, the NP9/NP10 zonal boundary would lie c. 41% up from the base of Zone NP9. The CIE would then lie at 17.46% and the BFE at 17.20% up from the base of Zone NP9. If we add 6 m to represent the uppermost part of Chron C24r (which correlates with Zone NPll), the NP9/NP10 zonal boundary lies approximately 38.75% up from the base of Chron C24r. We have little means to constrain the sedimentation rates for the NP9-1owermost NP10 zonal interval in Hole 690B. The age of 55 Ma for the NP9/NP10 zonal boundary is not applicable because level 149 mbsf in Hole 690B is not correlative with level 408 mbsf in Hole 550. The only tie points that can be used in Hole 690B are the boundaries of Chron C25n. The normal polarity interval between 185.47 and 195.94mbsf is interpreted by Spiess (1990) to represent Chron C25n. Considering that Chron C25n is 0.487 million years long in Cande & Kent (1995), the corresponding sedimentation rate is 2.14cm/ 103 years. This is a relatively high sedimentation rate for calcareous oozes (but comparable to the

UPPER PALEOCENE--LOWER EOCENE STRATIGRAPHY AND CARBON ISOTOPE EXCURSION

375

Table 3. Relative chronology of events in late Chron C24r (essentially correlative with Biochron NPIO) as deduced from the stratigraphic succession of events in DSDP Hole 550 Event

Subchron C24n. l-2n/Chron C23r Subchron C24n. 1-2n/C24n.2r Subchron C24n.2r/C24n.3n Subchron C24n.3n/chron C24r LAD S. velascoensis LAD T. contortus (Morph B) FAD T. orthostylus LAD M. aequa LAD T. bramlettei FAD T. contortus (Morph B) FAD M. lensiformis LAD T. contortus (Morph A) FAD T. contortus (Morph A) FAD M. formosa gracilis LAD M. acuta FAD A. wilcoxensis FAD T. bramlettei

Core interval

27-3, 146-149/27-4, 117-120 27-6, 12-15/27-6, 102-105 28-2, 138-140/28-3, 36-38 29-3, 27-30/29-3, 103-105 30-1, 49-53/29-6, 51-54 29-CC/30-1, 20-23 30-1 62-64/30-1,109-111 30-3 49-53/30-1,49-53 30-5 22-24/30-5,66-68 30-5 113-! 15/30-6, 7-9 31-2 65-67/31-2, 105-107 31-3 6-8/31-3, 39-41 31.5. 88-90/32-1,17-19 33-1 59-61/33-2,59-61 33-2 51-53/33-2,70-72 34-1 62-65/34-2,62-65

rate we have estimated for the post bentonite carbonate interval of Zone NP10 and N P l l in late Chron C24r in Hole 550), but which is compatible with the presence of fine grained terrigenous material in the section (Barker et al. 1988). We are concerned with the possibility that the normal polarity interval below 191.49 mbsf may not represent Chron C25n. Although well preserved, the calcareous nannofossil assemblages between 188.49 and 200.10 mbsf yield no species indicative of biostratigraphic levels younger than Zone NP5 (in particular no discoasters and no species of Heliolithus; Pospichal & Wise 1990, table 4; MPA, pers. observ.). However, unable at present to demonstrate stratigraphic disturbance around Core 23H in Hole 690B (although we note the alternation of lithologies, see Barker et al. 1988, core photograph p. 265), we tentatively accept a sedimentation rate of 2.14 cm/103 years for the NP9-NP10 zonal interval in Hole 690B, aware of the possibility of an increase in the rates immediately following the benthic foraminifera extinction (Thomas & Shackleton 1996). We thus have established a relative chronology for two segments of the composite section, but based on different sedimentation rates in each. The (average) sedimentation rate of 2.14 cm/103 years for the NP9 segment is 26.17% lower than the average rate of 2.92 cm/103 years for the NP10 segment. Thus if anything the NP9/NP10 boundary lies higher in Chron C24r than the estimate of 38.75% obtained from direct conversion thicknessduration. Since we cannot assess the importance of

Depth interval (mbsf)

Mean depth (mbsf)

Chronology (%)

341.46-342.67 344.60-345.02 349.38-349.86 359.27-360.03 366.03-363.94 365.50-365.72 366.13-366.60 366.03-368.99 371.73-372.17 372.64-373.08 375.17-375.5 378.07-378.40 381.89-384.68 394.51-396.09 396.02-396.21 403.72-405.22

342.06 344.81 349.62 359.65 364.99 365.61 366.36 376.51 371.95 372.86 375.36 378.23 383.28 395.05 396.11 404.47 408

0 11.07 15.91 17.72 18.33 32.84 35.26 37.98 49.78 61.88 81.24 83.06 94.55 100

-

the stratigraphic gap between Holes 550 and 690B, and considering the uncertainties on the contruction of the composite section, particularly on the rates at which the two component sections were deposited, we arbitrarily propose that the NP9/NP10 zonal boundary lies 40% up in Chron C24r. We emphasize that our study does not demonstrate that this is the position of the NP9/NP10 zonal boundary in Chron C24r, which differs from both Berggren et al. (1985) and Aubry et al. (1988), who placed it one-third and two-thirds, respectively, of the way up in Chron C24r. If the sedimentation rates that we have used for Holes 550 and 690B are essentially correct, our interpretation of the relative chronology over Chron C24r suggest that the NP9/NP10 zonal boundary lies closer to mid-way in Chron 24r than thought earlier. Obviously future work on P a l e o c e n e - E o c e n e sections must focus on verifying/improving the relative chronology that we suggest here. N u m e r i c a l c h r o n o l o g y o f e v e n t s in Chron C24r

Based on the relative chronology that we have tentatively established for Chron C24r, it should be a rather simple exercise to establish a numerical chronology for it based on age estimates for the Subchron C24n.3n/Chron C24r and Chron C24r/C25n boundaries in Cande & Kent (1995). Yet, incorrect assumptions in the GPTS of Cande & Kent (1992) prevent us from achieving our goal.

376

M.-P. AUBRYET AL.

An age estimate of 55 Ma for the NP9/NP10 zonal boundary constitutes one of the nine points that Cande & Kent (1992) used to calibrate their composite polarity sequence for the Late Cretaceous and Cenozoic. Swisher & Knox (1991) obtained laser total fusion 4~ ages on two characteristic volcanic ashes from Denmark. The -17 ash yielded a weighted mean age of 54.51 _+0.05 Ma, the +19 ash yielded a weighted mean age of 54.0 _ 0.53 Ma (Berggren et al. 1995). These ashes were identified in DSDP Hole 550 where they occur at c. 400 and 393 mbsf, respectively, i.e. 7 and 13 m above the NP9/NP10 zonal boundary following Mtiller (1985). The NP9/NP10 zonal boundary is used by many workers to characterize the Paleocene/Eocene boundary; thus, the occurrence of two dated ashes in a section which apparently contains the Paleocene/Eocene boundary was seen as providing the opportunity of estimating first hand its age. Extrapolating a calculated sedimentation rate of 1.4 cm/103 years between the two ashes, Swisher & Knox (1991) estimated an age of 55 Ma for the Paleocene/ Eocene boundary (as denoted by the NP9/NP10 zonal boundary), an age that Cande & Kent (1992) retained for the construction of their GPTS. Following placement of the Paleocene/Eocene boundary at c. two-thirds down in Chron C24r by Berggren et al. (1985), Cande & Kent located their calibration point at 0.66 in Chron C24r (= 1221.20km from the ridge axis in the South Atlantic). We now recognize several difficulties with the procedures followed by Cande & Kent (1992). First, as a minor point that we mention briefly, the age of the two ashes is controversial. Obradovich (in Berggren et al. 1992) obtained a bulk incremental heating 4~ plateau age of 55.07 _+ 0.16 Ma on the -17 ash. We have questioned above the low sedimentation rates implied by the age of the two ashes for the bentonite rich interval in DSDP Site 550. Underestimated sedimentation rates would have substantial consequences for the estimated age of the 408 mbsf level in Hole 550 taken by Swisher & Knox (1991) to correspond to the Paleocene/Eocene boundary. Of relevance here is the rather large uncertainty (___0.53 million years) on the +19 ash. Further problems relative to the age of the ashes but that are not directly pertinent to a numerical chronology in Chron C24r are discussed in Berggren & Aubry (1996). The main difficulty with the procedure followed by Cande & Kent (1992) resides in the fact that it was not recognized until now that the NP9/NP10 zonal boundary at 408 mbsf in Hole 550 is, in fact, an unconformable contact. The age estimate of 55 million years by Swisher & Knox (1991) is indeed the age estimate of the upper surface of the un-

conformity at 408 mbsf. It dates a level within lower Zone NP10 but of uncertain stratigraphic position with regard to the NP9/NP10 chronozonal boundary. We have seen that the lower surface of the unconformity is located low in Zone NP9, approximately coincident with the carbon isotope excursion that we have placed at c. 42% up from the base of Zone NP9. As a consequence, positioning of the NP9/NP10 zonal boundary approximately half way down in Chron C24r results in the Paleocene part of Chron C24r (which essentially correlates with Biochron NP9) being considerably misrepresented in the GPTS of Cande & Kent (1992, 1995). The Paleocene part of Chron C24r is represented by a 36.47 m thick interval in Hole 690B. Using the 2.14 cm/103 years that we have determined for the stratigraphic interval representing Chron C25n in the hole (see above), the duration of the Paleocene part of Chron C24r is 1.7 million years rather than 0.904 million years in Cande & Kent (1995). We recognize that this duration may be shorter if there is a local increase in productivity immediately following the benthic foraminiferal extinction, as suggested by Thomas & Shackleton (1996), but we note that the interval of Chron C24r below the carbon isotopic excursion alone represents 0.740 million years. Unfortunately, we cannot propose a corrected estimate for the position of the NP9/NP10 biochronal boundary in Chron C24r because that data required to do so (i.e. the age of the magnetic reversals, the sedimentation rates that we are using, and the duration of the Eocene part of Chron C24r) are contigent upon the use of the 55 Ma calibration point in Cande & Kent (1995). The rather frustrating consequence of this is that we cannot properly calibrate datums which occurred in the late part of Biochron NP9. It is obvious that an age of 55 Ma for the carbon isotope excursion (which we would estimate if Hole 550 was thought to be continuous across the NP9/NP10 boundary) is totally unrealistic. Mislocation of the NP9/NPI0 chronozonal boundary within Chron C24r results in compressing the succession of events which occurred during Biochron NP9 while, on the contrary, stretching the succession of events which occurred during Biochron NP10 (and NPll). However, we cannot relocate the NP9/NP10 zonal boundary in Chron C24r in the GPTS of Cande & Kent (1992, 1995) because it carries the age calibration of 55 Ma, upon which the numerical age of the boundaries of Chron C24r are dependent. In summary, what is to be done? The GPTS of Cande & Kent (1992, 1995) will necessitate revision based on repositioning of their '55 Ma' tie point and decoupling it from the Paleocene/Eocene (and NP9/NP10 zonal) boundary. This is needed independently of any confirmation or revision of

UPPER PALEOCENE--LOWER EOCENE STRATIGRAPHY AND CARBON ISOTOPE EXCURSION

the age of the early Eocene ashes dated in northwestern Europe. We recommend that the age of the ashes themselves be used rather than an estimated age derived from them. This revision implies modification of the magnetochronology between the next two calibration points (in Chron C29r and C21n) in Cande & Kent (1995), with decreasing age differences towards the next tie-points, and greater age differences in the Paleocene than in the Eocene. The current values in the GPTS on Chron C24r allow the establishment of a numerical magnetobiochronological framework for the late part of Chron C24r (between 53.347 and 55 Ma; Table 4). However, at present it is impossible to construct a firm numerical chronology for the early part of Chron C24r (between 55 and 55.904 Ma) because more than half of this time interval is not accounted for in the current GPTS (Cande & Kent 1995). Yet, without a sound GPTS, it is impossible to measure the rates at which changes proceeded at a time as critical in earth history as around the Paleocene/Eocene boundary. Until a revision to the GPTS is available, the only (admittedly unsatisfactory) solution is to construct the numerical chronology of events in early Chron 24r into the 0.904 million years allowed for the early part of Chron C24r (Table 4). The deadlock situation at which we have arrived in using the GPTS in geological history stems largely from the fact that Cande & Kent (1995)

377

used a calibration point which was also thought to correspond to an epoch boundary (and additionally a biozonal boundary). This has created a closed system from which there is no escape. In view of the tremendous progress currently achieved in understanding the relationship between the stratigraphic record and geological time, it would seem highly suitable that calibration points in the GPTS be chosen independant from epoch boundaries.

Summary Regional stable isotope (C, O) correlations consist of pattern matchings often within a weakly constrained biostratigraphic framework. In this exercise we have demonstrated the need for close sample spacing (similar to that commonly done in upper Neogene, Pliocene-Pleistocene) in studies in the Paleogene, as well as rigourous, integrated analysis of calcareous nannofossils and planktonic foraminifera to provide a relatively precise biostratigraphic framework within which stable isotopic curves can be placed. We have further shown that multiple unconformities with variable hiatuses occur at three deep sea sites in the North and South Atlantic Ocean. Simple pattern matching (observed in different biostratigraphic zones) would have led to a false interpretation of the carbon isotope record. Once a composite standard

Table 4. Numerical chronology of events in Chron C24r Chron C24r Late Chron C24r (53.347-55 million years)

NP9/NP 10 biochronal boundary Early Chron C24r (55-55.904 million years)

Datum events

Estimated ages (Ma)

LAD Tribrachiatus contortus (morphotype B) FAD T. orthostylus LAD T. bramlettei FAD T. contortus (morphotype B) LAD T. contortus (morphotype A) FAD T. contortus (morphotype A) FAD T. bramlettei FAD Rhomboaster cuspis < abundance (F to VR) Fasciculithus tympaniformis carbon isotope excursion benthic foraminifera extinction FAD G. australiformis

53.61 53.64 53.89 53.93 54.17 54.37 55 55.16 55.33 55.52 55.52 55.53

Age estimates for events in the early part of Chron C24r are provisional and established within the constraints of the current GPTS (Cande & Kent 1992, 1995) which allows only 0.904 million years for the interval comprised between the Chron C25n/C24r boundary and the NP9/NP10 biochronal boundary. Age (55 Ma) of T. bramlettei, estimated through extrapolation of sedimentation rates between -17 and +19 ashes, corresponds to the Chron C24.(0.66) calibration point in Cande & Kent (1992, 1995). The age of this datum is one of the constraints of the GPTS and is fixed. FAD of Rhomboaster cuspis is is tentative here because of preservation. There is a worldwide decrease in abundance of Fasciculithus tympaniformis during Biochron NP9, but its timing at different latitudes remains to be established.

378

M.-P. AUBRY ET AL.

was constructed within a biostratigraphic framework it became apparent that there are two, rather than one, decreases/excursions in the carbon isotope record in Chron C24. The chronology of these two events has been rendered difficult to determine/estimate because of the recognition that the NP9/NP10 zonal boundary at DSDP Site 550, chosen as one of the numerical calibration points in the recently revised Cenozoic geochronology of Cande & Kent (1992, 1995), is actually situated at an unconformity. This results in the impossiblity to derive a sound chronological framework for the early part of Chron C24r. This study shows the importance of detailed studies focusing on specific stratigraphic-temporal intervals aiming at defining GSSPs for epoch boundaries, and the need for rigorous interpretation of the stratigraphic record in geochronology as in geological history. Because they encourage the integration of data from very different disciplines, IGCP Projects provide an unmatched opportunity to m a k e f u n d a m e n t a l progress in geological sciences based on a continued and growing interest in the science of stratigraphy.

Postscript

Tribrachiatus contortus Morph. A is now described under the name Tribrachiatus digitalis Aubry, 1995 (Israel Journal o f Earth Sciences, in press). Its holotype is from level DSDP Site 550, Core 34, Section 3, 125-127 cm, and is illustrated in Fig. 3b in this paper.

We thank our colleagues in IGCP Project 308 (Paleocene/Eocene Boundary Events in Space and Time) for their continued support and collaboration. In particular stimulating discussions of an early version of this paper with R. M. Corfield, J. Hardenbol, D. V., Kent, R. W. O'B. Knox, K. G. Miller, R. Norris, D. Pak, B. Schmitz and E. Thomas have helped clarify the complex problems associated with determining the temporal and spatial sequence of events associated with the P/E boundary. MPA kindly thanks A. Boerma and E. Thomas for sharing with her samples from Hole 690B. We thank D. V. Kent, K. G. Miller, E. Thomas and B. Schmitz for reviewing the manuscript. This is a contribution towards IGCP Project 308. This is ISEM Contribution no. 95044 and Woods Hole Contribution no. 8856.

References

AUBRY, M.-R 1990. Late Paleogene calcareous nannoplankton evolution: A tale of climatic deterioration. In: PROTHERO, D. R. & BERGGREN, W. A. (eds) Eocene-Oligocene Climatic and Biotic Evolution. Princeton University Press, 272-309. -1991. Sequence stratigraphy: Eustasy or tectonic imprint? Journal of Geophysical Research, 96(B4), 6641-6679. 1995. From chronology to stratigraphy: assessing the continuity of the stratigraphic record. In: BERGGREN, W. A., KENT, D. V., AUBRY, M.-P. & HARDENBOL, J. (eds) Geochronology, Time Scales and Stratigraphic Correlations: Framework for an Historical Geology. Society for Economic Paleontologists and Mineralogists, Special Publication, 54. --, BERGGREN, W. A., KENT, D. V., FLYNN, J. J., KLITGORD,K. D., OBRADOVICH,J. D. & PROTHERO, D. R. 1988. Paleogene geochronology: An integrated approach. Paleoceanography, 6(3), 707-742. BACKMAN, J. 1984. Cenozoic calcareous nannofossil biostratigraphy from the northeastern Atlantic Ocean - Deep Sea Drilling Project Leg 81. In: ROBERTS, D. G., SCHNITKER, D., ET AL. Initial Reports of the Deep Sea Drilling Project, 81, 403-428. -1986. Late Paleocene to middle Eocene calcareous nannofossil biochronology from the Shatsky Rise, Walvis Ridge and Italy. Palaeogeography, Palaeoclimatology, Palaeoecology, 57, 43-59.

BARKER, P. E, KENNETT, J.-E, ET AL. 1988. Site 690. Proceedings of the Ocean Drilling Program, Initial Reports, 113, 183-292. BERGGREN, W. A. & AUBRY, M.-E 1996. A late Paleocene-early Eocene NW European and North Sea magnetobiochronological correlation network. This volume. & MILLER, K. G. 1988. Paleogene tropical planktonic foraminiferal biostratigraphy and magnetochronology. Micropaleontology, 34(4), 362-380. --, KENT, D. V. & FLYNN, J. J. 1985, Paleogene geochronology and chronostratigraphy. In: SNELLIN6, N. J. (ed.) The Chronology of the Geological Record, Geological Society, London, Memoir, 10, 141-195. --, --, SWISHER, C. C. III & AUBRY, M.-R 1995. A revised Cenozoic geochronology and chronostratigraphy. In: BERGGREN, W. A., KENT, D. V., AUBRY, M.-R & HARDENBOL, J. (eds) Geochronology, Time Scales and Stratigraphic Correlation: Framework for an Historical Geology. Society for Economic Paleontologists and Mineralogists, Special Publication, 54. BOERSMA, A., PREMOLI-SILVA, I. • SHACKLETON, N. 1987. Atlantic Eocene planktonic foraminiferal paleohydrologic indicators and stable isotope stratigraphy. Paleoceanography, 2, 287-331. BUKRY, D. 1975. Coccolith and silicoflagellate stratigraphy, northwestern Pacific Ocean, Deep Sea Drilling Project Leg 32. In: LARSON, R. L.,

UPPER PALEOCENE--LOWER EOCENE STRATIGRAPHY AND CARBON ISOTOPE EXCURSION MOBERLY, R., ET AL. Initial Reports of the Deep Sea Drilling Project, 32, 677-701. CANDE, S. C. & KENT, D. V. 1992. A new geomagnetic polarity time scale for the Late Cretaceous and Cenozoic. Journal of Geophysical Research, 97(B 10), 13917-13951. & -1995. Revised calibration of the geomagnetic polarity time scale for the Late Cretaceous and Cenozoic. Journal of Geophysical Research, 100 (B4), 6093-6095. CAVELIER, C. & POMEROL, C. 1986. Stratigraphy of the Paleogene. Bulletin de la Socidtd Gdologique de France, 8 (II, 2), 255-265. CO~IELD, R. M. & CARTLIDCE,J. E. 1992. Oceanographic and climatic implications of the Palaeocene carbon isotope maximum. Terra Nova, 4, 443-445. 9- - , PREMOLI SILVA, I. & HOUSEY, R. A. 1991. Oxygen and carbon isotope stratigraphy of the Palaeogene and Cretaceous limestones in the Bottaccione Gorge and the Contessa Highway sections, Umbrian, Italy. Terra Nova, 3, 414-422. ELLISON, R. A., JOLLEY, D. W., KING, C. & KNOX, R. W. O'B. 1994. A revision of the lithostratigraphical classification of the early Palaeogene strata in the London Basin and East Anglia. Proceedings of the Geologists' Association, 105, 187-197. DE GRACIANSKY,P. C., POAG, C. W., ET AL. 1985a. Site 549. In: Initial Reports of the Deep Sea Drilling Project, 80, 123-250. , ET AL. 1985b. Site 550. In: Initial Reports of the Deep Sea Drilling Project, 80, 251-355. FENNER, J. 1994. Diatoms of the Fur Formation, their taxonomy and biostratigraphic interpretation Results from the Harre borehole, Denmark. Aarhus Geoscience, 1, 99-163. HEATH, G. R., BURCKLE,L. H., ETAL. 1985. Initial Reports of the Deep Sea Drilling Project, 86. HOOKER, J. J. 1991. The sequence of mammals in the Thanetian and Ypresian of the London and Belgium basins. Location of the Paleocene-Eocene boundary. Newsletters on Stratigraphy, 25(2), 75-90. KENNETT, J. E & STO'rr, L. D. 1991. Abrupt deep-sea warming, paleoceanographic changes and benthic extinctions at the end of the Palaeocene. Nature, 353, 225-229. KNOX, R. W. O'B. 1984. Nannoplankton zonation and the Paleocene/Eocene boundary beds of northwestern Europe: An indirect correlation by means of volcanic ash layers. Journal of the Geological Society, London, 141, 993-999. - 1985. Stratigraphic significance of volcanic ash in paleocene and Eocene sediments at Sites 549 and 550. In: DE GRACIANSKY,P. C., POAG, C. W., ET AL. Initial Reports of the Deep Sea Drilling Project, 8 0 , 845-850. -& MORTON,A. C. 1988. The record of early Tertiary N Atlantic volcanism in sediments of the North Sea Basin. In: MORTON, A. C. & PARSON, L. M. (eds) Early Tertiary volcanism and the opening of the NE Atlantic. Geological Society, London, Special Publication, 39, 407-419.

379

KOCH, R L., ZACrtOS, J. C. & GINGERICH, R D. 1992. Correlation between isotope records in marine and continental carbon reservoirs near the Palaeocene/Eocene boundary. Nature, 358, 319322. LAUGHTON, A. S., BERGGREN, W. A., ET AL. 1972. Initial Reports of the Deep Sea Drilling Project, 12. MARTINI, E. 1971. Standard Tertiary and Quaternary calcareous nannoplankton zonation. In: FARINACCI, A. (ed.) Proceedings of the Second Planktonic Conference, Roma 1970, 739-785. MILLER, K. G., JANECEK,T. R., KATZ,M. E. & KEIL, D. K. 1987. Abyssal circulation and benthic foraminiferal changes near the Paleocene/Eocene boundary. Paleoceanography, 2(6), 741-761. MONECHI, S. 1985. Campanian to Pleistocene calcareous nannofossil stratigraphy from the northwest Pacific Ocean, Deep Sea Drilling Project Leg 86. In: HEATH, G. R., BURCKLE,L. H., Er AL. Initial Reports of the Deep Sea Drilling Project, 86, 301-336. MONTADERT, L. & ROBERTS, D. G. 1979. Sites 403 and 404. In: Initial Reports of the Deep Sea Drilling Project, 48, 165-209. MORTON, A. C., BAC~AN, J. & HARLAND, R. 1983. A reassessment of the stratigraphy of DSDP Hole 117A, Rockall Plateau: Implications for the Paleocene-Eocene boundary in N.W. Europe. Newsletters on Stratigraphy, 12(2), 104-111. MOLLER, C. 1979. Calcareous nannofossils from the North Atlantic (Leg 48). In: MONTADERT, L., ROBERTS, D. G., Er AL. Initial Reports of the Deep Sea Drilling Project, 48, 589-639. 1985. Biostratigraphic and paleoenvironmental interpretation of the Goban Spur Region based on a study of calcareous nannoplankton. In: DE GRACIANSKY,P. C., POAG, C. W., ET AL. Initial Reports of the Deep Sea Drilling Project, 80, 573-599. OKADA, H. • BUKRY, D. 1980. Supplementary modification and introduction of code numbers to the low-latitude coccolith biostratigraphic zonation. Marine Micropaleontology, 5, 321-325. PAK, D. K. & MILLER, K. G. 1992. Paleocene to Eocene benthic foraminiferal isotopes and assemblages: Implications for deepwater circulation. Paleoceanography, 7(4), 405-422. POSPICHAL,J. J. & WISE, S. W. 1990. Paleocene to middle Eocene calcareous nannofossils of ODP Sites 689 and 960, Maud Rise, Weddell Sea. In: BARKER, P. E, KENNETT, J. P., ET AL. Proceedings of the Ocean Drilling Project, Initial Reports, 113, 613-666. PROTODECIMA, E, ROTH, P. H. & TODESCO, I. 1975. Nannoplancton calcareo del Paleocene e dell'Eocene della Sezione di Possagno. Schweizerische Palgiontologische Abhandlungen, 97, 35-161. REA, D. K., ZACHOS, J. C., OWEN, R. M. & GINGERICH, R. D. 1990. Global change at the Paleocene/Eocene boundary: Climatic and evolutionary consequences of tectonic events. Palaeogeography, Palaeoclimatology, Palaeoecology, 79, 117-128. ROBERTS, D.G., SCHNITKER, D., ET AL. 1984. Initial Reports of the Deep Sea Drilling Project, 81.

380

M.-P. AUBRY ET AL.

ROMEIN, A. T. J. 1979. Lineages in early Paleogene calcareous nannoplankton. Utrecht Micropaleontological Bulletin, 22, 1-231. SINHA, A. & STOTT, U D. 1993. Recognition of the Paleocene/Eocene boundary carbon isotope excursion in the Paris Basin, France. (Abstract) In: Symposium on the Correlation of the Early Paleogene in Northwestern Europe, 1-2 December 1993. Geological Society, London. SNYDER, S. W. & WATERS, V. J. 1985. Cenozoic planktonic foraminiferaf biostratigraphy of the Goban Spur region, Deep Sea Drilling Project Leg 80. In: DE GRACIANSKY,P. C., POAG, C. W., ETAL. Initial Reports of the Deep Sea Drilling Project, 80, 439-472. --, MULLER, C., SIGAL, J., TOWNSEND, H. & POAG, C. W. 1985. Biostratigraphic, peloenvirnomental, and paleomagnetic synthesis of the Goban Spur region, Deep Sea Drilling Project Leg 80. In: DE GRACIANSKY, P. C., POAG, C. W., ET AL. lnitial Reports of the Deep Sea Drilling Project, 80, 1169--1186. SPIESS, V. 1990. Cenozoic magnetostratigraphy of Leg 113 drill sites, Maud Rise, Weddell Sea, Antarctica. In: BARKER,P. E, KENNETT,J. P., ETAL. Proceedings of the Ocean Drilling Project, Initial Reports, 113, 261-315. STOTT, L. D. 1992. Higher temperatures and lower oceanic pCO2: A climate enigma at the end of the Paleocene epoch. Paleoceanography, 7(4), 395-404. & KENNETT,J. P. 1990. Antarctic Paleogene planktonic foraminifer biostratigraphy: ODP Leg 113, Sites 689-690. In: BARKER, P. F., KENNETT, J. P., ET AL. Proceedings of the Ocean Drilling Program, Initial Reports, 113, 549-569. - - , SHACKLETON,N. J. & CORFIELD,R. M. 1990. The evolution of Antarctic surface waters during the Paleogene: Inferences from the stable isotopic composition of planktonic foraminifers, ODP Leg 113. In: BARKER, P. F., KENNETT, J. P., ET AL. Proceedings of the Ocean Drilling Program, Initial Reports, 113, 849-863. , SINHA, A., THIRY, M., AUBRY,M.-P. & BERGGREN, W. A. 1996. Global ~13C changes across the Paleocene-Eocene boundary: criteria for terrestrialmarine correlations. This volume. SWISHER, C. C. III& KNOX, R. W. O'B. 1991. The age of the Paleocene/Eocene boundary: 4~ dating -

-

of the lower part of NP10, North Sea Basin and Denmark. In: IGCP Project 308 (Paleocene/Eocene boundary events), International Annual Meeting and Field Conference, Brussels, 2-6 December 1991, Abstracts with Program. 16. THOMAS, E. 1990a. Late Cretaceous--early Eocene mass extinctions in the deep sea. In: SHARPTON,V. L. & WARD, P. D (eds) Global Catastrophes in Earth History. Geological Society of America, Special Paper, 247, 481-495. 1990b. Late Cretaceous through Neogene deep-sea benthic foraminifers (Maud Rise, Weddell Sea, Antarctica). In: BARKER,P. E, KENNETT,J. P., ETAL. Proceedings of the Ocean Drilling Program, Initial Reports, 113, 571-594. 1993. Cenozoic deep sea circulation: evidence from deep sea benthic foraminifera. In: KEr,~Err, J. E & WARNKE, D. (eds) The Antarctic Paleoenvironment: A Perspective on Global Change. American Geophysical Union Antarctic Research Series, 56, 141-165. -• SHACKLETON,N. J. 1993. The Paleocene benthic foraminiferal extinction: timing, duration and association with stable isotope anomalies. (Abstract) In: Symposium on the Correlation of the Early Paleogene in Northwestern Europe, Geological Society of London (1-2 December 1993). Geological Society, London. & 1996. The Paleocene-Eocene benthic foraminiferal extinction and stable isotope anomalies. This volume. --, BARRERA,E., HAMILTON,N., HUBER, T., KENNETT, J. P., O'CONNELL, S. B., POSPICHAL,J. J., SPIESS, V., STOTT, L., WEI, W. & WISE, S. W. JR. !990. Upper Cretaceous-Paleogene stratigraphy of Sites 689 and 690, Maud Rise (Antarctica). In: BARKER, E E, Kennett, J. E, ET AL. Proceedings of the Ocean Drilling Program, Initial Reports, 113, 901-914. TJALSMA, R. C. & LOHMAN, G. P. 1983. PaleoceneEocene bathyal and abyssal benthic foraminifera from the Atlantic Ocean. Micropaleontology, Special Publication, 4, 1-90. TOWNSEND, H. A. 1985. The paleomagnetism of sediments acquired from the Goban Spur on Deep Sea Drilling Project 80. In: DE GRACIANSKY,P. C., POAG, C. W., ETAL Initial Reports of the Deep Sea Drilling Project, 80, 389-421.

-

-

Global ~13C changes across the Paleocene-Eocene boundary: criteria for terrestrial-marine correlations LOWELL

D. S T O T T 1, A S H I S H

MARIE-PIERRE

S I N H A 1, M E D A R D

AUBRY 3 & WILLIAM

T H I R Y 2,

A. B E R G G R E N 4

1 Department of Earth Sciences, University of Southern California, Los Angeles, CA 90089-0740, USA 2 Ecole des Mines de Paris, Centre d'Information Gdologique, 35 rue Saint-Honord 77305 Fontainbleau cedex, France 3 Institut des Sciences de l'Evolution, Universitd Montpellier II, 34095 Montpellier, Cedex 5, France 4 Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA Abstract: The early Cenozoic marine carbon isotopic record is marked by a long-term shift from high 813C values in the late Paleocene to values that are 2 to 3 lower in the early Eocene. The shift is recorded in fossil carbonates from each ocean basin and represents a large change in the distribution of 12C between the ocean and other carbon reservoirs. Superimposed upon this long-term shift are several distinct carbon isotopic negative excursions that are also recorded globally. These carbon isotopic 'events' near the Paleocene-Eocene boundary provide stratigraphic information that can facilitate intersite correlations between marine and non-marine sequences. Here we present a detailed marine carbon isotopic stratigraphy across the Paleocene-Eocene boundary that is constrained by calcareous nannofossil and planktonic foraminifera biostratigraphy and magnetostratigraphy. We show that several distinct carbon isotopic changes are recorded in uppermost Paleocene and lowermost Eocene marine biogenic carbonate sediments. At least one of these isotopic changes in the ocean's carbon isotopic composition was transmitted to terrestrial carbon reservoirs, including plant biomass via atmospheric CO 2. As a consequence of this exchange of 12C between the ocean and terrestrial carbon reservoirs, it is possible to use carbon isotope stratigraphy to correlate the uppermost Paleocene and lowermost Eocene non-fossiliferous terrestrial sediments of the Paris Basin with marine sequences.

The transition from the Paleocene to the Eocene epoch was one of the most important environmental and biological transitions in earth history. It was at this time that many archaic terrestrial mammals became extinct and modern groups began to diversify (Hooker 1991). The biological changes that occurred during the transition between the Paleocene and Eocene were accompanied by dramatic warming of high latitude sea surface temperatures and deep waters throughout the oceans (Shackleton et al. 1984; Miller et al. 1987; Kennett & Stott 1990; Rea et al. 1990; Stott et al. 1990; Zachos et al. 1993; Pak & Miller 1992; Stott 1992). These changes also accompanied a longterm decrease in the carbon isotopic composition of the oceans (Fig. 1). Superimposed on the long-term climate, biotic and ocean chemistry changes was a short-term (c. 100 000 years) extreme warming of high latitude sea surface temperatures near the end of the Paleocene epoch and a rapid 3 negative

excursion in the carbon isotopic composition of dissolved inorganic carbon in the oceans (Fig. 2). These rapid oceanic changes coincided with an abrupt mass extinction of approximately 50% of deep sea benthic foraminifera (Tjalsma & Lohmann 1983; Thomas 1989, 1990; Pak & Miller 1992; Thomas & Shackleton 1993; Kennett & Stott 1995). These geochemical and biotic changes occurred within calcareous nannofossil Zone NP9 and provide a series of stratigraphic datums that can facilitate correlations between marine and nonmarine sections. For example, the large c. 3 carbon isotope change in the isotopic composition of oceanic CO 2 that occurred in association with the benthic foraminifera extinction would have been associated with a change of similar magnitude and direction in the isotopic composition of atmospheric CO 2. Plant biomass that utilized atmospheric CO 2 for photosynthesis should also record an isotopic change of similar magnitude. However, this excursion is not unique. Using

From Knox, R. W. O'B., Corfield, R. M. & Dunay, R. E. (eds), 1996, Correlation of the Early Paleogene in Northwest Europe, Geological Society Special Publication No. 101, pp. 381-399.

381

382

L . D . STOTT ET AL.

Plankton Magnetic Zonation Polarity Nanno Foram Chron Hist.

({~ 13 C I : ~ B -1.0

0,0

2.0

1,0

,

I

MP1t P7Ic~alm

~P11 '1c24 qPlOlP6a1 ~1PlllP6bI /

,,

3.0

I

'

i

I

[]

i

I

9O

~

Long.Term ,-,

O

oilJ o L ' ~ ~ ( ~ l t l ~ ~

1C24

qP1OI

BenthicExtincti~ O

, o-~

V-

O0 00

O,O..,,.v

9 DSDPSite 527

[] ODP Sites689/690 o

% Fig. 1. Paleocene to middle Eocene biostratigraphy, magnetostratigraphy and carbon isotope stratigraphy. The carbon isotopic record is based upon Atlantic benthic foraminifera NuttaUides (circles) and Cibicidoides (squares). Planktonic ~I3c values would be slightly more positive. The isotopic data is from deep sea sequences located in the Antarctic (Kennett & Stott 1990) and the South Atlantic (Shackleton et al. 1984). Note the marked increase in 513C from the C27n to the top of C25r. Note also the decrease in isotope values across the Paleocene-Eocene boundary.

detailed biostratigraphy to constrain the relative timing of carbon isotopic changes recorded at various oceanic sites we attempt to show that there must have been more than one isotopic excursion superimposed on the longer-term isotopic change outlined earlier. In order to apply isotope stratigraphy to marine-terrestrial correlations a continuous composite isotopic record for marine sections must first be developed. This paper represents our initial attempt to develop such a standardized composite record that can be used to

correlate between marine and terrestrial sections of northwestern Europe. In this paper we describe the systematic carbon isotopic changes across the P a l e o c e n e - E o c e n e boundary in three deep sea sites, two located within the Bay of Biscay (DSDP Sites 549 and 550) and one in the Antarctic (ODP Site 690). We use biostratigraphy and magnetostratigraphy to constrain the timing of the distinct carbon isotope changes for the Paleocene and Eocene. The carbon isotope stratigraphy from deep sea sections is used

GLOBAL PALEOCENE--EOCENE CARBON ISOTOPE CHANGES

383

1~13 CpDB

166

-2.0

-1.0

. . . .

I

0.0 . . . .

I

1.0 . . . .

I

2.0 . . . .

I

3.0 . . . .

t

4.0 . . . .

167

168 169

Depth (mbsf)

Benthic

170

Extinction

171 172 173174

1~18OpDB -2.5 166

. . . .

-2.0

-1.5

I

I

. . . .

-1.0 . . . .

t

. . . .

-0.5

0.0

I

I

. . . .

0.5 I..'..

. . . .

167

168 169

Depth 170 (mbsf)

,Benthic

Extinct ior

171 172 173 174

c .... 23

~ .... 21

~ .... 19

t .... 17

,'" 15

, 13

11

~

Fig. 2. The carbon and oxygen isotopic excursion in Antarctic ODP Site 690 that was associated with the mass extinction of deep sea benthic foraminifera at the end of the Paleocene (from Kennett & Stott 1991). Planktonic (Acarinina praepentacamerata, > 250 mm; Subbotina patagonica > 250 ram) and benthic (Nuttallides truempyi) are plotted together. The isotopic change was similar change in both planktonic and benthic values. During the excursion the surface (planktonic) to bottom-water (benthic) gradient was eliminated.

384

L.D. STOTT ET AL.

to develop a conceptual stratigraphic model for the terrestrial isotope record. This model predicts what the isotopic changes would have been in terrestrial sections across the Paleocene-Eocene boundary. We compare isotopic values from a Sparnacian sequence from the Paris Basin to this chemostratigraphic model to illustrate that the systematic carbon isotopic changes observed in marine carbonates are also preserved in terrestrial deposits. These systematic isotopic changes include a distinct carbon isotope excursion similar in magnitude and direction to that seen near the Paleocene-Eocene boundary in marine sediments. This excursion is used to correlate the PaleoceneEocene transition between these terrestrial sequences marine records.

Methods Organic carbon Organic carbon occluded within the carbonate nodules from the Paris Basin continental deposits was extracted for isotopic measurement in the following way. Cleaned nodules were crushed and placed into 250 ml precombusted Pyrex beakers. A dilute HC1 solution (10%) was added to each beaker (c. 50 ml). The beakers were then covered for 24 hours. When the samples were decarbonated, the beaker was poured over precombusted quartz fibre filters. The quartz fibre was placed into precombusted 9 mm quartz or Vycor tubes and placed on a high vacuum line and lyophilized. When dry, CuO, Cu and a small piece of Ag was added to each tube, re-evacuated and sealed. Each sample was combusted at 850~ for three hours. CO 2 was extracted cryogenically on a high vacuum line and sealed in 6 m m break seals until isotopic analysis was conducted on a VG Prism mass spectrometer. A total of 26 samples was analysed. Of these, six were duplicates. Reproducibility of the six duplicates is _+0.29 (1~). Additionally, ten organic standards (USC cellulose) were run to monitor analytical precision. Reproducibility of these ten samples is _+0.22 (lc~). Fossil c a r b o n a t e s All deep sea samples were dried, weighed and briefly soaked in sodium hexametaphosphate solution and then washed with tap water over a > 63 mm screen. The > 63 mm fraction was dried at 50~ and reweighed. Cleaned shells (c. 15-20) of size-specific planktonic foraminiferal species (e.g. Morozovella subbotinae and Subbotina patagonica) were isolated under a binocular microscope. The samples were briefly sonified in

methanol to remove adhering particles and dried at 50~ before being analysed in an automated carbonate preparation device connected to a VG Prism isotope ratio mass spectrometer. The isotopic results are presented in Tables 1 to 4.

The Paleocene-Eocene boundary problem in the Paris Basin The clay formations that comprise the stratigraphic succession across the Paleocene-Eocene boundary in the Paris Basin are generally designated the 'Sparnacian' (Dollfus 1880). The Sparnacian was once considered to be a chronostratigraphic unit (Stage). It is now more simply regarded as a facies (Laurain et al. 1983). The paralic Sparnacian deposits lie stratigraphically between marine deposits that are within mid-late calcareous nannofossil Zone NP9

Table 1. Isotopic composition of co-existing paleosol organic matter and pedogenic carbonate from the Paris Basin (Limay) (isotopic results expressed relative to the PDB standard) Depth (metres) 0.50 2.00 2.40 2.65 3.00 3.00 3.30 4.00 4.50 6.00 6.10 6.10 6.20 6.20 6.60 6.85 7.00 7.00 7.20 7.50 7.80 8.80 9.70 10.30

Sample

3229 3232 1 2 3 3* 3240 4 5 7 3156 3156* 10 10" 11 12 3160 3160* 22 14 15 16 17 18

11.10

19

11.50

20

* Duplicates

13C%e Organic

-23.2 -23.2 -23.4 -23.2 -24.6 -24.7 -24.1 -23.6 -23.6 -21.3 -22.6 -22.4 -22.0 -21.6 -21.2 -21.7 -22.1 -22.2 -26.9 -23.3 -22.1 -23.6 -23.1 -23.5 -23.0 -24.6

385

GLOBAL PALEOCENE--EOCENE CARBON ISOTOPE CHANGES Table 2. Carbon and oxygen isotopic results of planktonic foraminifera from DSDP Site 549 (isotopic results presented in per mil notation relative to the PDB standard) Core-section-cm

Depth (mbsf)

813C 8180 M. subbotinae

813C 8180 S. patagonica

16-1-22-24 16-1-71.5-73.5 16-1-102.5-104.5 16-1-136-138 16-2-42-44 16-2-72-74 16-2-102.5-104.5 16-2-137-139 16-3-13-15 16-3-23-25 16-3-34-36 ' 16-3-48-49 16-3-58-60 16-3-66~i8 16-3-78-80 16-3-100-102 16-3-117-119 16-3-144-146

331.73 332.21 332.52 332.86 333.43 333.73 334.02 334.38 334.64 334.74 334.85 334.99 335.08 335.17 335.29 335.50 335.68 335.95 336.09 336.24 336.41 336.73 336.85 336.98 337.11 337.67 337.73 337.83 338.05 338.05 338.21 338.34 338.42 338.47 338.70 338.98 339.15 339.68 339.97 340.40 340.60 340.72 340.95

3.21 2.94 2.37

-2.59 -2.51 -2.50

3.03 3.01 2.75 3.20 2.75 3.41

-2.34 -2.21 -2.31 -2.40 -2.30 -2.60

0.82 0.94 0.64 0.77 0.71 0.78 0.84 0.69

-1.36 -1.40 -1.18 -1.21 -1.21 -1.33 - 1.46 -1.12

2.98 2.89 3.01 2.95 3.12 3.82 3.52 3.91 3.52 3.65 3.78 3.76 3.83 3.39 3.67 3.42 3.28 2.82 2.99 2.35 3.03 2.59 1.69 0.22

-2.35 -2.28 -2.64 -2.05 -2.21 -2.01 -1.66 -2.12 -2.29 -2.31 -2.00 - 1.92 -2.09 -2.02 -2.28 -2.35 -2.12 -2.43 -2.30 -2.93 -2.45 -2.30 -3.29 -1.86

0.04 -0.40

-1.55 -2.23

4.31 4.71

-2.95 -2.34

3.82 3.88 3.49

-2.66 -2.00 -2.17

1.76 2.23 1.64

-1.20

16-4-08-10 16-4-23.5-25.5 16-4-40-42 16-4-72-74 16-4-84-86 16-4-97-99 16-4-110-112 16-5-16-18 16-5-22-24 16-5-32-34 16-5-54-56 16-5-54-56 16-5-70-72 16-5-83-85 16-5-91-93 16-5-103-105 16-5-119-121

16-5-147-149 16-6-14-16 16-6-67-69 16-6-96-98

16-6-139-141 16-7-009-011 16-7-21-23 16-7-44-46

(Sables de Bracheux: Aubry 1983, 1985; Woolwich & Reading Bottom Bed: Siesser et al. 1987) below, and mid Zone N P l l (Aubry 1983, 1985) above. Based on indirect magnetobiostratigraphic correlations, the Sparnacian beds were deposited during Chron 24R (Aubry 1983, 1985; Aubry et al. 1986). However, the precise stratigraphic position of the various Sparnacian f o r m a t i o n s is difficult to establish because the beds do not contain stratigraphic e l e m e n t s that allow for interregional correlations.

Carbon isotope stratigraphy DSDP Sites 549 and 550 are located on the Goban Spur in the northeast Atlantic, adjacent to the classical Paleocene and Eocene sections of the Paris and L o n d o n basins. At the present time these two deep sea sites, together with O D P Site 690, provide the most detailed chemostratigraphic and biostratigraphic records available for the Paleocene-Eocene boundary interval (Aubry et al. 1996; Berggren & Aubry 1996). D S D P Sites 550

Table 3. Carbon and oxygen isotopic results of planktonic foraminifera from DSDP Site 550 (isotopic results are presented in per mil notation relative to the PDB standard) Core-section-cm

Depth (mbsf)

813C

8180

M. subbotinae 30-1-21-23 30-1-62-64 30-1 - 109-111 30-1-141-143 30-2-28-30 30-2-63-65 30-2-10-104 30-2- 136-138 30-3-39-41 30-3-74-76 30-3-99-101 30-3-138-140 30-4-99-101 30-4-99-103 30-4-138-140 30-5-22-24 30-5-66-68 30-5-99-101 30-5-113-115 30-6-007-009 30-6-42-44 30-6-82-84 30-6-120-122 31-1-15-17 31 -1-47-49 31 - 1-91-93 31 - 1- 120-122 31-2-19-21 31-2-65~57 31-2-105-107 31-3-39-41 31-3-91-93 31-3-127-129 31-4-006-008 31-4-42--44 31-4-75-77 31-4-134-136 31-5-17-19 31-5-48-50 31-5-88-90 31-5-141-143 32-1-17-19 32-1-47--49 32-1-87-89 32-1-118-120 32-2-51-53 32-2-81-83 32-2-131-133 32-3-011-013 32-3-47-49 32-3-88-90 32-3-132-134 32-4-008-010 32-4-49-51 32-4-88-90 32-4-137-139 32-5-007-009 32-5-14-16 32-5-61-63 32-5-95-97 32-5-139-141 32-6-033-157 32-6-41--43 32-6-61-63 32-6-106-108 32-6-145-147 33-1-003-005 33-1-30-32

365.72 366.13 366.60 366.92 367.29 367.64 368,03 368.37 368.90 369.25 369.50 369.89 371.00 371.03 371.39 371.73 372.17 372.50 372.64 373.08 373.43 373.83 374.21 375.16 375.48 375.92 376.21 376.70 377.17 377.56 378.40 378.92 379,28 379.57 379.93 380.26 380.85 381.18 381.49 381.89 382.42 384.68 384.98 385.49 385.69 386.52 386.82 387.32 387.62 387.98 388.49 388.83 389.09 389.50 389.89 390.38 390.59 390.65 391.12 391.46 391.90 392.11 392.42 392.62 393.07 393.46 393.95 394.22

2.20 2.96 2.99 2.93 2.62 3.24 2.98 3.50 2.91 3.15 3.37

-1.95 -1.84 -1.76 -2.07 - 1.97 -2.27 -1.94 -2.11 -2.07

2.57 3.41 2.71

-1.80 -2.32 -1.83

3.10 2.94

-2.32 -1.96

3.05

-2.09

3.00 3.35 2.94 3.19 3.35 2.74

-2.36 -1.82 -1.64

3.68 3.13 3.36

2.86

813C

8180

S. patagonica

-1.48

0.78 0.73

-1.05 -1.18

0.56 0.65 0.73 0.84 0.77 0.58

-1.19 -1.43 -1.44 -1.18 -1.24 -1.68

-2.14

0.72

-1.52

-2.27

0.87 0.80 0.65 0.91 1.08 0,91

-1.6O -1.84 -1.59 -1.48 -1.37 -1.22

-2.25 -2.18 -1.98

1.28 1.56 1.30

-1.36 -1.59 -1.44

-1.91

1.24 1.01

-1.33 -1.28

1.22 0.99

-1.41 -1.12

1.05

-1.56

1.01

-1.41

1.30 1.43 1.25 1.29

-1.12 -1.37 -1.37 -1.24

1.53

-1.59

1.40 1.33 1.34 1.26

-1.56 -1.51 -1.63 -1.50

3.0l 2.77 3.13

-1.97 -l.88 -2.01

3.20

-2.09

3.35 3.51 3.06 3.35

- 1.93 -1.91 -2.41 -2.13

1.35

-1.91

3.55 1.14 3.13 3.40 4.04 3.64

-2.31 -1.56 -1.99 -2.00 -2.55 -2.13

Table 3. Continued Core-section-cm

Depth (mbsf)

ill3C

fi180

M. subbotinae 33-1-54-56 33-1-70-72 33-1-92-94 33-1-109-111 33-1-130-132 33-1-148-150 33-2-010-012 33-2-32-34 33-2-51-53 33-2-70-72 33-2-92-94 33-2-117-119 33-2-139-141 33-3-008-010 33-3-29-31 33-3-50-52 33-3-74-76 33-3-89-91 33-3-109-111 33-3-129-131 33-3-147-149 33-4-010--012 33-4-31-33 33-4-55-57 33-4-80-82 33-4-101-103 33-4-118-120 33-5-002~004 33-5-27-29 33-5-50-52 33-5-70-72 33-5-110-112 34-1-13-15 34-1-54-56 34-1-87-89 34-1-146-148 34-2-004-006 34-2-31-33 34-2-50-52 34-2-71-73 34-2-92-94 34-2-103-105 34-2-114-116 34-2-125-127 34-2-142-144 34-3-003-005 34-3-19-21 34-3-27-29 34-3-47--49 34-3-62-64 34-3-70-72 34-3-83-85 34-3-97-99 34-3-100-102 34-3-109-111 34-3-125-127 34-4-004-006 34-4-14-16 34-4-23-25 34-4-29-31 34-4-33-35 34-4-38-40 34-4-45-47 34-4-54-56 34-4-62-64 33-4-76-78 34-4-89-91 34-4-101-103 34-4-119-121 34-4-138-140

394.46 394.62 394.84 395.01 395.31 395.49 395.61 395.83 396.02 396.21 396.43 396.68 396.90 397.09 397.30 397.51 397.75 397.90 398.10 398.30 398.48 398.61 398.82 399.06 399.41 399.52 399.69 400.03 400.28 400.53 400.71 401.11 403.64 404.05 404.38 404.97 405.05 405.32 405.51 405.72 405.93 406.04 406.15 406.26 406.43 406.54 406.70 406.78 406.98 407.13 407.21 407.34 407.48 407.51 407.60 407.76 408.05 408.15 408.24 408.30 408.34 408.39 408.46 408.55 408.63 408.77 408.90 409.02 409.20 409.39

613C

6180

S. patagonica 1.36 1.24 1.44

-1.56 -1.50 -1.74

1.29 1.17

-1.49 -1.43

1.46

-1.88

0.95 1.04

-1.72 -1.67

-2.26 -2.35 -2.31 -2.20 -2.44 -2.58 -2.39 -2.29

1.40

-1.47

1.43

-1.79

1.38 1.31 1.21 1.20

-1.55 -1.64 -1.66 -1.84

4.05 3.84

-2.28 -2.43

1.24

-1.46

3.75

-2.04

3.88 3.84 3.90 3.98 3.85

-1.95 -2.01 -2.48 -2.12 -2.35

1.31 1.08 1.50 1.50 1.31 1.12 1.06

-1.44 -1.84 -1.55 -1.55 -1.64 -1.68 -1.54

3.81

-2.35

3.80 3.96

-2.05 -2.28

1.18 1.32

-1.25 -1.71

3.80 3.83 3.65 3.80 4.04 3.53 3.96 3.86 3.43 3.68 3.32

-2.57 -2.21 -2.06 -1.86 -2.23 -1.98 -~2.29 -2.44 -2.23 -1.94

4.04 4.10 3.39 3.36 3.21 3.30 3.23

-2.17 -2.30 -2.49 -2.45 -2.49 -2.52 -2.34

3.00 3.00 3.11 3.11 3.10 2.54 2.18

-2.16 -2.95 -2.39 -2.69 -2.54 -2.31 -2.52

0.54 0.52

-1.01 -1.94

3.38 3.51 3.37 3.93 3.87

-2.09 -2.20 -2.11 -2.48 -2.53

3.31 2.76 3.70 3.59 3.25 3.14 3.60

-2.17 -2.07 -2.10 -2.23 -2.34 -2.45 -2.35

3.47 3.94 3.88 3.82 3.95 3.57 3.77 4.01

388

L . D . STOTT ET AL.

Table 4. Carbon and oxygen isotopic results of planktonic foraminifera from ODP Site 690B (Kennett & Stott 1991; isotopic results presented in per mil notation relative to the PDB standard)

Core-section-cm

Depth (mbsf)

fi13C

~180

S. patagonica 19-1-15-19 19-1-27-30 19-1-45--49 19-1-55-59 19-1-64-68 19-1-73-77 19-1-86-88 19-1-109-112 19-1-135-139 19-I-147-150 19-2-6-9 19-2-15-19 19-2-45-49 19-2-58-62 19-2-73-77 19-2-86-89 l 9-2-94-99 19-2-109-112 19-2-115-119 19-2-125-129 19-2-137-141 19-2-145-149 19-3-4-5 19-3-15-18 19-3-26-30 19-3-45-47 19-3-55-58 19-3-64-68 19-3-72-76 19-3-86-90 19-3-94-97 19-3-109-112 19-3-115-119 19-3-127-130 19-3-136-140 19-3-148-150 19-4-5-9 19-4-15-19 19-4-26-28 19-4-45-49 19-4-52-55 19-4-65-68 19-4-74-77 19-4-85-89 19-4-93-99 19-4-109-112 19-5-5-9 19-5-15-19 19-5-26-28 19-5-44--47 19-5-56-59 19-5-64-68 19-5-73-77 19-5-86-89 19-5-94-96 19-5-109-112 20-2-110-112 23-6-36--40

167.05 167.17 167.35 167.45 167.54 167.63 167.76 167.99 168.25 168.37 168.46 168.55 168.85 168.98 169.13 169.26 169.34 169.49 169.55 169.65 169.77 169.85 169.94 170.05 170.16 170.35 170.45 170.54 170.62 170.76 170.84 170.99 171.05 171.17 171.26 171.38 171.45 171.55 171.66 171.85 171.92 172.05 172.14 172.25 172.33 172.49 172.95 173.05 173.16 173.34 173.46 173.54 173.63 173.76 173.84 173.99 176.8 199.06

-0.704 -0.852 -0.543 -0.818 -1.16 -1.191 -0.95 -1.004 -1.208 -1.465 0.117 0.173 0.299

0.360 0.037 -0.083 -0.118 -0.657 -0.757 -0.697 -0.356 -0.597 -0.451 -0.420 1.578 1.725 1.706

0.331 -0.01 0.001 0.27 -0.064 0.15 0.241 0.196 0.052 0.068 0.302 0.145 0.121

1.376 1.296 1.399 1.136 1.616 1.619 1.552 1.600 1.830 1.556 1.604 1.601 1.510

0.081 0.264 0.062 0.224 0.273 0.177 0.095 0.149 0.06

1.674 1.689 1.699 1.764 1.608 1.573 1.806 1.771 1.766

0.215 0.300

2.280 3.000

~13C

~180

M. praepentcamerata

~13C

~180

N. truempyi -0.2 -0.i12 -0.316 -0.161 -0.235 -0.308 -0.304 -0.36 0.026 -0.414 -0.315 -0.124 -0.464 -0.37 -0.052 0.582 -0.439 -0.83

0.561 0.626 0.470 0.533 0.551 0.409 0.411 0.330 0.518 0.246 0.318 0.337 0.114 0.100 -0.198 0.052 -0.470

-0.989 -1.063 -1.161 -1.316 -1.399

-0.853 -0.902 -0.999 -1.163 -1.150

1.565 1.558 1.453 0.930

-1.407

1.710

-1.657 -1.88

1.131 -0.t54

-2.159 -1.831 -1.667 -l.065 -1.11 -1.109 -0.26 -0.156

1.072 0.477 0.098 -1.052 -0.706 0.889 3.179 3.402

-0.600

3.476

0.151 0.270 0.078 -0.130

-0.581

2.995

0.262

1.230

-0.872 -0.751 -0.68 -0.329 -0.400 -0.333 -0.218 0.079 -0.642 -0.583

3.004 2.919 3.249 3.214 3.093 3.069 3.729 3.044 3.154 3.583

0.018 -0.378 0.043 0.206 0.073

1.134 0.803 1.580 1.621

0.066

1.573

-0.449 -0.574 -0.524 -0.285 -0.442 -0.581 -0.612 --0.345 -0.538

3.446 3.176 3.661 3.623 3.201 3.720 3.318 3.978 3.687

-0.100 -0.026 0.188

1.200 1.439 1.724

0.338 0.226 -0.034 0.105 0.086 0.018 -0.220

1.717 1.599 1.260 1.424 1.611 1.529 1.370

GLOBAL PALEOCENE--EOCENE CARBON ISOTOPE CHANGES and 549 are important because the low-latitude planktonic foraminiferal and calcareous nannofossil zonations can be recognized (Mtiller 1981; Snyder & Waters 1981). Furthermore, the sites are characterized by relatively high sedimentation rates across the Paleocene-Eocene boundary which affords an opportunity to sample at higher resolution to resolve more precisely correlations between isotopic and biotic changes. Site 550 also contains a series of ash layers, components of which have been found in the basal part of the London Clay Formation (now Harwich Formation: Ellison et al. 1992) of the London Basin (Knox 1984; Berggren & Aubry 1996). Knox (1984) showed that a series of over 40 ashes (correlated with the Phase 2 ashes of the North Sea area) are restricted to Zone NP10. The well characterized -17 and +19 ash layers that are recognized in the London Basin and the North Sea region also occur in Site 550. Both ash layers occur within the lower third of Zone NP10 and the lower half of the C24R, providing additional stratigraphic constraints on terrestrial-marine correlations across the Paleocene-Eocene boundary. The carbon isotopic stratigraphies for DSDP Sites 549 and 550 (Fig. 3) are combined here with the isotopic record from ODP Site 690 (Kennett & Stott 1991) following the stratigraphic interpretation for the upper Paleocene and lower Eocene by Aubry et al. (1996). These records are used to develop a composite carbon isotope record across the Paleocene-Eocene boundary and to estimate carbon isotope changes in terrestrial carbon reservoirs. We have chosen to use the planktonic foraminifera Subbotina patagonica for isotopic correlations. This species is inferred to be a deepdwelling planktonic foraminifera based upon its carbon and oxygen isotopic composition relative to other planktonic species (Stott et al. 1990; Corfield 1993). The reasons we chose this species to use in correlations are listed below.

Surface-productivity effect. The isotopic composition of surface-dwelling planktonic foraminifera varies in the ocean in response to differing productivity levels. This is because lZc is preferentially removed from surface waters during photosynthesis, leaving surface waters enriched in 13C. Consequently, the isotopic composition of surface waters in the modern ocean can vary by as much as 1.0 (Kroopnick 1985). These differences are recorded by surface-dwelling foraminifera. Such differences may have nothing to do with the ocean-wide changes that were occuring in the Paleocene-Eocene transition. The 12C that is extracted in surface waters is retumed to the water column by oxidation of

389

organic matter. In the modem ocean much of this return occurs in the upper 100-200 m. Using a deep-dwelling species such as S. patagonica that inhabited the these subsurface waters mitigates much of the surface water productivity effect. This is illustrated in Fig. 3 which shows the vertical profiles of dissolved inorganic carbon (DIC) in the modern ocean at GEOSEC stations in each of oceans. The inferred depth habitat of S. patagonica is indicated. The depth range occupied by S. patagonica in the ancient ocean is inferred to be where the isotopic compositon of the water column is the most similar between sites. This is the base of the nutricline where nutrients and carbon extracted from the surface waters by photosynthesis have largely been returned to the water column. We emphasize that this argument for use of S. patatgonica in isotopic correlations does not imply that the isotopic composition of S. patagonica will be exactly the same from site to site. It only means that we expect the pattern of isotopic variability across the Paleocene-Eocene boundary interval as recorded by S. patagonica will be more similar between sites than isotopic records of a surface-dwelling species. We also recognize that comparison of planktonic isotopic records must also take into account the possibility that upwelling influenced the isotopic composition of deep and shallow-dwelling species. We would not expect to find similar isotopic records between sites if one or more sites was influenced by upwelling since this results in relatively lower ~513C values in the DIC pool.

Biological effect. A number of modem surfacedwelling planktonic foraminifera contain photosymbionts which can affect the isotopic composition of the foraminiferal calcite (e.g. Spero et al. 1991). The deep-dwelling species do not contain photosymbionts. Depth effect. Subbotina patagonica is a ubiquitous species in the early Paleogene oceans. It has been observed in all of the oceans and occurs in marginal settings up to at least bathyal depths. Benthic foraminiferal species do not occur everywhere in the ocean. Nuttallides truempyi, for example, inhabited a wide range of depths but was restricted to the deep ocean. Even though the change in isotopic composition of deep waters is relatively small over a wide depth range a comparison of different sites may involve very different water depths and, perhaps, different water masses. The isotopic evidence from Maud Rise (Kennett & Stott 1990) suggests that even at closely adjacent sites there may be signficant depth-dependent changes in the water mass chemistry that can be attributed to different water masses.

390

L . D . STOTT ET AL.

GEOSEC 6 CPD8 -0.50 0

200

400

0.50

0.0

1.0

1.5

2.0

2.5

I-

i

N E

Inferred Depth of S. patagonica

600 depth(m) 800

1000

GEOSEC Station 1200

1400

[ ] 430 Antarctic --~--431 Antarctic --[~--448 N. Indian --~--441 S, eq. Indian ~251 S.W. eq. Pacific [~ 37 N.W. eq. Atlantic --0--26 N. Atlantic

1600

Fig. 3. Carbon isotopic composition of dissolved inorganic carbon in the modem ocean as measured at GEOSEC stations in each ocean basin. The inferred depth habitat of S. patagonica with respect to the modem 513C distribution in the water column is shown on the left. Note the differences between sites near the surface. This reflects the variable influences of productivity and thermodynamic exchange effects. Note how each of the sites, including those in the Antarctic are similar around 100 m water depth but diverge below. The divergence below reflects the 'aging effect'.

Basin-basin effect.

Using a deep-dwelling planktonic species such as S. patagonica mimimizes the water mass 'age' effect associated with benthic foraminifera. There is as much as 1.0 difference in the isotopic composition of deep waters between ocean basins today. This reflects the progressive incorporation of 12C into the dissolved CO 2 pool during its transit through the ocean basins (Kroopnick 1985). There may have have been significant changes in deep water circulation during the latest Paleocene and earliest Eocene that involved changes in sources of deep water (Kennett & Stott 1990; Pak & Miller 1992). Consequently, the isotopic compositon of deep waters at a particular site may have changed. This may have produced contrasting patterns of isotopic change between deep sea sites, depending on their proximity to the old and to the new deep water source. We believe that by using a deep-dwelling plank-

tonic foraminiferal species for carbon isotopic correlation we have minimized the large isotopic variability that might be expected between benthic foraminiferal records from sites located at different depths and in different ocean basins. We have also avoided using planktonic species for correlation that inhabited the surface waters in order to minimize the effects of variable productivity. However, as discussed below we have utilized surfacedwelling species to reconstruct atmospheric ~13C values. The reason for their use and the potential errors that this method introduces are discussed below. In Figs 4 and 5 the carbon isotope stratigraphies are shown for each of the sites. We have labelled portions of the isotope records in each site to facilitate the following discussion (Fig. 5). The carbon isotope stratigraphies for each of the sites look similar. Each contains a portion of a large negative excursion near the base of the record

GLOBAL PALEOCENE--EOCENE CARBON ISOTOPE CHANGES

.

.

.

.

i

.

.

.

.

r

9

391

9

i:i :::

'ill "~'

.~.~

"0"

:i:

':

,9: - ,

/% ,~

;:: 9 .

.~'

-~ ~9

s

",' 9

o

.

~

.~

.:. .., :::

~== (tsquj) q ldeQ

~

,

i!!

~ =

:i:

~: ~ i

:.':,

, II,

~

o

~

I!SSO t o u u ~ N snoeJ~OleO

..~

eJej!U!LUeJOd o!uo ~plUeld fJeAooet:l eJoo

~

~

'

'

"

'

"

,

'

..

9 i"

,

.

.

.

.

9

: : : : : ~ :

,

iiil/~

~

,

i!.~iiii

"

~

:it i=!: 9~..1~..

I

:1. 9 .~_.

"'~

9

~

~ .~

~ . ~

~

~ ' ~

1~

~.

:

:':':':

,

O

,

~ !:!.~:i ,

c,~

9

:::~:: -~

r~

,.

0

~

-I.:.: ....

.~

~.::::~

,~

,:,:. 9 9 ,,a.d

I

-,

i:i::,~ ,

.:.:.~

I

. . ,

, , , ,

~,

,I,

: . " 9"',"

,

,

,

,

,

,

:i.:.~

{~sqw) q IdeQ

l!$SOlOUUeN

I

Sno~J~31~

0 LdN

e~ej!u![ueJod o!uo ~,)lUeld

qgcl

I i ~

6dN Sd

0-

I

~t_~

~o^ooet:l

.~=

eJoo

-9

..~

. . ~

392

L. D . S T O T T ET AL.

eJe)!u!~eJo~ o!uo~u~d

leO!doJi

l!sso~ouue N snoeJeoeO

J~l

~ I =~

n

n

J

0 LdN

I

....., tI,D

,.: ....

I..

I

in

o u~

Q.

9 CO 13

I

9 ~t ~ ,

I

~

,

I

,

,

,

I

un

oo T--

'

'

'

9

'

I

9

I

~j~l

m 13

o

~

-

go

d

-

~

0

0

o

-

o 9

E

A ~

~o ._.

~o

-

-

"

~

.,,..-

..,-

"

"-

I!SSO~OUUeN snoeJeoleO eJej!u!uueJoB o~uol~ue~ o~;oJe~uv

0

I OLaN 1

P~IN

]

~/~sx/%~%~%/%/%/%s%s% /~/xsx/x/~/x/x/~/~sx

r

d ~o

.

.

.

.

.

.

s

s / / / / / p s z ~ s i%~1%~/~r162

.

i

,

u')

o0

...... 0,I CO

IrSSOIOUU~N s n o e J e O l ~O eJe;!u!weJ0-i o!uo]~lUeld

leo!doJj.

,lii!l......

l~....

'~"

CO

0LdN q9d

i I I

CO

6dN ~d

~

i.

.cJ 0

GLOBAL PALEOCENE-EOCENE CARBON ISOTOPE CHANGES

(label A) and an interval immediately above this where the values remain essentially constant (label B in Sites 549 and 690 and label D in Site 550). Each record also contains a decrease in carbon isotope values near the top of the analysed interval (label C in Site 690 and label E in Site 550). However, when the calcareous nannofossil biostratigraphies are included in this comparison it is clear that carbon isotope changes are not of equivalent age. The intervals labelled A to E represent different portions of the isotope stratigraphy of the Paleocene-Eocene transition. At sites 549 and 690, the distinct carbon isotope excursion (A) associated with the benthic foraminiferal excursion occurs within calcareous nannofossil Zone NP9. At Site 550, however, the excursion (A) and extinction occur within a zone of strong dissolution (Fig. 4). The calcareous nannofossil biostratigraphy indicates a hiatus at this level, marking the NP9-NPI0 zonal boundary (Aubry et al. 1996). Consequently, the interval immediately above the excursion in Holes 549 and 690B, which is characterized by essentially constant values of about 1.2%o (label B) is not the same interval above the excursion in Site 550 (label D) (Fig. 5). At Site 550, interval (D) is younger, occurring within Zone NP10. The hiatus at 408 m in Site 550 encompasses the upper portion of Zone NP9 and the lower portion of Zone NP10. At Site 690, the NP9-NP10 zonal boundary marks the base of a distinct negative shift in carbon isotope values (label C). At the base of this shift at 147.5 m the carbon isotope values of Subbotina patagonica are between 1.2 and 1.4 (Fig. 5). The values become progressively lower, reaching values of about 0.6 at 137.8 m where the calcareous nannofossil biostratigraphy indicates the presence of another hiatus (Fig. 5). At Site 550 an isotopic shift of similar magnitude and direction is observed between 380 m and 370 m (Aubry et al. 1996). Here too the values shift from between 1.2 to 1.4 to values of between 0.6 and 0.8 (Fig. 5). However, the calcareous nannofossil biostratigraphy indicates that this portion of Site 550 is within the upper Zone NP10 rather than the lower Zone NP10 (Fig. 5). At Site 549 a similar shift occurs between 336 m and 335 m. At this site a hiatus occurs at 335 m coincident with the NP9-NP10 zonal boundary. Therefore, the values above this hiatus appear to represent values typical of the upper Zone NP10 as seen at Site 550 between 380 m and 370 m (Fig. 5). A schematic representation of the carbon isotope stratigraphy across the Paleocene-Eocene boundary is shown in Fig. 6. On the basis of results presented here the Paleocene-Eocene boundary transition can be characterized by a series of distinct carbon isotope changes. There is a distinct

393

negative excursion that occurs within NP9 that is associated with the benthic foraminifera extinction. There is a second negative excursion that occurs in lower NP10. The upper portion of this second excursion is not represented in the records presented here because of a hiatus. We infer a smooth and linear trend between the most negative point in the excursion below the hiatus and the values directly above the hiatus. Each of these two excursions is superimposed upon the longer-term shift in carbon isotope values that encompassed much of the late Paleocene and early Eocene (Fig. 1). This is a different picture of the carbon isotope stratigraphy across the Paleocene-Eocene boundary from that previously presented, wherein only one major isotopic excursion had been well documented. Although several previous studies have indicated the presence of more than one isotopic excursion (Bah'era & Huber 1991; Seto et al. 1991; Lu & Keller 1993), the records used to describe them were too incomplete to derive conclusive evidence that they represent the same isotopic changes as described here. A picture of multiple isotope events or changes becomes apparent only when a relatively high resolution carbon isotope record from several sites is constrained with detailed biostratigraphy. It is important to recognize that although the duration of each of these isotopic changes differs, if a site contains hiatuses or changes in sedimentation rate across one of these isotopic changes it is possible to have similar appearing isotopic curves from stratigraphically different intervals.

The exchange of 12C between the ocean and terrestrial carbon reservoirs Ocean-atmosphere 13C02 exchange The large carbon isotope changes in the oceans across the Paleocene-Eocene boundary must have involved the exchange of 12C and 13C between the larger marine and the smaller terrestrial and atmospheric carbon reservoirs. This is because the ocean and atmosphere CO 2 reservoirs tend to maintain isotopic equilibrium on short time scales (mixing time of the ocean). The fractionation between the ocean and atmospheric CO 2 reservoirs is about -9.0. This means that if the surface ocean ZCO 2 ~13C is +2.0, the equivalent atmospheric CO 2 ~13C should be close to -7.0 (Fig. 7). Therefore, the large negative carbon isotopic changes observed in marine carbonates near the Paleocene-Eocene boundary would have been transmitted to terrestrial carbon reservoirs that were in contact with the atmosphere at that time. Koch et al. (1992) were the first to apply this reasoning to

394

L.D. STOTT ETAL.

u'~,

O~

O

9

z

I1

0,1

u"=-

~LL

AF~3b

> 0

~

~

.'=--

0.6

I'=

13 C

0.8

1

I

I

PDB

1.2 I

1.4 I

1.6 I

I:~b P6a NP10

AP6a P5

NP9

AP5 AP4

Fig. 6. Schematic composite curve of carbon isotopic changes across the Paleocene-Eocene boundary recorded by the plankonic foraminifer Subbotina patagonica. This species is plotted because it occurs in all of the sites. It is inferred to have been a deep dweller based upon its oxygen and carbon isotopic values. The portion of interval C indicated with ? is inferred, since this does not appear in any of the three records described here. This inference simply connects the two adjacent portions of the curve.

carbon isotopic studies of the terrestrial sequences of the Big Horn Basin, Wyoming. They showed how the long-term change in the isotopic composition of the ocean was also recorded by soil carbonates and mammal tooth enamel and how these changes could be used to correlate marine and terrestrial sections. Terrestrial plants, and freshwater carbonates are other potential carbon reservoirs that could also preserve the large isotopic signals across the Paleocene-Eocene boundary since their carbon is derived directly or indirectly from atmospheric CO 2. A conceptual model of carbon isotopic stratigraphy for plant-derived organic carbon across the Paleocene-Eocene boundary, similar to that developed by Koch et al. (1992) is developed here. In our model the available carbon isotope records from marine sections at DSDP Site 550, 549 and ODP Site 690B that have been correlated with calcareous nannofossil biostratigraphy and

magnetostratigraphy (Aubry et al. 1996) are used to estimate what the 513C of atmospheric CO 2 would have been across the Paleocene-Eocene boundary (calcarous nannofossil Zones NP9 to NPll). The carbon isotopic composition of the palaeoatmosphere and terrestrial plant organic carbon is estimated using the ~513C of surface-dwelling planktonic foraminiferal calcite and assuming a fractionation of lZc by the C3 photosynthetic pathway of-19.0%o -2.0%~ (Bender 1971). The magnitude of the photosynthetic fractionation differs between plants that utilize other photosynthetic pathways (i.e. C4 and CAM). However, during the Paleocene and early Eocene the floral communities did not contain elements with these other photosynthetic pathways (C4 and CAM). Therefore, the model need not contend with possible mixtures of C4 and C3 plants. Furthermore, Ceding (1992) showed that the isotopic composition of organic carbon buried in soils

GLOBAL PALEOCENE-EOCENE CARBON ISOTOPE CHANGES

Atmospheric CO2 (513C -- -7

-1

1 i

t

3 BIOMASS 813C = -

26

' A = -9

Surface MarineCarbonate ba3C = + 2

f Soilorganicmatter l

3C = -26

~"~+14.5 Soil Carbonate 813c = - n . s Fig. 7. Conceptual model of carbon isotope fractionations between marine, atmosphere and continental carbon reservoirs. Absolute values for each reservoir represent modern values characteristic of C3 ecosystems. Note that atmospheric CO 2 in equilibrium with the oceanic carbon reservoir is depleted in 13C by -9. A 4 change in the ocean as occurred at the end of the Paleocene would be transferred to the soil carbonate via the atmosphere and vegetation.

retains the isotopic composition of the original plant community from which the carbon was derived. The plant-derived organic carbon buried in the pedogenic sediments of the Paris Basin should only record isotopic variability related to changes in the isotopic composition of the atmosphere, with a predictable offset due to C3 photosynthesis. Even if the fractionation of 13C by C3 plants was, say, -21 rather than -19.0, we would still expect to see the temporal pattern of isotopic change from the mid Paleocene to early Eocene.

A terrestrial chemostratigraphic model With the considerations outlined above a conceptual model for ~13C changes in terrestrial C3 biomass in the latest Paleocene is shown in Fig. 7. This model uses the isotopic composition of the planktonic foraminifer Morozovella subbotinae from DSDP Sites 550 and 549 and Acarinina praepentacamerata from ODP Site 690 to estimate the 813C of atmospheric CO2. These species were chosen because the carbon and oxygen isotopic compositions of their shells suggest they inhabited near-surface waters. We assume that calcite precipitated in isotopic equilibrium with ECO 2 in the

395

surface ocean and that the surface ocean and the atmosphere were in approximate isotopic equilibrium. Additionally, we assume a constant isotopic fractionation between the surface ocean, the atmosphere and soil CO 2. This latter assumption is reasonable since changes in temperature and other hydrologic changes would have a relatively minor influence on the fractionation of 13C between CO 2 phases relative to the size of the isotopic shifts between the Paleocene and Eocene (Mook 1974). Perhaps the largest uncertainty in our model is the assumed constant -19 fractionation of 13C by C3 plants. The fractionation factor may have varied by one or two per mil, an effect that will produce an offset between the model and actual values. The effect of changing pCO 2 of the metabolic fractionation of 13C by plants is not well constrained. However, such changes should be small at pCO 2 levels at or above present day levels (Bender 1971). Furthermore, any small changes in fractionation would be superimposed on the longterm temporal changes associated with changes in the isotopic composition of the ocean. By analysing a large number of samples we hope to average-out shorter-term variations. A sustained change in the fractionation factor would, however, be seen as a systematic offset between the model values and the measured values. The temporal pattern of 813C change during the Paleocene-Eocene transition would still be evident. Another source of uncertainty in our model is the use of planktonic foraminiferal calcite from DSDP Sites 550 and 549 to reconstruct equilibrium isotopic values of dissolved CO 2 in the surface ocean. Because the planktonic foraminifera used here to estimate the isotopic composition of surface water are extinct we cannot be sure to what extent their calcite records isotopic equilibrium values. We recognize that many modern foraminifera display offsets between predicted equilibrium isotopic compositions and measured values. This is typically a relatively small offset in many modern species. We assume that the any offset from isotopic equilibrium among the Paleocene and Eocene species is similarly small and that the error associated with this assumption will produce a small and nearly constant offset between our model predicted values and actual values. The model predicts that the carbon isotopic composition of C3 plant organic carbon should have been approximately -24.0 in the latest Paleocene and become progressively more negative across the Paleocene-Eocene boundary, reaching values of (c. -25.0 to -26.0) in the earliest Eocene (Fig. 7). The isotopic excursions that occur within Zones NP9 and NP10 would be superimposed upon this long-term trend.

396

L . D . STOTT E T A L .

'"~

o

"F~

oS

N ~=N -~ ~ ' ~

N~o = ~ o 9

o

GLOBAL PALEOCENE-EOCENE CARBON ISOTOPE CHANGES

Preliminary carbon isotopic results from the Paris Basin

397

Conclusion

The initial results of our isotopic measurements of organic carbon from the Limay section is shown in Fig. 8. A comparison of the carbon isotope results from the Limay section matches the predicted pattern of isotopic change across the PaleoceneEocene boundary. All of the data falls within the uncertainty of the model predicted values across the Paleocene-Eocene boundary. In fact, the distinct carbon isotope excursion near the PaleoceneEocene boundary in the marine records is clearly evident in the Limay section (Fig. 8). Based upon a visual comparison between the model curve and Limay results we believe this portion of the Limay section encompasses the latest Paleocene (approximately equivalent to middle to upper NP9 Zone). We expect that our ongoing work in this basin will more fully resolve the isotopic record through the Sparnacian and allow a more definitive age determination. These results are compelling support for the use of carbon isotope stratigraphy as a unifying proxy for locating the Paleocene-Eocene boundary in terrestrial and marine sections. We believe that the position of the 813C excursion may provide the best criterion for pin-pointing the epoch boundary in non-fossiliferous sections. These results are therefore important to the goals of the IGCP 308 which is attempting to identify a standard section for the Paleocene-Eocene boundary and a criterion by which stratigraphers can correlate to it.

A detailed isotopic record from marine biogenic carbonates has been developed across the Paleocene-Eocene boundary. These marine records exhibit several distinct isotopic patterns, including the large negative excursion within Zone NP9 and P5, that can be recognized globally. The systematic isotopic patterns seen in ODP Site 690 and in DSDP Sites 550 and 549 serve as a reference for correlation to other sections that may not contain the requisite faunal or floral fossil elements needed for biostratigraphic correlation but can provide carbon isotope information. The isotopic changes recorded in the deep sea were transferred to the terrestrial carbon reservoirs via the atmosphere. A carbon isotope stratigraphy developed from terrestrial organic carbon extracted from the Limay section of the Paris Basin documents a distinct excursion and isotopic patterns that closely match changes in the marine isotopic record within Zone NP9. The recognition of these isotopic changes in the Limay section of the Paris Basin illustrates how carbon isotope stratigraphy can be used for marine-terrestrial correlations. Carbon isotope stratigraphy may provide the most appropriate criterion for locating the position of the Paleocene-Eocene epoch boundary. This work was supported by a grant from the NSF EAR9219093 to LDS. This paper represents a contribution to IGCP Projects 308 and 317. The comments and suggestions of D. Pak, K. Miller, E. Thomas, and J. Zachos are greatly appreciated. This is Woods Hole Oceanographic Contribution No. 8854 and ISEM Contribution No. 95048.

References AUBRY,M.-P. 1983. Biostratigraphie du Paldogbne dpicontinental de l'Europe du nord-ouest. Etude fond~e sur les nannofossiles calcaires. Documents des Laboratoires de Gdologie, Lyon, 89. 1985. Paleogene calcareous nannoplankton bitstratigraphy of northwestern Europe. Palaeogeography, Palaeoclimatology, Palaeoecology, 55, 267-334. --, HAILWOOD, E. A. & TOWNSEND, H. A. 1986. Magnetic and calcareous nannofossil stratigraphy of lower Paleogene formations of the Hampshire and London Basins. Journal of the Geological Society, London, 143, 729-735. ~, BERC~REN,W. A., STOTT,L. & SINHA,A. 1996. The upper Paleocene-lower Eocene stratigraphic record and the Paleocene-Eocene boundary carbon isotope excursion: implications for geochronology. This volume. BARRERA,E. & HUBER,B. T. 1991. Paleogene and early Neogene oceanography of the southern Indian Ocean: Leg 119 foraminifer stable isotope results.

In: BARRON, J. A., ~T AL. Proceedings of the

Ocean Drilling Program, Scientific Results, 119, 693-717. BENDER, M. M. 1971. Variation in the 13C/12C ratios of plants in relation to the pathway of photosynthetic carbon dioxide fixation. Phytochemistry, 10, 12391244. BERGGREN, W. A. & AUBRY, M.-P. 1996. A late Paleocene-early Eocene NW European and North Sea magnetobiochronological correlation network. This volume. CERHNG, T. E. 1992. Use of Carbon isotopes in paleosols as an indicator of the pCO2 of the paleoatmosphere. Global Biogeochemical Cycles, 6, 307-314. CORFIELD,R. M. 1993. Depth habitats and the Palaeocene radiation of the planktonic foraminifera monitored using oxygen and carbon isotopes. In: LEES, D. & EDWARDS, D. (eds) Evolutionary Patterns and Processes. Linnaean Society, London, 59-70. DOLLFUS, G. F. 1880. Essai sur l'extension des terrains tertiairs dans le Bassin Anglo-Parisien. Bulletin

398

L.D. STOTT ETAL.

de la Socigt~ G~ologique de Normandie, 6(1879), 584-605. ELLISON, R. A., KNox, R. W. O'B., JOLLEY, D. W. & KINr, C. 1992. A revision of the lithostratigraphical classification of the early Palaeogene strata in the London Basin and East Anglia. Proceedings of the Geologists' Association, 105, 187-197. HOOKER, J. J. 1991. The sequence of mammals in the Thanetian and Ypresian of the London and Belgium Basins: location of the Palaeocene-Eocene boundary. Newsletters on Stratigraphy, 25, 75-90. KENNEW, J. P. & STOT'r, L. D. 1990. Proteus and ProtoOceanus: ancestral Paleogene oceans as revealed from Antarctic stable isotopic results. In: BARKER, R E & KENNETT, J. P., ET AL. Proceedings of the Ocean Drilling Program, Scientific Results, 113, 865-880. &-1991. Terminal Paleocene deep-sea benthic crisis: sharp deep sea warming and paleoceanographic changes in Antarctica. Nature, 353, 225-229. & -1995. Terminal Paleocene mass extinction in the deep sea: association with global warming. In: STANLEY,S., KENNETT,J. R & KNOLL, A. (eds) The Effects of Global Change on Life. National Research Council, National Academy Press. KNox, R. W. O'B. 1984. Nannoplankton zonation and the Palaeocene/Eocene boundary beds of NW Europe: an indirect correlation by means of volcanic ash layers, Journal of the Geological Society, London, 141, 993-999. KOCH, P. L., ZACHOS, J. C. & GINGERICH, P. D. 1992. Coupled isotopic change in marine and continental carbon reservoirs near the Palaeocene/Eocene boundary. Nature, 358, 319-322. KROOPNICrr P. M. 1985. The distribution of C-13 of total CO 2 in the world oceans. Deep Sea Research, Part A, 32, 57-84. LAURAIN, M., Barta, L., Bolin, C., Guemier, C., GruasCavagnetto, C., Louis, P., Riveline, J. & Thiry, M. 1983. Le sondage et la coupe du Mont Bernon Epernay (Marne). Etude srdimentologique et palrontologique du stratotype du Sparnacien de la srrie 6oc~ne. G~ologie de la France, 3, 235-253. Lu, G. & KELLER, G. 1993. Climatic and oceanographic events across the Paleocene-Eocene transition in the Antarctic Indian Ocean: inference from planktic foraminifera. Marine Micropaleontology, 21, 101-142. MILLER, K. G., JANECEK,T. R., KATZ, M. E. & KEIL, D. J. 1987. Abyssal circulation and benthic foraminiferal changes near the Paleocene/Eocene boundary, Paleoceanography, 2, 741-761. MooK, W. G. 1974. Carbon isotope fractionation between dissolved bicarbonate and gaseous carbon dioxide. Earth and Planetary Science Letters, 22, 169176. Mt3LLEa, C. 1981. Biostratigraphic and paleoenvironmental interpretation of the Goban Spur region based upon a study of calcareous nannoplankton. In: GRACIANSKY,P. C. DE, POAG, C. W., ETAL. Initial Reports of the Deep Sea Drilling Project, 80, 573-597. -

-

-

-

PAK, D. K. & MILLER, K. G. 1992. Paleocene to Eocene benthic foraminiferal isotopes and assemblages: implications for deep-water circulation. Paleoceanography, 7, 405-422. REA, D. K., ZACHOS, J. C., OWEN, R. M. & GINGERICH, P. D. 1990. Global change at the Paleocene-Eocene boundary. Climatic and evolutionary consequences of tectonic events. Palaeogeography, Palaeoclimatology, Palaeoecology, 79, 117-128. SETO, K., NOMURA,R. & NIITSUMA,N. 1991. Data report: oxygen and carbon isotope records of the upper Maestrichtian to lower Eocene benthic foraminifers at Site 752 in the eastern Indian Ocean. In: PEmCE, J. W., WEtSSEL, J., Er AL. Proceedings of the Ocean Drilling Program, Scientific Results, 121, 885-889. SHACKLETON, N. J., HALL, M. A. & BOERSMA, A. 1984. Oxygen and carbon isotope data from Leg 74 foraminifers. In: Initial Reports of the Deep Sea Drilling Project, 74, 599-612. SIESSER, W., LORD, A. R. & WARD, D. 1987. Calcareous nannoplankton biozonation of the Thanetian Stage in the type area. Journal of Micropaleontology, 6, 85-102. SNYDER, S. W. t~ WATERS, V. J. 1981. Cenozoic planktonic foraminiferal biostratigraphy of the Goban Spur region, Deep Sea Drilling Project Leg 80. In: DE GRACIANSKV,P. C., POAG, C. W., ET AL. Initial Reports of the Deep Sea Drilling Project 80, 439-472. SPERO, H. J., LERCHE, I. & WILUAMS, D. E 1991. Opeing the carbon isotope "vital effect" black box. 2, Quantitative model for interpreting foraminiferal carbon isotope data. Paleoceanography, 6, 639-655. STOTT, L. D. 1992. Higher temperatures and lower oceanic pCO2: A climate enigma at the end of the Paleocene epoch. Paleoceanography, 7, 395-404. --, KENNETX, J. E, SHACKLETON, N. J. ~r CORFIELD, R. M. 1990. The evolution of Antarctic surface waters during the Paleogene: Inferences from the stable isotopic composition of planktonic foraminifers. In: BARKER,P. E, KENNETT,J. R, ETAL. Proceedings of the Ocean Drilling Program, Scientific Results, 113, 849-864 THOMAS, E. 1989. Development of Cenozoic deep-sea benthic foraminiferal faunas in Antarctic waters. In: CRAME, J. A. (ed.) Origins and Evolution of Antarctic Biota. Geological Society, London, Special Publication, 47, 283-296. 1990. Late Cretaceous-early Eocene mass extinctions in the deep sea. In: SHARPTON,V. L. & WARD, P. D. (eds) Global Catastrophes in Earth History Geological Society of America, Special Publication, 247, 481-495. & Shackleton, N. J. 1993. The Paleocene benthic foraminiferal extinction: timing, duration and association with stable isotope anomalies. (Abstract) Correlation of the early Paleogene in Northwest Europe. Geological Society, London. TJALSMA, R. C. & LOHMANN, G. P. 1983. PaleoceneEocene Bathyal and Abyssal Benthic Foramnifera from the Atlantic Ocean. Micropaleontology, Special Publication, 4. -

-

GLOBAL PALEOCENE-EOCENE CARBON ISOTOPE CHANGES ZACHOS, J. C., REA, D. K., SETO, K., NOMURA, R. & NIITSUMA, N. 1993. Paleogene and early Neogene deepwater paleoceanography of the Indian Ocean as determined from benthic foraminifera stable isotope

399

records. In: DUNCAN,R. A., Er AL. (eds) The Indian Ocean: A synthesis of results from the Ocean Drilling Program. Geophysical Monograph Series, American Geophysical Union, Washington, DC.

The Paleocene-Eocene benthic foraminiferal extinction and stable isotope anomalies E. T H O M A S

1' 2 & N. J. S H A C K L E T O N

3

1 Department of Earth Sciences, Downing Street, University of Cambridge, Cambridge CB2 3EQ, UK Present address: Center for the Study of Global Change, Department of Geology and Geophysics, PO Box 208109, Yale University, New Haven CT 06520-8109, USA 2 Department of Earth and Environmental Sciences, Wesleyan University, Middletown, CT 06459-1309, USA 3 Godwin Laboratory, Subdepartment of Quaternary Research, University of Cambridge, Cambridge CB2 3RS, UK Abstract: In the late Paleocene to early Eocene, deep sea benthic foraminifera suffered their only global extinction of the last 75 million years and diversity decreased worldwide by 30-50% in a few thousand years. At Maud Rise (Weddell Sea, Antarctica; Sites 689 and 690, palaeodepths 1100 m and 1900 m) and Walvis Ridge (Southeastern Atlantic, Sites 525 and 527, palaeodepths 1600 m and 3400 m) post-extinction faunas were low-diversity and high-dominance, but the dominant species differed by geographical location. At Maud Rise, post-extinction faunas were dominated by small, biserial and triserial species, while the large, thick-walled, long-lived deep sea species Nuttallidestruempyiwas absent. At Walvis Ridge, by contrast, they were dominated by long-lived species such as N. truempyi,with common to abundant small abyssaminid species. The faunal dominance patterns at the two locations thus suggest different post-extinction seafloor environments: increased flux of organic matter and possibly decreased oxygen levels at Maud Rise, decreased flux at Walvis Ridge. The species-richness remained very low for about 50 000 years, then gradually increased. The extinction was synchronous with a large, negative, short-term excursion of carbon and oxygen isotopes in planktonic and benthic foraminifera and bulk carbonate. The isotope excursions reached peak negative values in a few thousand years and values returned to preexcursion levels in about 50 000 years. The carbon isotope excursion was about -2%0 for benthic foraminifera at Walvis Ridge and Maud Rise, and about -4%0 for planktonic foraminifera at Maud Rise. At the latter sites vertical gradients thus decreased, possibly at least partially as a result of upwelling. The oxygen isotope excursion was about -1.5%o for benthic foraminifera at Walvis Ridge and Maud Rise, -1%o for planktonic foraminifera at Maud Rise. The rapid oxygen isotope excursion at a time when polar ice-sheets were absent or insignificant can be explained by an increase in temperature by 4-6~ of high latitude surface waters and deep waters world wide. The deep ocean temperature increase could have been caused by wanning of surface waters at high latitudes and continued formation of the deep waters at these locations, or by a switch from dominant formation of deep waters at high latitudes to formation at lower latitudes. Benthic foraminiferal post-extinction biogeographical patterns favour the latter explanation. The short-term carbon isotope excursion occurred in deep and surface waters, and in soil concretions and mammal teeth in the continental record. It is associated with increased CaCO 3dissolution over a wide depth range in the oceans, suggesting that a rapid transfer of isotopically light carbon from lithosphere or biosphere into the ocean-atmosphere system may have been involved. The rapidity of the initiation of the excursion (a few thousand years) and its short duration (50 000 years) suggest that such a transfer was probably not caused by changes in the ratio of organic carbon to carbonate deposition or erosion. Transfer of carbon from the terrestrial biosphere was probably not the cause, because it would require a much larger biosphere destruction than at the end of the Cretaceous, in conflict with the fossil record. It is difficult to explain the large shift by rapid emission into the atmosphere of volcanogenic CO2, although huge subaerial plateau basalt eruptions occurred at the time in the northern Atlantic. Probably a complex combination of processes and feedback was involved, including volcanogenic emission of CO 2, changing circulation patterns, changing productivity in the oceans and possibly on land, and changes in the relative size of the oceanic and atmospheric carbon reservoirs.

From Knox, R. W. O'B., Corfield, R. M. & Dunay, R. E. (eds), 1996, Correlationof the EarlyPaleogenein NorthwestEurope, Geological Society Special Publication No. 101, pp. 401--441.

401

402

E. THOMAS & N. J. SHACKLETON

During the late Paleocene and early Eocene important changes occurred in global climate, in plate tectonic processes and in the global carbon cycle. The discussion of these events has been complicated and confused by two facts: that there is no unequivocal definition of the 'Paleocene-Eocene boundary' because these words mean different things to different people (e.g. Berggren & Aubry 1996); and that different events occurred at different time scales, from millions of years to thousands of years. We will first discuss longer timescale events, occurring over several millions of years in palaeomagnetic chrons C25 and C24, at about 55-60Ma in the geomagnetic polarity time scale of Berggren et al. (1985), and 52.5-57.5 Ma in that of Cande & Kent (1995). The beginning of this interval has generally been placed in the Paleocene and the end in the Eocene, but the location of the boundary has varied. This transitional Paleocene-Eocene period witnessed a long-time warming of the deep oceans beginning in the earliest Paleocene, and a long-term decrease of carbon isotopic values in tests of benthic and planktonic foraminifera and bulk carbonate, beginning in early Chron C25 (Shackleton 1986, 1987; Zachos et al. 1993). Highlatitude areas experienced the highest temperatures of the Cenozoic as indicated by the presence of warm-water pelagic marine organisms (e.g. Haq et al. 1977; Premoli-Silva & Boersma 1984; Boersma et al. 1987; Stott & Kennett 1990; Aubry 1992; Berggren 1992; Ottens & Nederbragt 1992). Crocodiles, turtles and other thermophilic biota occurred at high northern latitudes (Estes & Hutchison 1980; McKenna 1980; Gingerich 1983; Markwick 1994) in the presence of vegetation and soil-types indicating warm climates (Kemp 1978; Nilsen & Kerr 1978; Wolfe 1978; Wolfe & Poore 1982; Schmidt 1991). Palynological data from the North Sea indicate peak warmth (Schroeder 1992). Clay mineral associations in oceanic sediments indicate high humidity and intense chemical weathering as indicated by high abundances of kaolinite (Robert & Maillot 1990; Robert & Chamley 1991; Robert & Kennett 1992, 1994); similar peaks in kaolinite abundance have been observed in sediments from the New Jersey margin (Gibson et al. 1993) and the North Sea region (Knox, written comm. 1993). Oxygen isotopic measurements show high temperatures and shallow temperature gradients from low to high latitudes (Shackleton & Boersma 1981; Oberh~insli & Hsti 1986; Stott et al. 1990; Barrera & Huber 1991; Seto et al. 1991; Stott 1992; Zachos et al. 1994; Bralower et al. 1995a, b). Dust concentrations in oceanic sediments reached very low levels in the upper part of Chron C24r (Janecek & Rea 1983;

Miller et al. 1987b; Rea et al. 1990; Hovan & Rea 1992; Rea 1994), suggesting very low wind strength. During this warm period terrestrial and shallowwater marine organisms did not suffer major net extinctions (e.g. Raup & Sepkoski 1986), but it was a time of major origination of species and high diversity on land (European mammals and flora: Hooker 1991; Collinson 1983; North American mammals: Butler et al. 1981, 1987; Wing 1984; Rea et al. 1990; Wing et al. 1991) and in the surface oceans (e.g. planktonic foraminifera: Kennett 1978; Boersma et al. 1987; Boersma & Premoli-Silva 1991; Corfield & Shackleton 1988; Berggren 1992; Corfield 1993; calcareous nannofossils: Romein 1979; Aubry 1992; dinoflagellates, Oberh~nsli & Hsfi 1986; McGowran 1991). An explanation of the several million year-long warm period has commonly been sought in elevated levels of atmospheric CO 2, caused by plate tectonic related processes (e.g. Williams 1986; McGowran 1991; Rea et al. 1990). During Chrons C24-C25 there was a worldwide plate-tectonic reorganization, involving the slow-down of the northward motion of the Indian subcontinent because of its collision with Asia (Klootwijk et al. 1991; Beck et al. 1995a). According to some researchers high-temperature metamorphism started in the Himalayas (Tonarini et al. 1993; Smith et al. 1994; Beck et al. 1995a) and delivered large amounts of CO 2 to the atmosphere by decarbonation (Touret 1992; Kerrick & Caldeira 1993, 1994), but this mechanism of CO 2 delivery has been doubted (Selverstone & Gutzler 1993). Increasing levels of CO 2 in the atmosphere resulting from the India-Asia collision could alternatively have been generated by erosion of sediments rich in organic matter (Beck et al. 1995b). In addition, subduction changed in direction in the North Pacific (Goldfarb et al. 1991). Continental break-up started in the North Atlantic (Roberts et al. 1984; Eldholm 1990; Larsen et al. 1992), accompanied by violent, partially subaerial plateau basalt eruptive activity (White 1989; White & MacKenzie 1989; Eldholm & Thomas 1993; Kaiho & Saito 1994). High hydrothermal activity along Pacific oceanic ridges may have contributed to increased atmospheric CO 2 levels (Owen & Rea 1985; Olivarez & Owen 1989; Kyte et al. 1993), although it is doubted whether increased spreading activity will cause a net increase in atmospheric pCO 2 (Staudigel et al. 1989; Varekamp et al. 1992). High concentrations of other greenhouse gases such as methane have also been invoked (Sloan et al. 1992, 1995). High atmospheric pCO 2 can not fully explain the global climate, however: how could high latitudes have warmed while the equatorial regions were not

PALEOCENE--EOCENE BENTHIC FORAMINIFERALEXTINCTION warmer than today (Shackleton & Boersma 1981; Zachos et al. 1994; Sloan et al. 1995)? The oceans could have transported more heat, involving a larger part of the ocean waters (Barton 1987; Sloan & Barron 1992; Barron & Peterson 1991), but it is not clear that even involvement of the whole ocean in heat transport can succeed in keeping the high latitudes at the temperatures of 15-18~ suggested by oxygen isotope studies (Crowley 1991; Walker & Sloan 1992; Sloan et al. 1995). The theory of increased oceanic heat transport appeared attractive because it had been long theorized that deep and intermediate waters in the oceans could have formed as high salinity, high temperature waters by evaporation in subtropical latitudes in the absence of very cold polar areas (Chamberlin 1906; Brass et al. 1982; Hay 1989). Surface oceans at low latitudes show the cooling expected if heat transport was more active than today (Zachos et al. 1993, 1994; Bralower et al. 1995b). The mechanisms for enhanced oceanic heat transport, however, remain undefined (Sloan et al. 1995). Evidence for the presence of such warm saline bottom water (WSBW) is not unequivocal. Some investigators suggest that WSBW was the dominant water mass over much of the Cenozoic (Matthews & Poore 1980; Prentice & Matthews 1988), or during at least the early Paleogene (Kennett & Stott 1990). Some models, however, predict that the deep oceans will turn anoxic over large regions when deep waters form largely at subtropical latitudes (Herbert & Sarmiento 1991), and this did not happen during the Cenozoic (e.g. Thomas 1992). Some investigators concluded that much of the oceans' intermediate and deep waters was formed at high southern latitudes during most of the Maastrichtian and Cenozoic (e.g. Barrera et al. 1987; Miller et al. 1987a; Katz & Miller 1991; Thomas 1992; Zachos et al. 1992, 1993). Major changes in the deep-water environments of the oceans would be expected to result from a reversal in deep water circulation (Kennett & Stott 1991), and these should be reflected in the composition of deep-sea benthic foraminiferal faunas. Paleocene deep-sea benthic foraminiferal faunas closely resemble Late Cretaceous faunas (Cushman 1946), and the major break in deep-sea benthic foraminiferal faunas occurred somewhere in Chron 24, i.e. between Paleocene and Eocene if seen at coarse time-scales (e.g. Beckmann 1960; von Hillebrandt 1962; Braga et al. 1975; Schnitker 1979; Tjalsma & Lohmann 1983; Boersma 1984b; Thomas 1990b; Bolli et al. 1994). In Chron 24r (corresponding to planktonic foraminiferal zones P5-P6, and calcareous nannofossil zone NP9) major faunal change occurred in bathyal and abyssal faunas in all the world's oceans (Miller et al. 1987b; Boltovskoy & Boltovskoy 1988, 1989;

403

Berggren & Miller 1989; Thomas 1989, 1990a, b, 1992; Katz & Miller 1991; Mackensen & Berggren 1992; Reynolds 1992, unpublished MSc thesis, Univ. of Maine; Pak & Miller 1992, 1995; Nomura 1991; Kaiho 1988, 1991, 1994a, b; Miller et al. 1992; Kaiho et al. 1993; Bolli et al. 1994. Benthic foraminifera underwent coeval extinction in neritic to upper-middle bathyal environments in land sections from Israel to Egypt, North Africa and Spain (Molina et al. 1992; Speijer 1994), in the North Sea shelf seas (King 1989; Charnock & Jones 1990), in New Jersey (USA, Gibson et al. 1993), and along the western Pacific margin from Japan (Kaiho 1988) to New Zealand (Hornibrook et al. 1989; Kaiho et al. 1993). At these depths the faunal change was associated with low oxygen conditions, as indicated by the presence of black or dark grey, commonly laminated sediments (North Sea, Japan, New Zealand, New Jersey, Spain, Israel, Egypt). Many authors have explained these conditions as resulting from local lack of circulation (e.g. for the North Sea Basin; Charnock & Jones 1990), but such local effects occurred during worldwide low oxygen conditions, which may have been at least partially responsible. Some authors suggested that relatively low oxygen conditions also occurred, at least locally, in the world's deep oceans (Thomas 1990b, 1992; Kaiho 1991). Most authors implied at least some form of change in deep-water circulation in the benthic foraminiferal extinction, which was seen as caused by changes in temperature as well as nutrient and oxygen content of the deep waters (Miller et al. 1987b; Thomas 1989; Katz & Miller 1991; Nomura 1991). Others suggested that oceanic productivity decreased drastically (e.g. Moore et al. 1984; Shackleton 1987; Shackleton et al. 1985; Coffield & Shackleton 1988; Rea et al. 1990; Corfield & Cartlidge 1992a, b; Corfield 1993), with possible effects on the benthic faunas. Yet others suggested that at least in some areas productivity increased (Thomas 1992; Speijer 1994). Only recently, however, has it become clear that the deep sea benthic extinction in the latest Paleocene was a unique event in its global extent and rapidity (Thomas 1989; Kennett & Stott 1991; Pak & Miller 1992; Kaiho 1994b; Robert & Kennett 1994). In the next paragraphs we will discuss events which happened on timescales of thousands to ten thousands and not millions of years, at some time early in the reversed part of chron C24, in nannofossil zone NP9, and in planktonic foraminiferal zone P5 (Berggren & Aubry 1996; Aubry et al. 1996). It had seemed reasonable to suppose that the benthic foraminiferal extinction was not synchronous world wide (Miller et al.

404

E. THOMAS 8~ N. J. SHACKLETON tope excursion was very short-lived, there may have been hiatuses at Site 577 (Berggren et al. 1995), and sampling by Pak & Miller (1992) was not so detailed as to ensure capture of the most extreme values. The explanation of these short-term changes in climate and in the carbon cycle is being actively debated. Oxygen isotope data suggest rapid warming (over less than 5000 years) of the deep ocean waters at high and low latitudes by 4-6~ but essentially no warming of surface waters in the tropical Pacific (Zachos et al. 1994; Bralower et al. 1995a, b), so that latitudinal temperature gradients of surface water were very low for about 50 000 years. Such a rapid, deep ocean-wide temperature change has been explained by a change in dominant oceanic circulation pattern, from dominant production of deep to intermediate waters at high latitudes to dominant production in subtropical regions (e.g. Kennett & Stott 1991; Thomas 1992). Major questions remain as to the exact nature of the upheaval in the carbon cycle, the feedback processes resulting from atmosphere-ocean interactions, and whether deep to intermediate waters did indeed form dominantly at subtropical latitudes, and for how long? In this paper, we present new benthic faunal and isotope data at high resolution from four sites at different depths in the southeastern Atlantic Ocean and the Weddell Sea (Fig. 1), to investigate the magnitude of the isotopic excursions in different areas and at different water depths. We use the new, detailed data to discuss

1987b) because a mechanism for globally synchronous extinction in such a large part of the earth's environment as the deep ocean is difficult to envisage. The observation, however, that the extinction was coeval with a very large, short-term negative excursion in the benthic as well as planktonic dl3C and dlsO records at ODP Site 690 (Kennett & Stott 1991) suggested that the extinction might have been caused by rapid warming at high latitudes, causing large-scale changes in oceanic circulation and upheaval in the global carbon cycle. Additional research demonstrated that the short-term, extremely negative shift in the carbon isotope values was, as expected by Kennett & Stott (1991), a more than local phenomenon: the event has been recognized in the Bay of Biscay (Pak & Miller 1992; Stott et al. 1996), Pacific Ocean (Pak & Miller 1992, 1995; Bralower et al. 1995a, b), and Indian Ocean (Thomas et al. 1992; Lu & Keller 1993). A large negative anomaly in carbon isotopes was also recognized in enamel of land-herbivore teeth and in carbonate concretions, clearly demonstrating that the atmosphere as well as the ocean was involved (Koch et al. 1992; Stott et al. 1996). The globally averaged magnitude of the shortterm shift is in question: it is about -2%0 in the deep Antarctic (Kennett & Stott 1991; this paper) and the southern Indian Ocean (Lu & Keller 1993), but only about -1%o in the Pacific and the Bay of Biscay (Pak & Miller 1992). The smaller values may not reflect a global average because the iso-

90 60

"

30

~

9" 5 2 7

.

.

~,

9 -'-~',

~

90

Fig. 1. Palaeogeographic map of the continents in the late Paleocene, after Zachos et al. (1994). Sites from which data are presented in this study are indicated by *; other sites mentioned in the text by +.

PALEOCENE--EOCENE BENTHIC FORAMINIFERAL EXTINCTION

the bathymetric and biogeographical patterns of the benthic foraminiferal extinction, and re-evaluate the existing database.

Material and methods Sites a n d stratigraphy: M a u d Rise

Sites 689 (64~ 03~ present water depth 2080 m) and 690 (65~ 1~ present water depth 2914 m) were drilled on Maud Rise at the eastern end of the Weddell Sea (Barker et al. 1988, fig. 1). Site 689 is on the northeastern side of the ridge near its crest, Site 690 is 116 km to the southwest on its southwestern flank. Depth estimates for the sites at the end of the Paleocene, based on faunal contents, agreed well with backtracking estimates, giving about 1100 m for Site 689, about 1900 m for Site 690 (Thomas 1990, 1992). At both sites sediments of Maastrichtian to Pleistocene age were recovered. Paleogene biomagnetostratigraphy was reviewed by Thomas et al. (1990). Upper Paleocene to lower Eocene sediments consist of chalks and calcareous oozes, with admixture of fine-grained terrigenous matter at Site 690. Recovery was good at Site 690, less so at 689. Core deformation was minimal, but the biostratigraphy was difficult to interpret because the high latitude of the sites caused the absence of many marker species of planktonic foraminifera, and opened the possibility of diachroneity for nannofossil markers (Pospichal & Wise 1990; Aubry et al. 1996). Major differences of opinion in the interpretation of the records centred on the lowermost Eocene. Stott & Kennett (1990) maintained that the section was basically complete over the upper Paleocene-lowermost Eocene, whereas SpieB (1990), Thomas et al. (1990) and Pospichal & Wise (1990) thought there was at least one, about 2 million year-long hiatus in Chrons 22 and 23. Aubry et al. (1996) concluded that an additional hiatus occurs in Chron C24n. We do not agree with Kennett & Stott (1991) that the records at Site 690 were undisturbed by bioturbation. In our opinion, the sediments show a different ichnofossil assemblage in the interval just after the extinction event, but no lamination. A few corroded specimens of Gavelinella beccariiformis are present two samples above the extinction, and presumably reworked. Therefore we think that the exact sequence of events in samples spaced only by a few centimetres is difficult to determine. At Site 689 the upper Paleocene-lower Eocene succession was very incomplete, but a section of sediment extending from in the lower part of Core 689B-22X to the poorly recovered Core 689B-24X contains the benthic foraminiferal extinction and the CP7/CP8 nannofossil zonal boundary (Thomas

405

1990; Pospichal & Wise 1990; Thomas et al. 1990). The section is bounded by hiatuses and sedimentation rates can not be determined with precision. Benthic foraminifera were studied by Thomas (1990a); in this paper we present additional data at higher resolution for Sites 689 and 690 (Appendices 1, 2). For Site 690 our data are from Cores 690B-16 to 690B-22, encompassing sediment deposited in Chron C24r according to Aubry et al. (1996). Our data from Hole 689B are from the interval described above (Fig. 2). Carbon and oxygen isotope data of benthic foraminifera were collected by Kennett & Stott (1990, 1991), for planktonic foraminifera by Stott et al. (1990) and Corfield & Cartlidge (1992b), and for bulk carbonate by Shackleton & Hall (1990). We present additional data at high resolution for both sites (Appendix 5). Sites a n d stratigraphy: Walvis R i d g e

Sites 525 (29.~ 02~ present water depth 2467 m) and 527 (28~ 01~ present water depth 4428 m) were drilled on the summit area and the lowermost western slopes of the Walvis Ridge, respectively (Moore et al. 1984, fig. 1). At both sites, Paleocene to Eocene calcareous oozes and chalks were recovered (Shackleton et al. 1984a). Palaeodepths for the end of the Paleocene as derived by backtracking (1600 m for Site 525, 3400 m for Site 527; Moore et al. 1984) are in agreement with benthic faunal data (Boersma 1984b; this paper). Carbon and oxygen isotope data from foraminifera as well as bulk carbonate were determined (Shackleton et al. 1984b; Shackleton & Hall 1984). Nannofossil stratigraphy was described at fairly low resolution by Manivit (1984) and in more detail by Backman (1986a, b). Planktonic and benthic foraminiferal biostratigraphy at low resolution was described by Boersma (1984a, b). In this paper, we present additional benthic faunal data and isotope data on benthic foraminifera and bulk carbonate over the interval of the benthic extinction. At both sites, despite the difference in palaeodepths, the extinction occurred within the lower few centimetres, but above the base of, an interval of dissolution of 40-50 cm thick, with a sharp lower boundary and a gradual upper boundary (Table 1). In this interval planktonic foraminifera were fragmented, but benthics showed good preservation. We took samples from the cores in which the extinction occurred only. N u m e r i c a l ages

Numerical ages for Paleocene sediments are in the process of being revised (e.g. Odin & Luterbacher

406

E. THOMAS & N. J. SHACKLETON 125 ~tm size fraction, and we analysed 12-20 specimens. Samples for isotope analysis of the small species Tappanina selmensis consisted of 60-80 specimens. Lenticulina spp. commonly show large fluctuations in isotope values; we excluded from analysis specimens larger than 500 ktm, which showed erroneous values, possibly because they resulted from reworking. All specimens were ultrasonicated to remove adhering

408

E. THOMAS • N. J. SHACKLETON

microfossils, then dried at 50~ Foraminifera were transferred to reaction vessels and roasted at 400~ Bulk samples, weighing a few milligrams, were taken from samples from Cores 525A-32 and 527-24, dried and vacuum roasted at 400~ to remove organic contaminants. The samples were then reacted with 100% orthophosphoric acid at 90~ using a VG Isotech Isocarb common acid bath system. The evolved carbon dioxide was analyzed in a VG isotech SIRA Series II mass spectrometer. The results were calibrated to PDB by repeated analysis of a carbonate standard. Analytical accuracy is better than 0.08%0 for both 5180 and ~513C. Benthic foraminifera for faunal analysis were picked from the > 63 lam size fraction, following Thomas (1990a). All specimens counted were picked and mounted in cardboard slides. All samples contained sufficient specimens for analysis (> 250), and counts are shown in the appendices. Taxonomy is as in Thomas (1990a) and largely follows Van Morkhoven et al. (1986). Thomas (1990a) misidentified Neoeponides hillebrandti as Neoeponides lunata and the reverse; that mistake has been rectified in the appendices of this paper. All species richness numbers were recalculated to 100 specimens using rarefaction (Sanders 1968).

Results Faunas We define the level of extinction as the interval where species richness declined most rapidly. Preextinction faunas at Walvis Ridge and Maud Rise are similar, with almost all species present at all 4 sites (Figs 4, 5, 6, 7), although at varying relative abundances. These faunas contain many cosmopolitan, Late Cretaceous through Paleocene taxa with a large depth range, such as Gavelinella beccariiformis, Gavelinella hyphalus, Neoeponides hillebrandti, Neoeponides lunata, Pullenia coryelli, Bolivinoides delicatulus, Neoflabellina semireticulata and agglutinated taxa such as Tritaxia paleocenica, Tritaxia havanensis and Dorothia oxycona. The agglutinated taxa are more common at Site 527, the deepest site studied. Aragonia velascoensis is rare or absent at Maud Rise, and rare at Site 525 (Fig. 6). Stilostomella spp. are more common at Maud Rise than at Walvis Ridge. Preextinction faunas have a very high species richness, with many rare uniserial lagenid species and unilocular taxa. Many common, long-lived species occur, such as Cibicidoides pseudoperlucidus, Oridorsalis umbonatus, Nonion havanense, Nonionella robusta, Anomalina spissiformis and Anomalinoides semicribrata. Overall, bi- and triserial species were less common at Walvis Ridge

than at Maud Rise. Pre-extinction, large peaks in relative abundance of species such as Bulimina thanetensis and S. brevispinosa are not observed at Walvis Ridge. Rectobulimina carpentierae occurs only at the two deeper sites, 527 and 690, with strongly fluctuating relative abundances at the latter site, and Bulimina thanetensis was more common at these sites. Gavelinella beccariiformis was more common at the shallower sites with highest relative abundances at Site 689, in agreement with Katz & Miller (1991) who consider this species typical for bathyal sites at high latitudes. Overall, the pre-extinction faunas were remarkably similar in species composition given the large differences in depth and geographical location, in agreement with Tjalsma & Lohmann (1983) and Kaiho (1988, 1991). This uniformity ended with the extinction. At Site 690, many of the typical Paleocene species had their last appearance in the highest sample with high diversity, 690B-19H-3, 72-74 cm. In sample 690B-19H-3, 66-68cm, the diversity decreased from 49 to 26 species per 100 specimens, and bitriserial species were much more abundant, although in our age model these samples differ in age by only 1000 years. Faunal abundance patterns thus changed at the same level where a high number of last appearances occurred. At Site 689, however, the last appearance of G. beccariiformis and other Paleocene species occurred in sample 689B-23X-1, 80-82 cm, whereas the major drop in species richness from 60 to 42 species per 100 specimens occurred between 23X-1, 80-82 cm and 23X-1, 87-89 cm. Samples 23X-1, 80-82 and 23X-1, 87-89 both contain common T. selmensis and A. aragonensis, typical species for the postextinction period, but the species richness is high in these samples and G. beccariiformis is present. These appearances might result from bioturbation or core disturbance. At both Maud Rise sites the extinction occurred after a decrease in carbonate content of the sediments to about 65% (O'Connell 1990; Thomas 1992), and after the first appearance of the keeled, warm-water planktonic species 'Morozovella' convexa at high southern latitudes (Stott & Kennett 1990; this paper). At Walvis Ridge a sharp drop in benthic foraminiferal species richness from 52 to 34 species per 100 specimens occurred between samples 525A-32-6, 145-147cm and 525A-32-6, 130132cm. At Site 527 species richness decreased from 53 to 24 species per 100 specimens between 527-24-2, 56-58 cm and 527-24-2, 38-40 cm. The level of most rapid decrease in species richness is within the lower part of a dark, low-carbonate layer (Table 1): the benthic extinction thus occurred after the onset of increased CaCO 3 dissolution. The preservation of benthic foraminifera in this layer of

PALEOCENE-EOCENE BENTHIC FORAMINIFERAL EXTINCTION

Fig. 4. Benthic foraminiferal relative abundances c-f most taxa at Site 689; see Appendix 1 for counts.

409

410

E. THOMAS •

N. J. SHACKLETON

Fig. 5. Benthic foraminiferal relative abundances of the taxa, Site 690; see Appendix for counts.

PALEOCENE--EOCENE BENTHIC FORAMINIFERAL EXTINCTION

411

Fig. 6. Benthic foraminiferal relative abundances of the most common taxa at Site 525; see Appendix 2 for counts.

412

E. THOMAS • N. J. SHACKLETON

Fig. 7. Benthic forarniniferalrelative abundances of the most taxa, Site 527; see Appendix 4 for counts.

low CaCO3-content and increased fragmentation of planktonic foraminifera is good. The samples in the interval with low diversity in the first 50 000 years after the extinction differ in faunal composition at the sites, with most differences between the Walvis Ridge and the Maud Rise

faunas, lesser differences between the sites at different depths in each area. These faunas occur during the interval with very low ~513C isotopic values at all 4 sites (see below), and thus can be considered coeval. At Site 690, the post-extinction faunas had high relative abundances of Eouvigerina

PALEOCENE--EOCENE BENTHIC FORAMINIFERALEXTINCTION sp., and at both Maud Rise sites Tappanina selmensis, Bolivinoides cf. decorata, Bulimina ovula, Bulimina simplex and Bulimina trinitatensis had high, but strongly fluctuating relative abundances. N. truempyi was rare or absent just after the extinction. Nuttallides umbonifera first occurred just after the extinction at Maud Rise, but is not present in the samples studied from Walvis Ridge. At the Maud Rise sites abyssaminid species (small, thin-walled, spiral species; Schnitker 1979; Tjalsma & Lohmann 1983) were more common after the extinction. At Walvis Ridge there was a much stronger increase in the relative abundance of these species, with the highest at the deeper Site 527, in agreement with Tjalsma & Lohmann (1983) and Katz & Miller (1991). In contrast to the patterns at Maud Rise, the post-extinction faunas at Walvis Ridge had very high relative abundances of Nuttallides truempyi, as described by Tjalsma & Lohmann (1983) and Katz & Miller (1991) for other Atlantic sites. A distinct drop in relative abundance of biserial and triserial species occurred at Site 527. Slightly higher in the section, however, the biserial species Tappanina selmensis and Aragonia aragonenis increased in relative abundance, with highest abundances at the shallower site. Tappanina selmensis never reached such a high abundance as at the Maud Rise sites (Figs 4, 5, 6, 7). Many rare species (such as lenticulinids, uniserial lagenids, unilocular taxa) were absent at all sites in the post-extinction faunas, at least partially causing the low species richness. Many of these species, however, reappeared higher in the section, and their absence may thus be apparent because of the decreased evenness of the faunas (Signor & Lipps 1982). At counts of about 300 specimens a lower number of species would be observed in an assemblage with higher dominance (Gage & Tyler 1991). Comparison of the faunas (Fig. 8) showed that at all sites the observed species richness dropped precipitously, with lowest values reached at the deepest Maud Rise site. The species richness appears to recover more quickly at the Walvis Ridge sites, especially at Site 525, if ages are determined according to Fig. 3 (fitting of the 813C curve of N. truempyi). Interesting is the strong increase in relative abundance of N. truempyi at the Walvis Ridge Sites, coeval with its strong decline at both Maud Rise sites. Abyssaminid species increased in relative abundance after the extinction at all sites, but much more so at the Walvis Ridge sites, and the shallower Walvis Ridge site has higher relative abundances of these species than the deeper Maud Rise site. Biserial and triserial species increased in relative abundance just after the extinction at both Maud Rise sites, but decreased

413

at the deeper Walvis Ridge site and showed little change at Site 525. A typical component of post-extinction faunas in all oceans are the species T. selmensis and A. aragonensis (Boersma 1984b), although their relative abundances vary from site to site. T. selmensis first appeared in the Late Cretaceous, but occurred at low abundances until the extinction event. A. aragonensis had its first appearance at the time of the extinction (Tjalsma & Lohmann 1983; Van Morkhoven et al. 1986; Bolli et al. 1994). G. beccariiformis last appeared at all sites at about the same time, although the species started a decline in relative abundance at Site 689 several hundred thousand years earlier (Fig. 4). Overall, however, this pattern does not agree with the statement by Tjalsma & Lohmann (1983) that the G. beccariiformis faunas became gradually restricted to shallower depths before becoming extinct. B. thanetensis became extinct in the late Paleocene at all sites, but always a few samples higher than G. beccariiformis.

Benthic foraminiferal isotopes We measured stable isotopes in several species of deep sea benthic foraminifera from the Maud Rise sites (Figs 9, 10). We wanted to derive an isotopic record of the extinction event that is as complete as possible, and N. truempyi and Cibicidoides spp. are not present in large enough numbers for isotope analysis just after the extinction. We also wanted to determine whether species that are presumed to be infaunal or epifaunal because of their morphology (Corliss & Chen 1988; Rosoff & Corliss 1992; Thomas 1990) show a different carbon isotopic signature, as observed in Recent and Neogene taxa (Woodruff & Savin 1985; Zahn et al. 1986; Altenbach & Sarnthein 1989; McCorkle et al. 1990). Infaunal species precipitate their tests in contact with pore waters, not sea water, and can thus be expected to have lower 813C values. Most species thought to be infaunal because of their morphology had indeed lower 813C values than species thought to be epifaunal, with the exception of Oridorsalis umbonatus. This longlived, cosmopolitan, trochospiral species plots in the field of infaunal species (Fig. 10). Its carbon isotope data should thus be interpreted carefully, especially because Rathburn & Corliss (1994) reported that Recent representatives of this taxon live infaunally. The benthic species show considerable scatter in their isotopic values, and there are differences between the patterns at Sites 689 and 690. At Site 690, Lenticulina spp., a presumed infaunal group, shows clear separation in 813C values from the epifaunal species, and has consistently lower values

414

E. THOMAS & N. J. SHACKLETON

%..., ,..-0 0

~E ~/)

e.-

0

0

,,~..

oo

r

O

o

~2

o

o

o

.

.

.

.

i

;

9

9

,

i

.

.

.

.

o

,=,

~

. . . .

i

. . . .

o

= ~

689 690 525 527

. ,,..~

/I

N

~9 o0,1' .~ E~

Age, Ma

Age, Ma

~ ~

PALEOCENE--EOCENE BENTHIC FORAMINIFERALEXTINCTION

415

03

r

~o

u

o

9

.

.

!

9

,

.

|

9

.

.

o o od

m.. "7'



~

'

1-=15

:

18

2

- t 7 2 . 1 6 55.542- ~'8 4 7 1 3 1 2 1 ] " ~ 173.31 55.573 66 48302' I 91

22H-2. ~ - 44

~7.~,.-. 5 5 ~ ~ . ~

22H-3, 42- 44

~88.~-58.o7o~,~3~

22H.cc

-~ -191.~ ~- - - - ~~ ; -~ i i

1" '

51 35 273 6 5,?: 37 298 7 73 48 342 t~ 1 65 42307 6 61. 37 343. 5. 1 52 34 308 12 81 41 297 8 87 38 334 8 3 62 44 316 23 5 55 3 6 ~ " ~ 15 54 36 2 6 1 ~ ~ 1,5 6 1 1 3 8 3 2 3 ~- - 7~ - - ~

~2~~

----1

t

-'~

1 ....

1~--

y ~

.....

~-

3-1-5

i ........

3

4

~26

55.50or eel 48 ~ e --- ~ - 1 - - - ~ - - ~ ] 170~65 55.501' 63 r 43 322' -171.05 55.511 71' 44 302' 11 13' 1-'

174.30 55.601 174.71 55.812 176.21 55.854 177.71 55.898 . 179.22 55.739 . 180.28 55,768 180.71 55.780 182.2--1 5 5 ~ L ~ 183.71 55.864 184.8-------'s 55.896 185.2~ 55.900 1185.71 55.915

18 23

t ---~----11 ~ ' - ~ . . . . ~

'"

-

~.~

19H,CC 20H-1.40- 42 20H-2.40-42 20H-3,40-42 20H-4, 41- 43 20H,CC 21H-1,40- 42 21H-2, 40- 42 21H-3, 40- 42 21H-4, 5 - 7 21H.CC 22H-I, 40-42

-

- - 3

.....

4

1 3 1 3 8 5 8 4 3 9

1

~ ~0 120,

~ 3 13i' ! 4' 51

4 2 2 4

t ,

i 7

I -4; ~IY 18,

I

I 3 , 2

9 8 9 4 6 7 7 5 2 8 10 3

li' 4" ~ J

1

1 2 3 1 I

4 3 2 - -

3 8 3 3 2

8 2 3 ~ 2 7 8, 1,

--~E

4

--~-~-~ [19-H:.4,7~76 119H-5,40-42

8

. -

---f . . . . . .

~-5 ~

~ ~

-

~

'=

~'~'~-~----4--3

21 11

--

-

,

o

5 2 3 28 --, - ~ _ --e-~ -~------~ ~ ..... ~ _ ~-9--'7~-~_ 1 28- :_- - - - - - - - - - : ~_: I ' r 13 8 8 2 14 ~ --, ' ' ' 1 1 ' - - 5 - - ~ .... ~ - ~ - ~

1-8"-

~7:26 55.175 41 32 278

6

.~

=

-9 ~ ~ I--2+9~-1~3~'~-~-~-~--I -~ ~ ~ ~---8~~-~22' ~ . ~ ' ~ - - ~ _ i ~ -- 1 ~ -~

~ ~ 8 _ I ~ - ~ ' i 6 ' - - ~ 8 j

/ !~.1~55.294_~7 29342 1~_._~.6155.330 40 28335

"

7.

...... 19 1~8 7' ~ ~ ~ ~ --" ---'---- 1--~L33- I ~ 1 { ' - 6

= . ,

= = . -

7

-~.91 54.937 58" 3 8 r ~ ~ -1~4-9:4154.974' 50 ~ 32.~337 I 150.91 55.022:48' 34 j 3 3 0 1 ~ ~

.

,

,

i

-~i

3 ' [ 2' i 3" 1

2~ 3 3 .... ~-~11 3 3 4

4

L 6L I u ~ 13 20 14 12 27 21 18 21 33 16 7 11 14 22 15

71

4" 8 14 24 3 5 6 14 1 10 12 1 5 - 3 11 5 4 3 3 9 4 2 6 1 8

15

"..... ~- 1 1

1-3

i

i 11

13

4

7

9 7

12

3,

I TO

1

! =3 ..... ~:--9

j 6 ~-~

.........

12 - - - - - ~ 9

~ ,

1

: :

~ - ~ .......... ,~-

11 8

1

7 2

I 3

2

-2~

~

2 4

9 8

1

9

9

2

5

~,

I

~--~ "-9

~, -6 ,

'

~

i

6

!

-1

435

PALEOCENE--EOCENE BENTHIC FORAMINIFERAL EXTINCTION

:

I

"

9

'

i

= 9~-~,I

e

9

~

~

~

~

,_~ 12

i

'

IJ

:

~

2

i - -- ? --1~

....

~

47 20

--+"

~

4

_2 _ ~ _ _ _ 2

4j

2 ++i . . . . . . .

P

. 5

~__._

~----+~ _ _ ~ 3+ _

1~

__~-_

[ -]+[---~-/ -,--

.....

_ ,L_

!

:

[

--++---4---++~-

,

.

1

13

1

62

2

1 114

1

!

~ -

-"++--+--I

Ii

46 94'

.

.

.

.

.

.

.

....

~--.-~L - i

.

.

.

.

.

.

. .

'

.

.

-~

- - -, r - - - - ~ - ~

.

_~__+_+~_

- + - : ++-+~ +._+

___.,

. . . .

u

:

,

24

4 41

..... + - - ~ - - -

+

3{+1-1 , - + - + --t--,4----

~

......

.....

'_J~:

q

I"-~V-~- 3--gt~l

1 1 1 + + I

+~._..._.,-~.._

~-.----4

11

i

3:

-; -+i 4 -

i

I

+-~; .... ~ -

-.-

~

21

i

i

~

~

~

_

_

~____+

_.__:+____, . . . . .

++ _ _ ~ _ _

ff :----+Z ~ + ++ ..J+_~ ,"

+ + . ~

F-2

+~-~-1

:-+ .... -+

--~---~'i -

i

=

+ 2 ~

44 i

, -

+ t3,

o!

~-+-1- . . . . . . .

-:_~:L+-/+ 9-~,--- 9.-r-- ...... __+_ , + _ ~ _ + +3~~ 5, ....... 4!

+

~ ,+3

'

a.~ - - - - ~ :

!

.....

--4-------~

......

-+-::~+4~-

.

! -

19'

4

i

=

4

--

,

-'~

I

.~ r

~

,

,

1

436

E. THOMAS •

N. J. SHACKLETON

Appendix 3. Faunal data, Site 525 i

[

9 ~.

i! sample 3~-4, 75,. 77 32-4,125-127 32-5, 25- 27 32-5, 48- 50 32-5. 77- 79

I

.-~,~,,

I

'

I

'

-o~

~,

=~

,=E

ii -'-~ 389.36 389.86 39OO8 390.38

55,38~ 55.394 55.404 55,412

3 47! 33 3221 42 31 312: 47 31 329!

32-5.',.~-11~ 32-8, 2,5. 27

! 3so.7e 55,419 55.427 39 34 32 25 33011 309 ! i ~el.33 55.441 52 40J 350

32-6, 5O- 52 32-6, 76- 78 32-6, 90- 92 32-6,113-115 33-6,130-132 32-6,145-147 32-7, 22- 24 32-7, 53-55

39101 ~ 391 87 ,! 392.01 392.24 392.41 , 382.56 392.81 3~3.11 I

1 4 10 5 ---5 2~9 14 17

55,447 43 = 3~. ~ 4 ] 55.454 49' 35 326 ! 55.457 53 40 313 55.475 54 3~ 343 55.491 34 27 328: 55,501 60 40 317 55.519 81 56 352 55.540 73 5 2 3 0 9

'

2 4 6 9 9 42 34 29 6 21 11

--3

1 7 _.. 4

1 7 44! 13 I 37 2 4

I I J

2 4 3

1 2 1

.

1

1 2 2 ~ 1 2 5 4 1

9 e 5

8

3 4 13 3 15 10 13 ~ 31 2,

2

5

2

I Ij I

1

i

3 1 3

26 7 18 14 16 17 17 13 13 13 19 3 9 20 13

66 51 6 56 86 59 30 27 1-8 11 13 14 16 6 3

1 11 7 4 5

11 t2~

1

4i 5t 1 10

1

17 11 8 3 8 1

1 1 3

I ~

i

sample 32-4, 75-

32-4.12~127 ! 389.38 55394 32-5, 25- 27 32-5,48- 50 32-5, 77- 79 32-6, 25- 27 32-6,50-52 32-6, 76- 78 32-6, 90- 92 32-6,113-115 32-6.130-132 32-6,145-147 32-7, 22- 24 32~ 53-55

389.86 390.08 390.38 390.76 391.38 391.61 391.87 392.01 392.24 392.41 392.56 392.81 393.11

55.404 55A12 55.419 55.427 55.441 55.44/ 55.454 55.457 5.5.47E 55.491 55.501 55.519 55.544]

= 32322 31 31 32 25 40 32 35 40 39 27 40 56 52

312 329 330 309 350 324 326 313 343 328 317 352 309

,.

1~ 11 17 19 12 14 18 13 12 10 10 4 12 14

,

1 11

2 1

8 8 6 15 7 14 12 15 10 4 e 5 4 4 1[- 6 i

= ,

4 2 1

1

= 6 5 6 1 4 4 6 8 7 5 8 5'

2 2 3 2 2 6 1

o

o,i~

2~

,,=

-~i

27 18 20 50 56 38 65 40 45 108 18i 8[ 8}

8 12 15 7 1 6 10 9 8 17 48 9 10

2 1 8

2 5 2 2

4 18

5 2 4

1 2 4 2 2 1 4

.

~

.

=1 20 1 24 1 28 12 2 1 27 22 2 23 1 23 3 35 1 22 1 4 14 1 7 20 [11 8

1

t I

437

PALEOCENE--EOCENE BENTHIC FORAMINIFERAL EXTINCTION

:+ ++:+

+ _,

+ +'++-++/+,~+, + ++, +:+ +

,- ~ ~ +

:+

m

+++,++ + o++ , + +, +, :: +++ '+ . +,+, ++ i ! +"

+++ .+.-+" .+::~' + 1 7+++++:+ ~ . ~+' + ++~+~-+ '6 j. t + ++" :+ +:++,-+++ .o +, !++ +

-.+

"r 9 ."r

~-~---

3 5 3 6 9 6 7 8 3 4 2

~

1

2

3 11.~ 61 161, ~24-11 17 17

9 13 18 18 ~

---

l 1i

1

t T. . . .

1

!

1 1

1

7 7

+

--

.,++

-

-~

2

1

.

,

.,,

,

.

4 1 11 4 1t 11 22 7 8 1 3

17

.

,

1 1 3 1 1

1 1 3

1

t,

.

,:'+

~

, I

E

~

|

u

_

2

~

. 24

_

2

-

3

1 1 1

3 1 I 4 1

3 1

1

-!!

11

3

9

2

-)0 ,5[

) ~

,41

/-~

7

1

1 2 1 5 2 2

,_. + ,

~ 'i+ .: ,, ..,. ,

-

4

1 2

++:8~_+

+ ~

4 1 3

1 2 1

+

"+ .~_

1 2

3

11 =

~

--4

- [ ~o I 14

.,++

438

E. T H O M A S & N. J. SHACKLETON

Appendix 4. Faunal data, Site 527

.

.

.

i' :

.

~m~ _

24.1, 62- 64 24-1, 85- 89 24-1,102-104 24.1,136-138 24.2, 5- 7 24.2, 20-22 24.2, 38- 40 24-2, 56-58 !24.2, 68- 70 r24-2,80-82 ~4-2, 88- 90 24-2,115-117 24-3, 28- 30 24-3,48-50 24-3, 75- 77

199.63 55.504 199.86 55.424 ~- 200.03 55.433 ~ 200.37 55.457 I 200.58 55,470 200.71 55.480 200.89 55.493 201.07 55.507 201.19 55.519 201.31 55529 201.39 55.538 201.86] 55.561 ~ 55,629 202.49 55.646 202.76:55.662 /

E sample 24-1, 62- 64 24.1,85-8g 24.1,102-104 24-1,136-138 24-2, 5- 7 24.2, 20-22 24-2, 38- 40 24-2.56-58 24.2, 68- 70 24-2,80-82 24-2.88- 90 24-2,115-117 24.3, 28- 30 24.3,48-50 24-3, 75- 77

19g,63 19g.8e 200.03 200137 200.58 200.71 200,89 201,07 201.19 201.31 201.39 201,66 202,29 202.49 202.76

31 37 39 32 31 19 24 53 62 571 63 60 55 60 62

24 320 28 313 27 324 27 304 22 315 15 342 20 216 37 '349 41 3 ~ 40!320 42 321 ~ 42 312 41 3,32~ 42 316 42 317

20 13 . 20 30 50 59 64 25 ' 10 18

---;-

1

i~ i i

1

2

i

31 37 39 32 31 19 24 53 62 57 83 80 55 60 62

24 320i 283131 27 324 27 3 0 4 - - - - 22 315 15 342 20 216 37 349 41 344 40 320 42 321 42 3 1 2 - 41 332 42 316 42 317

7 3 4

23 4 6 1 2

2

~ 11

8

!6

18 19 1 3 2

6 15 2 1

t

=

a

1

26 2o 12 3 17 3 3 I t 19

~ ~ i~415 20

1

~, 2

6

15 14 29 7 1 4

12 9 7

t e 2s ~--e

4

1 t

t

12 29 33

1

3 1

1

1 2 1 1

2 1

11 2

3 6

2 1 1

:

!-53

1

1

J I "i, i ! "~176 i

3 4

3 6 5 7

2

~,i--e

I ' ! 55.504 55.424 55.4,33 55.457 55.470 55.480 55.493 55.507 55.519 55.529 55.538 55,581 55.629 55,646 55.662

20 21

12

7e

2

_

3 36 4 12 5

10 12 7 8

5 8 4

5 12 3

15 8 65 15 5 61 1 8 12 63 6 13 54 11 2 51 5 26 67 3 7 50 13 33 1 8 37 5 35 4 28 5 16 6 28 9 14 11 28

181 14 _._ 7 13 3 il

5 8 30 14 1 6 9 7 4 3 7 8 4 15 6 16

2

2 2

8

10 -1-7 4 18 2--~4 2 6 3 5 3 10 5 1 5 8

39 22 38 23 26

2

2 1 2

1 11 7 10 15 9 17 7

6 9 7 . . . . . . 6 5 __ 6 8 1 10 !

5 5 2 10 5 4~ 2

439

PALEOCENE--EOCENE BENTHIC FORAMINIFERAL EXTINCTION

i~,

---L

i

.i i ! :

~

=

0i .ii; i 00i0 :

24 22 16 28 24 21

i

2t

3 5

1 -

5 2

1

1

1

i 1

7

4 3

6 2

3

22

3

t l

O,.

0,.

O,.

0.,

02

1

,

;

1

2 2

1

2 8 1 6

1 1

4 3

1

2

1

! 1

.~-~! .~1=, .-=...=~

1 2

1

4

~:

~

~

11 19 3 161

1

.=,1=

4 61 3 4 4 11

~

1 8 1 12

i ii '.r 9

19 35 23 33 20 1

2

"~

8,

i

20 14 12 11 4 8 34 1 2O 4

.~ ~

5 3 2 1

2

24 17 27 9 11 32

| ~! _

4 17 23 24 12

1

2

5 l

4

52 !

45

2

~

I I 3 37 25 21 49

I

1

1

~~

~-~ 1 2 11 i, 6 - - ,

......

_

2 1 4 3 12 4 5 5

4

6 19 13 16 14 9 13

1 1 1 4 2 23

1 4

9 9 4 3 4

!

1 5 4 4

1 1

1 1

15

2

-

21

3

51

8~.

42 ~33 6

.i_~., lo 11 1:@ ~

i

1

....

12

4 is --~ ~o

1

1

1

6[ 14

j__ el_! ~ [ 8[ 19

-

440

E. THOMAS •

~)cLQellep ')llnq O~q mlep ')llnq

O;,Lq (e-Ha) exe^uoo lq

3~t~ !euumlom "V

OeL~ e feuumlotu "V 0 ~ P !euueNotu V

OCt~ '.S~UJOp.pe~q OgL~'s.~o~H~ooeq "~) O~t~ 'tAdtuerul "N Om~ '.V~u~ml 'N

O~L~'snlmaoqu~! "0

~o~,,,.,~ Ogt9 "aas ee.P~Ol~,~

u~