The Deliberate Search for the Stratigraphic Trap (Geological Society Special Publication No. 254)

The Deliberate Search for the Stratigraphic Trap

The Geological Society of L o n d o n BOOKS EDITORIAL COMMITTEE Chief Editor: Bob Pankhurst (UK)

Society Books Editors John Gregory (UK) Jim Griffiths (UK)

John Howe (UK) Phil Leat (UK)

Mike Brown (USA) Reto Gier6 (Germany)

Jon Gluyas (UK) Doug Stead (Canada)

Nick Robins (UK) Jonathan Turner (UK)

Society Books Advisors Randell Stephenson (Netherlands) Simon Turner (Australia)

Geological Society books refereeing procedures The Society makes every effort to ensure that the scientific and production quality of its books matches that of its journals. Since 1997, all book proposals have been refereed by specialist reviewers as well as by the Society's Books Editorial Committee. If the referees identify weaknesses in the proposal, these must be addressed before the proposal is accepted. Once the book is accepted, the Society Book Editors ensure that the volume editors follow strict guidelines on refereeing and quality control. We insist that individual papers can only be accepted after satisfactory review by two independent referees. The questions on the review forms are similar to those for Journal of the Geological Society. The referees' forms and comments must be available to the Society's Book Editors on request. Although many of the books result from meetings, the editors are expected to commission papers that were not presented at the meeting to ensure that the book provides a balanced coverage of the subject. Being accepted for presentation at the meeting does not guarantee inclusion in the book. More information about submitting a proposal and producing a book for the Society can be found on its web site: www.geolsoc.org.uk.

It is recommended that reference to all or part of this book should be made in one of the following ways: ALLEN, M. R., GOFFEY, G. R, MORGAN, R. K. & WALKER, I. M. (eds) 2006. The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254.

CITRON, G. P., MACKAY,J. A. & ROSE, E R. 2006. Appropriate creativity and measurement in the deliberate search for stratigraphic traps. In: ALLEN, M. R., GOFFEY, G. P., MORGAN, R. K. & WALKER, I. M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 27-41.

G E O L O G I C A L SOCIETY SPECIAL PUBLICATION NO. 254

The Deliberate Search for the Stratigraphic Trap

EDITED

BY

M. R. A L L E N Shell UK Ltd. G. P. G O F F E Y Paladin Resources plc, UK R. K. M O R G A N Veritas DGC Ltd., UK and I. M. W A L K E R ConocoPhillips (UK) Ltd.

2006 Published by The Geological Society London

THE GEOLOGICAL SOCIETY

The Geological Society of London (GSL) was founded in 1807. It is the oldest national geological society in the world and the largest in Europe. It was incorporated under Royal Charter in 1825 and is Registered Charity 210161. The Society is the UK national learned and professional society for geology with a worldwide Fellowship (FGS) of 9000. The Society has the power to confer Chartered status on suitably qualified Fellows, and about 2000 of the Fellowship carry the title (CGeol). Chartered Geologists may also obtain the equivalent European title, European Geologist (EurGeol). One fifth of the Society's fellowship resides outside the UK. To find out more about the Society, log on to www.geolsoc.org.uk. The Geological Society Publishing House (Bath, UK) produces the Society's international journals and books, and acts as European distributor for selected publications of the American Association of Petroleum Geologists (AAPG), the American Geological Institute (AGI), the Indonesian Petroleum Association (IPA), the Geological Society of America (GSA), the Society for Sedimentary Geology (SEPM) and the Geologists' Association (GA). Joint marketing agreements ensure that GSL Fellows may purchase these societies' publications at a discount. The Society's online bookshop (accessible from www.geolsoc.org.uk) offers secure book purchasing with your credit or debit card. To find out about joining the Society and benefiting from substantial discounts on publications of GSL and other societies worldwide, consult www.geolsoc.org.uk, or contact the Fellowship Department at: The Geological Society, Burlington House, Piccadilly, London W1J 0BG: Tel. +44 (0)20 7434 9944; Fax +44 (0)20 7439 8975; E-mail: [email protected]. For information about the Society's meetings, consult Events on www.geolsoc.org.uk. To find out more about the Society's Corporate Affiliates Scheme, write to [email protected].

Published by The Geological Society from: The Geological Society Publishing House Unit 7, Brassmill Enterprise Centre Brassmill Lane Bath BA1 3JN, UK (Orders: Tel. +44 (0)1225 445046 Fax +44 (0)1225 442836) Online bookshop: www.geolsoc.org.uk/bookshop The publishers make 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 of London 2006. 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 W l P 9HE. Users registered with the Copyright Clearance Center, 27 Congress Street, Salem, MA 01970, USA: the itemfee code for this publication is 0305-8719/06/$15.00.

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. ISBN-10 1-86239-192-0 ISBN-13 978-1-86239-192-5 Typeset by Type Study, Scarborough, UK Printed by Cromwell Press, Wiltshire, UK

Distributors USA AAPG Bookstore PO Box 979 Tulsa OK 74101-0979 USA Orders: Tel. +1 918 584-2555 Fax +1 918 560-2652 E-mail [email protected] India Affiliated East-West Press Private Ltd Marketing Division G-l/16 Ansari Road, Darya Ganj, New Delhi 110 002 India Orders: Tel. +91 11 2327-9113/2326-4180 Fax +91 11 2326-0538 E-mail [email protected] Japan Kanda Book Trading Company Cityhouse Tama 204 Tsurumaki 1-3-10 Tama-shi Tokyo 206-0034 Japan Orders: Tel. +81 (0)423 57-7650 Fax +81 (0)423 57-7651 E-mail [email protected]

Contents

ALLEN, M. R., GOFFEY, G. P., MORGAN, R. K. & WALKER,I. M. The deliberate search for the stratigraphic trap: an introduction

1

BINNS, E E. Evaluating subtle stratigraphic traps: prospect to portfolio

7

CITRON, G. P., MACKAY,J. A. & ROSE, P. R. Appropriate creativity and measurement in the deliberate search for stratigraphic traps

27

DONNELLY,N., CAVE, K. R., WELLAND,M. & MENNEER,T. Breast screening, chicken sexing and the search for oil: challenges for visual cognition

43

ALLAN, J. R., SUN, S. Q. & TRICE, R. The deliberate search for stratigraphic and subtle combination traps: where are we now?

57

ATKINSON, C., RENOLDS, M. & HUTAPEA, O. Stratigraphic traps in the Tertiary rift basins of Indonesia: case studies and future potential

105

GOOD, T. J. Identification of stratigraphic traps with subtle seismic amplitude effects in Miocene channel/levee sand systems, NE Gulf of Mexico

127

STOKER,S. J., GRAY,J. C., HALLE,P., ANDREWS,I. J. & CAMERON,T. D. J. The importance of stratigraphic plays in the undiscovered resources of the UK Continental Shelf

153

MILTON-WORSSELL,R. J., STOKER, S. J. & CAVILL,J. E. Lower Cretaceous deep-water sandstone plays in the UK Central Graben

169

MOORE, R. M. & BLIGHT, R. D. The geological exploration techniques applied by BG in evaluation of the Buzzard Field prior to discovery

187

CORCORAN,J. Application of a sealing surface classification for stratigraphic related traps in the UK Central North Sea

207

LOlZOU, N., ANDREWS,I. J., STOKER,S. J. & CAMERON,D. West of Shetland revisited: the search for stratigraphic traps

225

MCINROY, D. B., HITCHEN,K. & STOKER,M. S. Potential Eocene and Oligocene stratigraphic traps of the Rockall Plateau, NE Atlantic Margin

247

GARDINER, A. R. The variability of turbidite sandbody pinchout and its impact on hydrocarbon recovery in stratigraphically trapped fields

267

HURST, A., CARTWRIGHT,J. A., HUUSE, M. & DURANTI, D. Extrusive sandstone (extrudites): a new class of stratigraphic trap?

289

Index

301

The deliberate search for the stratigraphic trap: an introduction MATTHEW

R. A L L E N 1, G R A H A M E G O F F E Y 2, R I C H A R D & I A N M. W A L K E R 4

K. M O R G A N

3

1Shell UK. Ltd. (e-maik matthew.allen@shelLcom) 2paladin Resources plc 3Veritas D G C Ltd. 4ConocoPhillips U.K. Ltd. Abstract: This Special Publication draws upon contributions to a similarly titled conference 'The Deliberate Search for the Stratigraphic Trap - Where Are We Now?' held at the Geological Society in London during 2004. Observations in this introductory paper have been drawn from the authors' experience, talks given at the conference and papers within this volume. Specifically it is noted that by analogy to basins which are perceived to be mature for structural traps, stratigraphic traps can have substantial remaining potential. Additionally, current exploration for stratigraphic traps seems rather restricted to areas where seismic data allow the direct assessment of fluid fill and reservoir development. It is argued that the industry is probably not doing enough to learn from established stratigraphic traps to guide future exploration for such traps. Looking forward, it is suggested that the industry faces two key challenges. Firstly, the use of all available data to assess fluid type and reservoir presence in areas of unfavourable rock physics, and secondly, the development of sufficiently sophisticated predictive models of stratigraphic trap development.

Twenty four years have elapsed since the original A A P G Memoir entitled 'The Deliberate Search for the Subtle Trap' (Halbouty 1982). Since that time, the technologies employed in hydrocarbon exploration have in many respects become extraordinarily sophisticated. Seismic imaging and interpretation tools have seen significant development, wireline logging has substantially improved, digital interpretation and rapid manipulation of vast quantities of data are the norm, and interpretive approaches such as sequence stratigraphy and quantitative analysis of seismic attributes have become prevalent. This Special Publication records a number of the papers given at the conference titled 'The Deliberate Search for the Stratigraphic Trap - Where A r e We Now?', organized by the Petroleum Group of the Geological Society and held in London from May 11 th to 13 th, 2004. The conference posed the question 'Where Are We Now?' in order to examine current industry perceptions of stratigraphic trap exploration and the technologies, tools and philosophy employed in such exploration, given the changing industry environment. It was felt timely to be assessing the current state of exploration for stratigraphic traps given both the increasing exploration maturity of many onshore and shallow water basins, and a subjective perception amongst the convenors

that this maturity was leading to greater emphasis on stratigraphic traps as remaining exploratory targets. Also, the industry has moved into exploration and development in appreciably more challenging and costly environments, in particular the deepwater basins. A developed understanding of stratigraphic trapping arising from deepwater exploration programmes in seismically well-imaged deepwater sediment gravity flow deposits, and the prevalent use of seismic direct hydrocarbon detection techniques in this setting, seemed likely to offer new insights.

H o w do we define stratigraphic traps? In the opening talk of the conference, Binns (2006) references the Levorsen (1966) characterization of oil and gas fields according to three trap dimensions, namely hydrodynamic, structural and stratigraphic. This appears to offer a useful conceptual approach, but as Binns notes, there are a wide variety of unconventional traps such as basin-centred gas accumulations and sand injectites (tturst 2006) which are not readily classified by these attributes. A complementary view is that of Charpentier & Cook (2004), who characterize trapping as a spectrum from discrete 'conventional' traps through to continuous traps, such as basin-centred gas

From: ALLEN,M. R., GOFFEY,G. P., MORGAN,R. K. & WALKER,I. M. (eds) 2006.

Stratigraphic Trap. Geological Society, London, Special Publications, 254,1-5. 0305-8719/$15.00. 9 The Geological Society of London 2006.

The DeliberateSearchfor the

M.R. ALLEN ETAL.

2

accumulations. Another approach by Corcoran (2006) uses a seal based classification from Milton & Bertram (1992), whereby stratigraphic traps are characterized as poly seal traps, in which closed contours at the reservoir/seal interface do not exist or do not explain the trap, thus demanding one or more base or lateral seals. Clearly a number of definitions can be employed. We feel that the key concept is the recognition that there is a continuum between a number of end-member trapping mechanisms. Any hydrocarbon accumulation which is less than entirely dependent on structural closure, be it due to some degree of depositional pinchout, facies change, erosional truncation, diagenesis, hydrodynamics, dynamic fluid flow, or other mechanism, is likely to lead to traps with a greater or lesser degree of subtlety and hence is relevant in the context of exploration for non-structural traps.

The current state of the industry In the process of canvassing individuals and companies for papers for this conference, compiling the conference schedule, and through the conference itself, the convenors gained a degree of insight into how the exploration industry currently perceives exploration for stratigraphic traps and how it is behaving in respect of exploration for such traps. These insights can be characterized into a number of themes:

Stratigraphic traps are seen to have the most remaining potential in mature basins. In mature or maturing basins, where all but the smallest or most difficult structural traps have been identified and drilled, stratigraphic traps are often seen as holding the largest remaining prospectivity. For example, Stoker et al. (2006) believe that 50% of the UK's undiscovered resources lie in stratigraphic traps. Perceptions of the maturity of an exploration play may well be misleading if the supporting data are biased towards structural traps. Moore & Blight observed the lack of relevance of play creaming curves in describing play maturity in a Jurassic play in the UK Moray Firth, where structural traps had dominated historic drilling in that play A time lag of stratigraphic trap exploration behind exploration of structural traps was well demonstrated by Atldnson et aL (2006) with reference to the Powder River Basin

(USA). Here, a successful phase of exploration for stratigraphic traps followed earlier phases of exploration for structures based initially on surface geology and subsequently on seismic data. Similar patterns were also demonstrated by Macgregor & Miele (unpublished conference paper), with respect to the UK North Sea and deep-water West Africa. An obvious explanation of this time lag is the typically greater exploratory risk attached to such traps as exploration targets owing to greater difficulty in both accurate trap definition and in assessing sealing potential. Such traps tend only to be drilled once more simple structural traps have been exhausted. Allan et al. (2006) observed that 80% of discovered hydrocarbons in stratigraphic traps reside in North America, an observation attributed by the authors solely to a greater density of drilling in North America.

The majority o f current stratigraphic trap exploration b raking place in Tertiary rift basins and passive margins, driven largely by seismic direct fluid indications. No statistics are available to support this assertion, but it is felt by the authors to be representative of much of current stratigraphic trap exploration. Allan et al. (2006) note how some two thirds of deep-water discoveries in their studied database of fields rely on stratigraphic or combination structural-stratigraphic trapping. Such discoveries are typically based on seismic direct hydrocarbon indications. It is possible that these discoveries may be demonstrating that certain trap types in sediment gravity flow deposits are more common than previously perceived. The updip trapping of coarse clastic reservoirs in marine, lowstand canyons seems to be more prevalent than the authors previously suspected. Such traps were reviewed by Freer et al. in Mauritania and Liu et al. in Cameroon (unpublished conference papers). The Cameroon example was of an initially unpromising monoclinal slope, in which oil had been discovered in Palaeocene age channel thalweg and sheet-like turbidites contained in a 3 to 5 km wide belt and entrapped by up-dip pinchout within the channel. The Mauritanian example stressed the importance of understanding the mechanisms of sand delivery in pinpointing the areas of best reservoir development, but again heavily supported by amplitude analysis.

DELIBERATE SEARCH FOR THE STRATIGRAPHIC TRAP Assisted by outstanding 3D-based seismic imaging, a high level of geophysical sophistication can often be achieved, as demonstrated by Fervari et al. in the use of multi elastic seismic attributes to quantitatively define reservoir properties in the East Nile Delta. This unpublished conference paper showed how, in a gas sand below seismic resolution, careful integration of well and seismic data allowed quantitative prediction of in-place hydrocarbon volumes. Given the very heavy dependence on seismic data in such exploration, the paper by Donnelly et al. (2006) is a ground-breaking review of how geoscientists may be making interpretation of geophysical data more difficult through the use of common display and data search techniques. Using established principles derived from theories of visual cognition, the authors showed how interpretation performance could be improved. Given the great reliance placed by the industry on the use of colour displays to portray spatial variation in seismic attributes, Donnelly et aL's paper represents a unique assessment of the appropriateness of techniques used in a routine fashion by companies and academia. Stratigraphic trap exploration without seismic direct fluid indications is of course still taking place. Atkinson et al. (2006) demonstrate a deliberate and measured search for stratigraphic traps in Tertiary back-arc basins in Indonesia, based on the occurrence of a number of required regional indicators to localize the search for candidate traps. These regional indicators are favourable hydrocarbon charge, basin and reservoir architecture, seal quality and low stratal dips in the trap area. By contrast, Moore & Blight (2006) review a wide range of geological and geophysical techniques which were employed prior to the drilling of a single, seismically mapped stratal wedge which proved to be the North Sea Buzzard Field. These two papers serve to demonstrate that in areas of unfavourable rock physics, there is little substitute for regional and local geological understanding through a play based approach coupled with high quality seismic data.

Geoscientists have far greater enthusiasm for stratigraphic traps than do decision-makers. Perhaps it was ever thus. The authors question whether there is a communication

3

gap between on the one hand the geoscientists, who recognize that in mature basins, stratigraphic traps often offer the largest remaining potential, and on the other hand the decision-makers, who have yet to fully appreciate the advancing maturity of many basins. Alternatively, perhaps decisionmakers are rightly suspicious of perceived high risk stratigraphic traps. Citron et aL (2006) note that explorers are required to serve three main, sometimes conflicting, roles. These roles involve firstly the creative conceptualization and identification of subtle traps, which explorers must then accurately measure, and finally they must communicate the uncertainty and probability aspects associated with their characterization of the opportunity. Citron et al. review the techniques available to allow explorers to fully and accurately characterize stratigraphic prospects, and to clearly convey conclusions to decision-makers.

The industry is probably not learning enough, or attempting to learn enough, from established stratigraphic traps. Whether the techniques of Citron et aL allow decision-makers to overcome suspicions of high risk associated with stratigraphic traps is another matter. In convening the conference, it proved impossible to persuade companies to describe what has been learned from developed fields contained in stratigraphic traps. This is despite the existence in such fields of enviable datasets comprised of many wells, often multiple seismic datasets and a detailed understanding of internal reservoir architecture and limits. We speculate that this is because, at least in NW Europe, developed fields are commonly managed by teams with limited resources or limited briefs, and disconnected from individuals exploring in the same basin. We also question whether sufficient, or sufficiently detailed or appropriate work (e.g. Play based exploration) is routinely undertaken to understand the regional and local setting within which stratigraphic traps may reside. Godo (2006) demonstrated how extremely detailed work to understand every discovery (and failure) in a Miocene deepwater channel/levee system in the NE Gulf of Mexico was valuable in the discovery of a large portion of some 2 trillion cubic feet of gas in an area perceived initially to be one of relative unprospective monoclinal dip.

4

M.R. ALLEN E T A L .

Overall, industry's sophistication in manipulating and employing seismic data in stratigraphic trap exploration is relatively high, particularly where seismic data can assist in fluid identification and in assessment of reservoir distribution and quality. However, it is not clear that geological techniques have reached or aspire to comparable levels of sophistication in terms of understanding and predicting stratigraphic traps. It is possible that techniques have largely been forgotten during the 'amplitude chasing' years and are now having to be relearned. The in-depth understanding of analogue fields and relevant outcrop examples, coupled with deep insights into basin evolution and reservoir deposition that the convenors might have expected to be instrumental in successful exploration for stratigraphic traps were generally not well demonstrated at the conference.

Looking forward... In the future, it seems clear that stratigraphic trap exploration will become increasingly predominant in the worlds' mature and maturing basins. However, the industry must address two major challenges: (1) The use of all available data in a play based approach to develop deep insights which allow explorers to reduce risk on trap (seal), reservoir and charge where seismic data does not lend itself either to the ready differentiation of hydrocarbon from water, or reservoir from non-reservoir. With respect to the UK West of Shetlands basin, Loizou et al. (2006) addressed the limitations of seismic data in areas where the rock characteristics do not lend themselves to ready detection of fluid type. Further improvements in seismic data, but also a much better understanding of the geological building blocks that form hydrocarbon plays, are seen as important elements of exploration in such settings. (2) The development of sophisticated, predictive geological models that guide exploration for stratigraphic traps. The authors believe that the industry needs to adopt a more sophisticated level of geological insight before geoscientists can match their enthusiasm for stratigraphic prospects and leads with a predictive understanding, which demonstrates to decision-makers that exploration funds are being wisely spent. A good understanding of trap analogues, both subsurface and outcrop, the rigorous application of sequence stratigraphic concepts and closer

integration of well and seismic data seem in general to be areas of relatively deficient analysis at the moment. Academia has a major part to play in developing these themes, particularly in developing more sophisticated geological models. Academia and joint industry consortiums can very usefully dissect and understand well-drilled analogue traps to provide improved understanding of trap geometries and place these in an appropriate regional context. Coreoran et al. (2006) show how an understanding of stratigraphic trapping configurations in a basin can assist ongoing exploration. Similarly, Haughton & McCaffrey (unpublished conference paper) demonstrated with reference to outcrop observations and an in-depth understanding of depositional mechanisms, the range in possible style of lateral termination of turbidites against confining slopes. Gardiner (2006) expands upon this and demonstrates how pinchout variability affects reservoir behaviour though the use of reservoir models. The world's readily accessible basins are becoming rather mature for exploration, yet history shows that the challenge of basin maturity can sometimes present an opportunity, where favourable geology leads to the stratigraphic trapping of commercial hydrocarbons. Technology, good science or serendipity can allow the realization that this trapping potential exists. To meet the challenge of successfully exploring for such traps, the authors believe the geology and evolution of a basin needs to be fully unravelled. This is only likely to arise if the exploration effort is sufficiently well resourced in terms of skills, data, technology, funds and time for the rigorous integration of information and creativity to generate such insight.

References ALLAN, J.R., SUN, S.Q. & TRICE, R. 2006. The deliber-

ate search for stratigraphic and subtle combination traps: where are we now? In: ALLEN,M.R., GOFFEY, G.P., MORGAN, R.K. & WALKER, I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 57-104. ATKINSON, C.E., RENOLDS, M. & HUTAPEA,0. 2006.

Stratigraphic traps in the Tertiary rift basins of Indonesia: case studies and future potential. In: ALLEN, M.R., GOVEEY, G.E, MORGAN,R.K. & WALKER,I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 105-126.

DELIBERATE SEARCH FOR THE STRATIGRAPHIC TRAP BINNS,P.E. 2006. Evaluating subtle stratigraphic traps: prospect to portfolio. In: ALLEN,M.R., GOFFEY, G.P., MORGAN, R.K. & WALKER, I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 7-26. CITRON,G.E, MACKAY,J.A. & ROSE,ER. 2006. Appropriate creativity and measurement in the deliberate search for stratigraphic traps. In: ALLEN, M.R., GOFFEu G.E, MORGAN, R.K. & WALKER, I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 27-42. CORCORAN, J. 2006. Application of a sealing surface classification for stratigraphic related traps in the UK Central North Sea. In: ALLEN,M.R., GOFFEY, 6.19., MORGAN, R.K. & WALKER, I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 207-224. CHARPENTIER,R.R. & COOK,T. (2004). Conventional and Continuous Accumulations: a Spectrum, Not a Dichotomy. American Association of Petroleum Geologists. Annual Convention, Dallas, Texas. DONNELLY,N., CAVE, K., WELLAND,M. & MENNEER, T. 2006. Breast screening, chicken sexing and the search for oil; challenges for visual cognition. In: ALLEN, M.R., GOFFEY, G.E, MORGAN, R.K. & WALKER,I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 43-56. GARDINER, A.R. 2006. The variability of turbidite sandbody pinchout and its impact on hydrocarbon recovery in stratigraphically trapped fields. In: ALLEN, M.R., GOEFEY, G.P., MORGAN, R.K. & WALKER,I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 267-288.

5

GODO, T.J. 2006. Identification of stratigraphic traps with subtle seismic amplitude effects in Miocene channel/levee sand systems, NE Gulf of Mexico. In: ALLEN,M.R., GOFFEY,G.P., MORGAN,R.K. & WALKER,I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 127-153. HALBOUTY, M.T. 1982. (ed.) The Deliberate for the Search Subtle Trap. Memoir 32, American Association of Petroleum Geologists, Tulsa, OK. HURST,A., CARTWRIGHT,J., HUUSE,M. & DURANTI,D. 2006. Extrusive sandstones (extrudites): a new class of stratigraphic trap? In: ALLEN, M.R., GOFFEY, G.P., MORGAN, R.K. & WALKER, I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 289-300. LEVORSEN, A. 1966. The Obscure and Subtle Trap. Bulletin American Association of Petroleum Geologists, 50, 10, 2058-2067. LoIzou, N., ANDREWS,I.J., STOKER,S.J. & CAMERON, D. 2006. West of Shetland revisited: the search for stratigraphic traps. In: ALLEN, M.R., GOFFEY, G.P., MORGAN, R.K. & WALKER, I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 225-246. MILTON, N.J. & BERTRAM, G.T. 1992. Trap styles, A new classification based on sealing surfaces. The American Association of Petroleum Geologists Bulletin, 76, 983-999. STOKER, S.J., GRAY, J.C., HAILE, P., ANDREWS, I.J. & CAMERON,T.D.J. 2006. The importance of stratigraphic plays in the undiscovered resources of the UK Continental Shelf. In: ALLEN,M.R., GOFFEY, G.E, MORGAN, R.K. & WALKER, I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 153-168.

Evaluating subtle stratigraphic traps: prospect to portfolio P. E. B I N N S

Consultant, The Old Farmhouse, Broomlee Mains, West Linton, Edinburgh EH46 7BT, UK (e-mail: [email protected], co. uk) Abstract: A compilation of 85 stratigraphic traps demonstrates the variety of trapping mechanisms and the scope for developing new concepts by matching geological models with features in 3D seismic volumes. However, aspects of quantitative evaluation may discourage exploration. Investors require assurance in the form of probabilistic evaluations of risk and value but information critical to the evaluation of new stratigraphic concepts is likely to be lacking. As estimates of risk and uncertainty vary with information, prospects evaluated with radically different levels of information must be ranked with care. The requirements for quantitative project ranking and portfolio optimization have to be reconciled with the need to 'venture into the unknown'. The character of stratigraphic prospects dictates different evaluation methods from those used to evaluate structural prospects. This, together with the high degree of sensitivity of value to evaluation methodology, can also lead to inconsistencies in ranking. Within the context of a company's overall strategy and risk tolerance, organizational and cultural factors may influence prospect selection. In particular over-emphasis on quantitative methods may not have the intended effect. A common understanding, amongst technical and commercial disciplines and decision makers, of the background to quantification is essential. Factors which encourage the progression of stratigraphic prospects include a dedicated geoscience effort, a separate 'growth' portfolio of new concepts, a formal structure for progressing these and a stable organization.

Stratigraphic traps may contain significant reserves but their seismic responses may be subtle and, if the play concept is new, information critical to the accurate estimation of risk and uncertainty may be lacking. Investors, however, require systems to be in place to realistically evaluate prospects and to produce predictable returns from a portfolio. New stratigraphic concepts are commonly evaluated as high risk and are o u t r a n k e d by structural prospects in competition for funding. This paper attempts to draw together various aspects of stratigraphic trap evaluation and suggests approaches which will result in more stratigraphic prospects being drilled. It draws on a compilation of 85 proven stratigraphic traps in 40 sedimentary basins (Table 1 & Fig. 1). Many of the discoveries have been made when drilling for other objectives ('serendipity'). A very high proportion of the discoveries have been made in North America, suggesting unrealized potential elsewhere. The compilation shows that 'subtlety' is largely due to low gross reservoir thickness; low acoustic impedance contrast does not seem to be a common cause of subtle seismic response. High volumes in subtle stratigraphic traps are thus achieved through areal extent (Fig. 2). Area is the only control on volume which has

the scope to increase it significantly without creating a feature, clearly visible on seismic data. Giant fields such as East Texas (Ultimate Recovery 5.4 billion barrels; Halbouty 1991, 2003) and D a u l e t a b a d - D o n m e z (Ultimate Recovery 27.9 TCF; Clarke & Kleshchev 1992; Halbouty 2003) have proven areas of 534 km 2 and 2503 km 2 respectively. By showing the great variety of stratigraphic trapping mechanisms and their interaction with structural and hydrodynamic controls, the compilation demonstrates the scope for prospect generation based on wellresearched geological models. The basic techniques for evaluating prospects are well established (Newendorp 1975; Mackay 1996; Rose 2001). However, the critical dependence of risk and uncertainty on available inform a t i o n has received less a t t e n t i o n recently, although thoroughly discussed in the past (Knight 1921; Keynes 1936). This aspect is discussed after a review of trends in stratigraphic trap exploration. Special characteristics of stratigraphic traps which pertain to evaluation are discussed next, followed by portfolio aspects. A company's culture and internal communications have a critical impact on evaluation and these are discussed before summarizing approaches likely to lead to the maturation and drilling of more stratigraphic prospects.

From: ALLEN,M. R., GOFFEY,G. P., MORGAN,R. K. & WALKER,I. M. (eds) 2006. The DeliberateSearchfor the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 7-26. 0305-8719/$15.00. 9 The Geological Society of London 2006.

8

P.E. BINNS

g

~9~

~


500 M M B O E (Millions of barrels of oil equivalent)) discovered. However, there still appear to be a healthy proportion of moderately-sized discoveries being made. In the 1980s and 1990s about 35% of the 'elephants' were stratigraphically trapped, up from about 15% in the 1960s and 1970s (Binns 2006).

The median (P50 value) field size remained fairly constant since the 1950s at about 5 MMBOE, while the mean (average) size of fields discovered dropped precipitously from the heyday of the 1960s, but shows an encouraging upturn in the last decade to about 60 M M B O E (Fig. 3). Note also that the number of giant discoveries - those of a billion barrels or more - peaked in the 1960s, and has been declining steadily to about the same levels as we saw in the 1920s and 1930s.

Demand Yet amidst the decline in petroleum resources discovered, the demand for world energy (Fig. 4) is forecast to rise continually into the 22nd century, with oil and natural gas probably filling more than 50% of that demand over the next few decades. This strong world demand for crude oil and natural gas should likely continue as far ahead as we foresee. In fact, you might say that petroleum geoscientists, engineers and managers are 'buying time' for alternate energy sources to come on stream. If there is an extended supply shortfall, the negative consequences to the world economy and geopolitical stability are sobering indeed.

Challenge Thus, the stage is set. Amidst a declining resource base and robust onward demand, our capital, creativity and

Fig. 1. Discovered resources by decade (courtesy IHS). (BBOE, Billions of barrels of oil equivalent.)

APPROPRIATE CREATIVITY AND MEASUREMENT

29

Fig. 2. Histogram of discovery size by decade (courtesy IHS).

Global Field Sizes and Numbers

1000 "0 41 U.. imm

Jltm

q..

0 L-

41

100

E z L.

0

t."

10

0 m :S

1 1920

1930

1940

1950

1960

1970

1980

1990

Fig. 3. Mean and median discovered field size in relation to the number of giant fields discovered by decade (courtesy IHS). Note that the y-axis scale is logarithmic.

30

G.P. CITRON E T A L .

Fig. 4. Global energy demand is forecast to increase steadily, with about half needed to be supplied by oil and natural gas (from Edwards 1997).

ability are challenged to convert ideas to profitable oil and gas production. In meeting this challenge, exploration is buying precious time for the development of alternative sources. We have the capital and the ability. The main questions are: 9 9

Can we mobilize the creativity and the will to do so? Can we explore and discover efficiently?

Conferences such as this one are important generators of ideas, creative inspiration, and motivation for our business of exploration. Think of that business as a series of decisions under (hopefully) decreasing uncertainty, designed to profitably grow a company's reserve base. To make the best decisions, we need to clarify the exploration role.

Reaffirmation of the exploration role There is a growing reaffirmation t h a t exploration is clearly one of the crucial functions required to meet the challenges ahead. Let's take a look at how recent business performance indicates that reinvestment in exploration is vital to the health of our industry.

Wood Mackenzie (2003) and Deutsche Bank (2003) recently collaborated to report on the business performance of companies' exploration and acquisition efforts over the threeyear period 2000-2002. Figure 5 plots profitable production replacement by exploration discoveries (x axis, 'organic') against annualized replacement of production via profitable acquisition (y axis) during the three-year period. The bold vertical and horizontal lines indicate 100% production replacement by organic exploration or acquisition, respectively. For example, company A replaced about 150% of its annualized production during 2000-2002 from acquisitions, but only about 30% via organic exploration in 2002. On the other hand, two companies plotted adjacently (B) replaced about 60% of their production through acquisitions, and about 120% through organic exploration. Of the 11 companies shown, three failed to replace production by either route. Only one company replaced production via acquisition. However, seven companies replaced production through organic exploration. This signals the need for more companies to renew their exploration efforts.

APPROPRIATE CREATIVITY AND MEASUREMENT

31

Fig. 5. Production replacement by exploration versus production replacement by the 2000-2002 annualized acquisition efforts for 11 major oil companies (from Deutsche Bank AG 2003).

As Wood Mackenzie (2003) notes, spending more on acquisitions may not be the answer to the call for increased earnings year after year, especially with oil prices now near a 15-year high. There appears to be a limit on the viability of profitable acquisition-dominated strategies. JPMorgan (2003) shares this concern (Fig. 6). The mid-capitalization companies they routinely follow that have dominant acquisition-led strategies showed that their valuecreation rarely outpaces their cost of capital (which we infer to be about nine percent). Marko (2000), in a wake-up call report, expressed concern that the industry has so overfocused on cost cutting that value-creation may have now become a lost art. As flattering (or perhaps as ominous) as it sounds, sustained exploration - that is, building petroleum resource growth at a profit- has been a major contributor to the improving standard of living of most of the world during the 20th century. According to past A A P G President Ray Thomasson: 'Industry is recognizing that in order to grow, they need to re-focus on exploration.' For this endeavor to be successful,

profitable and sustained, we need to practice three key exploration values: creativity, measurement, and respect for the investor. Here is how those values translate to the roles we accept as E&P professionals: First, we need to be creative in generating the new exploration concepts and opportunities, involving pattern recognition, perceptive geoscience and informed intuition. Those are mostly right-side brain functions. Second, but simultaneously, we must practice responsible measurement and preservation of our estimates for effective portfolio management and the long-term benefit of our shareholders. These tasks are generally considered to be left-brain functions. There is a healthy dilemma created by the first two roles. It's hard to have the left and right side of the brain simultaneously work in concert. This is one reason why many companies have chosen to engage prospect review teams at the portfolio level. These teams are chartered to provide feedback, mentoring and suggestions on measurement techniques (reality checks, appropriate analogs, techniques applied by other teams, failure-mode analyses) and post-audit reviews based on the creative work presented.

32

G.P. CITRON ETAL.

Value Creation 10

.

tO i

to

L)

(1)

-co

o

.g=

2

to et/1

ll

LU

>. -2 ,

n

ii

"The industry has been in a cost-cutting mode for so long that it tends to be the way everything is viewed, and that needs to change." Marko (2000)

-6

Company

)"

Fig. 6. Annualized shareholder value creation for large independent oil companies. The y-axis is defined as current share price percentage above (or below) the 5 year volume weighted daily average price, reported on a compound annual growth rate (CAGR) basis (from JP Morgan 2003).

Third, in order to have informed decisionmaking and clear portfolio processes to deliver on our promises, we need to demonstrate better communication to our decision-makers and leadership of both the creative and measurement aspects of our work.

Lessons learned We have a few lessons, or insights that may help in this regard. We review these insights in the hope that they make work more interesting and productive. They come from the fields of play analysis, economics and risk aversion, and complex trap analysis. W h y play analysis? Play analysis is the consistent, systematic fullcycle economic evaluation of groups of genetically-related leads, prospects and producing fields in a geological trend. Plays can be objectively assessed for reserves, chance of success and net present value, just as prospects. While incredibly valuable in helping companies differentiate better from lesser plays, there has been,

in the past, an alarming lack of discipline in applying this type of analysis. Because of this lack of discipline, almost every major company has its share of horror stories. For example, in one large company, the chairman once thought out loud in his conference room, 'Wouldn't it be nice if we owned acreage on the Arabian Peninsula?' While only brainstorming to generate discussion, his words were misinterpreted as a directive. Within weeks, this company had its best negotiators in the capitols of Oman and Yemen doing what they do best: negotiating concession terms! As a result of the reversal and distortion of the exploration work process, they returned with very favorable terms on largely nonprospective acreage - as well as long-term commitments that set this company's profitmaking potential back several years! Most seasoned explorationists can relate analogous disastrous play campaigns that arose from superficial, poorly sequenced analysis or misplaced incentives. Because of the lack of a valid play-analysis process, many companies have suffered from poor return on invested time and money; missed opportunities to participate

APPROPRIATE CREATIVITY AND MEASUREMENT in good plays, typically because of the lack of a comprehensive comparative analysis; and the advantage gained by competition by not having you as a competitor in those better plays, thus solidifying their position as preferred partner by the host countries for repeat business.

Selecting plays is a key exploration decision Not surprisingly then, selecting 'plays to pursue' is the key exploration decision - not 'which prospect to drill.' Figure 7 shows rough estimates provided by Conoco in the mid 1990s, and plotted on a logscale, of the capital necessary to detail a prospect with seismic data, drill various types of prospects, and properly execute exploration of a new play (shooting seismic programs, concession costs, overhead, drilling, etc). The vast of amount of money, time and staff potentially committed means that plays deserve rigorous analysis, hopefully tied to company strategy. We note that companies with sustained good exploration performance tend to manifest the key attributes of broad-gauge geological basin and trend analysis, in other words, distinctive use of play analysis. By this, we mean integration of regional data, and interpretations of those data that are tied to the risk elements of the petroleum system. Skilled play analysts utilize recent discovery data, arrayed as field size distributions, as a predictive tool. Furthermore, they are able to translate the engineering

33

efforts necessary to explore the trend into a coherent economic evaluation of the trend. Disciplined estimation of costs, timing, and prices means that plays that they undertake have a high probability of performing much as anticipated. Skilful estimation of uncertain parameters is equally as important for engineers as it is for geoscientists!

Binomial probability in the petroleum system A powerful take-away from play analysis is the communication that arises in determining how many consecutive dry holes your company would tolerate in exploring a play before abandoning that effort. That number of trials is referred to as the dry hole tolerance (DHT). As a company is willing to accommodate more chances (and test the independent factors) to drill a discovery that covers sunk and pointforward costs (i.e. an economic discovery), the probability of achieving that threshold actually increases, but can never exceed the probability that the play actually exists (e.g. the play chance). There are, of course, limits to this formula as previous company forecasts of prospect success rates (PSR) less than 20% have rarely been proven to be calibrated (Citron et al. 2002). Thus, in cases where the PSR is less than 20%, a company should always ask 'has this population of prospects in our last few years' portfolios actually delivered a success rate of

Fig. 7. Range of monetary investment needed to properly define a prospect from a seismic program, drill a prospect and properly explore a Play. Note that the x-axis is logarithmic.

34

G. E CITRON

about 10% (the mid-point of the 0-20% range)?' This communication is designed to generate a key metric in play analysis, the program Pe (Fig. 8), defined as the probability that your exploration program will yield at least one full-cycle economic discovery after a certain number of trials, specified by the D H T (and also referred to as the dry hole program). Because of the success and failure probabilities contained in the formula, the equation falls within the family of binomial probability expresssions. Play chance is the chance the play exists, dealing largely with dependent chance factors that may be shared by all prospects in the play. Prospect success rate (PSR) deals with the prospect-specific characteristics in the play, and represents the percent of prospects in the play that would yield discoveries that would flow hydrocarbons, given that the play actually does exist. The diagonal line on the cumulative log-probability graph in Figure 8 represents the forecast rcsourcc . , ~ . u size distribution (FSD) for the play, and the m i n i m u m economic field size (MEFS) needed that would generate a PV > O, when fully burdened by the costs associated with the discovery well (such as seismic surveys, overhead, and concession costs). The projection of the MEFS from the FSD to the cumulative

ETAL.

probability y-axis determines the PMEFS, or the percent of that FSD that would be considered 'economic' in a full-cycle sense. This portion of the FSD is used to generate the success case value distribution of a play. Note how Program Pe has the D H T as an exponent, which dictates that this n u m b e r contributes a strong influence in the calculation. Indeed, this represents a powerful communication tool that can help your m a n a g e m e n t appreciate how many trials are appropriate in your play. Continually remind your staff that play exploration should be leveraging and updating the play analysis with the geotechnical learnings that arise from drilling within the 'dry hole program.' Let's use a simple example illustrated in Figure 9. When we graph the Program Pe versus the DHT, with the Play Chance at 0.6 and the PSR at 0.2, we note two points on the lowest curve. First, we can communicate to decision makers the dramatic increase in program Pe from 11% to 28% - with just a small increase in DHT, say between 2 a^-. u"~6. W e al~u ~'- ^ note that the curve becomes asymptotic to the play chance since the program Pe can never exceed the chance that play actually exists. Perhaps with some effort in studying a shared element indicated by hydrocarbon shows in this play, we might either c o n d e m n the play, or

Fig. 8. The program Pe relates the probability that a play will yield a full-cycle economic discovery after a prescribed number of independent trials. Play chance, prospect success rate (PSR), the number of dry holes tolerated (DHT), and the probability location of the minimum economic field size (MEFS) are required inputs. The MEFS is posted as a vertical line on the forecast field size distribution (FSD) for the play, and is read off the y-axis as the Pmefs.

APPROPRIATE CREATIVITY AND M E A S U R E M E N T

35

Fig. 9. Program Pe versus the number of dry holes tolerated, illustrating the benefit of sequential learning in a

play.

improve play chance, thereby demonstrating enhancement to the program Pe (as represented by the middle curve), well before we address notoriously prospect-specific factors such as reservoir, closure, or containment. Then, after perhaps some consideration of the independent (or local) aspects in the play, such as improvements forecast from anticipated seismic resolution for the trend, it is possible to drive the program Pe even higher. Unlike prospect analysis, in which all probabilities record a 'snapshot in time' - the knowledge that justified the decision to drill - there is a temporal aspect to play analysis. It is important to note that, with each additional item of significant new data, chance of success and reserves potential can be expected to change. While every trend is unique, at each step, the exploration process typically requires a series of decisions: 'Shall we add or reduce our equity position in the play area?' and/or 'Shall we acquire additional information?' Note how conducive this type of analysis is to examining exit strategies.

'Power of thought' exercises There is a figure in American folklore known as Rip Van Winkle, an early Dutch homesteader in the Catskill Mountains, who in the 1770s fell asleep for 20 years. When he awoke, with flowing white hair and long scraggly beard, he was amazed at all the changes he observed. Ed Capen utilized this concept when he introduced power-of-thought exercises to Arco and the industry. Imagine trying this approach with your Play scenario: It's 20 years hence, and you are flying over the Play fairway you analysed, amazed by the mature development of producing infrastructure - well heads, tank batteries, flow lines, equipment yards - built to accommodate the fabulous success of your play. With such a thought-provoking scenario, ask your team: 'What elements of the geology or geotechnology were operative to contribute to that SUCCESS?'

Now, turn the situation around; go through a contrasting scenario of abject failure. Ask the team to explain the catastrophic outcome as well. Can they imagine these extreme

G.P. CITRON ETAL.

36

outcomes? If so, where do they fit into their probabilistic assessment? Are their probabilistic envelopes wide enough? Are yours? Also remember to place your patternrecognition skills, as well as your analytical thinking, to work on the analysis of often subtle oil or gas shows, which, in effect, represent the footprint, or migration pathways, of hydrocarbons moving in the host rocks of the play. This is always a great starting point to help analyse the petroleum system, paying attention to the kitchen, and its exit routes to obvious or subtle traps.

Balanced risk behaviour Play analysis helps companies get things about right, so that stupid mistakes can be avoided. A balanced approach has best served companies to profitably grow (Fig. 10), since it tends to avoid overly passive or simplistically aggressive approaches. Overly passive behaviour can 'bleed' and diffuse a company's efforts. In this mode, very few people are assigned to cultivate new trends, those people that are assigned to regional efforts typically have too many trends to examine, and the review of such efforts tend to be micromanaged. As a result, few commitments are made, resulting largely in the loss of valuable staff time. At the other end of the spectrum, we see 'make or break' risk-seeking behaviour. In this

mode, there tends to be an over-reliance on too few trends, possibly overstaffed, with very few pre-determined milestones or management reviews. Unfortunately with this attribute combination, there is little attention to an appropriate dry hole tolerance and a tendency towards over-commitments without substantive learning. Play analysis, when conducted with distinction, is balanced, and offers the greatest probability for profitable growth. There is a concerted effort to evaluate a number of plays in a reasonable time frame that can potentially contribute to the company's goals. Reviews are timely and commitments are phased and commensurate with the company's stated risk tolerance. Rather than becoming repeated victims of the same old mistakes, companies with a balanced approach to play analysis capitalize on information archival and integrate company learnings to determine which trends need to be expanded and which plays need to be exited.

Economics and risk aversion In Figure 10 we reference overly risk-averse and risk-seeking behavior in play analysis. Understanding risk aversion leads to better management of the needed capital exposure, the determination of the appropriate discount rate for your company, and ways to diversify company portfolios with appropriate interest in joint ventures.

Characteristics of Different Risk Behaviours Passive Balanced Aggressive One on all (watch and wait)

Basket of selected opportunities

All on one

No commitment (each step contingent)

Commitment in phases optimized # of tries

Commitment to as many tries as it takes

Review/decision at each step

Review/decision at each phase

No review

Minimize cost to fail

Minimize cost to succeed (plan to win)

Minimize tries (spend to win)

(plan to fail)

Fig. 10. Balanced risk behaviour, as compared to risk-averse behaviour in the left column and risk-seeking behaviour in the right column, seeks to grow a company at a sustainable rate.

APPROPRIATE CREATIVITY AND MEASUREMENT Whenever possible, always select the lowest possible discount rate for your company. While this number is usually pre-determined by the CFO's office, rather than the E&P senior VP, be aware that discount rates that exceed a company's average weighted cost of capital actually discriminate against the exploration needed to grow a company because they disproportionately jeopardize the value of 'out-year' cash flows. High discount rates are the enemy of successful play analysis. Elevated discount rates are not an acceptable proxy for risk which is properly handled via expected value, portfolio diversification, and risk-aversion methodologies. The under-utilized field of risk aversion teaches us that there are ways to determine, for every venture, the appropriate equity position to target for your company, depending upon c o m p a n y budget. Believe it or not, this is superior to the 'feels about right' determination that often occurs in high level m a n a g e m e n t decisions and negotiations. The optimum working interest (OWI) is that share in a venture which maximizes the riskadjusted value of that opportunity (Fig. 11). As you can see from the formula, there is nothing exotic about the inputs, except for a company's risk tolerance (RT). This number, reported in millions of currency units is typically closely related to a company's exploration budget. To get an idea of your

37

company's RT, imagine that anticipated capital cost at which your management says, 'Let's get a partner for this, it's too risky for us at 100%.' For many risk-balanced companies, RT turns out to be roughly a fifth of their annual exploration budget. C o m p l e x trap analysis: d e p e n d e n c y a n d variance Our last set of insights comes from analysis of complex prospects, where we focus on geological p h e n o m e n a that could create dependency between prospect areal extent and reservoir thickness. Imagine a stratigraphic trap in cross-section, perhaps with a slight roll-over near the structural high point (Fig. 12). Here, the exploration team can make a clear case that, as the prospect area increases, the effective net reservoir thickness almost certainly increases as well. If each dimension increases proportionately, we say that reservoir thickness is to some degree dependent upon area. Let's examine the impact on the reserves distribution. First, let's examine the input distributions of area and net reservoir thickness (Fig. 13), and note that the P10/P90 ratio of each is 10. Multiplying these input distributions gives us an impression of the reserves potential (Fig. 14). Dependence always increases the variance, or dispersion of the product output distribution,

Fig. 11. Graph of risk adjusted value (RAV) versus working interest for a venture. RAV is the chance weighted value for an opportunity discounted by a company's or decision maker's utility, or risk aversion. RAV = -RT*In[Pc*e(-ewRT) + (1--Pc)*e(C/RT)]where RT = company's risk tolerance, in millions of dollars; PV = net present value of the opportunity, in millions of dollars; Pc = probability of commercial success (decimal), c = cost of the opportunity, in millions of dollars (from MacKay 1995).

38

G. R CITRON ETAL.

Fig. 12. Cross-section of a hypothetical stratigraphic trap, with areal extent and thickness axes.

relative to the same calculation without any dependence in the system. In the case where the thickness distribution is totally independent of the area distribution, the resulting product (labelled 0% dependent) has a P10 of about 50 000. In the case where the thickness distribution is totally dependent on the prospect area distribution, the resulting product has a P10 of 100 000, and is shown by the line with the greater slope. On each, the circle represents the average or mean (often referred to as the expectation) of each line. The mean in the dependent case is more than twice as great as the mean of the i n d e p e n d e n t case, which arises from the

increased dispersion expressed here with the flatter slope. In other words, in the totally dependent case, your computer software takes the P10 of the thickness and forces it to be multiplied by the P10 of the area and places the product at the P10 of the output distribution. In the independent case, you never know, a priori, what percentile of the thickness can be matched with the P10 area. Mathematically, in the independent case, the product of two P10 values falls at about the P4 of the product, leaving little room for higher upsides or lower downsides. Any specified partial dependency between area and thickness would result in a product distribution line with the same median, falling somewhere between the lines representing the ends of the dependency spectrum (total or zero dependence). Dependence always increases the variance, or dispersion, of the product output distribution, often dramatically, relative to the same calculation without any dependence in the system. Thus, we need to clearly articulate the geological p h e n o m e n a to properly characterize the reserves distribution.

Complex trap analysis: when chance elements change with prospect scope Our last insight concerns the chance assessment of this stratigraphic trap as we investigate how large it can be (Fig. 12). Let's assume that the reserves associated with the small four-way

Fig. 13. Plot of the thickness and area distributions of stratigraphic trap in Figure 12, plotted on a cumulative log-probability graph.

APPROPRIATE CREATIVITY AND MEASUREMENT

39

Fig. 14. Distribution of the area x thickness product, considering both 100% dependence and 0% dependence between the input distributions. Note how the slope and mean of the dependent case are greater than the independent case. closure falls short of the current commercial threshold. Moreover, the chance elements change with the scope of the prospect. In this case, as the prospect gets bigger, we are concerned that another nearby sand body may shingle onto the prospect, creating a possible seal breach. If the chance elements change across the full range of prospect outcomes, as they do here (larger outcomes have a decreased probability of appropriately sealing next to the adjacent sand body), we need to account for the overall chance of the entire prospect. Perhaps the clearest way to demonstrate this approach is with a decision tree. Here (Fig. 15), the analyst has illustrated the chance factors associated with the small (P90), median (P50) and large (P10) reserves outcomes. The chance factors in each row are multiplied to represent the percent probability (in bold print) of meeting or exceeding the reserves sizes of P90, P50 and P10, respectively. The reserves values are shown in the white cells beneath the MMBOE label. The product of the lower row of chance numbers (with shaded background) represent the traditional, base case probability of geological success, where the reservoir and closure chance factors are typically assigned to meeting the P90 of the

average net pay and productive area distributions, respectively. In other words, to better characterize the chance of this complex prospect, we take the specific, representative size outcomes, chanceweight those values, and add them up to generate the prospect's Expected Mean Reserves (EMz), shown on the far right column. In this example, the EMz is 6.80 MMBOE. The specific non chance-weighted (i.e. success case) reserves outcomes cannot be added. Rather, the mean of this reserves distribution can be calculated via Swanson's approximation (Mz). Swanson's approximation adds 0.3 • P90 reserves plus 0.4 x P50 reserves plus 0.3 x P10 reserves, in this case yielding a mean of 34.30 MMBOE. The size-weighted chance for a complex prospect (Pg) is represented by ratio of the EMz to the Mz, or EMz / Mz, which here is 20%. This percentage approximates the proper chanceweighting of the success case value in any expected monetary value calculation. When chance elements change with the scope of the prospect, the probability numbers need to reflect that change, to properly characterize the opportunity in relation to other projects considered for inclusion in the portfolio, and for proper prediction of portfolio outcomes.

40

G. E CITRON E T A L .

Fig. 15. Decision tree approach to demonstrate the chance of achieving successively larger outcomes for the stratigraphic trap introduced in Figure 12. The size-weighted chance is the sum of the chance weighted outcomes, EMz (here 6.80 MMBOE), divided by the Swanson's mean (here 34.30 MMBOE), defined as 0.3 (PIO + P90) + 0.4(P50). Condusion

We believe creativity, m e a s u r e m e n t and communication are three attributes needed in the deliberate search for stratigraphic traps. We have learned a number of lessons to cultivate these attributes, and note that they admirably serve the i m p o r t a n t function of exploration: 9

9

9 9

Execution of a play is all about learning. Be sure to engage in a balanced approach that includes thought experiments, and recognize as early as possible the footprints of hydrocarbon migration established by the geographic and stratigraphic distribution of hydrocarbon shows. In play analysis, your dry hole tolerance strongly influences the probability your initial program will yield an economic discovery. Low corporate discount rates encourage new-play exploration. Determine the optimal working interest to appropriately diversify your efforts across several plays.

9

9

Geological phenomena that create dependency always disperse the reserves outcomes such that greater dependency increases the mean of the reserves distribution. Understand and communicate the impact of stated dependencies. W h e n faced with prospects where the chance elements change across the scope of reserves sizes, a decision tree approach can help to clarify and communicate the complexity. In such cases, a size-weighted chance is a more appropriate probability weighting of the success case value in an expected monetary value calculation.

We hope these insights serve your roles of creativity, measurement and communication in the deliberate search for the stratigraphic trap. We wish you every success towards that end. References

BINNS, P. 2006. Evaluating subtle stratigraphic traps: prospect to portfolio. In: ALLEN,M.R., GOFFEY, G.P., MORGAN, R.K. & WALKER, I.M. (eds) Deliberate Search for the Stratigraphic Trap.

APPROPRIATE CREATIVITY AND MEASUREMENT Geological Society, London, Special Publications,

254, 7-26. CITRON,G.E, COOK,D.M. & ROSE,ER. 2002. Performance Tracking as a Portfolio Management Learning Tool. AAPG 2002 Annual Meeting, March 10-13, 2002, eposter, pg. A32. DEUTSCHE BANK. 2003. Global Oils: Sustainability Costs Money (Part 1: Analysis), Global Equity Research report. EDWARDS,J.D. 1997. Crude oil and alternative energy production forecasts for the twenty- first century: the end of the hydrocarbon era. AAPG Bulletin, 81, 1292-1305.

41

JPMORGAN. 2003. Long-Term E&P Value Creation. North American Equity Research report (September 4, 2003), 10 pp. MACKAY, J.A. 1995. Utilizing Risk Tolerance to Optimize Working Interest. SPE paper 30043, HEE Symposium, Dallas, 103-109. MARKO,W.A. 2000. 2000 Global Competitive Assessment. SPE paper 62886. WOOD MACKENZIE. 2003. Value Creation Through Acquisitions. Horizons Energy Issue 14, October 2003.

Breast screening, chicken sexing and the search for oil: challenges for visual cognition N. D O N N E L L Y 1, K. R. C A V E 2, M. W E L L A N D 3 & T. M E N N E E R 1

1Centre for Visual Cognition, School of Psychology, University of Southampton, UK (e-maik n. [email protected], uk) 2Department of Psychology, University of Massachusetts, USA 30rogen Limited, 175 Southwark Bridge Road, London SE10ED, UK Abstract: Interpretation of images of the Earth's subsurface is a process whereby humans perceive and categorize visual features derived from seismic data. The seismic data are presented in the form of vertical slices showing points of change in some variable being measured (e.g. acoustic impedance) and horizontal slices showing surfaces interpolated between values at a particular time or horizon across multiple vertical slices. These images are usually highly complex and their nature has been determined largely by the technical capabilities of the hardware and software of the imaging technology. Because of these constraints, we argue, images do not convey information as readily as they could. We believe that these images could be more informative if they were constructed and tailored with known properties of the human visual system. Furthermore, little or no consideration has been given to the training and selection for image interpretation vis-h-vis the fundamental psychological skills that distinguish good from poor interpreters. In this paper we argue that tailoring images to the human visual system and developing working practices that eliminate biases will improve the detection of subtle features related to hydrocarbon traps. Furthermore, establishing training procedures that enhance the visual system's ability to detect and encode hydrocarbon traps, and creating selection procedures that select individuals with excellent visual imagery skills will also facilitate performance.

The apparently effortless act of seeing and interpreting information coming into our eyes hides a fact. Visual processing is one of the most complex tasks achieved by humans. Large areas of our brains are devoted to the task of processing visual information so that we can see and comprehend the visual world about us. Our visual skills include not only recognition, but also the guidance of actions and the creation and manipulation of visual images. In forming visual images and the mental tools for their manipulation, we create the basis for visual problem solving. The goal of this paper is to understand the processes involved in image interpretation for hydrocarbon deposits, given the basic visual data presented to image interpreters using imaging (for example seismic i n t e r p r e t a t i o n ) software. This understanding will be framed using basic principles of the human visual system that have been gleaned from many years of psychological experimentation. In exploring the problems of image interpretation, we will examine two related issues. First, why is the process of image interpretation so hard given the image data available? Second, how might image interpretation be improved? In bringing together the two

disparate areas of visual cognition and image interpretation for hydrocarbons, we accept at the outset that the treatment of both will be simplistic. Nevertheless, by showing how image interpretation might be advanced by greater consideration of the nature of h u m a n visual processing, we will demonstrate how technical advances in imaging need to be tuned to the humans interpreting the images. We contend that it is possible to improve the accuracy, reliability and speed of image interpreters by considering the visuo-cognitive elements of the job they perform. To our knowledge, nobody has made such an attempt to do so previously, and so this paper is the first attempt to understand issues of visual cognition in respect of image interpretation in geophysics.

Why is image interpretation so hard? In the search for h y d r o c a r b o n deposits, the primary data are derived from the collation and processing of artificially-generated acoustic waves that have been reflected by inhomogeneities in rocks below the Earth's surface (Brown 1999). Vertical and lateral changes in the properties of the rocks and the fluids

From:ALLEN,M. R., GOFFEY,G. E, MORGAN,R. K. & WALKER,I. M. (eds) 2006. The DeliberateSearchfor the StratigraphicTrap.Geological Society, London, Special Publications, 254, 43-55. 0305-8719/$15.00. 9 The Geological Society of London 2006.

44

N. DONNELLY ETAL.

contained within them influence the character of the reflected seismic energy; these changes are displayed through colour transitions and variations in the image. Image interpretation is typically based on two different image types, vertical profiles and horizontal or 'topographic' surfaces (representing slices through the 3dimensional seismic data volume.) Vertical slices typically contain more or less continuous coloured lines of high spatial frequency representing boundaries of change in subsurface acoustic properties as recorded in the seismic data. Horizontal surface images (e.g. time and horizon slices, maps) contain coloured regions of common data values (amplitude, travel-time, etc.) that vary irregularly over space. Crudely speaking, the vertical images enable the detection of geological features (faults, folds, stratigraphic discontinuities and so on) that have diagnostic spatial properties (Selley 1998). In addition, the patterns of seismic reflections surrounding these spatial features provide significant information through the spatial relationships, positions of lines, or discontinuous bands displayed in different colours. Increasingly, interpretation requires identification of extremely subtle geological relationships imaged through equally subtle, and ambiguous, changes in the seismic data. There is, however, no simple visual process that will allow detection of these subtleties. In order for humans to detect and classify significant features of the types linked to the presence of hydrocarbon traps, they must construct specialized visual routines (Ullman 1985). These routines will perform basic search and analyses of images. We begin by describing some of the properties of these routines.

Searching for targets Interpreting a seismic image is to some important degree a visual search task. Visual search has been a popular topic within experimental psychology, and this work has led to a distinction between search for a target that is defined by a single feature, and search for targets defined by conjunctions of features (Treisman & Gelade 1980). Features are simple visual properties, such as orientations, colours, etc., whereas conjunctions are combinations of these basic features; for example, lines that are both red and horizontal. A long series of carefully constructed experiments in visual search have been conducted to uncover the psychological processes associated with detecting features versus conjunctions. In the most widely used visual search paradigm, participants are asked

to report as quickly as possible whether a specific target has been presented amongst a set of distractors. Two things are manipulated in this type of experiment: the relationship between the target and distractors, and the number of distractors. The evidence shows that if the target can be readily discriminated from distractors on the basis of at least one feature, then search for the target is fast and the speed of target detection increases little with the number of distractors (Treisman 1986; see Fig. la). In contrast, if targets and distractors can only be differentiated by conjoining features (see Fig. lb) then the time necessary for target detection increases with the number of distractors. One conclusion drawn from these experiments is that the presence of a simple feature is detected across the visual field without visual attention, and that visual attention is required to detect conjunctions of features. However, visual attention is limited in the sense that it cannot be applied to the whole visual field simultaneously. The requirement to use attention to detect conjunctions of features places a fundamental restriction on all complex visual tasks, because spatial attention can only be allocated to a limited number of locations at a time and takes a finite time to move to other spatial locations (Cave & Bichot 1999). Consequently, perceiving conjunctions of features takes time and requires active exploration of the visual field. The kinds of structures being searched for by image interpreters in vertical images are conjunctions of features. With respect to the search for hydrocarbons, we can understand part of the problem faced by image interpreters as resulting from the need to move visual attention around images in order to detect conjunctions of features that might indicate important subsurface relationships.

Colour categorization Considering only the basic properties of these vertical images and how they must be encoded by the human visual system does not, however, come close to providing a full understanding of why these images are visually so difficult to interpret. A further reason why vertical images are so difficult to interpret relates to the arbitrary use of colour in images (for some discussion of how colour should be used, see Brown 1999, Chapter 2.). By arbitrary, we do not mean that colour is applied without reference to a colour scale, or that some thought has not been given by designers into the colour scales themselves (Ware 1988). However, colour perception is a complex issue and scales should

VISUAL COGNITION AND IMAGE INTERPRETATION

45

Fig. 2. Colour (Hue) is a linear construct but is perceived in a non-linear fashion such that colour categories are perceived with points of transition between categories. The graph at the top of the figure illustrates the difference (in frequency) required between two colours before they are perceived as being different. Note how the ease of discrimination changes within and across categories. Reprinted with permission, Clark University Press.

Fig. 1. Example displays of visual search experiments designed to show how targets defined as a simple feature difference relative to distractors 'pop-out' without attention (a), while targets defined as a conjunction of basic features require attention (b). (a) the target (black horizontal bar) is detected as easily in the top and bottom panels whereas this is not the case in (b) (light grey vertical bar). be designed with full consideration to all issues of perceived similarity and dissimilarity of the colours they contain. A cursory understanding of how humans encode colour i n f o r m a t i o n demonstrates why this is so problematic. For the m o m e n t we will satisfy ourselves with demonstrating a single point: while colour is a continuous variable with the visible wavelengths stretching from around 300 to 700 nm, it is perceived in distinct categories (red, green, blue, yellow etc) separated by apparent boundaries. Furthermore, two colours that are separated by, for example, 10 nm will be easily discriminated if there is a colour boundary between them, but very difficult to discriminate if they are both within the same colour category (Fig. 2). In other words, the discriminability of pairs of colours is not a linear function of the difference

in their wavelengths, even for typical viewers with perfectly n o r m a l colour vision. Hence, colour scales that do not take account of this basic fact, for example seismic workstation colour scales, risk making some differences imperceptible when other differences of equal magnitude are clearly perceptible (for a further discussion of this point, see Della Ventura & Schettini 1993). For this reason, the unprincipled use of colour adds visual noise to images that, even without that noise, would be difficult to interpret. With respect to horizontal surface images, the issues are somewhat different. These images are used to reveal the spatial extent of geological systems whose presence has been identified initially in vertical profiles. The issues relating to these images are primarily to do with determining a story through which all aspects of the image fit within a geologically consistent account. We will turn to this shortly, but before doing so, a further impact of colour categorization on performance needs to be described. The horizontal surface images typically used by image interpreters demonstrate an invidious aspect of the failure to appreciate the effects of colour category boundaries. The problem we refer to is the inappropriate segmentation of images into figure and ground.

46

N. DONNELLY ETAL.

Whenever humans perceive an object (figure) it appears in some context (ground; see Fig. 3). We perceive only the figure as having shape, whereas the 'hole' in the ground covered by the figure is 'shapeless'. A n o t h e r important fact is that only a single object, or group of related objects, can be perceived as figure at any one moment in time. Finally the boundary distinguishing figure from ground is of particular interest to the visual system, as it represents the limit to the spread of attention at any particular moment. Putting these facts together, we can state that attention acts as if it spreads across the surface of objects to define its shape and extent, being delimited by the perceived edges of figures. When imaging surfaces to establish topography, one thing that should be avoided is the arbitrary segmentation of images by boundaries that do not reflect large differences in the real topographical data. Unfortunately, the wrong choice of colour scale for an image can lead to the perception of sharp boundaries where they do not belong. A n example of this effect is shown in Figure 4. In studying this image, note how attention is drawn to the red, blue and white regions sequentially, and how difficult it is to see these regions as being part of the single

continuous surface that the data, in reality, represent. Furthermore, note how difficult it is to discriminate within the blue, red and white regions, and how easy it is in contrast to discriminate between blue and red or between red and white regions. The difference between two points that are shown as two different shades of red may be just as great as the difference between a red point and a blue point, but the colour scale attaches much more salience and importance to the red/blue difference. By using the red-blue-white cotour scale, the continuous surface described by the data has been transformed into a surface of multiple figures and grounds in the image, making it difficult to visualize what the original data must have been. The location of real boundaries is critical to understanding the architecture and history of a geological system; the introduction of artificial boundaries through artefacts of colour coding has the potential to bias or distort the interpretation process.

Fig. 3. The Rubin's face/vase figures demonstrate how we cannot perceive the two faces at the same time as the vase.

Fig. 4. Example of an image of a continuous surface coded to create segmented regions that do not exist in the data. (Courtesy Amerada Hess)

The influence of top-down knowledge on perception It has long been known that experience and expectancies influence image interpretation, most obviously in the viewer deciding where to move their eyes next. However the influence of top-down knowledge on the visual routines used in image interpretation is much more pervasive. This can be demonstrated using pictures that

VISUAL COGNITION AND IMAGE INTERPRETATION can be interpreted in two different ways. The ambiguity in some of these stimuli is created by visual noise, while in other cases the interpretation can be changed by changing figure-ground assignment. A particularly good example of this is seen in the Salvador Dali picture presented in Figure 5. This figure can be interpreted in two ways. At first, the viewer will probably interpret it as a scene with various women and men set amongst a pair of arches. Under the second interpretation, the outer parts of the figure are similar to the first, but there is a large face set into the right hand arch. The assignment of what is figure and ground serves to determine whether the face (a 'Bust of Voltaire') is detected in the image or not. The fact that figure can emerge from ground is an interesting p h e n o m e n o n in its own right. However, it is a second aspect of this effect that is of primary interest in the current context. Having detected the 'Bust of Voltaire' in the image, now try to interpret the scene as if it is not there, i.e. as if the face was not seen when first viewing the image. It is difficult, if not impossible, to know that the 'Bust of Voltaire' is in the image but to treat the elements of the face as belonging to different people and so to not see Voltaire. The point of this demonstration is to show that one cannot readily let go of hypotheses, once formed, concerning the perceptual organization of a scene; 'unlearning' is virtually impossible. Therefore, when interpreting ambiguous images, one must avoid being drawn towards

47

interpretations of scenes based on hypotheses that c a n n o t be supported by real features present in images. Of course, it might be supposed that the ability of expectations to change image interpretation only arises in the context of situations designed to play perceptual tricks. Unfortunately this is not the case, as illustrated by a case from radiology. Berlin (2000) reports on a radiologist convicted in court for negligence for missing a visually-identifiable tumour on a chest x-ray. The accusation was based on the fact that a patient was diagnosed with a lung tumour three years after having a chest x-ray. When the patient's lawyers showed the original chest xrays to a number of 'expert' radiologists and asked the question 'do you see the tumour?', they confirmed that the tumour was present and visible. These experts had the benefit of knowing about the tumour's presence before ever viewing the image, and thus were not able to view the image with the uninformed mindset of the original radiologist when he first examined the image. The experts' judgments could reflect a bias from hindsight. That is to say that the clarity of the tumour in the original chest x-ray became more striking in the context of the subsequent x-ray taken three-years later. Hindsight of this sort could potentially bias observers to indicate the presence of a target on the basis of minimal evidence. One would expect this response bias to raise the number of correct detections ('hits') while also raising the

Fig. 5. The 'Bust of Voltaire' by Salvador Dali. Original is in colour. Reprinted with permission.

48

N. DONNELLY E T A L .

number of times that a target is reported without actually being present ('false alarms'). Recent studies, however, suggest that in some circumstances hindsight may increase the accuracy of perception without increasing the number of errors. In a study of medical students detecting visual features relevant for determining diagnoses, Brooks et al. (2000) showed participants images of patients whose diseases left marks or discoloration of the skin. The task was twofold: first, to report the number of clinically-relevant visual features present in each image, and second to report which of a list of six pre-specified features were present in each image. The list of six features composed two features that were present, two that might have been present given the condition but were not, and two features not present and unrelated to the correct diagnoses. Brooks et al. (2000) split participants into three groups. Group I was told the correct diagnoses, Group 2 was given five possible diagnoses, and Group 3 was given no information at all. The results of the experiment showed that having diagnoses in mind influenced the total number of clinically relevant features reported. Furthermore, the certainty of diagnoses increased the reporting of clinically-relevant visual features that were actually present, but did not increase the reporting of potential features that might have been present but were not actually shown. In other words, knowing the diagnosis increased the sensitivity of perception without also increasing any response biases. Apparently, some types of knowledge can influence the ratio of signal to noise in image interpretation. A study by Sowdon et al. (2000) confirms the basic finding that experience influences basic perceptual sensitivity. In their experiment they took radiographers with experience interpreting mammograms (breast x-rays), and compared their ability to discriminate subtle differences in shades of grey in pairs of dots presented on xray film (though not in the context of mammograms). Note here that these radiographers have spent many years studying grey-scale images for abnormalities indicated by changes in value along a grey-scale. The radiographers demonstrated much more sensitivity to subtle differences in shades of grey than a control sample of younger adults. The fact that a correct hypothesis can influence the signal-to-noise ratio of perception is a real asset if the working hypothesis is correct. But what happens when interpretation proceeds under an incorrect hypothesis? One might suppose that hypotheses are commonly formed prematurely in the process of seismic image

interpretation, and potentially valid alternative geological models discarded early on, or, more likely, never identified. Furthermore, this issue is exacerbated by the typical nature of conversations between interpreters: 'you see this incised channel here?' is a more common form of question than 'what do you make of this feature?' Assuming that the working hypotheses are based on some perceptual evidence available in images, then those features consistent with the hypothesis will be amplified relative to those that are inconsistent with the hypothesis. Obviously, the impact of this adjustment in features will be to further bias interpretation along the lines of the initial incorrect hypothesis. In summary, the images being interpreted in the search for hydrocarbons tax basic human visual processing systems. One set of difficulties arise because 'targets' in vertical slice images are conjunctions of colour and form that require focal attention to encode. Another set of difficulties are introduced by inappropriate segmentation from the colour scales applied to horizontal or 'topographic' images. Finally, the working practices of image interpreters leads to geological stories being created which, whether right or wrong, accentuate some features over others. The manipulation of basic sensory signals is good if working hypotheses are correct, but extremely damaging if they are incorrect.

Improving image interpretation The issues outlined above do, we contend, lead to the task of image interpretation being even more difficult, and less objective, than it might otherwise be. In this section we propose a number of modifications to aspects of the image interpretation process. We make no claim that the issues discussed represent an exhaustive list or that the conclusions reached are necessarily correct. It is our contention, however, that these issues should be examined and that a good case exists for their consideration in a review of how image interpretation is conducted. Image construction

Important information can be obscured in vertical and horizontal image slices because the colours used to encode the data have not been carefully chosen. A colour scale for displaying values on a continuous variable (e.g. time, depth, amplitude etc.) should be designed so that equally perceptible differences in colour correspond to equal differences in the data, irrespective of the ranges of the scale being

VISUAL COGNITION AND IMAGE INTERPRETATION compared. In principle this is easy to achieve, and colour spaces have been defined which attempt to rectify for the non-linearity of human colour vision so that perceived differences between colours map onto the distance between colours in some metrical space. Examples of these colour spaces are the Munsell color set (Munsell 1929) and C I E L A B space (see CIE 2004). As long as monitors are properly calibrated, these colour spaces might form the basis of colour scales that are calibrated for the non-linearity in h u m a n colour perception (Della Ventura & Schettini 1993; Ware 1988).

49

However, b o t h the Munsell color set and C I E L A B are not especially good at accounting for human colour perception across some types of images, especially those in which colours appear in high spatial frequency patterns. Perceived saturation is generally reduced as spatial frequency increases, and this reduction is not equal for all colours, as studies from our laboratory have shown (Fig. 6). Furthermore, human colour perception is influenced by other contextual effects such as the set of colours actually shown in any specific image (Fig. 7; see Robertson 1988). In summary, we cannot just

Fig. 6. Data from two conditions of an experiment run in the University of Southampton laboratory in which participants were asked to judge the similarity of pairs of colours, taken from a set of forty colours. The spatial frequency of images was varied and multi-dimensional scaling used to construct a 2-dimensional solution shown in the polar graph for each condition. Note how the shape of the curves differ across spatial frequency, with the spatial frequency being 4 cycles per degree in high spatial frequency condition (Graph A) but varying between 0.5 to 2 cycles per degree in the low spatial frequency condition (Graph B). These data show that spatial frequency influences colour similarity (and hence discrimination) judgements. Moreover the impact of spatial frequency is not the same on red-green and blue-yellow axes.

50

N. DONNELLY E T A L . place using a correctly-calibrated workstation monitor. In the 'rainbow' image shown in Figure 8a, note how (1) attention is drawn to sub-regions of the image without regard to data values but to colours, (2) gradual changes in data values appear as categorical shifts in colours, (3) a sense of topographic variability is present despite continuous changes in data values, (4) some subtle changes in data values are transformed into categorical shifts so that the reality of the subtle changes is lost.

Fig. 7. Example of how colour context can affect perceived colour. The Figure contains only two colours. assume that it is possible to take a standard colour space and use it to create psychologically unbiased coiour scales that will suit all situations. We have been conducting experiments on human colour perception to collect the data necessary to create psychologically unbiased colour scales suitable for use in the kinds of images generated from seismic data. In these experiments, we establish similarity matrices for sets of colours that are frequently used in colour bars and presented in situations designed to replicate those in which they appear in geophysical images. These data are then analysed using multi-dimensional scaling to form a representation of how similar/dissimilar colours are from each other. These data can then be used, along with methods of interpolating between measured data points, to construct colour scales that have a number of desirable properties. First, equal perceptual differences equate to equal differences in whatever value is being represented in images. Second, they do not contain artificially-created distinct colour boundaries: points of segmentation will only occur in images when justified by the data. We are only at the beginning of understanding how the colour similarity data might assist image interpretation but some examples of images coded in traditional colour scales and the same images coded using psychologically unbiased colour bars are shown in Figure 8a-d. Figures 8a and 8c are coded using a traditional colour bar taken from an industry standard package. Figures 8b and 8d are coded using colour bars derived from psychological colour space experiments. It should be noted that proper inspection of such images should take

In contrast, the same data represented in Figure 8b appear as a continuous but changing surface where subtleties in the data are preserved and attention is free to explore the surface unbiased by the influence of the choice of colour scale. It is recognized that an initial inspection of Figure 8b may create a sense of 'discomfort' in a viewer accustomed to the more traditional appearance of the image (Fig. 8a); however, if this can be overcome, the ability of the visual system to discriminate subtle colour changes in the coding in this image enables detection of subtle features in the data represented. In Figure 8c note how the metrical values representing intensity are lost in the traditional image but can be interpreted in Figure 8d. It should also be noted that, as with the 'discomfort' factor recognized above, the initial reaction to Figure 8d might be that data variations visible in Figure 8c are not so readily discernible; arguably, closer inspection and familiarization with the image reveal that this is not the case. Of course, there are many different kinds of data imaged in the process of seismic interpretation, and some displays rely more than others on establishing subtleties in colour space. It is intended that colour coding experiments will be conducted on a variety of display types (for example, seismic attributes, seismic facies maps) that would appear to be particularly sensitive to colour coding. We contend that the proper imaging of at least some types of data will be facilitated by ensuring the colours presented in images are calibrated to the human visual system. In this section we have discussed the role of colour in improving image interpretation, and we have done so partly grounded in our own studies. However, we believe other aspects of the interpretation process are open to investigation and improvement although, as yet, we have not studied them in detail. In the final two sections we provide some introductory remarks

VISUAL COGNITION AND IMAGE INTERPRETATION

51

Fig. 8. Examples of 'surface' images coded using a traditional colour scheme (a) and a colour scheme derived from psychological colour space (b). Examples of vertical profile images coded using a standard colour scheme (c) and a colour scheme derived from psychological colour space (d). Note that precise reproduction of images on monitors and in print is dependant on calibration of devices. on these issues, while acknowledging that they await full investigation.

Training and expertise Image interpreters are highly-trained individuals who bring substantial experience to the process of interpretation. We have yet to discuss how this experience might moderate the task of image interpretation, given the images being interpreted and the architecture of the human visual system. Two issues are of interest in this regard. First, given that detection of relevant features is reliant on having knowledge of the characteristics of such features, can knowledge be imparted to trainees more quickly and with

greater impact than at present so that they produce more accurate interpretations earlier in their careers? Second, can training influence the speed and accuracy of the detection of conjunction targets in crowded visual fields, as found when subtle geological relationships are displayed in vertical profile images? Currently, little consideration is given to how new interpreters can be most effectively trained. Beyond the basics, learning how to interpret images, is in large part implicit. This raises the question of how efficient and objective this implicit learning is and whether it might be improved by explicit training. An extreme example from a different type of visual discrimination serves to illustrate the point. Until the

52

N. DONNELLY E T A L .

late 1920s, sexing day old chicks was considered an impossible task as sex-specific characteristics do not emerge in chickens until around one month. However, there is an incentive to identify male chicks early, as they are of no economic value and eat vast amounts of food before reaching one month old. In short, early chicken sexing is of value because it allows male chicks to be identified and destroyed. In the late 1920s a Japanese delegation visiting the US demonstrated a technique for sexing day-old chickens. Many of the details of the technique are of no interest to geophysicists, but in 1930s America they caused such a stir that chicken sexing schools began to develop, whose sole reason for existence was to train chicken sexers. It is commonly accepted that after training it takes chicken sexers two to three years to reach the full extent of their skills. In the normal course of events, chicken sexers operate with high accuracy, but some chickens are very difficult to categorize. Two perceptual psychologists (Biederman & Shiffrar 1987) tested some of these chicken scxcrs, who had a lifetime experience of sexing chickens, on eighteen cases known to be extremely difficult to categorize; they performed at a level of 72% correct. Biederman & Shiffrar noted that the sex of chickens in these images could be classified according to a very simple perceptual principle called the minima rule, and wrote a one-page description of how the minima rule could be used to discriminate the sex of day-old chickens. They then took two groups of na'fve participants and asked them to distinguish male from female chickens. Both groups attempted the task without instruction and scored around 60%. After this, one group was shown the instructions, taking around 1 minute to read them, and then reclassified the images, scoring 84% correct. The second group repeated the classification task without instruction and scored 54%. Therefore, this experiment demonstrated that if a simple visual rule can be used to define targets, brief explicit instructions can lead to superior performance compared to many years of implicit learning based on instances. The chicken sexing example demonstrates that visual discrimination tasks that might be considered to be the fruits of distilled knowledge can in some cases rely on simple visual features that distinguish targets from distractors. For these tasks, relative novices can achieve excellent performance with the right type of training on detecting the key features. In the search for hydrocarbons, of course, it may or may not be possible to identify simple visual

features that distinguish targets from distractors. Nevertheless, it raises the possibility of a role for training via explicit identification Of basic image features that map onto those whose presence is readily detected by the human visual system. It also emphasizes the importance of understanding 'proven' examples of the features being sought in order to try to define relevant visual features, before embarking on the data search itself. Other experiments show that search for some targets, while initially difficult, can improve qualitatively with practice. Repeated exposure to specific stimuli can lead to the development of special skills for finding a particular target. However, efficient search only develops when the task is consistent enough to allow a particular type of practice. For example, Schneider & Shiffrin (1977) used letter and digit displays to investigate changes in the speed of search for targets with practice as a function of whether the items (targets and distractors) were drawn from the same or different categories (letters ~ digits). Their results showed that if highly familiar items (letters and digits) were reliably separated into target and distractor items, then efficient search eventually developed. In contrast, if targets could be either letters or digits, then the search remained relatively inefficient. These data suggest that if some kinds of items are always search targets, while other kinds of items are consistently distractors, then visual search can become very efficient. Whether the search for hydrocarbon traps might similarly be enhanced through training is another question to be answered by research. The chicken sexing example described above serves to illustrate how categorization of image features can be improved by providing clear explicit instructions. Despite our comments regarding how basic image features might, with experience, come to be detected efficiently, it is apparent that the search for traps containing hydrocarbons is more complex than chicken sexing, and if performance is to be improved from current levels, a more carefully designed training procedure will probably be necessary. One complication comes from the wide variation across different exemplars in any given class of geological formations. No two geological features are identical, and their rendition in seismic data is ambiguous and dependent on a multitude of factors; the seismic method remains a crude tool, often incapable of imaging geological features of the scale and subtlety that is desired, yet at the same time giving the appearance of doing so. Traditionally learning of visually confusing items has employed simple

VISUAL COGNITION AND IMAGE INTERPRETATION prototypes; for example, in wartime aircraft identification or bird-watching. These training programmes are based on the assumption that once trainees have learned to make easy discriminations, they will be better able to learn the more difficult discriminations. The alternative is to start from the beginning with practice at difficult discriminations. Under this approach, trainees would practice with real instances of difficult or ambiguous seismic geometries from the very start. The issue is exactly the same as that faced by those training X-ray baggage interpreters who must identify threatening items without knowing the exact form and orientation of any threat item. In fact this issue has only just begun to receive the attention it deserves and so it is hard to give a simple recommendation. However, studies with X-ray stimuli in our own laboratory show little advantage of starting training with highly discriminable exemplars, and suggest that better outcomes are reached if difficult exemplars are shown in training from the beginning. We would argue that exactly the same is true in the interpretation of geophysical data, and that training of novices should be based on difficult exemplars, as these will eventually allow generalized representations of target features to develop despite being difficult to learn at the outset.

Mental imagery The final issue we consider in improving interpretation of geological images is visual imagery. Image interpreters often begin their process of analysis by working with sequences of 2-dimensional images which represent slices through the 3-dimensional data volume. Interpolation across these vertical slices is used to create surfaces. The representations formed in the minds of interpreters are, however, 3dimensional, with visual imagery skills used to relate image features in one slice to those in another. Only through visual imagery can one begin to understand the 3-dimensional structure intended to be communicated by successive 2dimensional slices. Visual imagery is, therefore, fundamental to the process of image interpretation. Studies of mental imagery in humans have a long history, and it is worthwhile reviewing some classic experiments to reveal some fascinating insights into the hidden world of visual imagery. In one task, participants were given two similar objects and asked to judge whether the two objects were the same or different (Shephard & Metzler 1971). Figure 9 illustrates conditions from an experiment which reveals a

53

surprising pattern in the time necessary to compare the two shapes. Specifically, there is a linear relationship between response time (RT) and angular difference between images. It is as if one object is being mentally rotated to the orientation of the other object before making a decision. The speed of rotation of the mental image is apparently constant; i.e. it takes time to change orientation in mental space with time being scaled linearly by change in orientation as in the physical world. The fact that visual imagery obeys lawful rules of the physical world turns out to be true across a number of different types of experiments. For example, Kosslyn et al. (1978) asked participants to learn a map containing seven named locations. Once they were confident they had learnt the map, they were then told to imagine the map and focus on a specific location. They were then instructed to imagine a spot moving from that location to some other location named by the experimenter and then to press a response button. The crucial result was that the time to respond was scaled by the physical distance travelled by the imagined spot. These studies form part of a voluminous literature. Taken as a whole, these experiments indicate that visual imagery is not a single undifferentiated skill, but has separable components relating to broad categories that we might call image generation, manipulation and storage. Furthermore, it is apparent that there are individual differences in the component imagery skills (Kosslyn et al. 1984). This research raises the intriguing possibility that the use of imagery skills in solving the problem of image interpretation might be used to enhance performance either by selecting interpreters who will be good at the task via some form of psychometric testing, or by training interpreters to enhance specific imagery skill components that are less well developed than others. Both of these proposals are untested at the present, but deserve attention in the near future.

Summary In this paper we have investigated the human dimension to image interpretation. Given the images available, interpretation is difficult because the necessary discriminations challenge the abilities of human visual processing mechanisms. Critical features do not automatically signal their presence and so risk being missed. Furthermore the arbitrary use of colour in colour scales creates two distinct problems. First, images contain segmentation cues where colours abut each other. The segmentation cues

54

N. DONNELLY ETAL. colour space and the difference between data values on whatever dimension is being imaged. This arbitrary mapping makes it impossible to understand values in images without explicit reference to colour bars, and this makes interpretation unnecessarily complicated. Both of these problems in using colour for data coding can be solved by using colour bars derived in psychological colour space. The p e r f o r m a n c e of image interpreters should be improved if training is provided in which features to be searched for are made explicit. The exemplars used for this training should reflect the full complexity normally found in geological images, and should not be artificially simplified. Furthermore, trained interpreters should conduct their interpretation for as long as possible in a 'hypothesis free' state. Of course, interpreters will be bound to be driven by contextual knowledge of previous successes etc. However, each interpreter should approach a new image without any suggestion or biasing from other interpreters. The expect,at; . . . . . . . t~r] hy th,~,,o,, . . . . .~"~b~, . . . . . "~o---',,~t;'~'~will change perceptual signal-to-noise ratios, blocking the likelihood of discrepant findings being reported and different (equally valid) interpretations being made. Finally, we raised the prospect that not all are born to interpret, and that differences in visual imagery skills might be used to select good from poor interpreters. The authors would like to acknowledge the help given by all at Ikon Science in the preparation aspects of this paper.

References

Fig. 9. Example of three images in a visual imagery study. Are the images in a--e the same? In experiments of this type, reaction times to make decisions about the sameness or difference of figures increase linearly with the angular difference between pairs of figures. So, deciding a and b are the same takes proportionately less time than deciding a and r are the same. The results hold whether rotations are made in 2- or 3-dimensions. attract attention to locations that are likely to be of no greater real interest than others that arbitrarily lack such cues. Second, the arbitrary nature of colour bars means that there is no relationship b e t w e e n perceived distance in

BERLIN, L. 2000. Malpractice issues in radiology: Hindsight bias. American Journal of Radiology, 175, 601. B1EDERMAN, I. & SHIEFRAR,M.M. 1987. Sexing dayold chicks: A case study and expert systems analysis of a difficult perceptual learning task. Journal of Experimental Psychology: Learning, Memory and Cognition, 13, 640-645. BROOKS,L.R., LEBLANC,V.R. & NORMAN,G.R. 2000. On the difficulties of noticing obvious features in patient appearance. Psychological Science, 11, 112-117. BROWN A.R. 1999. Interpretation of ThreeDimensional Seismic Data. 5th edn. American Association of Petroleum Geologists, Memoir 42, Tulsa, Okla. CAVE, K.R. & BICHOT,N.E 1999. Visuo-spatial attention: Beyond a spotlight model. Psychonomic Bulletin and Review, 6, 204-223. CIE. 2004. Colorimetry. CIE Publication No. 15, 3rd Edn. Commission Internationale de LEclairage, Vienna.

VISUAL COGNITION AND IMAGE INTERPRETATION DELLA VENTURA,A. & SCHETI'INI,R. 1993. Computer aided color coding. In: THALMANN, N.M. & THALMANN, D. (eds) Communication with Virtual Worlds. Tokyo, Hong Kong, Springer-Verlag, 62-75. KOSSLYN,S.M., BALL,T.M. & REISER, B.J. 1978. Visual images preserve metric spatial information: Evidence from studies of image scanning. Journal

of Experimental Psychology: Human Perception and Performance, 4, 47-60. KOSSLYN, S.M., BRUNN, J., CAVE, K.R. & WALLACH, R.W. 1984. Individual differences in mental imagery abilities: A computational analysis. Cognition, 18, 195-243. MUNSELL, A.H. 1929. The Munsell Book of Colour. Munsell Color Company, Inc., Baltimore, Maryland. ROBERTSON, EH. 1988. Visualizing color gamuts: A user interface for effective use of perceptual colour spaces in data displays. IEEE Computer Graphics and Applications, 8, 50-64. SCHNEIDER,W. & SHIFFRIN,R.M. 1977. Controlled and automatic human information processing: 1. Detection, search, and attention. Psychological Review, 84, 1-66.

55

SELLEY, R.C. 1998. Elements of Petroleum Geology. Academic Press, San Diego. SHEPHARD, R.N. & METZLER,J. 1971. Mental rotation of three-dimensional objects. Science, 171, 701-703. SOWDON, ET., DAVIES, I.R.L & ROLING, P. 2000. Perceptual learning of the detection of features in X-ray images: A functional role for improvements in adults' visual sensitivity. Journal of

Experimental Psychology: Human Perception and Performance, 26, 379-390. TREISMAN, A. 1986. Features and objects: The fourteenth Bartlett memorial lecture. Quarterly Journal of Experimental Psychology, 40A, 201-237. TREISMAN,A. & GELADE, G. 1980. A feature integration theory of attention. Cognitive Psychology, 12, 97-136. ULLMAN, S. 1985. Visual routines. Cognition, 18, 97-159. WARE, C. 1988. Color Sequences for Univariate Maps: Theory, Experiments and Principles. IEEE Computer Graphics and Applications, 8, 41-49.

The deliberate search for stratigraphic and subtle combination traps: where are we now? J. R. A L L A N 1, S. Q. S U N 2 & R. T R I C E 2

1C&C Reservoirs, Inc., 1038 East Bastanchury Road, #183, Fullerton, CA 92835, USA (e-mail: Jack.Allan@ccreservoirs. corn) 2C&C Reservoirs Ltd., Arcadia House, 15 Forlease Road, Maidenhead, Berkshire SL6 1RX, UK Abstract: Stratigraphic and subtle combination traps have a well-documented track record

as significant producing hydrocarbon resources. The majority of these success cases come from onshore USA where unprecedented drilling densities combined with a long history of hydrocarbon exploration provide a large portfolio of stratigraphic and subtle combination traps. By comparing USA-based examples with other global cases it is evident that numerous basins still have the potential for exploration success associated with these traps. This paper attempts to raise the awareness of the exploration potential of stratigraphic and subtle combination traps through four approaches. Firstly, a summary of key global statistics related to stratigraphic and subtle combination traps is provided with the intention of demonstrating that they have historically represented a key hydrocarbon resource. Secondly, a classification scheme is introduced and acts as a reference for the observations and case studies presented. The third component of this paper is to present a range of case studies, which serve to demonstrate the exploration history behind successful discoveries associated with stratigraphic and subtle combination traps. The final component is to consider the general exploration history and from this long experience highlight the key techniques necessary for the development of a successful exploration strategy.

In structural traps, closure is created by folding and faulting, while in stratigraphic traps, closure is created by stratigraphic, lithologic or hydrodynamic variations. Most stratigraphic traps are subtle, in that detection is problematical and lateral closure is difficult to prove. Some stratigraphic traps, most noticeably organic buildups (reefs and mounds) and buried hills, have fourway dip closure, are generally easy to image on seismic, and are therefore not subtle. These are not discussed in this paper. Combination traps contain elements of both structural and stratigraphic entrapment (Levorson 1954). Combination traps that occur on or near the crests of four-way structural closures are not subtle and are also not considered in this paper. Most were misidentified as structural prospects and discovered by drilling the structural crests. However, stratigraphic trapping mechanisms may combine with open structural noses and small anticlinal closures to produce subtle combination traps. O p e n structural noses provide the lateral structural seals to subtle combination traps, which are sealed up-dip by lateral depositional or onlap pinchout, subcrop trucation, or diagenetic cementation of reservoirs. Depositional pinchout of reservoirs on the flanks of small anticlinal closures can create

subtle combination traps with hydrocarbon columns far larger than could be predicted from the size of the closure. Subtle combination traps of this sort are considered along with pure stratigraphic traps in the discussions that follow. In a broad sense, stratigraphic and subtle combination traps include: (1) Traps that lack obvious four-way closure and would not have been discovered using exploration strategies designed for structural traps. (2) Traps which, if associated with a major structure, occur in an unexpected place (e.g. in a down-flank location). Building on these definitions we consider stratigraphic straps to be those traps arising from the processes of: (a) deposition, (b) erosion, (c) intrusion, (d) diagenesis, and, (e) fluid variation. Although most types of stratigraphic and subtle combination trap have been known for decades they represent an underexplored resource in comparison to structural traps. The challenges that they present to the explorationist can be summarized as detection and quantification. In the context of detection, stratigraphic and subtle combination traps can be harder to find than

From:ALLEN,M. R., GOFFEu G. P., MORGAN,R. K. & WALKER,I. M. (eds) 2006. The DeliberateSearchfor the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 57-103. 0305-8719/$15.00. 9 The Geological Society of London 2006.

58

J.R. ALLAN E T A L .

structural traps since they are more difficult to image seismically. From the quantification perspective, once identified, they are more difficult to assess with respect to their closure risk, seal integrity, potential hydrocarbon column, and hydrocarbon reserves. The combination of these challenges adds to the difficulty of selling the concept of exploring for stratigraphic and subtle combination traps to management and investors. Despite these challenges, the opportunity presented by stratigraphic and subtle combination traps is significant. An insight into this global exploration potential can be gained by reference to Figure 1 and Figure 2, which are based on a review of 174 fields worldwide that are associated with stratigraphic and subtle combination traps as portrayed in Figure 3. Figure 1 includes the in-place hydrocarbons for the combination structural-stratigraphic accumulation of Prudhoe Bay ( > 1 2 B B O (Billion barrels of oil)), the stratigraphic trap of East Texas field (6 BBO), and the huge Hugoton accumulation, which is the largest US gas field (27 TCFG (Trillion Cubic Feet of Gas)). Figure 2 demonstrates that hydrocarbon column height may attain values in excess of 2000 ft. Such summary figures are important tools in challenging the common perception that stratigraphic and subtle combination traps are

associated only with small accumulations with limited hydrocarbon columns. Figure 3 shows the global distribution of stratigraphic and subtle combination traps that form the basis of this paper. It can be clearly seen from this Figure that the bulk of stratigraphic and subtle combination traps reside in North America. However, from the perspective of trapping mechanism potential there is nothing geologically unique about North America. The disproportionate number of stratigraphic and subtle combination traps is attributed to the extremely high drilling density, which has resulted from an extensive and long lived US domestic oil industry. From these observations it is clear that significant opportunity must exist in basins outside North America. This statement can be further appreciated by recognizing that it is within mature basins that stratigraphic and subtle combination traps represent real exploration opportunities. This opportunity arises in mature areas as most structural traps have been drilled, an infrastructure exists, and a knowledge base has been accumulated, from which cieative exploration ideas can be generated. A further important consideration is that, as part of the research for this paper, an analysis of 119 global deepwater discoveries showed that two-thirds of these are stratigraphic and subtle combination traps. In order to further develop the concepts, the

Fig. 1. In-place hydrocarbon volume distribution for the stratigraphic and subtle combination traps examined in this paper. Only traps for which reliable data were available are shown.

STRATIGRAPHIC AND SUBTLE COMBINATION TRAPS

59

Fig. 2. Total hydrocarbon column height (oil and gas) distribution for stratigraphic and subtle combination traps examined in this paper. Only traps for which reliable data were available are shown.

Fig. 3. Global distributions of the studied 174 stratigraphic and subtle combination traps examined in this paper. following section is intended to summarize prior trap definitions and introduce the definitions applied by the authors of this paper in assessing the exploration potential of stratigraphic and subtle combination traps.

Classification of stratigraphic traps Previous classifications A trap is a geometric configuration of permeable reservoir and less permeable seal rocks which,

60

J.R. ALLAN E T A L .

when combined with favourable physical and chemical properties of subsurface fluids, can allow hydrocarbons to accumulate (Vincelette et al. 1999). The trap is limited by several intersecting surfaces, which include boundaries between solids (e.g. reservoir and seal) and boundaries between fluids (e.g. oil-water and gas-water contacts). The geometric configuration of these boundary surfaces defines the trap. The petroleum trap concept originated in 1844, when William Logan recognized that oil seeps in the Gaspe region of eastern Canada were associated with anticlines. However, it was not until 1885 that the application of the trap concept to petroleum exploration began, when I.C. White stated the anticlinal theory of oil accumulation in print and others began to use the idea to look for oil in surface anticlines (Levorsen 1941, 1954). Since then the concept of trap and trapping mechanisms has evolved (Clapp 1929; McCotlough 1934; Wilhelm 1945; Levorsen 1954; Rittenhouse 1972; North 1985; Biddle & Wielchowsky 1994; Vincelette et al. 1999). It was not ,,.,~1 l o ~ ~r,,~ tevorsen drew special attention to stratigraphic traps. The concept of stratigraphic traps was built on by Sanders (1943) who described several types of structural-stratigraphic combination trap. The next decade saw the development of two seminal trap classifications. Wilhelm (1945) divided petroleum reservoirs into five major categories and Levorsen (1954) integrated the work of many of his predecessors into a simple yet elegant classification scheme. Leverson's scheme has influenced geological thinking up to the present day as most modern trap classifications recognize Levorson's categories. The key to Levorsen's classification is that he defined two trapping end-members: structural traps, in which closure is created by local deformation (folding and faulting) and stratigraphic traps, in which closure is created by stratigraphic, lithologic or hydrodynamic variations (e.g. facies change, depositional pinchout, unconformities, diagenetic changes). Between these two endmembers lie a continuum of combination traps, in which both structural and stratigraphic elements are necessary for hydrocarbon entrapment. Levorsen further divided stratigraphic traps into primary (depositional) traps, secondary (unconformity) traps, and fluid (hydrodynamic) traps. A later, more comprehensive attempt at classifying stratigraphic traps was undertaken by Rittenhouse (1972). uxxLxx

a_3Ju

LXXO. L

Table 1. Organization o f major subdivisions (systems and regimes) in the trap classification o f Vincelette et al. (1999) Structural trap system

Stratigraphic trap system

Fluidic trap system

Fold regime Fault regime Fracture regime Depositional regime Erosional regime Diagenetic regime Pressure regime Temperature regime Fluid-composition regime

Vincelette et al. (1999) developed a classification that combines Levorsen's ideas with the organizational structure of the biological classification scheme, which subdivides all life on Earth into kingdoms, phyla, classes, orders, genuses, and species (Curtis 1983). Employing this approach, structural, stratigraphic and . . . . . . . . . . r, ~ , ~ m s are ~u~,dv~lu,~d into three trap regimes (Table 1), which are further subdivided into classes and families based on geological processes, trap geometry and composition and genesis of the traps. Combination traps of any sort can be described simply by combining elements from the various categories. Although complicated, Vincelette et al.'s classification scheme enjoys the advantage of a logical yet flexible structure.

Classification applied in this paper In this paper, we apply a classification scheme for stratigraphic and fluidic traps that is modified from Vincelette et al. (1999). A few major differences between this classification and Vincelette et al.'s (Table 1) classifications are apparent. Specifically we combine stratigraphic and fluidic traps into one system. We also include fracture traps, which lack structural expression, as part of the diagenetic regime of the stratigraphic-fluidic trap system rather than part of the structural trap system. Traps produced by fracturing occur where oil or gas are entrapped in fractures in low permeability lithologies such as shale, chalk, chert or coal. They should not be confused with fractured reservoirs within structural or stratigraphic traps. These concepts are summarized in Figure 4, which provides an overview of the trapping

Fig. 4. Stratigraphic trap classification as applied in this paper. Combination traps can be described by combining any of the stratigraphic or fluidic elements here with various structural elements.

S T R A T I G R A P H I C A N D SUBTLE COMBINATION TRAPS

61

62

J.R. ALLAN E T A L .

system, regime, class and family as well as schematic cross-sections of each trap type. As in Vincelette e t a l . ' s classification, combination structural-stratigraphic traps can be described by combining structural elements with any of the stratigraphic or fluidic elements shown in Figure 4.

History of strafigraphic and subtle combination trap exploration The 174 stratigraphic and subtle combination traps in this report were discovered between 1885 and 1996. This section represents an attempt to summarize this history and from that summary distill some key learning points that can be used to better focus exploration efforts. Between 1885 and 1929 only 12 stratigraphic and subtle combination traps where discovered (Fig. 5). All were found in North America and Latin America and were very shallow and quite large (Figs 6 & 7). They were discovered using surface geological mapping, by drilling adjacent to oil or gas seeps, or by accident while drilling for other objectives (Fig. 8). It must, of course, be born in mind that the stratigraphic trap concept had not yet been developed at that time. Thirteen of the studied traps were discovered during the 1930s (Fig. 6). Once again, all were in North America and Latin America and most were very large (Fig. 7). During the 1940s 15 of the studied traps were discovered, again, all in North America and Latin America (Fig. 6). Smaller traps (Fig. 7) were found at greater depths, as seismic and subsurface geology began to be used more frequently (Fig. 8). However, more than half of the traps were found by accident, indicating that a systematic search for stratigraphic and subtle combination traps had not yet begun. In fact it was not until the 1950s that a boom in the deliberate search for stratigraphic and subtle combination traps occurred (Figs 5 & 6). During this decade 36 were discovered, however, the average trap size was noticeably smaller than in most previous decades (Fig. 7). This boom was the result of a wide variety of techniques (Fig. 8), being applied intentionally to search for lateral facies-change, lateral depositional pinchout, and channel-/valley-fill traps in the foreland basins of North America. Exploration activity remained high in the 1960s, with 32 discoveries, as North American foreland basin exploration continued at a rapid pace. Cretaceous valley-fill traps in the Rocky Mountain basins emerged as an important play

at this time. Most discoveries occurred at depths of 5000-10 000 ft and, although the average trap size was about the same as in the 1950s, several large accumulations were discovered (Fig. 7). It was during the 1960s that for the first time, discoveries were made in Asia and Australasia (Fig. 6) and exploration for stratigraphic and subtle combination traps moved offshore, with discoveries made at Arenque in Mexico and Halibut-Cobia in Australia. The 1970s was another active decade. Thirtysix discoveries were made using a wide variety of exploration techniques (Figs 5 & 7). The search for stratigraphic and subtle combination traps finally went worldwide, with discoveries in Brazil, the North Sea, Libya, Central Asia, India, China and Australia (Fig. 6). It was during this time that basin-centre gas and coal bed methane plays first came to the world's attention. Although the trap-size distribution remained about the same as in the 1960s (Fig. 7), deeper discoveries became more abundant. Another interesting trend was that offshore exploration for .tUlt, . . . '-:-~: . . . .buau~ldr, . . . : . . . . l":lUlt~ lll~, and subtle combination traps became active in the North Sea during the 1970s (Fig. 6). The number of stratigraphic and subtle combination traps discovered in the 1980s and 1990s is lower than that in each of the previous three decades (Fig. 5). In part, this reflects a decrease in North American onshore exploration activity as the price of oil plummeted in the mid-1980s and stratigraphic and subtle combination trap plays in Rocky Mountain foreland basins reached maturity. However, the apparently small number of traps discovered in the 1990s is largely an artifact of the data used in research for this paper, which relies on released data as its source material. The authors are aware that there have been hundreds of stratigraphic and subtle combination trap discoveries in offshore Brazil, Gulf of Mexico, U K North Sea and offshore West Africa. However, as information from these fields remains largely proprietary, their inclusion in research for this paper was not possible. During the 1980s and 1990s, the locus of exploration activity for stratigraphic and subtle combination traps moved from North America to offshore Brazil, offshore West Africa, the North Sea and onshore and offshore Asia. Fewer than half of the traps discovered since 1980 were located in North America. More than half of the discoveries were offshore turbidite traps, principally in the North Sea, the Atlantic margins of Africa and South America, and the US Gulf of Mexico. The trap-size distribution remained unchanged from the 1970s (Fig. 7).

STRATIGRAPHIC AND SUBTLE COMBINATION TRAPS

63

Fig. 5. Frequency of onshore and offshore stratigraphic and subtle combination trap discoveries plotted against year of discovery. The small number of discoveries shown for the 1990's reflects the fact that much of the data for recent discoveries remains proprietary and could not be included in this paper.

Seismic eclipsed all other exploration techniques (Fig. 8) as 3D seismic became a routine exploration tool and amplitude, attribute, AVO (Amplitude variation with off-set) and D H I (Direct hydrocarbon indicator) analyses came into general use offshore. During the 110 years of exploration for stratigraphic and subtle combination traps, a wide variety of techniques was used to make the discoveries. Seismic was involved in only 33 % of the discoveries with 87% of the traps being located onshore (Fig. 5). Active exploration for stratigraphic and subtle combination traps moved into deepwater areas, where costs are higher and risks are greater, at about the same time that 3D seismic became available. As a result, 95% of the offshore discoveries were made using seismic or seismic in conjunction with other techniques (Fig. 8).

Case studies The understanding and application of classification schemes aids in the exploration for stratigraphic and subtle combination traps and knowledge of past exploration approaches allows us to learn from our successes and failures. However, it is also important to consider analogue fields, which can be used to provide benchmarks for geological thinking and

to provide the creative seed for new exploration opportunities. With this objective in mind the following section provides case studies, which can be used as analogue material in the exploration for stratigraphic and subtle combination traps. Five case studies are presented. These cover a diverse selection of discoveries, which have been made by, (a) accident, (b) step-out drilling, (c) facies mapping, (d) change in play concept, and (e) seismic anomaly. By comparing and contrasting these case studies it is possible to examine some of the key elements necessary to successfully explore for stratigraphic and subtle combination traps.

East Texas Field The East Texas Field is located on the flank of the Sabine High in the east of the East Texas Basin, U S A (Figs 9 & 10). It has a STOIIP (Stock tank oil initially in-place) of 7326 MMBO (Million barrels of oil) with ultimate recoverable reserves estimated at 5992 MMBO. The East Texas Field is a classic stratigraphic trap, consisting of a very gently dipping erosionally truncated wedge situated between two unconformities (Fig. 11). The field is unfaulted and the reservoirs consist mainly of high-permeability, strand-plain and beach-ridge sandstones deposited in a wave-dominated

64

J.R. ALLAN E T A L .

Fig. 6. Number of stratigraphic and subtle combination traps discovered by decade for: (a) North America; (b) Latin America; (c) Europe.

STRATIGRAPHIC AND SUBTLE COMBINATION TRAPS

65

Fig. 6. (cont.) Number of stratigraphic and subtle combination traps discovered by decade for: (d) Africa; (e) Asia; (f) Australia.

66

J.R. ALLAN E T A L .

Fig. 7. Year of discovery versus original in place oil and gas volumes for stratigraphic and subtle combination traps. delta. The field was discovered in October 1930 and within eight months was producing at a peak rate of 900 000 BOPD (Barrels of oil per day) from 1500 wells. To date, more than 33 000 wells have been drilled on this accumulation. By the late 1920s the area to the east of Dallas, east Texas, was widely regarded within the oil industry as unprospective. A total of 24 wildcats had been drilled in Rusk County and Gregg County and all had been dry. The East Texas Field was discovered by accident. An amateur geologist, Dr A. D. Lloyd, recommended the drilling location to a promoter, Columbus Joiner, on an apparently misconceived geological model. Joiner drilled his first exploration well (Joiner Bradford No. 1) in Rusk County, in the SE of the eventual field area, but the well was abandoned as a result of mechanical and financial difficulties. Undeterred, Joiner drilled his second well only 100 ft from the first, but once again the well was junked and abandoned, although encouragement was drawn from an oil show. The discovery well (Joiner Bradford No. 3), was drilled in October 1930 only 300 ft south of the original well, and tested 300 BOPD (Minor & Hanna 1933). During the next six months, other operators drilled three more wildcats, which tested at rates of up to 10 000 BOPD. However, these wells were up to 15 miles from the discovery and initially it was not clear that they had all pene-

trated the same pool. An oil bonanza was nonetheless certain and impoverished local landowners and speculators bought up acreage and began to drill at an unprecedented rate. By end-May 1931, approximately 3000 wells had been drilled, and 500 000 B O P D were being produced from 1000 completions. As production reached some 900 000 B O P D in August 1931, the East Texas Field was recognized as the largest field in the USA, if not the world. The East Texas Field is approximately 67 km long and 8 km wide and covers an area of 528 km 2 (Fig. 9). The reservoir is contained in a regional subcrop trap in the Upper Cretaceous Woodbine Formation. The Woodbine Formation thins up-dip onto the flanks of the Sabine Uplift and truncates against the base-Austin Chalk unconformity (Fig. 10). The oil is trapped in an unconformity wedge-out with the reservoir occurring between the top-Washita Group unconformity and the Upper Cretaceous baseAustin Chalk unconformity (Fig. 11). The unconformities probably resulted from uplift of the Sabine Uplift. An alternative interpretation of the upper surface of the reservoir was presented by Gussow (1972, 1973), who argued that the up-dip pinch out is not erosional, but reflects depositional thinning and onlap onto the Sabine High. This latter interpretation is not widely accepted and hence the erosional truncation model is the preferred model in this

STRATIGRAPHIC A N D SUBTLE COMBINATION TRAPS

67 O

O

O,.~

O

,_~

O

O

~,..,,,,i

q...-i

68

J.R. ALLAN ETAL.

Fig. 9. Gross isopach map of the Woodbine Sandstone in the East Texas Basin, USA. Note cross section depicted in Figure 10 (after Halbouty 1991; AAPG 9 1991, reprinted by permission of the AAPG, whose permission is required for further use). summary. The top of the reservoir dips to the west at 90%) are contained in the J Sand unit, in particular the J2 Sand, which has an estimated hydrocarbon column height of 1500-1600 ft (Holman & Robertson 1994; Westrich et aL 1995). The J Sands are sealed laterally and vertically by bathyal shales. On the basis of PVT analyses of the fluids, the structure at the J2 Sand level is considered to consist of two reservoir compartments separated by a stratigraphic permeability barrier.

Summary data for 174 stratigraphic and subtle combination traps The five case studies discussed above are a small part of a wide spectrum of stratigraphic and

Fig. 19. ENE-WSW seismic line CNST-82-3 crossing the Nelson Field through the first well location on the block, 22/11-1. Nelson Field is a channel-fill subtle combination trap formed by lateral deposition pinchout of sand-filled channels on a low-relief anticline. The well is located between two prominent Upper Forties channels, shown by local thickening (Whyatt et al. 1992b 9 Springer Verlag).

82

J.R. A L L A N E T A L .

STRATIGRAPHIC AND SUBTLE COMBINATION TRAPS

Fig. 21. WNW-ESE 3D seismic cross-sections across the Bullwinkle Field (O'Connell et al. 1993). Hydrocarbon related seismic amplitude anomalies are shown in brackets, with the limits of the J Sand amplitude anomaly arrowed,

Fig. 20. Hydrocarbon occurrence in the Green Canyon Block and the location of the Bullwinkle Field (Weimer et al. 1998; A A P G 9 1998, reprinted by permission of the AAPG, whose permission is required for further use).

83

84

J.R. ALLAN ETAL.

Fig. 22. Early NW-SE 2D seismic cross-section over the Bullwinkle Field, showing northwestward onlap of Upper Pliocene section and pinchout of the 'J' Sand onto the basin-margin unconformity (Holman & Robertson 1994) 9 SEPM, Society for Sedimentary Geology. The penetrations of well GC 109 and its sidetrack are also shown.

subtle combination trapping mechanisms in our much larger study of 174 fields, which together contain over 2009 BBOE. The range of trapping mechanisms is summarized in Figure 23, which illustrates the diversity of stratigraphic and subtle combination traps discovered since 1885. From the case studies and exploration history review, it is apparent that the successful exploration for stratigraphic and subtle combination traps requires a more intimate knowledge of the geological history and stratigraphic relationships within a basin than exploration for structural traps. As an aid to achieving a successful exploration strategy, several key considerations relevant to stratigraphic and subtle combination traps are noted. These considerations are intended to act as a framework to help develop robust exploration strategies and consist of the following non-sequential steps, (a) establish basin setting, (b) establish probable trapping mechanism, (c) estimate the trap size probability distribution.

Basin setting By far the greatest number (95) and largest percentage (55%) of stratigraphic and subtle combination traps occur in foreland basins (Fig. 24). Intracratonic, rift and passive margin basins place a distant second, third and fourth. Each of the remaining six structural settings accounts for no more than 4% of the traps. Intrashelf basins on passive margins and delta/salt mini-basins (located on the US Gulf of Mexico continental slope) both occur within larger passive margin structural settings. When traps located within intrashelf basins and delta/salt mini-basins are grouped with passive margin traps, they still account for only 13% of the stratigraphic and subtle combination traps in this report. This, of course, is due in part to the difficulty in obtaining proprietary data on recent deepwater discoveries, many of which occur in passivemargin settings. Foreland basins contain the greatest number

STRATIGRAPHIC AND SUBTLE COMBINATION TRAPS

85

Fig. 23. Distribution of principal trap categories for stratigraphic and subtle combination traps. of stratigraphic and subtle combination traps and the largest in-place volumes and recoverable reserves. However, large reserves do not always correspond to a large number of traps in other

structural settings. For example, wrench, salt and intrashelf basins contain larger in-place volumes and recoverable reserves than intracratonic basins, even though they possess far fewer traps

Fig. 24. Distribution of stratigraphic and subtle combination traps by the main structural settings of the basins in which they are located. About 55 % of the traps occur in foreland basin settings.

86

J.R. ALLAN E T A L .

Fig. 25. (a) Distribution of original in-place oil and gas volumes in stratigraphic and subtle combination traps amongst main structural settings. (b) Distribution of ultimate recoverable oil and gas reserves in stratigraphic and subtle combination traps amongst main structural settings. Foreland basins contain the greatest number of stratigraphic traps and the largest in-place and recoverable volumes.

(compare Figs 25a & b to Fig. 24). This is because wrench, salt and intrashelf basins each contain a multi-billion barrel field (Kern River Field in the San Joaquin wrench basin, Marlim

Field in the Campos salt basin, and East Texas Field in the East Texas intrashelf basin), while no fields of this size occur in intracratonic basins.

STRATIGRAPHIC AND SUBTLE COMBINATION TRAPS

87

Trapping mechanism

Trap sizes

Whilst such statistics are interesting in their own right it is perhaps more useful to look at accumulation volumes in light of the trapping mechanism (Fig. 4). From the studied data it is evident that nearly 40% of the unconformity-truncation and onlap traps have recoverable reserves >600 MMBOE (Million barrels of oil equivalent) (Fig. 26d). In contrast, 93 % of the channelfill and valley-fill traps have recoverable reserves of 200 M M B O E than do lateral facies change traps (compare Fig. 26b & h). Most of the lateral depositional pinchout traps in the 200-500 MMBOE size range are clastic delta and turbidite fan-lobe traps that produce from numerous reservoir sands. The many productive intervals in each trap explain their larger size compared to lateral facies change traps, which tend to produce from fewer zones. Fracture traps tend to be regional in scale. Because of their large productive areas, many contain large recoverable reserves (Fig. 261). Finally, different trapping mechanisms within the same trap regime may result in very different reserves distributions. Hydrodynamic and basin-centre gas accumulations are both fluidic traps. However, hydrodynamic traps, which are local in scale, tend to have limited productive areas, low net pay, and small reserves, while basin-centre gas traps, which are regional in scale, generally contain huge reserves (Fig. 27b). A few trapping mechanisms are responsible for a disproportionate amount of the in-place volumes and recoverable reserves found in stratigraphic and subtle combination traps. Lateral depositional pinchout traps are the most abundant trap type and as a group contain the largest in-place volumes (Fig. 27a). Subcrop and onlap traps, although far fewer in number, are much larger in size and contain the largest recoverable reserves (Fig. 27b). Six of the subcrop traps and four of the onlap traps have ultimate recoverable reserves >1 BBOE. Basin-centre gas and coal bed methane traps are regional traps with huge productive areas and also have large reserves (Fig. 27b).

Major oil companies have, in recent years, targeted their exploration efforts to locate 'elephants' as it is only through their discovery that the promised company growth can be realized. Medium-sized oil companies and independents can of course afford the luxury of exploring for smaller accumulations as the net growth from a medium or even small find is significant. In considering stratigraphic and subtle combination traps as potential targets for companies of any size, it is worth remembering that, like structural traps, stratigraphic and subtle combination traps can occur as single accumulations (e.g. East Texas, Nelson and Jay fields) or they can occur in trends and clusters, which add up to significant volumes (e.g. Ansai Field). From analysis of the data included for this paper it can be demonstrated that stratigraphic and subtle combination traps are associated with a wide range of trap sizes as depicted in Figures 26 and 27. Thirty-six of the stratigraphic and subtle combination traps have original inplace oil and gas volumes >1 BBOE. Ninetyone traps have original in-place oil and gas volumes 1 BBOE. However the modal frequency occurs in the field-size range of 50-300 MMBOE. It is worth reflecting that amongst all these statistics the largest conventional oil field is Prudhoe Bay Field, USA, a regional subcrop trap with an original in-place volume of 30.8 B B O and ultimate recoverable reserves of 12.3 BBO.

Heavy oil traps Light oils are obviously attractive and it is not worth expanding on them from the perspective of stratigraphic and subtle combination trapping mechanisms. What is worth further consideration is heavy oils and gas, as these fluids are often associated with stratigraphic and fluidic trapping mechanisms, which have large areal distributions and contain large volumes. They are often not given appropriate focus due to their classification as 'unconventional'. Heavy oil, often dismissed because of low value per barrel, can be present in large volumes in stratigraphic and subtle combination traps. For example, the largest heavy oil field is the Athabasca Oil Sands, Canada, a lateral depositional pinchout trap which contains an in-place volume of 902 BBO of heavy oil and tar. Its ultimate recoverable reserve is as yet unknown. Cerro Negro and Machete, two producing areas

88

J.R. ALLAN E T A L . ,~

rA

~

~

2=

"&"6 90,..~

9~

~.~

.-~

N'~

~9

~

~2

~8

~g

STRATIGRAPHIC A N D SUBTLE COMBINATION TRAPS

89

O

x.....,-

t9

o~

90

J.R. ALLAN

ETAL.

8

.~

O

O

.~

.m

= ,,.-,,

STRATIGRAPHIC AND SUBTLE COMBINATION TRAPS

91

Fig. 27. Distribution of hydrocarbon volumes in stratigraphic and subtle combination traps shown by main trapping mechanism; (a) in-place volumes; and (b) ultimate recoverable reserves. Regional subcrop and onlap traps, although far fewer in number than lateral pinchout traps, are much larger and contain greater recoverable reserves.

92

J.R. ALLAN E T A L .

which produce from onlap traps in the Orinoco Heavy Oil Belt, Venezuela, have the largest recoverable oil reserves, estimated at 25.6 and 20.0 BBO respectively. The potential of discovering volumes of such magnitude is a powerful argument for targeting stratigraphically trapped heavy oil or tar accumulations, particularly in today's economic climate where heavy hydrocarbons are economically recoverable due to improved technology (Alberta Energy and Utilities Board 2000). Gas and condensate traps

Gas and condensate are also important from the perspective of stratigraphic and subtle combination traps, as they are associated with basincentre gas and coal bed methane accumulations, both of which have been proven to contain significant volumes of in-place hydrocarbons (Fig. 27a & b). Many of these traps contain huge reserves, as exemplified by the ElmworthWapiti basin-centre gas trap in Canada (800 TCFG in-place, 20 TCFG recoverable) and the Bowen Basin coal bed methane trap in Australia (>60 TCFG in-place). Basin-centre gas accumulations may result from hydrostatic entrapment caused by lateral capillary pressure changes from gas-saturated clean sandstones down-dip into water-wet shalier sandstones updip (Elmworth-Wapati and Blanco Mesaverde fields, North America), by hydrodynamic entrapment on the flank of a regional high (Dauletabad-Donmez Field, Turkmenistan), or by a combination of the two (Hoadley Field, Canada). The current economics of world hydrocarbon resources make gas an increasingly valuable commodity, as quoted from the Petroleum Technology Transfer Council's Year 2000 Symposium on Basin-Centred Gas: 'It is becoming critical that the US. industry more fully develops the Nation's large "unconventional" gas resources. A t present about 15% o f the US. total gas production is from "unconventional" basin-centred (synclinal) gas resources. Basin-centred gas production is in its infancy, but will increase as more explorationists become aware o f this enormous gas resource. More effective exploration and exploitation strategies will be necessary in the near future to meet the increasing emphasis on this resource'.

Whilst there are clear analogues for reviewing stratigraphic and subtle combination trapping mechanisms from the perspective of basin setting, trapping mechanism and trap size the successful exploration for such plays requires not only an understanding of relevant analogues but also the application of appropri-

ate geoscience techniques. Techniques relevant to the exploration for these traps are outlined in the following section.

Exploration techniques Whilst numerous successful discoveries of stratigraphic and subtle combination traps can be documented, it is important to note that dry hole risk is greater than in structural traps. This situation arises from difficulties in locating updip depositional pinchouts or facies-change seals, particularly if deterioration in reservoir quality has created a large waste zone at the pinchout. Consequently, several dry holes are usually drilled prior to a discovery. Additional costs and risks are incurred through the necessity of drilling numerous appraisal wells to locate lateral seals and to establish the size and commercial viability of a trap. Failure to recognize a stratigraphic or subtle combination trap must also be considered a significant exploration risk. Several large accumulations included in this study (e.g. Nelson, Miller, Harding, Gryphon fields in the North Sea, Upper Valley Field in the USA) were discovered by the second or third leaseholder on relinquished acreage because the original leaseholder failed to recognize the trap. Most of these missed opportunities were caused by one or more of the following reasons: (1) Poor seismic resolution, which made it difficult to identify the updip or lateral seals to a trap and prove closure. As a result, the prospect was considered too risky to drill. (2) Encountering a thin pay zone, failing to recognize that it covered a large productive area and dismissing the prospect as noncommercial. (3) Misinterpreting a large combination trap as a small structural closure because of failure to recognize a stratigraphic component to entrapment. (4) Failure to recognize a stratigraphic or subtle combination trap because of unfamiliarity with unconventional trapping mechanisms (e.g. basin-centre gas, subtle fracture, hydrodynamic traps). These often occur in unexpected locations (e.g. in basincentre synclines or down-flank on large structures) and sometimes have unusual well-log responses. Although stratigraphic and subtle combination traps carry a few more risks than structural traps, by using the exploration techniques most appropriate for each trap type, many of these risks can be minimized. Whilst a detailed

STRATIGRAPHIC AND SUBTLE COMBINATION TRAPS analysis of all available geoscience techniques that have application to the exploration for stratigraphic and subtle combination traps is beyond the scope of this paper it is worth noting some of the key geological, geophysical and geochemical techniques. Geological techniques

Arguably the most important starting point in exploring for stratigraphic and subtle combination traps is a detailed sequence stratigraphic framework. A comprehensive discussion of sequence stratigraphic principles and techniques is beyond the scope of this paper and the reader is referred to excellent publications by Bally (1987); Mitchum et al. (1977); Payton (1977); Sarg (1988); Vail (1987); Van Wagoner et aL (1988, 1990) and Weimer & Posamentier (1993). An excellent five-step approach to stratigraphic trap exploration, which relies heavily on sequence stratigraphic analysis, is described by Dolson et al. (1999). In summary these stages include: (1) Using seismic and log cross-sections, identify all unconformities in a basin or area of interest, subdivide the sedimentary section into genetically related sequences, and construct a sequence stratigraphic framework. Identify all maximum flooding surfaces, condensed sections and transgressive surfaces within each sequence. The locations of stratigraphic and subtle combination traps are generally controlled by third-, fourth- and fifth-order sequences. Third-order sequences represent major sea-level cycles and consist of highstand systems tracts (HST), transgressive systems tracts (TST) and lowstand systems tracts (LST). Different types of stratigraphic and subtle combination trap are found within each systems tract (Fig. 28). (2) Interpret seismic facies within each sequence, using reflection-pattern geometries and calibrate the reflection patterns to lithology using well data (Mitchum et al. 1977; Brown 1999). (3) Using seismic, lithologic, biostratigraphic, chronostratigraphic and palaeoenvironmental data, construct facies maps and cross-sections for all areas of interest. A good sequence stratigraphic framework is essential, because it provides the means for constructing palaeogeographic maps of facies belts at precisely defined intervals of time. (4) Within each sequence stratigraphic framework, use the facies maps and cross-

(5)

93

sections that have been constructed to identify those sequences that appear most likely to contain good reservoirs and viable seals. Within each of these sequences, identify and map the shoreline trends, shelf breaks, regional facies transitions, and local unconformities. Different trap types will be associated with each feature. Using analogue traps as a guide, examine each of these features for evidence of stratigraphic and subtle trapping configurations using the maps, cross-sections and seismic data.

Although the optimal exploration strategy for each trap type described in this paper (Fig. 4) is somewhat different, similar strategies often apply to traps located in the same systems tract or in similar palaeogeographic settings. For example, although trapping configurations may be quite different, the same basic approach is used to search for lateral depositional pinchout and channel-/valley-fill traps that occur in marginal marine and continental settings. Turbidite lateral depositional pinchout traps, turbidite channel-/canyon-fill traps, and carbonate debris-flow traps belong to different trap categories. However, because all occur in the same palaeogeographic setting (the outer shelf and slope), exploration strategies for these three trap types have many similarities. In contrast exploration for subcrop and onlap traps, which are both associated with unconformities, involves a set of exploration approaches that differ from those for depositional pinchout traps. Delineation and development approaches and strategies differ for combination traps as opposed to pure stratigraphic traps. Structure maps are useful for evaluating spill points and estimating locations of hydrocarbon-water contacts in combination traps, particularly those located on structural noses. However, they provide little useful information for evaluating pure stratigraphic traps. 3D seismic technology now offers the potential to define stratigraphic traps (see the following section). Before 3D seismic was available, depositional pinchout, lateral facies change, channel-/valley-fill, subcrop and onlap traps were delineated by first drilling step-out wells up-dip and down-dip. Analogue fields, particularly earlier discoveries in the same play, were often used to predict trap geometry and guide step-out drilling. Analogues are particularly useful when the trap has an unusual geometry, such as the narrow, elongate shapes characteristic of incised shoreface and barrier island lateral depositional pinchout traps, channel- and valley-fill traps and

94

J.R. ALLAN E T A L .

Fig. 28. Schematic cross-section illustrating typical locations of stratigraphic and subtle combination traps within a third order sequence. Different types of stratigraphic and subtle combination trap are found within different systems tracts (Baum & Vail 1988; Sarg 1988; Dolson et al. 1999) 9 SEPM, Society for Sedimentary Geology. dolomitization/dissolution traps. Careful welllog analysis is of the utmost importance. The sequential log analysis approach described in Rider (1996) can be used along with image logs to define key unconformity surfaces and to aid in delineating facies geometries. Accurate log evaluation is also important when evaluating tight sands in basin-centre gas traps to determine whether they will be capable of commercial flow (Shanley et al. 2004). Special

maps and cross-sections may be required to evaluate certain reservoir types. For example, a potentiometric map is the best tool for predicting the direction in which an oil accumulation has been displaced by hydrodynamic flow in an aquifer. Maps of bulk permeability calculated from log porosity-permeability transform equations are helpful for identifying fractured high-productivity 'sweet spots' in basin-centre gas traps and subtle fracture traps, but only

STRATIGRAPHIC AND SUBTLE COMBINATION TRAPS really have true value when integrated with dynamic data, which is a significant limitation when historic data are unavailable.

Geophysical techniques 3D seismic technology has changed the way in which stratigraphic and subtle combination traps are delineated and developed, dramatically lowering dry-hole risk and enhancing the effectiveness of well placement. It is particularly effective for imaging and delineating young, poorly consolidated siliciclastic reservoirs with high porosity and sandstones that pinch out into shale, since the pronounced acoustic impedance contrast between sandstone and shale allows the edges of reservoirs to be precisely defined. Seismic reflection amplitude is now used to accurately map the wedge-out edges of porous reservoirs in stratigraphic traps, which previously had been difficult or impossible to detect. Amplitude-derived attributes and acoustic impedance contrasts can, when calibrated to core and well logs, be used to map rock properties within reservoirs. Examples that illustrate the application of these techniques to delineation and development of stratigraphic and subtle combination traps are presented in the following section. For a more exhaustive treatment of 3D seismic interpretation techniques, the reader is referred to Brown (1999). Amplitude displays of 3D seismic data are extremely useful for identifying locations of reservoir pinchouts and estimating reservoir thickness and quality. Seismic reflectors dim as a reservoir thins and disappear at the pinchout edge, providing a means for precisely determining the positions of up-dip and lateral pinchout seals. This greatly diminishes the risk of drilling dry holes along the edges of stratigraphic and subtle combination traps. The amplitude of a seismic bright spot generally becomes higher with increase in N/G ratio, porosity, or hydrocarbon saturation. Horizon-slice maps can be used to display this lateral variation in seismic amplitude within a reservoir. Once seismic amplitude has been calibrated to well data (Fig. 29a), seismic amplitude maps can be used to identify areas with the best porosity (Fig. 29b) or highest N/G ratio (Fig. 30). Using this information, development wells can be sited to target areas of highest reservoir quality thus maximizing flow rate and ultimate recovery. It is now possible to define the spatial distribution of amplitude anomalies so accurately that they can be used to plan trajectories of horizontal wells. Because of the acoustic impedance contrast between sandstone reservoirs and enveloping

95

siltstone and shale seals, seismic lithologic velocity modeling can be used to precisely define the boundaries of lateral depositional pinchout, onlap pinchout, and channel-/valleyfill traps. Seismic AVO modelling (amplitude variation with offset) can be used to identify oiland gas-bearing reservoirs, image facies changes in carbonates, and map net pay in sandstone traps. Stacking patterns of seismic reflectors are now regularly used to decipher the internal geometry of reservoir sand bodies, which gives geologists and geophysicists the ability to identify and map depositional facies using seismic data. Once the reservoir distribution has been mapped from seismic amplitude, acoustic properties can be correlated to rock properties and used to create maps of important reservoir parameters. Fractured 'sweet spots' within reservoirs can often be identified through the use of 3-component seismic data or by break-up of seismic reflectors. An example of this phenomenon can be seen in Figure 31. In highporosity, poorly consolidated reservoirs, direct hydrocarbon indicators (DHIs) can often be used to identify hydrocarbon-water contacts. Hydrocarbon accumulations may appear as bright or dim spots that terminate at hydrocarbon-water contacts, while fluid contacts may appear as flat spots on seismic profiles. The ability to directly detect and map fluid contacts seismically lowers down dip dry hole risk during delineation drilling and allows trap size to be appraised using fewer wells. Recently, 4D seismic has come into use as successive time lapse 3D seismic surveys have been utilized to monitor steam flow and sweep efficiency in reservoirs undergoing steam flood tertiary recovery. This new 3D seismic technique will be of particular importance in the exploitation of the world's heavy oil accumulations, which occur mostly in stratigraphic and subtle combination traps. Although 3D seismic technology has significantly lowered delineation and development risk in stratigraphic and subtle combination traps, it does not work equally well for all trap and reservoir types. 3D seismic is very effective for imaging poorly consolidated siliciclastic reservoirs and sandstones that pinch out into shale. It is therefore most successful in evaluating lateral depositional pinchout, channel-/ valley-fill, and subcrop/onlap traps. 3D seismic is less effective in imaging well-indurated sandstone reservoirs, carbonate facies-change traps, traps in which the lithologic contrast between reservoir and seal is gradational and thin reservoirs that are below seismic resolution. Furthermore, unless DHIs are present, seismic can only

96

J.R. A L L A N E T A L .

STRATIGRAPHIC AND SUBTLE COMBINATION TRAPS

97

Fig. 30. Seismic amplitude map demonstrating high amplitude related to high net/gross ratio from the Halibut Cobia Field, a regional subcrop trap in Australia (Hinton et al. 1994). tell the explorationist that a trap is present, not indicate whether it is filled with hydrocarbons. Thus, other techniques must be integrated with seismic in order to delineate the boundaries of many types of stratigraphic and subtle combination trap. Some of the greatest advances may yet come from i m p r o v e m e n t s in seismic interpretation techniques gained by matching features on seismic data with known geological analogues.

Geochemical techniques

Surface geochemical analysis, an emerging technology that in the past was used mainly for exploration purposes, also holds great promise for delineating stratigraphic and subtle combination traps. W h e n used correctly, surface geochemical analysis has dramatically increased the success rate and decreased dry hole risk in hundreds of d o c u m e n t e d cases around the

Fig. 29. (a) Cross-plot calibrating seismic amplitude to porosity using average porosity data from logs, Cretaceous Kharaib limestone reservoir, Idd AI-Shargi North Dome Field, a lateral facies change combination trap in Qatar. (b) Structure contour map, top Kharaib B limestone. The porosity cutoff of 26% corresponds to a permeability of 2 mD. Potentially productive areas with average porosity >26 % and tight areas with average porosity n :> t c ~ h O ZLI N

~ Stratigraphlc Traps ~ 3D Seismic ~ PHASE III & Geological PHASEII ~ C. . . . pt 2D & 3D Seismic ~

kStruct .... PHASE I ~

sU"~u s

~

,

\

~0 TIME

(YEARS)

Fig. 1. Phases of exploration within a typical mature hydrocarbon bearing basin. Note the tendency for discovery sizes within each phase to decrease through time until a new phase is adopted.

Fig. 2. Basin type location of discovered hydrocarbons in the decade 1980-1990 (Amoco Production Company, pers. comm.).

seen to be attributed to the successful drilling of subtler structural/stratigraphic and pure stratigraphic traps in basins with well defined petroleum systems. The three cycles of exploration are clearly evident in most of the large onshore basins of North America. For example in the Powder River Basin, in NE Wyoming, USA oil was first found in the late 1800s. The early prospects were obvious surface structures known as 'sheepherder anticlines' because geologists located them simply by describing what they were looking for to the local sheepherders. Through 1930, the oil companies drilled progressively smaller features until no more viable prospects existed. Only one stratigraphic trap was found during this period, in a channel sandstone reservoir, near to a surface oil seep. With the development of seismic techniques, the Powder River Basin saw a resurgence of activity. On the deformed flanks of the basin, many more structural traps were drilled with success whereas in the central part of the basin seismic data revealed only homoclinal dip and little drilling was conducted. One more stratigraphic trap was found on the flank of a structural prospect but by 1960 all the seismic structural anomalies were drilled and the basin was largely dormant once again. In 1967, Bell Creek, a large stratigraphically trapped oil field was found in the central portion of the basin. The field was found in an area of homoclinal dip by an operator with the conviction to chase a stratigraphic concept. By 1980, the entire central portion of the basin was developed, with over 100 stratigraphic traps discovered. Somewhat ironically, the Hartzog Draw Field, discovered in 1975, is a 250 million barrel field which was found after nearly 100 years of exploration in the basin. It is a common theme in the discovery of subtle or stratigraphic traps that many have been drilled by accident. In many cases

STRATIGRAPHIC TRAPS IN INDONESIA prospects were drilled which would have been economic if they resulted in a small discovery. However, when the discoveries were appraised and developed they resulted in larger than expected fields, mainly due to important stratigraphic components in the trap set-up. If a strategy can be developed to maximize this upside potential, the drilling of subtle/stratigraphic traps becomes much easier to justify. Also, the example of the Powder River Basin, demonstrates that vast areas with high economic potential can be missed if one assumes that stratigraphic traps will be found as accidental by-products while drilling for pure structural closures. Clearly, in order to find stratigraphic traps you must pursue an exploration strategy which believes that they will exist.

The petroleum system approach Many structurally defined prospects consist of a very well documented closure with a speculative source. Historically in the exploration business these features have been fairly easy to get drilled because it is usually possible to devise a plausible scenario of source and migration to justify the risk. The fact that a large proportion

107

of the structural closures drilled worldwide are not charged by hydrocarbons demonstrates that source and migration should be considered a major concern. This approach to ranking generative basins on the basis of their charge or 'petroleum system' was first devised by Demaison (1984) and expanded upon by Demaison & Huizinga (1991). In Indonesia the recognition of a functioning petroleum system is key since migration distances from kitchen to trap are always relatively short (,>

~

^ ~

" ~ - ~ ..c~ "

~.=

~"~

o

o

~

~

~

~

~

o

~

~

~ ~-) 0

~ , S . ~ . ~ m ~ ~.= ~/. ~, ~= g :~

~..~..~

s

~ .~

~

"

o

~V

~

~

~a

=~

o

o

~

o

~_~o

-e

0"~

~, .~ h~ ..~ ~ ~

%

~

~

o

o

~

.~.~

~

o

214

JOEL

~9

CORCORAN

'~.

N

_

o

o=

~ ~ Z

~

~ , ~ ,~

,~

~'

~

~,,u ~

0

o

~

~

0

~

,..= ,,.-, ~

r~

~,

~ ~ ='S

~

~

= ~"~

~

~ o ~

~

~

~~

~.,~

-,-'

o

~:

,-,

~.., ~

~,

,~=

"d

= = o

_

~ ~

a o ~ ,~ o

~

~.~.u.=.

~ ~ ~ ~:.S ~ . :

=

~o

8

i

-~

.~ ~ "%

0

-7~.~ ~o=~

,.-,

o g,=~

I~

~

o

,u.

2

~=

,_I

.,,-,

t~

8 e4 J

, ~

"4

~-u..

SEALING SURFACE CLASSIFICATION

=o

~

~9

~

=o ~ -

~

~=~

~

~..

F,~

.~,

~.~

~.~ ~_~,~ ~

~ ~.~ = ~ ~.=~.~

:~

~

n~ 0

~,-,~

-~

~ ~ ~, o

o~ ~

o

~.~

=~

:~

~ x~

~ ~

~_~,~

~ ~

~-~:~

,-~o o ~

er~

~o_~

~

~

0

0

o

0

~

~o=~o

~ ~, ~ . . . . .

~:.~ o

-=

~o~o~

~.o ~ ~

~ ~

215

0

~ ~

= o

~ o_~ ~ o ~.~

216

JOEL CORCORAN

._.

o.

.--,~ o

~ o: ~ . ~

.. ~ ~ = ~ ~ o~-' ~

~:

~'~"

.u "o o

~

~

~

~

~

= ~

~.o=~:.~ ~

@

~, .

""'~

o E

-

0

~

~

"~ '~

~.~.~ ~= ~

~

.~ ~

.o~o=~:~

~"~'~

.~,.~

~

~9

~

~,.oo

~

# ~

~

-.-~ _~.-~

-

~

......

_~:~

~

~'~

~ooo~

~

. ~

""

0

0

0

0

>

ii

0 ~

~, o

~j *"

~

~ 0

~: ~

~.~

e ",'~' O . r ~

;., ~ ~

o

- -

:~ ~ . z :

ii

.t

o ~ a=

~ o~

~ ~,

~.:=

SEALING SURFACE CLASSIFICATION

217

I3, Salt induced truncation antiform (Fig. 3) One-seal trap with an unconformable top seal. The basal seal is not significant to the trapping mechanism. Pinchout and/or truncation of the reservoir in the lateral sense may be locally important. Examples include Kittiwake (Glennie & Armstrong 1991)

U(C)T, Footwall trap with unconformable/ conformable top seal (Fig. 4) Footwall related poly-seal trap with a variable truncated top seal. A conformable partial seal is not critical to the trap success although may aid pooling additional upside. Note that the lateral fault seal is required where the subcrop edge branchline cuts the fault. Examples include Saltire (Casey et al. 1993), Petronella (Sgiath & Piper Sandstone) (Waddams & Clark 1991) and the Piper F i e l d s (Schmitt & Gordon 1991; Maher 1981).

Fig. 7. (U/U) Truncation enveloped onlap trap, (U/U) Truncation enveloped onlap, (U) Unconformable top seal, (U) Unconformable bottom seal, Polyseal.

C, Compaction anticline (Fig. 5) Four-way dip closure induced through differential compaction. The closure mechanism is chiefly structural, however the geometry is a result of differential compaction imparted by stratigraphy and depositional set-up. Examples include West Brae.

U/T, Downthrown fault closure with truncation trap (Fig. 6) Poly-seal trap with an unconformable top seal and fault bounded side/bottom seal. Examples include the Saltire field (Casey et al. 1993), Galley (Moseley 1999) and Petronella field (Waddams & Clarke 1991).

U/U, Truncation enveloped 'onlap' trap (Fig. 7) Poly-seal trap with an unconformable top and bottom seal. The geometry is such that an apparent onlap/pinchout against the elevated basal/side seal may be evident on seismic; though the onlapping nature of the side and basal seals would give evidence of erosional nature of the reservoir package. Examples include the Piper sandstone reservoirs of the Highlander field (Whitehead & Pinnock 1991).

C/C, Onlap trap (Fig. 8) Poly-seal onlap trap with conformable top and bottom seal required for closure.

Fig. 8. (C/C) Onlap with conformable top and bottom seal trap, (C/C) Onlap with conformable top and bottom seal (conformable in terms that, only one bedding surface in contact with bottom and top reservoir respectively), (C) Conformable top seal, (C) Conformable base seal, Poly-seal (two independent risks).

U/C, Subcrop trap (Fig. 9) Poly-seal trap with the reservoir subcropping beneath an unconformable top seal. The basal seal remains conformable e.g. local regions of Captain field (Rose et al. 2000).

218

JOEL CORCORAN It is worth noting the subtle differences between the C/C onlap trap (Fig. 8) and the C/U onlap trap (Fig. 10). In the former trap the basal seal has only one bedding plane in contact with the base of the reservoir. This could prove a key component in risking the Lower Cretaceous prospects in the outer Moray Firth. Here a thin drape of Kimmeridge Clay is occasionally sufficient enough to blanket the palaeotopography such as to remove the reservoir from an otherwise high risk unconformable basal seal.

U(C)/TC, Footwall trap with unconformable/conformable top seal (Fig. 11)

Fig. 9. (U/C) Truncation Trap, (U/C) Truncation trap (below), (U) Unconformable top seal, (C) Conformable bottom seal, Poly-seal (Milton & Bertram 1992).

C/U, Onlap trap (Fig. 10) Poly-seal trap created by onlap onto an eroded high. The basal seal is unconformable in comparison to the top seal. A key risk is the potential for hydrocarbon re-migration from the reservoir through the unconformable basal seal sequence via the bedding planes or along permeable beds.

Poly-seal footwall trap with variable truncated top seal. The conformable seal is not critical to trap success although may aid pooling additional upside. The extent of the stratigraphic component of the trap is often minimal in areal extent but critical to trap success in crestal regions. A basal seal is required when the reservoir pool does not extend into fault plane as illustrated in Fig. 11. Examples include local regions of the Rob Roy (Parker 1991; Fraser et al. 2003), Scott (Fraser et al. 2003) and Piper fields (Schmitt & Gordon 1991; Maher 1981).

Fig. 11. (U(C)/I'C) Footwall related, variable Fig. 10. (C/U) Truncation Trap (above), (C/U) Truncation trap (above), (C) Conformable top seal, (U) Unconformable bottom seal, Poly-seal (Milton & Bertram 1992).

truncated top seal with basal seal trap, (U(C)/TC) Footwall related, variable truncated top seal with basal seal, (U(C)) Unconformable top seal with less critical conformable seal, (C) Conformable bottom seal, part of bottom/lateral seal potentially fault induced, Poly-seal.

SEALING SURFACE CLASSIFICATION

219

U/U, Differentially compacted channel trap (Fig. 12) An example of this trap is the Alba field. Here the top reservoir is distinguished by the differentially compacted base Oligocene unconformity. The basal seal is also unconformable resulting from rapidly infilled channel incision. In the case of Alba the presence of injectites introduce a high risk element to the seal risk since multiple top seal surfaces will interact with an otherwise low risk structurally closed top reservoir. The severity of risk will be dependent on the thickness and lithology of the sealing sequences. Injectite structures are post depositional and complicate trap classification as they fall into neither of the generic stratigraphic or structural schemes.

C/U, Mounded channel trap (Fig. 13) These are predominantly structural traps with a single low risk seal at the top reservoir (NB. top seal closure mechanism analogous to 'C, compaction anticline', Fig. 5). Additional stratigraphic upside and the polyseal trapping mechanism is presented by the basal erosional surface. The trapping mechanism in the one-seal (i.e. top reservoir) case is clearly structural if closed uniquely by closed contours. However the overall geometry is both depositionally and compactionally induced resulting in a structural/stratigraphic classification. In addition a further stratigraphic element is implicated when a basal seal is required e.g. The MacCuUoch field (Gunn et al. 2003).

Fig. 12. (U/U) Differentially compacted 'channel' with unconformable top and base seal trap, (U/U) Unconformable top and bottom seal, (U) Unconformable top seal, (U) Unconformable base seal, Poly-seal.

C/F, Mounded channel trap (Fig. 14) A poly-seal trap with a conformable top seal over a structural 'mound' and lateral seal(s) provided by facies change into non reservoir channel flank/overbank deposits. A possible example is the Brenda discovery (inferred from Jones et al. 2004).

C/CF, Marine pinchout trap (Fig. 15) Two independent seals are required (base and top) to create this trap. Both are conformable with the reservoir bounding surfaces. It is likely that both seals will be of similar, if not identical lithology. In this case top and bottom seal will carry the same risk (Milton & Bertram 1992). This poly-seal category is the most prolific stratigraphic closure mechanism identified within the

Fig. 13. (C/U) Erosive mound trap, (C/U) Erosive base mound, (C) Conformable top seal, (U) Unconformable bottom seal, Poly-seal. study region (Tables 1-3), example include closure within the Britannia (Oakman & Partington 1998), Everest and Fleming fields (O'Connor & Walter 1993).

220

JOEL CORCORAN

Fig. 14. (C/F) Channel with flank facies closure trap, (C) Sedimentary mounding enhanced by differential compaction, (F) Facies change induced by non reservoir/sealing levees, Poly Seal.

Fig. 16. (C/F) Facies change/waste zone traps, (C) Conformable top seal, (F) Facies change, Poly-Seal (modified after Milton & Bertram 1992).

moving proximal to the fault scarp e.g. Brae fields (Fraser et al. 2003; Harms et al. 1981).

Risk

Fig. 15. (CICF) Pinchout/Shale out traps, (C) Conformable top seal, (CF) Conformable base seal with facies change, Poly-Seal (Milton & Bertram 1992).

C/F, Depositional facies change/Waste zone trap (Fig. 16) A poly-seal trap with closure provided by a conformable top seal along a single surface and a lateral seal occurring over a waste zone created by a degradation in reservoir quality;

Table 4 describes in qualitative terms the relative risking of the trapping models illustrated in this paper. A risk value of high to low has been applied to each of the models described in addition to a brief comment on the justification behind the risk rank. For the purpose of this illustration we are concerned only with the geometrical nature of the sealing surface with respect to the reservoir bounding surfaces, remaining purposely generic. Lithologies and properties of the seals have been neglected, although clearly the nature and combination of the seal lithology associations will have a strong control on sealing efficiency, providing an additional insight into sealing integrity when considered in relation to trapping mechanism.

Conclusions The sealing-surface classification concept and scheme initially proposed by Milton & Bertram (1992) provides a suitable framework for describing proven hydrocarbon accumulations which can be applied to similar un-drilled traps. Furthermore by considering the nature and

SEALING SURFACE CLASSIFICATION

221

,o

~,8{

~..~

8

~

=

~

O

OS6

~,~=-d

z

=~

=:

a:

S ~g~55

=:

~o=

~

~

~

~.-

~

O~

~

~e ~

#

0{5oo

o

~e

~

~F=:~

~

.~

~

~=~

~,-a 4..,6

a

a

a .~

o

o~

~

0

~

~..~ ~j o ~

~9' ~

{

,.~

222

JOEL CORCORAN

g e o m e t r y of the sealing surface in relation to the reservoir/seal interfaces an assessment of risk is possible. Such a s c h e m e warrants revisiting as a useful tool for the c o n t i n u e d successful exploration for stratigraphic and stratigraphic-related traps in similar extensional settings. Thanks to Milton and Bertram for compiling the original scheme and concepts upon which this publication is based. Gratitude is extended to J. Argent and C. Oakman for their useful review and contributions which have greatly improved this manuscript. The paper represents work that contributed partly towards an MSc thesis on the 'Significance of Stratigraphic Trapping in the Britannia Satellites and Beyond, UK CNS' carried out over the summer of 2003 in fulfilment of the requirements for the MSc in Petroleum Geoscience (Imperial College, London). Mentorship by T. Evans (Imperial College) and ConocoPhillips UK individuals is gratefully acknowledged. Final thanks to S. Emberson for aid with references.

References AHMADI, Z.M., SAWYERS, M., KENYON-ROBERTS, S., STANWORTH,C.W., KUGLER, K.A., KRISTENSEN,J. & FUGELL1,E.M.G. 2003. Paleocene. In: EVANS, D., GRAHAM, C., ARMOUR, A. & BATHURST, P. (eds) The millennium atlas, petroleum geology of the central and northern North Sea, Geological Society, London, 235-260. ARMSTRONG, L.A., TEN HAVE, A. • JOHNSON, H.D. 1987. The geology of the Gannet Fields, Central North Sea, UK Sector. In: BROOKS,J. & GLENNIE, K.W. (eds) Petroleum Geology of North West Europe, Graham and Trotman, London, 533-548. BIDDLE, K.T. ~: WE1LCHOWSKY, C.C. 1994. Hydrocarbon Traps. In: MAGOON, L.B. & Dow, W.G. (eds) The petroleum system-from source to trap, American Association of Petroleum Geologists, Memoir 60, 219-235. CASEY, B.J., ROMANI, R.S. & SCHMITT, R.H. 1993. Appraisal geology of the Saltire Field, Witch Ground graben, North Sea. In: PARKER,J.R. (ed.) Petroleum geology of northwest Europe: proceedings of the 4th conference, Geological Society of London, 507-517. COPESTAKE, P., SIMS, A.E. CRITrENDEN, S., HAMAR, G.R, INESON,J.R., ROSE, RT. & TRINGHAM, M.E. 2003. Lower Cretaceous. In: EVANS,D., GRAHAM, C.,ARMOUR, A. & BATHURST,R (eds) The Millennium Atlas: Petroleum geology of the Central and Northern North Sea, Geological Society of London, 191-212. COWARD,R.N., CLARK,N.M. & PINNOCK,S.J. 1991. The Tartan Field, Block 15/16, UK North Sea. In: ABBOTTS, I.L. (ed.) United Kingdom oil and gas fields, 25 years commemorative volume, Geological Society, London, Memoir, 14, 377-384. DIXON, R.J. & PEARCE, J. 1995. Tertiary sequence stratigraphy and play fairway definition, Bruce-

Beryl Embayment, Quadrant 9, UKCS. In: STEEL, R.J., FELT, V.L., JOHANNSSEN,E.P. & MATHHIEU, C. (eds) Sequence stratigraphy on the Northwest European margin. Geological Society, London, Special Publications, 94, 443-469. DORE, G. 2002. The Buzzard Field - A n overlooked North Sea giant, PETEX 2002 Exploration Session, Extended abstracts: PETEX 2002 CD-Rom, mk:@Store:D:\petex 2002.chm: /file0022.htm. FLETCHER, K.J. 2003. The South Brae Field, Blocks 16/07a,16/07b, UK North Sea. In: GLUYAS,J.G. & HICHENS, H.M. (ed.) United Kingdom oil and gas fields, commemorative millennium volume, Geological Society, London, Memoir 20, 211-221. FRASER, S.I., ROBINSON,A.M., JOHNSON,H.D., UNDERHILL, J.R., KADOLSKY, D.G.A., CONNELL, R., JOHANNESSEN, P & RAVNAS, R. 2003. Upper Jurassic. In: EVANS,D., GRAHAM,C., ARMOUR,A. & BATHURST, E (eds) The Millennium Atlas: Petroleum geology of the Central and Northern North Sea, Geological Society, London, 157-190. GAMBARO, M. & DONAGEMMA,V. 2003. The T-Block Fields, Block 16/17, UK North Sea. In: GLUYAS, J.G. & HICHENS, H.M. (ed.) United Kingdom oil and gas fields, commemorative millennium volume, Geological Society, London, Memoir 20, 369-382. GARLAND,C.R. 1993. Miller Field: reservoir stratigraphy and its impact on development. In: PARKER, J.R. (ed.) Petroleum geology of Northwest Europe: Proceedings of the 4th conference, Geological Society, London, 401-414. GLENNIE, K.W. & ARMSTRONG,L.A. 1991. The Kittiwake Field, Block 21/18, UK North Sea. In: ABBOTTS, I.L. (ed.) United Kingdom oil and gas fields, 25 years commemorative volume, Geological Society, London, Memoir 14, 339-345. GLUYAS, J. 2001. Upper Jurassic play fairways of the south Viking graben, Gluyas Petroleum Geoscience (commissioned for ConocoPhillips). GUNN, C., MACLEOD,J.A., SALVADORP. & TOMKINSON, J. 2003. The MacCulloch Field, Block 15/24b, UK North Sea. In: GLUYAS,J.G. & HICHENS, H.M. (eds) United Kingdom oil and gas fields, commemorative millennium volume, Geological Society, London, Memoir 20, 453-466. GUSCOTT, S., RUSSEL, K., THICKPENNY,A. & PODDUBIUK, R. 2003. The Scott Field, Block 15/21a, 15/22, UK North Sea. In: GLUYAS, J.G. & HICHENS, H.M. (ed.) United Kingdom oil and gas fields, commemorative millennium volume, Geological Society, London, Memoir 20, 467-482. HALBOUTY,M.T. (ed.) 1982. The deliberate search for the subtle trap, American Association of Petroleum Geologists, Memoir 32. HARMS, J.C., TAKENBERG,E, PICKLES, E. & POLLOCK, R.E. 1981. The Brae Oilfield area. In: ILLING,L.V. & HOBSON, G.D. (eds) Petroleum geology of the continental shelf of North-West Europe, Institute of Petroleum, London, 352-357. HARKER, S.D., GREEN, S.C.H. & ROMANI, R.S. 1991. The Claymore Field, Block 14/19, UK North Sea.

SEALING SURFACE CLASSIFICATION In: ABBOTFS, I.L. (ed.) United Kingdom oil and gas fields, 25 years commemorative volume, Geological Society, London, Memoir 14, 269-278. JONES, E., JONES, B., EBDON, C., EWEN, D., MILNER, P., PLUNKETT, J., HUDSON, G. & SLATER, R 2003. Eocene. In: EVANS,D., GRAHAM,C., ARMOUR,A. & BATHURST,P. (eds) The millennium atlas, petroleum geology of the central and northern North Sea, Geological Society, London, 261-277. JONES, I.E, CHRISTENSEN,R., HAYNES,J., FARAGHER,J., NOVIANTI,I. & MORRIS, H. 2004. The Brenda field development: a multi-disciplinary approach, E A G E First Break, 22, 85-91. OAKMAN,C.D. 8~;PARTINGTON,M.A. 1998. Cretaceous. In: GLENNIE,K.W. (ed.) Petroleum geology of the North Sea, basic concepts and recent advances, 4th edn, Blackwell Scientific, Oxford, 294-349. O'CONNOR, S.J. & WALKER,D. 1993. Paleocene reservoir of the Everest trend. In: PARKER, J.R. (ed.) Petroleum geology of northwest Europe: proceedings of the 4th conference, Geological Society, London, 145-160. MAHER, C.E. 1981. The Piper Oilfield. In: ILLING,L.V. & HOBSON, G.D. (eds) Petroleum geology of the continental shelf of North-West Europe, Institute of Petroleum, London, 358-370. MCGANN, G.J., GREEN, S.C.H., HARKER, S.D. & ROMANI, R.S. 1991. The Scapa Field, Block 14/19, UK North Sea. In: ABBOTrS, I.L. (ed.) United Kingdom oil and gas fields, 25 years commemorative volume, Geological Society, London, Memoir 14, 369-376. MILTON,N.J. & BERTRAM,G.T. 1992. Trap styles, A new classification based on sealing surfaces, The American Association of Petroleum Geologists Bulletin, 76, 983-999. MOSELEY, B.A. 1999. Downthrown closures of the Outer Moray Firth. In: FLEET, A.J. & BOLDY, S.A.R. (eds) Petroleum geology of Northwest Europe: Proceedings of the 5 th Conference, Geological Society, London, 861-878. MUNNS, J. & STOKER, S. 2003. UKCS: The future is stratigraphic!, Sharp IOR Newsletter, 2003, http://ior.rml.co.uk/issue5/articles/DTI_strat plays/ strat_plays.htm. NEWTON,S.K. & FLANAGAN,K.E 1993. The Alba Field: Evolution of the depositional model. In: PARKER, J.R. (ed.) Petroleum geology of northwest Europe: proceedings of the 4th conference, Geological Society, London, 161-171. OAKMAN,C.D. & PARTINGTON,M.A. 1998. Cretaceous. In: GLENNIE,K.W. (ed.) Petroleum geology of the North Sea: Basin concepts and recent advances, Blackwell Science, Oxford, 249-349. PARKER, R.H. 1991. The Ivanhoe and Rob Roy Fields, Blocks 15/21a-b, UK North Sea. In: ABBOTTS,I.L. (ed.) United Kingdom Oil and Gas Fields, 25 Years Commemorative Volume, Geological Society, Memoir 14, 331-338.

223

PARTINGTON, M.A., COPESTAKE, P., MITCHENER, B.C. UNDERHILL, J.R. 1993. Biostratigraphic calibration of genetic stratigraphic sequences in the Jurassic-lowermost Cretaceous (Hettangian to Ryazanian) of the North Sea and adjacent areas. In: PARKER, J.R. (ed.) Petroleum geology of Northwest Europe: Proceedings of the 4th Conference, Geological Society, London, 371-386. RITTENHOUSE, G. 1972. Stratigraphic-trap classification. In: KING, R.E. (ed.) Stratigraphic oil and gas fields - classification, exploration methods and case histories, American Association of Petroleum Geologists Memoir 16, SEG Special Publication, 10, 14-28. ROOKSBURY,S.K. 1991. The Miller Field, Blocks 16/7B, 16/8B, UK North Sea. In: ABBOTrS, I.L. (ed.) United Kingdom oil and gas fields, 25 years commemorative volume, Geological Society, London, Memoir 14, 159-164. ROSE, RT.S., MANIGHETrl,A.A., REGAN,K.J. & SMITH, T. 2000. Sand body geometry, constrained and predicted during a horizontal drilling campaign in a Lower Cretaceous turbidite sand system, Captain Field, UKCS Block 13/22a, Petroleum Geoscience, 6, 255-264. SCHMITT, H.R. & GORDON, A.E 1991. The Piper Field, Block 15/17,UK North Sea. In: ABBOTrS, I.L. (ed.) United Kingdom oil and gas fields, 25 years commemorative volume, Geological Society, London, Memoir 14, 361-368. STEPHENSON,M.A. 1991. The North Brae Field, Block 16/7a, UK North Sea. In: ABBOTTS, I.L. (ed.) United Kingdom Oil and Gas Fields, 25 years Commemorative Volume, Geological Society, London, Memoir 14, 43-48. TURNER, C.C. ~; ALLEN, P.J. 1991. The Central Brae Field, Block 16/7a, UK Field North Sea. In: ABBOTrS, I.L. (ed.) United Kingdom oil and gas fields, 25 years commemorative volume, Geological Society, London, Memoir 14, 49-54. WADDAMS, E ~; CLARK, N.M. 1991. The Petronella Field, Block 14/20b, UK North Sea. In: ABBOrrS, I.L. (ed.) United Kingdom oil and gas fields, 25 years commemorative volume, Geological Society, London, Memoir 14, 353-360. WHITEHEAD,M. & PINNOCK, S.J. 1991. The Highlander Field, Block 14/20, UK North Sea. In: ABBOTTS, I.L. (ed.) United Kingdom oil and gas fields, 25 years commemorative volume, Geological Society, London, Memoir 14, 323-329. WRIGHT, S. 2003, The West Brae and Sedgewick Fields, Blocks 16/06a,16/07a,UK North Sea. In: GLUYAS, J.G. & HICHENS, H.M. (ed.) United Kingdom oil and gas fields, commemorative millennium volume, Geological Society, London, Memoir 20, 223-231.

West of Shetland revisited: the search for stratigraphic traps N. L O I Z O U 1, I. J. A N D R E W S 2, S. J. S T O K E R 2 & D. C A M E R O N 2

1Department of Trade and Industry, 1 Victoria Street, London S W 1 H OET, UK (e-mail: [email protected], uk) 2British Geological Survey, D TI Core Store, 376 Gilmerton Road, Edinburgh EH17 7QS, UK Abstract: The West of Shetland area has scope for the stratigraphic entrapment of hydrocarbons at various Jurassic to Palaeogene stratigraphic levels. Mapping and identification of such traps requires a fundamental understanding of the regional geology, the study of analogues and source kitchens, and a thorough approach to trap validation. Since 1982, 47 exploration wells have been positioned on Paleocene prospects with a significant stratigraphic component, but few have found hydrocarbons - many failing as a result of poor trap definition and overconfidence in the predictive use of amplitude anomalies. Hydrocarbon sourcing of many of the failed prospects was also poorly constrained. Few amplitude-related stratigraphic features could be tied with confidence to a viable source kitchen. The presence of a regional seal is a prerequisite ingredient for a successful Paleocene play. Many remaining undrilled, subtle prospects rely on a stratigraphic trapping component, and high-quality 3D seismic data are seen as an essential search tool. Examples of undrilled prospects are presented from the Paleocene of the northern Faroe-Shetland Basin and the Mesozoic of the East Solan Basin and Corona Ridge.

Between 1972 and 2003, 138 exploration wells were drilled in the West of Shetland area, U K Continental Shelf (UKCS) (Fig. 1), with an overall technical success rate of about 1 in 6 (Loizou 2003b). Unlike in the North Sea, where most wells have had structural targets, 47 exploration wells drilled West of Shetland are recognized to have targeted Paleocene traps with a significant stratigraphic component. For these wells, the success rate has been better than 1 in5. A prerequisite for a true stratigraphic trap is a porous and permeable reservoir, which passes laterally on one or more sides into a non-permeable rock by facies changes or pinchout. A classic regional setting for such a trap involves lateral pinchout of a sand facies at the margin of channel deposition. Pure stratigraphic traps are relatively rare, as some degree of structural closure is often evident. The angle of dip of the reservoir relative to the overlying top seal is an important factor in the trapping of significant hydrocarbons (Allan et al. 2006). The successful traps in the Foinaven, Schiehallion and Laggan fields (Fig. 1) have their reservoirs dipping up to 7 degrees steeper than their top seals. Using the exploration techniques available in the past, stratigraphic traps have proved extremely difficult to predict. Historically, most stratigraphic traps on the UKCS have been found serendipitously while drilling structural objectives. By analysing the results of the 47 Paleocene targeted wells, we can obtain a better

understanding of why the success rates for this play appear to have been relatively poor so far. The key question is, by using hindsight, how many wells can be said to have actually drilled valid traps? Furthermore, what is the success rate for wells drilled on valid stratigraphic traps? Can this lead to better expectation for the future? With on-going improvements in seismic technology, and a better u n d e r s t a n d i n g of what represents a valid stratigraphic trap, greater volumes of stratigraphically-trapped hydrocarbons will undoubtedly be discovered West of Shetland. This analysis uncovers some promising areas where the ingredients for potentially successful stratigraphic traps appear to come together.

Key elements for a Paleocene stratigraphic trap, West of Shetland

Trap definition The most important prospect-specific success factor is the presence of a reliable trap model (Loizou 2003a), in particular requiring the accurate prediction of the pinchout of reservoir sands. The ingredients that produce a good stratigraphic trap include the clear identification of reservoir, seal and source. When all these combine favourably with good quality seismic and other key data, then they produce a robust trap model with the optimum chance of success.

From: ALLEN,M. R., GOFFEV,G. R, MORGAN,R. K. & WALKER,I. M. (eds) 2006. The Deliberate Searchfor the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 225-245. 0305-8719/$15.00. 9 The Geological Society of London 2006.

226

N. LOIZOU E T A L .

Fig. 1. Structural elements and main Paleocene depocentres, West of Shetland. The 47 wells used in the analysis of stratigraphic traps are shown in either blue (failure) or red (success). Figure 2 shows a simplified stratigraphic trap model for the Flett Sub-basin. Significant advances in 3D seismic technology in the 1990s have improved trap definition. However, trap definition could be further improved by enhanced processing techniques or the availability of new, purpose-designed, 'high resolution' 3D seismic data. There are, for example, potential sandstone reservoirs beneath the Paleocene T35-T36 regional seal (Ebdon et aL 1995; Fig. 3) that are almost sub-parallel to or have a low angle of dip relative to the seal, but these are difficult to interpret using early to mid 1990s 3D lower-resolution seismic data. Improved, high-resolution 3D seismic datasets should enable a more precise pinchout edge to be interpreted for these sandstone units.

Fig. 2. Simplified model of a West of Shetland stratigraphic trap in Vaila Formation sandstones.

STRATIGRAPHIC TRAPS, WEST OF SHETLAND

227

Reservoir presence In the early stages of West of Shetland exploration, when well control was limited, the first major challenge was to predict sandstone distribution (geographically and vertically) within the Paleocene Faroe-Shetland Basin. Now, with a portfolio of 138 wells (of which 129 have been released by 2004) and several regional studies (published and proprietary), the risk associated with this element of the trap is much reduced (e.g. Ebdon et al. 1995; Mudge & Bujak 2001). The Paleocene Vaila play in the Faroe-Shetland Basin consists predominantly of turbidite sandstone reservoirs in combination structural/ stratigraphic traps. All discovered Paleocene pools have been found in slope turbidites derived from the Scottish hinterland. R e s e r v o i r quality Good quality reservoir sandstones occur in many of the Paleocene sequences in the West of Shetland area. In the Flett Sub-basin, porosities range from less than 10% to greater than 30%, and permeabilities from 0.1 mD to 2 D (Fig. 4). Although there is an overall reduction of reservoir quality with depth of burial, certain sandstone units have retained high porosities (>20%) and permeabilities (10-100 mD) at burial depths below 3 km (Sullivan et al. 1999). Sandstones in the Paleocene T35 Vaila Formation in Laggan Field wells 206/1-2 and -3 have porosity and permeability preservation (or enhancement) at depth (Fig. 4). Although showing the same composition as the older sandstones here, the T35 sandstones are much better sorted, with ubiquitous chlorite grain coating (Sullivan et al. 1999). The presence of this coating appears to have prevented further quartz diagenesis and led to locally preserved, anomalously high porosities. In the adjacent Torridon area, wells 206/2-1 and 214/27-3 have poorer quality T35 sandstones, which are devoid of chlorite. Furthermore, between 150-200 m of tight, non-reservoir quality T25-T28 Lower Vaila Formation sandstones were also penetrated by these wells. The prediction of areas where reservoir quality is best preserved is a major challenge for continuing exploration, particularly in the deeper parts of the Faroe-Shetland Basin. Using the Laggan case, there is a strong relationship between high porosities and high seismic amplitudes; therefore true amplitude preservation is certainly an important element for predicting reservoir quality prior to drilling here.

Fig. 3. Summary of mid-Paleocene stratigraphy, West of Shetland, showing the Kettla Tuff Member and the regional seal.

Fig. 4. Simplified porosity-depth trends in Paleocene sands, West of Shetland (modified after Lamers & Carmichae11999).

Seal p r e s e n c e a n d effectiveness In the Flett Sub-basin, shales within the T35-T36 Vaila Formation sequence combine with the overlying Kettla Tuff Member to form an effective, basin-wide pressure seal (Lamers & Carmichael 1999). The Kettla Tuff is typically 10-50 m thick, while the underlying shales add up to a further 200 m to seal thickness (Fig. 5). A seal of equivalent age is also present in the Foinaven Sub-basin, but by and large it is less well defined there and much thinner (Fig. 3).

228

N. LOIZOU E T A L .

The distribution of hydrocarbons within the Vaila Formation sandstones strongly relates to the extent of the T35-T36 regional seal. In general, an increase in aquifer pressure of 350-650 psi can be observed across the T31-T35 sequence over most of the Flett Sub-basin as in well 214/27-2 (Fig. 6). However, in both the 205/14-1 and -2 wells, where the Kettla Tuff is absent, the Paleocene sequence was normally pressured. The composition of the Kettla Tuff Member varies across the area, and S. Linnard (pers. comm.) interprets it as an influx of basalt outwash material rather than an airfall deposit. The gamma and velocity log responses for the Kettla Tuff are typified by well 206/1-3 (Fig. 5), whereas on the composite log section for well 205/9-1 (Fig. 7) the same sequence is partially described as 'coarse sandstone'. A map illustrating the extent of the Kettla Tuff (Fig. 8) has been constructed as a proxy for the T35-T36 regional pressure seal, and superimposed onto the underlying T34-T35 sand play fairway as an aid to understanding whether the . . .~. . . . .a va.u 1;a ..at,. ... we !!s were o pti mfi !!yl~. . ~. . . a .~u

Fig. 5. Summary of log responses and stratigraphy for the T35-T36 interval in the Laggan 206/1-3 appraisal well.

Fig. 6. Formation pressure data from well 214/27-2 in the Flett Sub-basin, providing an example of the raolnnnl

T'~lq n r a ~ ] l r ~

~onl

Fig. 7. Summary of log responses and stratigraphy for the T35-T36 interval in the 205/9-1 well.

STRATIGRAPHIC TRAPS, WEST OF SHETLAND

229

Fig. 8. Interpreted limits of the area with potential stratigraphic T34-T36 prospectivity. Pale shading colour key as in Figure 1. The map demonstrates that there is indeed a strong relationship between the extent of this seal and of hydrocarbon occurrences in the Faroe-Shetland Basin. However, this is only one stage in predicting the location of subtle hydrocarbon accumulations below that regional seal.

Source rocks and charge The UK Atlantic Margin is part of a passive continental margin that formed as a result of multiphase extension. This extension generated a complex assortment of rift basins during the Mesozoic and Tertiary. Because source rocks have been encountered in only a few of the wells, identifying and extensively mapping the

source rocks on seismic data remains problematic. As a result there have been no realistic estimates of the volumes of hydrocarbons generated and expelled in the area prior to the rifting phase. Nonetheless, the presence of source rocks is not a key geological constraint for the West of Shetland area. The Foinaven Sub-basin is underlain by source rocks of both Middle and Late Jurassic age (Fig. 9). Well data and geochemical modelling suggest that the pre-Tertiary strata initially reservoired oil and gas, but these traps were subsequently breached by later overpressuring caused by rapid burial in the Tertiary. Fields such as Foinaven and Schiehallion, which directly overlie pre-Tertiary fault blocks and lie

230

N. LOIZOU ETAL.

Fig. 9. Geological cross-section across the Foinaven and Flett Sub-basins. This section highlights the importance of the Westray Ridge and Westray Inversion in providing the regional charge/migration focus for the Quadrant 204 Paleocene oil accumulations. The prevalence of gas in the Flett Sub-basin is attributed in part to the presence of a thick Late Cretaceous basin, which developed above the Upper Jurassic source rock interval NE of the Westray Transfer Zone. Location of cross-section on Figure 8. Modified after Lamers & Carmichael (1999). on an inversion anticline, received multiple phases of charge (Iliffe et al. 1999). In the Flett Sub-basin there are only three notable Paleocene gas discoveries - 206/1-2 (Laggan), 214/27-1 (Torridon) and 214/30-1 (Laxford) mainly because a large number of wells have been positioned on invalid traps (Fig. 8). Since the work of Lamers & Carmichael in 1999, the understanding of source rock distribution and pre-Tertiary burial history still remain somewhat speculative. However, based on a number of wells that also encountered 'minor' gas shows within the Vaila Formation (Fig. 10), gas charge in the Flett Sub-basin appears to be persistent.

Direct hydrocarbon indicators (DHIs), amplitude anomalies (AAs), amplitude variations with offset (AVOs) and related geophysical features Given that true stratigraphic traps have little or no structural control, the location of drilling targets that contain convincingly predicted hydrocarbons has relied heavily on additional geophysical techniques, such as DHIs, flat spots,

AAs and AVO technology. AVO technology was introduced in the 1980s and became a primary c o m p o n e n t of seismic exploration West of Shetland throughout the 1990s until the present. Considerable financial investment has been put into AVO studies, and there has been much confidence in its ability to detect the presence of hydrocarbons in reservoirs (or at least to reduce prospect risk). A number of wells were drilled mainly on geophysical anomalies (Table 1). The AVO studies to date indicate that conventional D H I anomalies (soft amplitude anomalies conforming with structure) should be represented in typical hydrocarbon-bearing sands above 2000 m sub-sea-bed (Smallwood & Kirk 2005). Their detection should reduce the level of risk of any shallower prospect. The same studies suggest that D H I anomalies should not be expected below 2500 m (sub-sea-bed) in typical oil-bearing sandstones, or below 2700 m (sub-sea-bed) in typical gas-bearing sandstones. When so-called D H I anomalies are seen at depths of less than 2700 m, it could indicate particularly favourable conditions (e.g. anomalously high porosity reservoir as in the Laggan gas accumulation, the presence of gas, very good signal-to-noise ratio data, or very uniform rock

STRATIGRAPHIC TRAPS, WEST OF SHETLAND

231

Fig. 10. Flett Sub-basin hydrocarbon occurrence. Key: red = tested gas, yellow = gas shows, orange = interpreted gas, blue -- water-bearing, green = sand absent or not penetrated. properties), or it could indicate secondary effects associated with the presence of hydrocarbons (e.g. cementation contrasts near the hydrocarbon-water contact). In these circumstances of seismically invisible pay, the trap must be very well defined, as the level of risk will be much higher than for amplitude-supported targets. Amplitude anomalies are influenced by other factors, such as lithology, porosity, anisotropy, and also fluids. It would be incorrect to infer a direct link between amplitude anomalies and the presence of hydrocarbons.

Post-drill analysis of 47 West of Shetland wells A total of 44 exploration and three appraisal wells West of Shetland are considered to have targeted Paleocene traps with a significant stratigraphic component (Fig. 1, Table 1). In this postdrill analysis, wells were classed as a success if significant volumes of hydrocarbons were discovered. The description 'success' is defined here as a h y d r o c a r b o n accumulation that if tested would flow to surface. It does not necessarily indicate the commercial potential of the discovery. Analysis shows that all of the successful wells are located close to or at the basin margins, with seven discoveries located in the Foinaven Sub-basin (Foinaven, SE Foinaven, Schiehallion, Loyal, Alligin, Cuillin and Arkle) and a further three located in the Flett Sub-basin (Laggan, Torridon and 214/30-1). The Flett SubBasin discoveries all lie immediately west of the Flett Ridge (Fig. 1). Most of these discoveries have a northwesterly structural dip and are sealed up-dip by an E - W or N E - S W fault in combination with stratigraphic pinchout of the Vaila Sandstones.

The post-drill analysis of the failed wells forms the basis of this study (Table 1). Each well has been assessed in terms of the key stratigraphic trap elements (i.e. trap, reservoir, seal and charge). The key reason for most failures in both the Foinaven and Flett Sub-basins has been poor trap definition. However, many wells failed on a combination of geologic components (trap, reservoir, seal, and source). For this analysis, if the trap constituted more than 50% towards the well failing to find hydrocarbons then trap is assigned as the key element for failure. The majority of wells (84%) are deduced to have failed as a result of a poorly defined trap; 8% of the wells failed as a result of thin or absent reservoir, and 8% failed due to the seal being either thin or absent. Intriguingly, none of the wells specifically failed as a result of source rock absence. However, many poorly defined traps could also have failed due to lack of migration. Lamers & Carmichael (1999) published a similar analysis of the Foinaven Sub-basin wells, in which they showed the primary reasons for failure were 74% trap, 13% reservoir and 13% charge. Of the 37 failed wells, around 73 % were positioned too far up-dip to trap hydrocarbons, and about 27% were positioned down-dip of any trapping potential (Fig. 2). Quite surprisingly, none of the failed wells are considered to have tested what constitutes a valid stratigraphic trap (Fig. 8). A p p r o x i m a t e l y 39 wells were positioned on an amplitude or AVO a n o m a l y (Table 1). Of these, nine encountered notable hydrocarbons. Following post-mortem studies, the majority of the 30 wells that failed to find h y d r o c a r b o n s could be shown to r e p r e s e n t poorly u n d e r s t o o d amplitude anomalies (various lithologies including igneous), AVO artefacts and spurious DHIs (which also include

N. L O I Z O U E T A L .

232

Table 1. Post-drill analysis of west of Shetland wells, which targeted Paleocene, traps with a significant stratigraphic component Well number

204/14-2 204/18-1 204/19-2 204/19-3 A 204/19-5 204/19-6 (Appr) 204/20-1 204/20-3 204/20-4 (Appr) 204/22-2 204/24-2Z (Appr) 204/25-1 204/25b-4 204/25b-5 204/26-1A 204/27a-1 205/8-1 205/9-1 205/10-2B 205/10-3 205/10-4 205/10-5A 205/12-1 205/14-1 205/14-2 205/14-3 205/17a-1 205/17b-2 206/1-1A 206/1-2 208/15-2 208/17-1 208/17-2 208/19-1 208/21-1 208/22-1 208/23-1 208/24-1A 208/27-2 214/24-1 214/27-1 214/27-2 214/27a-3 214/27a-4 214/28-1 214/29-1 214/30-1

Year

1998 2001 1991 1994 1995 1995 1993 1994 1995 1994 1992 1991 1995 1995 1995 1990 1996 1989 1984 1985 1997 1997 1995 1990 1996 1997 1995 1995 1985 1986 1995 1985 1995 1983 1985 1986 1983 1986 1982 1998 1985 1986 1997 2000 1984 1985 1984

Amplitude anomaly on 2D or 3D seismic

Success (with name) or failure

3D 3D 2d 3D 3D 3D 2d 3D 3D 3D 2d 2d 3D 3D 3D 2d 2d 2d

Failure Failure Arkle Cuillin Failure Alligin Schiehallion Loyal Failure Failure Foinaven Failure Failure SE Foinaven Failure Failure Failure Failure Failure Failure Failure Failure Failure Failure Failure Failure Failure Failure Failure Laggan Failure Failure Failure Failure Failure Failure Failure Failure Failure Failure Torridon Failure Failure Failure Failure Failure Laxford

2d 3D 3D 3D 2d 3D 3D 3D 3D 2d 2d 3D 2d 3D 2d 2d 2d 3D 2d 2d 3D 3D

Reason for failure (key reason = X) Trap

multiples). A large n u m b e r of failed wells w e r e p o s i t i o n e d o n i n t e r p r e t e d A V O or high amplit u d e f e a t u r e s b e l i e v e d to c o i n c i d e with t h e t e r m i n a t i o n or up-dip l i m i t / p i n c h o u t e d g e of a s a n d s t o n e interval. F u r t h e r m o r e , w o r k c a r r i e d out by most c o m p a n i e s on these features

Reservoir

Seal

Charge

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

i m p l i e d that a h y d r o c a r b o n a c c u m u l a t i o n was p r e s e n t . For a n u m b e r of failed cases, the cause o f t h e A V O or high a m p l i t u d e f e a t u r e s was misinterpreted. C o m p l i c a t i o n s in A V O r e s p o n s e d u e to overlying c o n d e n s e d sections or variations in r o c k

STRATIGRAPHIC TRAPS, WEST OF SHETLAND property can significantly reduce or even destroy AVO responses. For example, Margesson & Sondergeld (1999) deduced that dry well 208/17-2 had drilled a manifestation of anisotropy and not a predicted hydrocarbonsrelated AVO anomaly. Hence AVO studies cannot be used as the only key measure of prospect risk, but they must be combined with other techniques. Foinaven displays a classic and easily understood Class 3 AVO response (E. Liu, pers. comm.), which, if seen within an exploration prospect, would be significant in reducing prospect risk. Foinaven is also an excellent example of a soft/negative acoustic response that increases with offset angle. However, in other cases in which hard shales overlay hard sands, the far offset is usually a negative response that actually dims with offset. Whilst explorationists normally appreciate the presence of higher porosity reservoirs, the downside for the Flett Sub-basin is that seismic anomalies generated by normal porosity sandstone containing hydrocarbons are indistinguishable from anomalously high porosity sandstone that is brine-filled. No pattern has yet been detected either in the geographical or stratigraphical distribution that would allow significant risk reduction of this 'False D H I ' phenomenon in the Flett Sub-basin. It is therefore difficult to separate out amplitudes associated with gas from those related to high porosities. In the right structural/stratigraphic context the A V O / D H I approach can be powerful, even without much geophysical understanding. Unfortunately, the majority of the failed well prognoses were more heavily influenced by geophysics-based deductions than by actual geology. There has been a proliferation of AVO analyses with too little focus on determining how the AVO anomaly is located with respect to receiving and trapping hydrocarbons. AVO methods can in certain cases add reliable constraints to quantitative reservoir characterization if underlying concepts and how to apply the technology is understood. Much of the AVO work was carried out on 3D seismic datasets that had angle offsets of up to 35 ~ (realistically reliable for AVO analysis to a sub-sea depth of approximately 2.1 km), which are not ideal for robust AVO studies. At least 75 % of the drilled AVO anomalies were at depths greater than 2.1 km subsea. For increased accuracy and confidence in AVO analysis, there is a need to acquire seismic data with offsets longer than 5 km. Despite the pitfalls, the use of AVO/DHI has been fairly widespread West of Shetland since the early 1990s. The optimal setting for the technique is for gas detection in shallow, porous,

233

poorly consolidated clastic rocks of Eocene and Paleocene age. A good example where AVO has worked effectively is the Foinaven area, mainly because the burial depth here for the T31-T35 reservoir sandstones is generally less than 2500 m below sea-bed.

Examples of successful Paleocene stratigraphic traps It is significant to note that all the wells that have encountered hydrocarbons are associated with amplitude anomalies that, at least partly, conform to structure, e.g. Foinaven (Lamers & Carmichael 1999).

Foinaven Oil Field (discovery well 204/24a-2) The Foinaven/Schiehallion geological setting is unique in terms of hydrocarbon charge history, reservoir quality and trapping style (Figs 9 & 11). All the traps in the Foinaven Sub-basin are combined structural/stratigraphic traps (Cooper et al. 1999; Leach et al. 1999). The Vaila T35-T36 sequence, which includes the Kettla Tuff, provides an effective top seal across the subbasin, and all the significant discoveries have been made in the T31-T35 fairway directly underlying it. The hydrocarbon-saturated sandstones generate strong seismic amplitude anomalies, which help to define the extent of the traps (Lamers & Carmichael 1999), and hence the trapping mechanism, with a high degree of confidence (Fig. 12). On the seismic line through the 204/24a-2 discovery well, all the reservoir sandstones appear to pinchout in a similar position, but this is not generally true moving away from this location. As with all depositional systems, there is, of course, a structural control on the extent of the sands, which changes through time.

Laggan Gas Field (discovery well 206/1-2) Located in block 206/la within the Flett Subbasin, the Laggan gas accumulation represents an unusual example of a stratigraphic trap. Laggan was discovered in 1986 by well 206/1-2 (positioned on 2D seismic data), which encountered pay in T35 Vaila Formation sandstones. Ten years later, appraisal well 206/1-3 also found gas within the same reservoir sandstones 4 km to the SW. During 2004, Total drilled two successful appraisal wells (206/la-4A and 206/la-4Z) to further evaluate the potential Laggan gas accumulation. The 206/la-4Z well

234

N. LOIZOU ETAL.

Fig. 11. Geoseismic section showing the position of the 204/18-1 Assynt prospect well (poor trap) and the up-dip Foinaven area 204/19-3A Cuillin well (good trap). Location of section shown on Figure 8.

tnc UlO-UqOl l J J l J t tu~ U I C ~aggan ~,~ accumulation is interpreted to be a pinchout of T35 Vaila Formation sandstones almost against a NE-trending set of growth faults (Figs 13 & 14). The faults appear to have influenced the geometry of sandstone deposition. Not surprisingly, the high amplitudes displayed on seismic data represent the extent of the high porosity sandstones, which terminate quite close to the west of the growth faults. Although a gas-water contact was established by well 206/1-3, giving the down-dip limits of the Laggan accumulation, the high amplitudes associated with the sands extend beyond the gas-water contact. This suggests that the amplitudes at least partly indicate the extent of high porosity sands, and not exclusively the occurrence of gas.

Examples of poorly defined traps and lack of reservoir

Fig. 12. (a) 3D seismic line through the Foinaven 204/24-2A discovery showing that hydrocarbonsaturated sands generate a strong seismic amplitude (coloured inversion of full stack in depth) courtesy of BP, (b) Geoseismic interpretation. Note that in other areas of the Foinaven Field, T35-T36 sandstones also contain hydrocarbons. tested at a rate of approximately 36 mmscfd, whilst the original discovery well flowed at 25 mmscfd.

The analysis of prospect failures highlights the inadequacy of trap definition in many cases. Two examples are presented below. Q u a d r a n t 205 N o r t h

In the southern Flett Sub-basin, wells 205/8-1, 205/9-1, 205/14-2 and 205/14-3 (Figs 7 & 15) were drilled on amplitude/AVO features predicted, pre-drill, to indicate the presence of hydrocarbons, but all of these features turned out to be lithology-related. Wells 205/8-1 and 205/9-1 were positioned using 2D data, while the

S T R A T I G R A P H I C TRAPS, WEST OF S H E T L A N D

235

Fig. 13. Amplitude and structure map over the Laggan gas discovery area, block 206/la. The highlighted area shows Total's current interpretation of sand extension. Courtesy of Total.

Fig. 14. Example 3D seismic line across the Laggan appraisal well 206/1-3.3D seismic line courtesy of Total. Location of section on Figure 13.

236

N. LOIZOU ETAL.

Fig. 15. Geoseismic section through dry wells 205/9-1,205/8-1,205/14-3 and 205/14-2. Location of cross-section on Figure 8.

other two wells were located on the basis of 3D seismic data. Importantly, none of these wells were situated on what can in hindsight be described as a valid robust trap. There has been some ambiguity in the well correlations in the area, particularly of the T35-T36 sequence between 205/9-1 and 205/14-3. Well 205/9-1 encountered 425 m of good quality Paleocene Vaila T35-T36 Formation sandstones, whereas, well 205/8-1, located 8.5 km up-dip of this well only encountered 27 m of Vaila T35 sandstones. Well 205/8-1 is reported to have 'dubious minor oil shows' within the T36 and T38 sands (Smallwood et al. 2004). Further up-dip, well 205/14-3 failed to encounter any Vaila T35-T36 sandstones beneath what is described on the composite log as the Kettla Member. The lowermost 15-18 m of this 'Kettla' interval represents a tight sequence, which is considered as part of the regional seal. Interestingly, the T34-T35 Vaila Formation in the 205/91 well is overpressured by 363 psi (Lamers & Carmichael 1999), and the 205/8-1 well is also mildly overpressured beneath the regional T36 pressure seal. Well 205/14-3 had several RFTs taken from a 'sandstone unit' within the T36 Kettla Member interval, which not surprisingly indicate normal pressure. Intriguingly, the pinchout of the T35 Vaila Sandstones is inferred to be approximately 2 to 3 km down-dip of the 205/14-3 well. In contrast, the 205/14-2 (also 205/14-1) well, which lies beyond the regional pressure seal, is normally pressured. All four wells shown in

Figure 15 failed to locate hydrocarbons, but more importantly, none are positioned on a valid trap. Well 205/14-2 was unwisely located 3.5 km from and marginally up-dip of 205/14-1 and drilled the same play but on a brighter amplitude feature created by 'lithology tuning' (Smallwood et al. 2004). Not surprisingly, both wells were dry as, realistically, supplying these localized sands with hydrocarbons would be virtually impossible; vertical migration through more than 2500 m of underlying Cretaceous mudstones would be required. Additionally, they are cut off from the more likely migration of the main Vaila sandstone fairway encountered down-dip.

The A s s y n t prospect (well 204/18-1) The Assynt prospect in the Foinaven Sub-basin was largely based on amplitude analysis and was proven dry by well 204/18-1 in 2001. Pre-drill, the Assynt prospect was interpreted as comprising stacked sandstone intervals in the Upper and Lower T35 Sequence deposited as slope turbidites in three main channels orientated N-S, parallel to the structural dip (Fig. 16). The prospect was interpreted to be a direct fairway analogue to discoveries such as Foinaven. Compared to Foinaven, however, there was no evidence of true amplitude conformance with depth. The predominantly stratigraphic nature of the Assynt prospect relied heavily on the definition of a sealing mechanism. At the Foinaven and Schiehallion Fields, the existence of a thick

STRATIGRAPHIC TRAPS, WEST OF SHETLAND

237

Fig. 16. (a) Example 3D seismic line through the 204/18-1 Assynt well. Seismic courtesy of Veritas DGC. (b) Assynt amplitude anomaly map (red/yellow = high amplitude; blue/purple = low amplitude). (e) Contours in depth (m) to top Assynt amplitude anomaly.

and dominantly mud-prone T35 lowstand wedge provides a ubiquitous top seal. The location of Assynt suggested that it is downslope of, or even within the basinward equivalent of, this package. The T36 sequence would thus be required to provide the ultimate top seal to the Assynt prospect. Post-drill AVO analysis (E. Liu, BGS, pers. comm.) of the Assynt amplitude anomaly shows a fundamental difference to the operator's AVO analysis, which pre-drill suggested the presence

of hydrocarbons (Class 3 type AVO). On the near and mid offset stacks (375-2241 m) the Assynt amplitudes are quite strong; however, on the far offsets (2241-3174 m) the amplitudes are much weaker. The AVO and various attribute analyses conclusively show no evidence of hydrocarbon presence. More significantly, postdrill analysis of the Assynt amplitude anomaly identifies it as a Class 1 type AVO. In a geologic/geomorphologic context, the strong amplitudes are mainly confined to channels that

238

N. LOIZOU ETAL.

show evidence for incision into the underlying strata. They are thought to record a significant contrast in rock properties between the highporosity channel fill and the surrounding sediments, but without supporting geological evidence this should not have been interpreted as conclusive proof of hydrocarbon charge in the prospect. Comparison of the regional setting of the Assynt prospect to the nearby Foinaven Field (Fig. 11) reveals a fundamental problem for the validity of the trap. Essentially, there is very limited scope for stratigraphic trapping potential. In order to trap the sizeable reserves that were anticipated for the Assynt amplitude anomaly, a significant sealing fault would have been required to prevent migration up-dip from Assynt along T34-T36 sandstones and siltstones directly into the Foinaven Field area. Furthermore, the sourcing for the Assynt prospect is less straightforward than at Foinaven, where large basin-marginal faults also provide a direct migration route from the underlying Middle to Upper J . . . . . . ;. . . . . . . . rocks (Figs 9 & 11).

basin-bounding faults, such as that defining the north-west of the Flett Ridge, acted as a focus for migrating hydrocarbons. Generally, the seal for all of the identified traps requires the pinchout reservoir sandstone to be encased in mudstone. The overlying Kettla Tuff and T35-T36 mudstones provide a regional seal to many oil and gas accumulations in the Faroe-Shetland Basin (e.g. Foinaven Field and Laggan gas accumulation). The nature of the underlying strata is less certain, but these are required to provide a bottom seal to prevent leakage up-dip of any mapped prospect to the SE. The quality of the bottom seal is thus the principal risk for this type of play. As previously discussed, the Paleocene regional seal causes the underlying succession to be overpressured by up to 650 psi above the hydrostatic gradient (Fig. 6). By and large, the seal thins gradually towards the SE onto the Flett Ridge; it also thins and onlaps the Corona Ridge to the NW of the basin. Analysis suggests that where the Kettla Tuff ~ - ;~ ~'" ~ , ~ is present, ,,,~'1,1~'r,ho~,,.l~ . . . . . . and Upper Cretaceous sediments below are likely to be overpressured. Overpressured rocks are prone to hydro-fracturing, providing potential

Identification of new stratigraphic concepts There can be little doubt that stratigraphic traps remain an attractive proposition West of Shetland, but finding and de-risking such traps requires improved, high quality, targeted 3D seismic data and a more comprehensive understanding of the local geology and rock physics. The examples reviewed below are based on the interpretation of mid-1990s 3D seismic data, which were less than optimally acquired to aid identification of subtle stratigraphic traps.

Paleocene stratigraphic traps A study area comprising fifteen UK blocks within the Flett Sub-basin was evaluated to investigate the potential for Vaila Formation sandstone stratigraphic plays beneath a regional intra-Paleocene unconformity. An example of an undrilled Vaila pinchout prospect is located on block 214/25 (Figs 17 & 18) and is described in more detail elsewhere (DT12004). Significant gas discoveries were noted in wells 206/1-2 and 214/27-1, and gas shows in several wells (Fig. 10) indicate that there is a strong likelihood of further gas accumulations in the Flett Subbasin. The majority of these gas indications lie on or immediately west of the Flett Ridge (Fig. 1). It appears that this Ridge and large

Fig. 17. Example of a potential untested Paleocene stratigraphic trap, located in block 214/25 in the Flett Sub-basin. Depths are in feet to a Paleocene intraVaila Formation event. The subcrop of this event beneath the regional seal has defined the limit of the prospect.

STRATIGRAPHIC TRAPS, WEST OF SHETLAND

239

Fig. 18. 3D seismic line showing a potential untested stratigraphic trap beneath the T36 regional seal in block 214/25. The target trap is defined by the pinchout up-dip of a Paleocene intra-Vaila Formation reservoir unit beneath the regional seal. Location of section shown on Figure 17. Released Shell seismic data, available from PGS Geophysical.

pathways for vertical migration from Jurassic source rocks into the sub-Kettla sandstone bodies. In this scenario, there may be no need for significant faults to be present. In the Flett Sub-basin study area, the NW trending Clair Transfer Zone (Rumph et al. 1993) could also have acted as a migration conduit. A good example is the 214/27-1 gas discovery, which confirms that migration of gas into closures distant from the Flett Ridge can take place. Individual gas accumulations, such as those in the lower Vaila Formation T25 sands in well 214/27-1, have been effectively sealed by intraformational claystones (Fig. 9), in this case forming a four-way closure over a NE trending shale diapir. Where prospects rely wholly or partly on stratigraphic trapping, the risk of

leakage through seismically unresolved sands is always present (even with high resolution 3D seismic data). Whilst only a few valid combination structural/stratigraphic traps have been drilled, the success rate of more than 50% for these has been relatively high. In the Flett Sub-basin, reservoir sandstones can occur throughout the Paleocene section. Their presence is not a critical factor for the study area, as reservoir is generally well calibrated and proven in nearby wells. The Upper Vaila F o r m a t i o n is relatively sand-prone in nearby wells, with up to 100 m of net sandstones. The anticipated reservoir depths for the Vaila sequence in the study area range from 2100 m to about 3200 m. Based on well control for this interval, the porosities are expected to range

240

N. LOIZOU ETAL.

from 10-25% depending on the burial depth (Fig. 4). Not surprisingly, all reservoirs containing gas shows and significant reserves occur below the Kettla Tuff. These reservoirs range from lower Vaila Formation T22 sandstones that form the reservoir in the Torridon discovery well (214/27-1) to younger Vaila Formation T28-T35 sands that partly form the reservoir in the previously discussed Laggan discovery. Geochemical and well data indicate that the chief source rocks in the Faroe-Shetland Basin are Jurassic in age (Iliffe et al. 1999). These source rocks are predicted to be over-mature for oil generation in this part of the Faroe-Shetland Basin. They have generated gas, e.g. in the nearby Laggan discovery to the south, but cannot be readily mapped on seismic data. What is particularly interesting is that all the identified leads (not presented in this paper) are near down-dip faults, which have also served as conduits for the hydrocarbons.

Cretaceous stratigraphic traps The likelihood of finding commercial hydrocarbon accumulations within the Cretaceous interval relies heavily on identifying and accurately defining significant traps and good reservoir quality sandstones. Generally, three types of Cretaceous traps are recognized: 4-way dipclosed structures, fault-bounded three-way dip closures, and stratigraphic pinchouts. Unfortunately, there are no available analogues for Cretaceous stratigraphic pinchout traps, as none have been drilled West of Shetland. Nonetheless, potential does exist for Turonian Commodore Formation stratigraphic plays within the Faroe-Shetland Basin, with the main risks being reservoir presence and effectiveness. An example of a stratigraphic trap has been mapped immediately west of the Corona Ridge (DTI 2004), located on block 213/20 and adjacent blocks (Figs 19 & 20). This trap is interpreted to comprise basin-floor sandstones encased within basinal mudstones. The predicted reservoir sandstones are within a wedging unit towards the base of the Upper Cretaceous Shetland Group. Turonian sandstones are present in both of the nearest wells, 214/9-1 and 213/23-1, but it is not known whether these were derived ultimately from Greenland to the NW or from the UK landmass to the SE. The hydrocarbon source is expected to be mature Upper Jurassic Kimmeridge Clay Formation mudstones, which are predicted from seismic interpretation to occur down-dip to the NE. Overall, the structural configuration is conducive to hydrocarbon migration and

Fig. 19. Example of a potential untested Upper Cretaceous stratigraphic trap, west of the Corona Ridge and located mainly in block 213/20. Amplitude extraction map from the top of the Upper Cretaceous wedge, superimposed on depth contours (ft). High amplitude = blue/green, low amplitude -orange/brown; contour interval = 250 ft.

focusing towards the Corona Ridge. Top seal is provided by thick Upper Cretaceous (Shetland Group) mudstones. The nature of the up-dip fault seal is uncertain, but the presence of reservoir remains the principal risk for this prospect. Interestingly, there is a strong amplitude anomaly at the up-dip culmination of the mapped prospect, possibly implying the presence of gas, and providing a degree of confidence in the validity of the trap. Furthermore, there is a brightening of amplitudes above in the Eocene Balder Formation (Fig. 20), which could signify further evidence of an active hydrocarbon system. However, there are no obvious gas chimneys on the seismic data to suggest that the vertical seals above the Turonian target reservoir have been breached, either by fracturing associated with an episode of Oligocene inversion, or by capillary failure. The depth of the prospect (5360 m) is fairly significant in terms of reservoir quality. The 204/19-1 well penetrated an Upper Cretaceous reservoir with porosities ranging from 11-21% at 4000 m depth. With normal burial conditions, porosities ranging

STRATIGRAPHIC TRAPS, WEST OF SHETLAND

241

Fig. 20. Interpretation of 3D seismic line showing an example of a potential untested Cretaceous stratigraphic trap west of the Corona Ridge and mainly located in block 213/20. AA = amplitude anomaly. Location of seismic line on Figure 19. Seismic data courtesy of PGS Geophysical.

between 8-15% are expected for the 213/20 prospect. However, overpressuring is possible in this deep part of the basin that may have preserved better-quality porosity and permeability. Two further basin-floor fan prospects have been identified from seismic interpretation in a shallower depth setting (DTI 2004) within the Lower Cretaceous strata of the East Solan Basin (Figs 21 & 22). An upper fan unit is interpreted to be sourced from the Rona Ridge to the NW (Fig. 1), as suggested by the thickness distribution of the fan, and evidence of downcutting on the seismic data. Interestingly, the fan geometry is closely matched with low RMS amplitudes for this 'upper fan' interval (Fig. 21); a similar response is observed for the Upper Jurassic reservoir interval at the Solan oil discovery on the SW flank of the basin. Not surprisingly, because of the up-dip slope setting in the area, there are no Lower Cretaceous sandstones encountered within the Solan Field

wells. A lower and slightly deeper basin floor fan unit is also interpreted. The potential reservoirs in both fans are anticipated to be locally sourced from the Kimmeridge Clay Formation, with the principal risk being their lateral pinchout seal.

Potential Jurassic Stratigraphic Traps Overlying the eastern portion of the Strathmore Lower Triassic oil accumulation in block 205/26a (Fig. 1) is the oil-bearing basin-floor Solan Sandstone that sits within the Upper Jurassic Kimmeridge Clay Formation; the latter thickens and dips northeastward into the East Solan Basin. At the Solan Field (Figs 21 & 22), the Solan Sandstone forms a stratigraphic trap, onlapping and pinching out southwestwards against an intrabasinal high created by the Judd Transfer Zone, which separates the East Solan Basin from the South and West Solan Basins (Fig. 1; Herries et al. 1999). The oil in both the Strathmore and Solan

242

N. LOIZOU E T A L .

Fig. 21. RMS amplitude map for a Lower Cretaceous 'upper fan' interval showing the Solan Field and Lower Cretaceous 'upper fan' prospect in the East Solan Basin (low amplitude = red/yellow, low amplitude = green/blue). Faults are shown in red. accumulations was generated in the East Solan Basin from the Kimmeridge Clay Formation. B o t h accumulations have similar oil-water contacts, and both share a heterogeneous oil

column that becomes richer in asphaltine with depth, possibly reflecting two h y d r o c a r b o n charges. The Solan Field forms a self-sourcing, self-sealing U p p e r Jurassic accumulation (Herries et al. 1999). Reservoir extent and prospectivity is predicted elsewhere in the East Solan Basin from the total Late Jurassic isochron. A potential Solan analogue prospect is interpreted in adjacent, unlicensed acreage (Figs 23 & 24). However, exploration for potential Solan analogues requires the presence of thick, Late Jurassic section as a possible indication that the thickening relates to the presence of Solan Sandstone. A combination structural/stratigraphic closure has been mapped at top Upper Jurassic level at the northeastern corner of the East Solan Basin in open block 205/27 (Figs 23 & 24; DTI 2004). The reservoir is prognosed to be composed of basin-floor fan sandstones of the Solan Sandstone. The wells in the Solan Field have encountered up to 30 m of reservoir, which is difficult to resolve on seismic data through normal interpretation. However, RMS amplitude extraction of the Upper Jurassic interval (Fig. 23) reveals an area of low amplitudes crossing the East Solan Basin that, by analogy

Fig. 22. 3D seismic line showing a basin-floor, Lower Cretaceous stratigraphic trap in block 205/27 in the East Solan Basin. Location of seismic line on Figure 21. Released BP seismic data.

STRATIGRAPHIC TRAPS, WEST OF SHETLAND

243

Fig. 23. (a) Two-way time map to top Upper Jurassic, showing the Solan Field and an undrilled Jurassic prospect in the East Solan Basin (shallow depth = yellow/red/green). (b) RMS amplitude map for the Upper Jurassic interval (low amplitude = red/yellow, high amplitude = blue/purple).

Fig. 24. 3D seismic line across an Upper Jurassic, Solan analogue prospect in the East Solan Basin. Location of seismic line on Figure 23a. Released BP seismic data.

with a comparable response at the Solan Field, is interpreted to indicate the presence of the Solan Sandstone. Interestingly, well 205/27-2, located 2 k m from

the identified prospect (Fig. 23), encountered a 10 m-thick basal U p p e r Jurassic sandstone with minor oil shows that was incorrectly ascribed to the Solan S a n d s t o n e on the c o m p o s i t e log.

244

N. LOIZOU E T A L .

Instead, this unit is the R o n a Sandstone of probable shelfal origin, which typically has relatively poor reservoir quality. In contrast to the Solan Field and the prospect, the area around this well has high RMS amplitudes.

Conclusions The analysis of 37 failed wells shows that 84% were located on unreliable traps. Not surprisingly, those wells located on more reliable, robust structures achieved a higher success rate of approximately 60%. Therefore, the sound mapping of a valid trap is viewed as the key component to increasing exploration success West of Shetland. A key observation from the analysis of the wells indicates that many were not optimally positioned to test a valid stratigraphic trap. With this in mind, exploring for valid stratigraphic structures requires a great deal more care and an improved understanding of specific trap ingredients than are necessary to generate a successful structural trap. So what makes a valid hydrocarbon trap? A valid trap can be defined as a robust structural closure or a combination structural/stratigraphic or purely stratigraphic feature that can be m a p p e d with high confidence utilising good quality seismic and other key data. Without doubt, many of the failed Paleocene wells record a general lack of understanding of the occurrence of sandstone pinchout plays relative to the basinal setting and regional seal. The majority of the 37 unsuccessful wells failed to find hydrocarbons because there was no valid trap. Bearing this in mind, correctly identifying and confidently mapping robust stratigraphic traps should result in a much improved success rate. Evaluation of proven examples of successful P a l e o c e n e traps like Foinaven and Laggan, which have a strong stratigraphic component, can add to the understanding of why a large n u m b e r of stratigraphic wells have failed. A fundamental awareness of the key ingredients that constitute a successful stratigraphic trap will contribute to the success of future exploration. Utilizing the appropriate data, robust stratigraphic traps in Paleocene and older successions can be successfully mapped with a high degree of confidence. If all the ingredients that contribute to the making of a stratigraphic trap are present, then future exploration should be viewed more optimistically. The senior author (NL) would like to thank BP and Shell for giving permission to publish NL's postmortem analysis of the Assynt prospect. The authors

gratefully acknowledge Total for provision of the amplitude map and seismic line across the Laggan discovery. This paper is published with the permission of the Director of Oil and Gas Licensing and Exploration, Department of Trade and Industry and the Executive Director, British Geological Survey (NERC). The views expressed in this paper are mainly the opinions of the authors and are not necessarily those of the DTI.

References ALLAN, J., ROSEWAY,J. & SUN, S.Q. 2006. Evaluating risk factors and exploration/development strategies in stratigraphic and subtle traps. In: ALLEN, M.R., GOFFEY, G.E, MORGAN, R.K. & WALKER, I.M. (eds) The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 57-104. COOPER, M.M., EVANS, A.C., LYNCH, D.J., NEVILLE, G. & NEWLEY, T. 1999. The Foinaven Field: managing reservoir development uncertainty prior to start-up. In: FLEET,A.J. & BOLDY,S.A.R. (eds) Petroleum Geology of Northwest Europe: Proceedings of the 5th Conference. Geological ~.)UblK;Ly,bUIIUUII, U/ J--UOZ~. DEPARTMENT OF TRADE AND INDUSTRY.2004. Promote

United Kingdom 2004: Petroleum potential of the United Kingdom Continental Shelf. CD-ROM. EBDON, C.C., GRANGER, P.J., JOHNSON, H.D. & EVANS, A.M. 1995. Early Tertiary evolution and sequence stratigraphy of the Faeroe-Shetland Basin: implications for hydrocarbon prospectivity. In: SCRUTTON, R.A., STOKER,M.S., SHIMMIELD, G.B. & TUDHOPE,A.W. (eds) The Tectonics, Sedimentation and Palaeoceanography of the North Atlantic Region. Geological Society, London, Special Publications, 90, 51-69. HERRIES, R., PODDUBIUK, R. & WILCOCKSON, P. 1999. Solan, Strathmore and the back basin play, West of Shetland. In: FLEET,A.J. & BOLDY,S.A.R. (eds) Petroleum Geology of Northwest Europe." Proceedings of the 5th Conference. Geological Society, London, 693-712. ILIFFE, J.E., ROBERTSON,A.G., WARD, G.H.E, WYNN, C., PEAD, S.D.M. & CAMERON, N. 1999. The importance of fluid pressures and migration to the hydrocarbon prospectivity of the FaeroeShetland White Zone. In: FLEET,A J. & BOLDY, S.A.R. (eds) Petroleum Geology of Northwest Europe: Proceedings of the 5th Conference. Geological Society, London, 601-611. LAMERS,E. & CARMICHAEL,S.M.M. 1999. The Paleocene deepwater sandstone play West of Shetland. In: FLEET,A.J. & BOLDY,S.A.R. (eds) Petroleum Geology of Northwest Europe: Proceedings of the 5th Conference. Geological Society, London, 645-659. LEACH, H.M., HERBERT, N., Los, A. & SMITH, R.L. 1999. The Schiehallion development. In: FLEET, A.J. & BOLDY,S.A.R. (eds) Petroleum Geology of Northwest Europe: Proceedings of the 5th Conference. Geological Society, London, 683-692.

STRATIGRAPHIC TRAPS, WEST OF SHETLAND LoIzou, N. 2003a. Post-well analysis of exploration drilling on UK Atlantic Margin provides clues to success. First Break, 21, 45-49. LoIzou, N. 2003b. Exploring for reliable, robust traps is a key factor to future success along the UK Atlantic Margin. AAPG International Conference & Exhibition, Extended Abstracts with Program. MARGESSON, R.W. t~z SONDERGELD, C.H. 1999. Anisotropy and amplitude versus offset: a case history from the West of Shetlands. In: FLEET,A.J. & BOLDY, S.A.R. (eds) Petroleum Geology of Northwest Europe: Proceedings of the 5th Conference. Geological Society, London, 634-643. MUDGE, D.C. & BUJAK, J. 2001. Biostratigraphic evidence for evolving palaeoenvironments in the Lower Paleogene of the Faeroe-Shetland Basin. Marine and Petroleum Geology, 18, 577-590. RUMPH, B., REAVES,C.M., ORANGE,V.G. & ROBINSON, D.L. 1993. Structuring and transfer zones in the Faeroe Basin in a regional tectonic context. In"

245

PARKER, J.R. (ed.) Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference. Geological Society, London, 999-1009. SMALLWOOD, J.R., KIRK, W.J. & PRESCOTt, D. 2004. Alternatives in Paleocene exploration West of Shetland: a case study. Scottish Journal of Geology, 40, 131-143. SMALLWOOD,J.R. • KIRK, W.J. 2005. Paleocene exploration in the Faroe-Shetland Channel: disappointments and discoveries. In: DORI~,A.G. & VINING, B. (eds) Petroleum Geology: North-West Europe and Global Perspectives: Proceedings of the 6th Petroleum Geology Conference, Geological Society, London, 977-991. SULLIVAN,M., COOMBES,Z., IMBERT,P. t~zAHAMDACHDEMARS, C. 1999. Reservoir quality and petrophysical evolution of Paleocene sandstones in the West of Shetland area. In: FLEET,A.J. & BOLDY, S.A.R. (eds) Petroleum Geology of Northwest Europe: Proceedings of the 5th Conference. Geological Society, London, 627-633.

Potential Eocene and Oligocene stratigraphic traps of the Rockall Plateau, NE Atlantic Margin D. B. M C I N R O Y , K. H I T C H E N

& M. S. S T O K E R

British Geological Survey, Murchison House, West Mains Road, Edinburgh EH9 3LA, UK (e-mail: [email protected], uk) Abstract: Following thermal uplift during the late Paleocene to early Eocene, the denudation of the subaerial hinterland provided a massive sediment supply that led to the development of a number of large, prograding sedimentary wedge systems flanking the Hatton and Rockall basins. Regional seismic data mapping and borehole data indicate that the wedges are Eocene in age and have a high percentage of coarse clastic material typical of highenergy, fluvial or near-shore marine environments. The prograding wedges have been mapped and can be viewed as large, clastic fairways within which trapping at a number of scales exists. Seismic interpretation suggests that the wedges are present at various stratigraphic levels within the Eocene and are locally separated by unconformities. However, all pre-date the margin-wide late Eocene unconformity (C30), which resulted in subsidence and deepening of the Rockall and Hatton basins. A marine transgression inundated most former land areas, and a marked change occurred in basinal facies; a change from fluvial/near-shore clastic sedimentation to deep-water mud and ooze deposition influenced by bottom-currents. These conditions persisted throughout most of the Oligocene and Neogene and hence provided a seal for potential hydrocarbon-bearing sand-prone Eocene reservoirs internal to the wedge-systems. Additional sealing potential may be provided by shale layers internal to the wedges. Buried Eocene pinchout lobes, submarine fans at the base of basalt scarp faces and Oligocene slump deposits also provide potential trapping mechanisms. High, and probably unacceptable, risks include biodegradation and poor seal development due to the typically shallow depth of burial of the wedges. However, the majority of the wedges should be treated as analogues, with some of the deeper examples providing some scope for consideration as exploration targets. The scale of the prograding wedge play fairway is massive, with volumes measured in tens of cubic kilometres.

Cenozoic post-rift sands are currently important targets for h y d r o c a r b o n exploration in the Atlantic Margin region of the United Kingdom Continental Shelf (UKCS). These targets have become more attractive as the U K sector of the North Sea oil province matures, and opportunities to discover large accumulations of oil and gas there decrease. Exploration of older and deeper plays is hindered by the presence of extensive Iceland hot-spot-related late Paleocene/early Eocene volcanics, which obscure the pre- and syn-rift geology on seismic records across large parts of the margin. Consequently shallower, and often subtle, stratigraphic traps have become favoured exploration targets in the A t l a n t i c Margin region. P a l e o c e n e and Eocene basin-floor-fan reservoirs are the principle target around the proven Foinaven and Schiehallion fields in the Faroe-Shetland Basin, and are potentially sealed by lowstand and highstand mudstones (Brooks et al. 2001). Currently attractive hydrocarbon targets in the Rockall and H a t t o n basins are Paleocene-Eocene postrift plays, in addition to Mesozoic tilted faultblock plays. In this paper we present examples

of potential stratigraphic traps identified in the H a t t o n - R o c k a l l area. The H a t t o n - R o c k a l l area is situated in the N E Atlantic Ocean between 450 and 1000 km west of the Scottish mainland (Fig. 1), and is comprised of the R o c k a l l P l a t e a u and the Rockall Trough bathymetric features. The crust underlying the area is continental, and is highly a t t e n u a t e d across the Mesozoic H a t t o n and Rockall basins. Thicker crust exists beneath the intervening H a t t o n and Rockall highs and at the inner continental shelf, while other highs are formed by several C r e t a c e o u s and Paleocene igneous centres. This configuration of basins and highs reflects the rifting and m a g m a t i s m t h a t occurred in the H a t t o n - R o c k a l l area throughout the Palaeozoic, Mesozoic and early Cenozoic, which ultimately led to the separation of Greenland and Europe along an axis further to the west in the early Eocene. Seaward dipping reflectors (SDRs) in the basalt sequence on the western flank of H a t t o n High sit above the approximate location of the continental margin. West of the S D R sequence, the crust thins to n o r m a l

From: ALLEN,M. R., GOFFEY,G. R, MORGAN,R. K. & WALKER,I. M. (eds) 2006. The DeliberateSearchfor the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 247-266. 0305-8719/$15.00. 9 The Geological Society of London 2006.

248

D.B. M C I N R O Y E T A L .

~o~ O

r~

~ , . = _=

~
5 m laterally to the adjacent cliff (see b). Scale is 0.3 m. (b) Top right Exposure on the wave-cut platform between the exposure in a and the main cliff. A network of sand-filled fractures, now tar saturated, forms elongate fracture vents that feed the overlying sand extrusion. A crudely orthogonal pattern of dykes, c. NE/NNE-SW/SSW (light grey) and c. N/NNW-S/SSE (black) feed the lower and upper extradites, respectively in the composite extrudite (shown in Fig. 9b). Scale bar (tape) is 0.3 m. (e) Bottom right A sandstone dyke feeding a 0.25--0.3 m thick sand extrusion in the aeolian Hopeman Sandstone (Late Permian). The dyke is planar and extends into the laminated sandstones within the field of view. The measuring tape is 0.4 m long (modified from Glennie & Hurst 2006).

diverse in extrudites, while intrusive sandstones may display discordant tops and bases and internal deformation bands (Purvis et al. 2002; Jonk et al. 2003), shale-clast breccias, microscale conjugate fault sets (Huuse et al. 2005), and margin-parallel (shale) clast alignment (Duranti & Hurst 2004). Intrusive sands are very unlikely

295

to have any trace of b i o t u r b a t i o n w h e r e a s bioturbation is c o m m o n in extrusive sands. Extrudites provide potentially valuable stratigraphic correlation surfaces within sand injectite systems, which otherwise have discordant relationships with biozones. The chrono- and event-stratigraphic significance of extrudites has, however, not b e e n investigated in any detail. Their value as stratigraphic markers is likely to be enhanced in subsurface studies as their bedding-parallel lithological character is likely to form impedance contrasts that will be imaged by seismic data (the background for Fig. 2). If core or borehole images are available the presence of small intrusive bodies above

296

A. HURST ETAL.

Fig. 8. Interfingering packages of low-angle laminae and beds that have stacked together to form an approximately 1 m thick sand extrudite (Miocene, Majors Creek Beach, Santa Cruz, California). Each package represents a period of flow from a vent in the palaeo-sea-floor. At the base a thin sub-parallel sill feeds the extrudite and local detachment of fiat mudstone clasts occurs. The top of the unit is bioturbated so forming a gradual transition into the overlying Santa Cruz Mudstone Member). All the sands are tar saturated.

approximately bedding-concordant sills provides strong evidence for a non-extrudite origin (Fig. 10), unless these formed during a later phase of sediment remobilization (cf. Huuse et al. 2005).

E x t r u s i v e vs. ' n o r m a l ' d e p o s i t i o n a l sandstones Both on seismic and at outcrop the association between sand injectites in the underlying section and extrudites is critical (Table 1). It is feasible that depositional sands erosionally truncate and overlie sand injectites but typically the presence of sandstone dykes and sills below a sandbody will be a record of a genetic relationship. If the disruption of the underlying strata is substantial this may be revealed on seismic and is certainly visible at outcrop. In core it is highly likely that extrudites have been and will

continue to be confused with 'normal' depositional sandstones as both may contain sedimentary structures; the current paucity of core data from known extrudites limits this comparison. Low-angle lamination and/or bedding in extrudites (Figs 7a, 8 & 9b) may be confused with tabular cross-bedding. Lamination and bedding in extrudites is unlikely to approach unidirectional because of the radial or crudely elliptical flow around point and fracture vents, respectively. It is quite possible that extrudites have already been cored but misinterpreted as dunescale cross-bedding. Deep-marine and alluvial environments are those in which extrusive sandbodies are known to occur. Extrudite sandbodies in terrestrial environments are best k n o w n from recent examples (e.g. Obermeier 1998; Leeder 1999; Gallo & Woods 2004) with limited documentation of ancient examples (Netoff 2002; Chan et al. 2006). Deep-marine examples of extrusive

EXTRUSIVE SANDSTONES (EXTRUDITES)

Fig. 9. (a) An example of highly disrupted strata underlying a sandstone extrusion on Majors Creek Beach (Santa Cruz). The extruded sand varies in thickness dramatically to the left (NE), probably associated with proximity to vents in the sea-floor. Below the extrudite, and in particular in the right field of view, the underburden Santa Cruz Mudstone is brecciated and intruded by a complex series of sandstone dykes and sills (all tar saturated). (b) An extruded sand sheet in the Santa Cruz Mudstone (Miocene, California). The sand is medium grained and tar saturated. The base of the sand sheet is irregular (> 1 m relief in places) and fed by numerous sandstone dykes that cut through the fractured, porcellanous mudstone. Large (some >2 m length) rafts of mudstone (M) are common. The upper surface has a very gentle slope and is conformably overlain by mudstone of the Santa Cruz Mudstone Member (Thompson et al. 1999). Cross bedding and burrows are common in the upper part of the sand sheet.

sandstone bodies are represented in the rock record by sand volcanoes (Gill & Kuenen 1957; Jonk et al. 2006) and submarine extrusive sand sheets ( B o e h m & M o o r e 2002). Given the limited knowledge of intrusive traps, particu-

297

larly on a global basis, it is quite likely that extrusive sandbodies that are associated with them have been encountered but not recognized as of extrusive origin. We are unaware of exploration wells that have deliberately targeted the extrudite stratigraphic elements of intrusive traps. Development drilling on fields known to be affected by sand injection may have drilled through sandrich units that extruded onto a palaeo-sea-ttoor but, although candidate sandbodies exist (Figs 2, 3, 4 & 5), none are proven. As such the prospectivity and reserve potential of extrusive sandbodies is untested. F r o m the dimensions of known extruded sandbodies it is probably rare that isolated extrusive sandbodies will reservoir major reserves, but they are likely to be interesting secondary targets when exploring in areas associated with intrusive traps, or where they merge laterally with similar features (cf. Obermeier 1989, figures 14 to 18). In the Santa Cruz area the extrudites have a large areal extent and have been quarried for their tar. Estimates of gross rock volume of large extrudite bodies may be of the order of 105-108 m 3, equivalent to a possible pore volume of these often poorly cemented sandstones of the order of 105-108 barrels.

Conclusions Extrudite sand sheets are entirely stratigraphic traps associated with sand injectites and intrusive traps. Extrudites have four-way dip closure and typically overlie and underlie lower permeability strata. A l t h o u g h sand injectites are increasingly recognized, particularly in deep-water systems, they are a new play in terms of hydrocarbon prospectivity beyond the N o r t h Sea. As extrudites are even less widely

A. HURST E T A L .

298

Table 1. Guidance for differentiation between extrudites, sills and depositional sands

Seismic Injectite association Max thickness individual group Boundaries top . base Internal structures

Bioturbation Connectivity with underlying units

Extrudites

Sills

Depositional (mounds)

bedding parallel

bedding parallel (+ local discordance)

mounded

yes

yes

yes

30 m

>20 m

concordant (graded) concordant

discordant discordant

concordant (graded) concordant &/or erosional

(i) low-angle lamination/bedding (ii) soft sedimentary deformation (iii) cross bedding

(i) deformation bands (ii) margin-parallel clast alignment (iii) conjugate micro-faults

diverse

common

unknown

present

highly connected porous networks

highly connected porous networks

rarely connected

(a)

(b)

Fig. 10. Cross-sections of possible (a) sill and (b) extrudite geometries that demonstrate the overall similar macro-scale geometry but fundamentally different relationships with adjacent strata. (a) is bedding parallel but has a series of small intrusions above the main sill and has similarities to the geometry inferred in Figure 5. In (b) the extrudite is bedding parallel but shallower intrusions are not present. The sand body has macro-scale similarities to Figure 2 and internal characteristics similar to Figures 8 & 9b. recognized than injectites their d o c u m e n t a t i o n is presently very sparse. However, we believe that they will be recognized m o r e commonly in

the future as the awareness of 'unconventional' s a n d s t o n e occurrences grows in the E & P community. I n d e e d we think that the main reason why extrudites are so rarely characterized is that they have been misinterpreted as depositional units in some subsurface studies. Examination of analogue data suggests that extrudites h a v e uniformly high reservoir quality except for where mudstone clasts are present. Extrusive sand sheets are fed by sand injectite systems, which provide potential pathways for h y d r o c a r b o n migration and aquifer support, and they should thus be an attractive and straight-forward play type, once it is realized that their occurrence cannot be predicted using paleogeographic maps for the stratigraphic interval in which they are located. Some of this work was supported by the sponsors of the Injected Sands Research Consortium, ChevronTexaco UK, Enterprise Oil Norway, Kerr McGee UK, Norsk Hydro, Shell UK, Statoil and TotalFinaElf, to whom we are grateful. Review comments by L. Richmond and R. Fitzsimmons were most helpful. Seismic and well data from the Chestnut Field were kindly provided by WesternGeco. JAC and MH acknowledge the generous software support to the 3DLab at Cardiff University by Schlumberger Information Solutions.

EXTRUSIVE SANDSTONES (EXTRUDITES)

References BOEHM, A. & MOORE, J.C. 2002. Fluidized sandstone intrusions as an indicator of paleostress orientation, Santa Cruz, California. Geofluids, 2, 147-161. CHAN, M., NETOFF, D., BLAKEY, R., KOCUREK, G. & ALVAREZ,W. 2006. Syndepositional deformation structures associated with Jurassic eolian deposits; Examples from the Colorado Plateau. In: HURST, A. & CARTWRIGHT, J.A. (eds) Sand Injectites: Implications for Hydrocarbon Exploration and Production. American Association of Petroleum Geologists, Memoir 87, Tulsa, Oklahoma, in press. DE BOER, W., RAWLINSON, P. & HURST, A. 2006. Successful exploration of a sand injectite complex: Hamsun prospect, Norway Block 24/9. In: HURST, A. & CARTWRIGHT, J.A. (eds) Sand Injectites: Implications for Hydrocarbon Exploration and Production. American Association of Petroleum Geologists, Memoir 87, Tulsa, Oklahoma, in press. DURANTI,D. & HURST,A. 2004. Fluidisation and injection in the deep-water sandstones of the Eocene Alba Formation (UK North Sea). Sedimentology, 51, 503-531. DURANTI, D., HURST, A., BELL, C. & GROVES, S. 2002. Injected and remobilised sands of the Alba Field (UKCS): sedimentary facies characteristics and wireline log responses. Petroleum Geoscience, 8, 99-107. GALLO, E & WOODS, A.W. 2004. On steady homogeneous sand-water flows in a vertical conduit. Sedimentology, 51, 195-210. GILL, W.D. & KUENEN,P.H. 1957. Sand Volcanoes on slumps in the Carboniferous of County Clare, Ireland. Quarterly Journal of the Geological Society of London, 113, 441460. GLENNIE, K.W. & HURST, A. 2006. Fluidisation and associated soft-sediment deformation in eolian sandstones: Hopeman Sandstone (Permian), Scotland, and Rotliegend, North Sea. In: HURST, A. & CARTWRIGHT, J.A. (eds) Sand Injectites: Implications for Hydrocarbon Exploration and Production. American Association of Petroleum Geologists Memoir 87, Tulsa, Oklahoma, in press. HURST, A. 2004. Sedimentology of seafloor sand extrusions: an example from the Miocene of central California. British Sedimentological Research Group Annual General Meeting, Manchester, 19th-21 st December, Abstract. HURST, A., CARTWRIGHT, J.A., DURANTI, D., HUUSE, M. & NELSON, M. 2005. Sand injectites: an emerging global play in deep-water clastic environments. In: DORE, A. & MINING,B. (eds) 6th Petroleum Geology Conference: North West Europe & Global Perspectives, Geological Society, London, 133-144. HUUSE, M., DURANTI,D., STEINSLAND,N., GUARGENA, C., PRAT, P, HOLM, K., CARTWRIGHT, J.A. & HURST, A. 2004. Seismic characteristics of largescale remobilised and injected sand bodies in the Paleogene of the South Viking Graben (North

299

Sea): steep-sided mounds, wings and Vs. In: DAVIES, R.J., CARTWRIGHT, J.A., STEWART, S.A., LAPPIN, M. & UNDERHILL,J.R. (eds) 3-D Seismic Technology: Application to the exploration of sedimentary basins. Geological Society, London, Memoir, 29, 263-277. HUUSE, M., CARTWRIGHT, J., GRAS, R. & HURST, A. 2005. Km-scale sandstone intrusions in the Eocene of the Outer Moray Firth (UK North Sea): migration paths, reservoirs, and potential drilling hazards. In: DORI~, A.G. & MINING, B. (eds) Petroleum Geology: North-West Europe and Global Perspectives - Proceedings of the 6th Petroleum Geology Conference, Geological Society, London, 1577-1594. JOLLEu J.H.R. & LONERGAN, L. 2002. Mechanisms and control on the formation of sand intrusions. Journal of the Geological Society, 159, 605~17. JONK, R., DURANTI, D., PARNELL, J., HURST, A. & FALLICK,A.E. 2003. The structural and diagenetic evolution of injected sandstones: examples from the Kimmeridgian of NE Scotland. Journal of the Geological Society, 160, 881-894. JONK, R., CRONIN,B.T. & HURST,A. 2006. Sand extrusion at the sediment-water interface: sand volcanoes from the Namurian of County Clare, Ireland. In: HURST,A. & CARTWRIGHT,J.A. (eds) Sand Injectites: Implications for Hydrocarbon Exploration and Production. American Association of Petroleum Geologists Memoir 87, Tulsa, Oklahoma, in press. LAWRENCE, D.A., SANCAR, B. & MOLYYEUX,S. 1999. Large-scale elastic intrusion in the Tertiary of Block 24/9, Norwegian North Sea: origin, timing and implications for reservoir continuity. American Association of Petroleum Geologists Bulletin, 83, 1324. LEEDER, M.R. 1999. Sedimentology and sedimentary basins: from turbulence to tectonics. Blackwell Science, Oxford. LOWE, D.R. 1975. Water escape in coarse-grained sediments. Sedimentology, 22, 157-204. MACLEOD, M.K., HANSON, R.A., BELL, C.R. & McHuGO, S. 1999. The Alba Field ocean bottom cable seismic survey: Impact on development. The Leading Edge, 18, 1306-1312. NETOFF, O. 2002. Seismogenically induced fluidization of Jurassic erg sands, south-central Utah. Sedimentology, 49, 65-80. OBERMEIER, S. 1989. The New Madrid earthquakes: an engineering-geologic interpretation of relict liquefaction features. US Geological Survey Professional Paper 1336-B. PURVIS, K., KAO, J., FLANAGAN,K., HENDERSON,J. & DURAYrI, D. 2002. Complex reservoir geometries in a deep-water clastic sequence, Gryphon Field UKCS: injection structures, geological modelling and reservoir simulation. Marine and Petroleum Geology, 19, 161-179. RITTENHOUSE, G. 1972. Stratigraphic trap classification. In: KING, R.E. (ed.) Stratigraphic oil and gas fields - classification, exploration methods and case histories. American Association of Petroleum Geologists, Memoir 16, Tulsa, Oklahoma, 14-28.

300

A. HURST E T A L .

SHOULDERS,S.J. & CARTWRIGHT,J.A. 2004. Constraining the depth and timing of large-scale conical sandstone intrusions. Geology, 32, 661-664. THOMPSON,B.J., GARRISON,R.E. & MOORE, C.J. 1999. A late Cenozoic sandstone intrusion west of Santa Cruz, California: fluidised flow of water and hydrocarbon-saturated sediments. In: GARRISON, R.E., AIELLO, I.W. & MOORE, C.J. (eds) Late Cenozoic fluid seeps and tectonics along the San Gregorio fault zone in the Monterey Bay region,

California. American Association of Petroleum Geologists Pacific Section, Volume and Guidebook GB-76, Tulsa, Oklahoma, 53-74. THOMPSON, B.J., GARRISON,R.E. & MOORE, C.J. 2006. A giant Miocene sandstone injectite near Santa Cruz, California. In: HURST, A. & CARTWRIGHT, J.A. (eds) Sand Injectites: Implications for Hydrocarbon Exploration and Production. American Association of Petroleum Geologists Memoir 57, Tulsa, Oklahoma, in press.

Index Page numbers in italic indicate figures, those in bold denote tables.

accidental discovery 68, 71 see also serendipity Alba Field, UKCS 163,164, 212, 219, 292 analysis, post-drill 232 Ansai Field, China 73, 74, 75, 76, 87 Aptian-Albian play 172-174,174, 178-184 Arbroath Field, UKCS 163 Ardjuna Basin, Indonesia 110 Asahan Offshore, Indonesia 115-116,122,123 Asri Basin, Indonesia 116,117, 119 Assynt prospect, UKCS 234, 236-238 Athabasca Oil Sands, Canada 87 Auk Field/High, UKCS 176,178,183 Australia 92 back-stripping 203 basalt 176-178,183,252, 253,260, 263 basin setting for stratigraphic traps 84-86 bathymetry 197-198,199, 203,251 Beatrice Field, UKCS 190 Bell Creek Field, Wyoming 106 Biliton PSC, Indonesia 116,119,124,125 biodegradation 263 Blake Field 200, 202 Bonga Field, Nigeria 13-14,15 Bowen Basin, Australia 92 Brae Field, UKCS 212, 214, 215, 217 Brenda Field, UKCS 163,212, 219 Britannia Field, UKCS 172, 182, 183,213,219 Buchan Basin, South 173,175,176,177,179,180, 184 Bud Field, Gulf of Mexico 138,145 Bullwinkle Field, case study 81, 82, 83, 84 burial 227,229, 230, 263 Buzzard Field, UKCS 154,159,187-204,215 Canada 8, 9, 87, 92,100,101 Captain Field, UKCS 173,190 carbonate facies 68, 95,128, 158 Carboniferous play, UKCS 156, 165 case histories Bullwinkle Field 81, 82, 83, 84 Buzzard Field, UKCS 187-206 deep water sands 144-150 East Texas Field 63, 66, 68-70, 87 Indonesia 111-124 Jay Field 68, 70, 71, 72, 87 Central Limit Theorem 21-22 Central North Sea 207-222 trap classification 181-183 Central North Sea Graben, UKCS 158,159, 161,163, 164 Cretaceous plays in 169-185 chalk play, UKCS 162-163 Chalufy, France 268,273 channel/levee sand system, Gulf of Mexico 127-150

channel-fill reserves 87, 89, 93 channel-fill sandstone 197, 199-200, 238 Chanter Field, UKCS 215 charge volume 203, 229-230 evaluation of 193-195 Chestnut Field, UKCS 212, 291,292 China 73, 74, 75, 76, 87 Cimmerian unconformity 191,192, 195, 197 classification of traps in Central North Sea Graben 181-183 combination trap 24, 61 deep-water sand 135-141 stratigraphic 59-62 Claymore Field, UKCS 161,213, 216 coal bed methane 87, 92 colour categorization 4446, 48, 5/ column height 18,108,110 combination trap, classification 24, 61 communication, technical-commercial24, 32, 40 compaction 199,217,219 constructional trap 135,143,150 creaming curves 203 Cretaceous play, UKCS 160-163,165,178-180 Central North Sea Graben 169-185 see also Ryazanian-Barremian Cretaceous trap 240-241 Decision Tree Analysis 19-20, 39, 40 decompaction 203 deep-water sand 134-135, 159,160, 162, 163,164, 165 plays 127-150,169-185 seismic interpretation 143-144,145 Denmark, Halfdan Field, 163 depositional model 134-135, 192-193 destructional trap 135-143, 145,150 diagenetic trap 90 dip 108,109, 113,130, 133,144, 225 in extrudites 289,291,298 impact on oil recovery 279, 280, 282, 284, 285 discount rate 21, 37, 40 dry hole risk 33, 34, 35, 40, 92, 97, 99 dynamic behaviour modelling 271-285 East Central Graben, UKCS 179, 183 East Solan Basin UKCS 161,162 East Texas Field, case study 63, 66, 68, 69, 70, 86, 87 Efficient Frontier Technique 21 Elmworth-Wapiti trap, Canada 92 energy-demand forecast 30 Enoch Field, UKCS 212 Eocene 164, 240, 247-265,291, 292,293 subsidence 250, 253,263 Ettrick Field, UKCS 190 evaluating prospects 7-25,202-204

From: ALLEN,M. R., GOFFEY,G. P., MORGAN,R. K. & WALKER,I. M. (eds) 2006. The Deliberate Search for the Stratigraphic Trap. Geological Society, London, Special Publications, 254, 301-304. 0305-8719/$15.00. 9 The Geological Society of London 2006.

302 evaluation of technology 187,190, 190-200 evaluation of well data 191,231 Everest Field, UKCS 212, 219 exploration 27-40,106-111,180-185 and investment 31-33, 37 for stratigraphic traps 1-4, 28-33 exploration history 77, 81, 84,106 exploration industry, review of 2-4 exploration techniques 92-100, 187-204 extrusive sandstone (extrudite) 289-298 failure 231-232, 234, 236-238, 244 fan 164,173,190, 248 basin floor 161,241,242, 247,263 detached basin floor 182,184, 185 hanging-wall trap 181,184 Faroe-Shetland Basin 225-244, 247,289 field size 7,11, 37-39,106 Fife Field, UKCS 163 Foinaven Field UKCS 163,225,227-238,244 Fischschiefer Bed 170,171,173,176,177,180 Fisher Bank Basin 173,177, 184 Fleming Field, UKCS 212, 219 Flett Sub-basin, UKCS 226-233,238, 239 Flora Field, UKCS 163 fluid identification 4 fluvial sandstone plays 158,163 foreland basin 84, 86 Forties Field, UKCS 163 Forties Volcanic Province 183 fracture trap 87, 90 fractures in seismic response 98 France 268,273 Frigg Field, UKCS 163 Galley Field, UKCS 214 Gannet Field, UKCS 212 gas and condensate trap 92 gas show 175-176,178,233-234, 238, 239, 240 geochemical techniques 97-100 Geologically Driven Integration 195,196, 197 Giddings Field, USA 98 Guillemot Field, UKCS 163,212 Gulf of Mexico 100, 127-151,268 gull-wing 133, 134, 135,137 Halfdan Field, Denmark 163 Halibut Field, South, UKCS 183,185,187, 190 Hartzog Draw Field, Wyoming 106 Hatton-Rockall, UKCS 247-265 heavy oil trap 87, 92, 95 Highlander Field, UKCS 159,213 hydrocarbon indicator anomalies 15, 230-233 hydrocarbon reserves 14 Iceland hot spot 251-252 Idd A1-Shargi Field, Qatar 96 igneous rocks 176-178,183 image interpretation 43-54 Indonesia 10,105-126 injectite 164, 219,289-298 Interval Probability Theory 13 Irish Sea Basin 158

INDEX Jay Field, UKCS 68, 70, 71, 72, 87 Joanne Field, UKCS 163 Jurassic play, UKCS 158-159,165,241-244 Kaji Semoga, Indonesia 113,118,119,120-121 Kimmeridge Clay seal 159, 218 Kimmeridge Clay source rock 183-184 West of Shetland 240, 241,242 Buzzard Field 187, 192,193 Kingfisher Field, UKCS 215 Kittiwake Field, UKCS 158,159,214, 217 Laggan Field, UKCS 225,228, 233-234,235, 244 lithologies of stratigraphic traps reviewed 8--10 Lyonesse Field, UKCS 259, 260 MacCulloch Field, UKCS 212, 219 mass flow deposits 2,176,179,248,254, 260-262, 263 sandstone in hydrocarbon fields 173,174 Mexico 10 Michelob Field, Gulf of Mexico 140,141, 143,145 Miller Field, UKCS 215 Minas field, Indonesia 111 Miocene 111,251 Miocene extrudites 292-296 model design 271-272,272, 273, 274 modelling, impact of shale erosion 280-282, 283 Montrose Field, UKCS 163 Moray Firth 160,161,181,182,185,187 mud fan, Mississippi 127-150 Nelson Field, UKCS 74, 78, 79, 80, 87 subtle combination trap 77 Neogene tectonics 251 North Sea 153, 159,252, 253 see also Central North Sea Norway Block 24/7 293 oil and gas 101,175-176, 190 oil recovery 275-279, 281, 282-285 oil seeps 73, 98-99,100, 106, 111 Oligocene inversion 240 Oligocene traps, Rockall Plateau 247-265 pinchouts 225,232, 234,239, 240 Oman 10, 13,14, 15 onlap 200, 208,218, 268-269 model 272, 273, 274, 275 trap 135, 88 onlap dip, impact on oil recovery 279-280, 282, 284, 285 Orinoco Heavy Oil Belt, Venezuela 92 Pabst Field, Gulf of Mexico 140,141, 142, 145 Palaeogene 164, 165,251-254 palaeogeography, Rockall-Hatton 251-254 palaeomagnetic study, Buzzard Field 195 Paleocene prospect 225-245 passive margin basin 2-3, 84 permeability 227,274, 280, 285 Permian play 157,165 petroleum-system approach 107-108, 111 and rift basins 107-108,112, 116, 124 Petronella Field, UKCS 216, 217

INDEX Pilot Field, UKCS 212 pinchout 107,159,160, 163,182,208, 220 Buzzard Field 187, 193,195,200, 202 Rockall Platform 225,232, 234, 239, 240, 259-260 trap 87, 88, 93, 95, 99 turbidite sandbody 267-285 pinchout and onlap 271-285 classification of 269-271 Piper Field, UKCS 216, 217 play analysis 32-36 see also risk play groups, UKCS 155, 156 play history, deep-water sands 130-132, 144-150 play in rift basins 156-165 Pliocene uplift 251 polarity reversal 140, 143-144 ponded fill 181-182, 184 porosity 96, 131-132, 227, 256 visible 256, 262,263 Powder River Basin, Wyoming 106,107 pressure analysis, Buzzard Field 193,228 probability 33, 34, 35, 38-39, 40 and risk 12-13,18,19, 22 profitable production 30-31 prograding wedge 253-255, 256, 259 prospectivity 262-263 prospect evaluation 7-25, 31 Prudhoe Bay Field, USA 87,101 Qatar 96 Quadrant 205, UKCS 236-238 reserves, recoverable 85, 86, 88, 89, 90, 91 reservoir 161,171-180,182-183,227, 247 in beach sandstone 63, 68, 73, 77 sands 131-135,137 in turbidite 15,100,130,135,187,190 resources discovered 28, 29 rift basin 84,105-125 risk 124,165,185,200, 202-203,220, 231 analysis 12-23, 31,211 behaviour 36-37 dry hole 92, 97, 99 estimation of 7,16-17, 18, 100, 108 RMS amplitude 241-242, 244 Rob Roy Field, UKCS 216 Rockall Plateau 247-265 Ross Field, UKCS 190 Ryazanian-Barremian play 171-173, 175-177, 181-184 Safah Field, Oman 13,14, 15 salt 81, 85, 86,159, 212 Saltire Field, UKCS 172, 213, 215, 217 sand injectite 164,219,289-298 sand volcano 289-290, 292, 297 Sandarro igneous centre 258 sandbody pinchout and onlap 269, 270-285 sandstone, extrusive 289-298 sandstone (beach ridge) reservoir 63, 68, 73, 77 Santa Cruz 289,290, 292, 294-296, 29& Scapa Field, UKCS 161,169,182,183,213 Schiehallion Field, UKCS 225,229 Scott Field, UKCS 216

303

seal 108,110,159,183-184, 203,263 Kettla Tuff 227-228, 233,236, 238 one-seal 210-211 poly-sea1211-220 sealing surface classification 207-222 in Tertiary fields 212-216 sealing surface and risk evaluation 16,18, 221 seismic 3D data 169,170,189, 195,200, 202 trap definition 226,238,239, 241-243 seismic amplitude anomaly 83, 84, 99, 140, 260 West of Shetland 230-236, 240 seismic amplitude maps 136-138, 141, 142 seismic amplitude variations with offset 230-234, 237 seismic data 44, 50,125, 248-249 and colour choice 44, 45, 46, 48-50, 51 seismic interpretation 191 Buzzard Field 198-200 deep-water sands 143-144,145 seismic interpretation and visual cognition 43-55 seismic interpreters, training for 51-54 seismic profiles 3D Corona Ridge 241 3D Solan Basin 242, 243 3D trap concept 189 Assynt, UKCS 237 basalt scarps 260 Buzzard Field 199 channel sandstone 134 extrudite 291, 292, 293 George Bligh Bank, UKCS 258 igneous centres 258 Laggan 235 Lyonesse 259 mass slump deposit 261 pinchout structure 99, 259 rift basin 114,117, 120-123 West Central Graben 172 wet sand with gas pay 146-150 seismic techniques, 3D, 4D 95-97 sequence stratigraphy 93, 94, 95,129-131 Central North Sea 209 serendipity 7,12, 24,125,159,163,169 accidental discovery 68, 71,165 shale erosion, in modelling 280-282,283 Solan Field, UKCS 241,242, 243, 244 Southern North Sea Gas Basin 157,158 statistics on discoveries, UKCS 153-156, 157, 160, 161, 165 on stratigraphic traps 64, 65, 66, 67 Strathmore Field, UKCS 158 stratigraphic trap classification 61 defined 1-2, 57 and extrudites 289-298 location 11 review 4, 8-10 statistics 64, 65, 66, 67 summary data 81-92 stratigraphic modelling 192-193,194 stratigraphy Buzzard Field 191 Central North Sea Graben 171 Paleocene 227

304 Rockall Basin 249-254 see also sequence stratigraphy subtle combination trap 57-103 Swithin igneous centre, UKCS 253,258 syn-rift play, UKCS 159-160,165 targets, search for 44, 48 Tartan Field, UKCS 159,215-216, 273 T-block Field, UKCS 214 Teal Field, UKCS 216 technology evaluation 187,190-200 tectonostratigraphy, Rockall-Hatton 249-254 Tertiary, rift and passive margin basins 2-3 transmissibility 280-281 trap 12, 85, 87, 88, 89 analysis 7,11, 37-38, 39 classification 207-222 definition 225-226 exploration history 62-63 global distribution 59 seal 210-220 see also under seal trap, Eocene 247-265 trap, subtle combination 57-103, 81-92 trap, volume distribution 58 trapping mechanism 8--10, 87, 91, 92,221 Tree Field, UKCS 214

INDEX Triassic play, UKCS 158-159 turbidite pinchout, onlap surface 268-269 turbidite reservoir 15,100,130, 135,187,190 UK Continental Shelf 153, 153-167 statistics on discoveries 153-156, 157,160,161, 165 Vaila play, UKCS 226-230, 233,234, 236,238,239 Venezuela 8, 92,101 Victory Field, UKCS 161 Viking Graben, North Sea 158,150 visual cognition 43-55 visual images 46, 47, 54, 292 volume of resource 85, 86, 87, 88-91, 184, 262 well data, evaluation 191,231 well location selection 200 well position and oil recovery 275-277, 278,285 wells, Central North Sea Graben 175, 177,178, 180 West Buchan Graben 187,192, 290 West Central Graben 179, 181 West of Shetland 161,163,164,225-245 Widuri Field, Indonesia 116, 117, 119 Witch Ground Basin 173 wrench basin 85, 86 Wyoming 106,107

The Deliberate Search for the Stratigraphic Trap Edited by M. R. Allen, G. R Goffey, R. K. Morgan and I. M. Walker \

"

Twenty-four years have elapsed since the publication of Halbouty's AAPG Memoir of 1982, The Deliberate Search for the Subtle Trap. Since then, the technologies employed in hydrocarbon exploration have become extraordinarily sophisticated, yet current exploration for stratigraphic traps is to some extent restricted to areas where seismic data simplifies exploration by allowing direct inference of fluid fill and reservoir development. This Special Publication draws upon contributions that examine current industry perceptions of stratigraphic trap exploration and the technologies, tools and philosophies employed in such exploration, given the changing industry environment. iP,

This book contains a collection of papers examining a number of themes related to exploration for stratigraphic traps, rangi~qg from play and risk assessment, through regional assessments of stratigraphic trapping potential, specific exploration programmes targeted at stratigraphic traps to specific working traps and plays where stratigraphic trapping is prevalent. Visit our online bookshop: http://www.geolsoc.org.uk/bookshop Geological Society web site: http://www.geolsoc.org.uk

Cover illustration: AVO-basedseismicinversion showing reservoircomplexity in a deep water channel offshore Nigeria. Image supplied by R. K. Morgan (VeritasDGC Limited)