GEOLOGICAL SOCIETY SPECIAL PUBLICATION NO. 169
Forced Folds and Fractures
EDITED BY
JOHN W. COSGROVE Imperial College of Science Technology & Medicine UK
MOHAMMED S. AMEEN Saudi Aramco Saudi Arabia
2000
Published by The Geological Society London
THE GEOLOGICAL SOCIETY The Geological Society of London was founded in 1807 and is the oldest geological society in the world. It received its Royal Charter in 1825 for the purpose of Investigating the mineral structure of the Earth and is now Britain's national society for geology. Both a learned society and a professional body, the Geological Society is recognized by the Department of Trade and Industry (DTI) as the chartering authority for geoscience. able to award Chartered Geologist status upon appropriately qualified Fellows. The Society has a membership of 8600, of whom about 1500 live outside the UK. Fellowship of the Society is open to persons holding a recognized honours degree in geology or a cognate subject and who have at least two years' relevant postgraduate experience, or not less than six years' relevant experience in geology or a cognate subject. A Fellow with a minimum of five years' relevant postgraduate experience in the practice of geology may apply for chartered status. Successful applicants are entitled to use the designatory postnominal CGeol (Chartered Geologist). Fellows of the Society may use the letters FGS. Other grades of membership are available to members not yet qualifying for Fellowship. The Society has its own Publishing House based in Bath, UK. It produces the Society's international journals, books and maps, and is the European distributor for publications of the American Association of Petroleum Geologists (AAPG), the Society for Sedimentary Geology (SEPM) and the Geological Society of America (GSA). Members of the Society can buy books at considerable discounts. The Publishing House has an online bookshop (http://bookshop.geolsoc.org.uk). Further information on Society membership may be obtained from the Membership Services Manager. The Geological Society. Burlington House, Piccadilly, London W1V 0JU (e-mail:
[email protected]: tel: + 44 (0)171 434 9944). The Society's Web Site can be found at http://www.geolsoc.org.uk/. The Society is a Registered Charity, number 210161. Published by The Geological Society from: The Geological Society Publishing House Unit 7, Brassmill Enterprise Centre Brassmill Lane Bath BA1 3JN, UK (Orders: Tel. +44 (0)1225 445046 Fax +44 (0) 1225 442836) Online bookshop: http://bookshop.geolsoc.org.uk First published 1999 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. ©The Geological Society of London 1999. All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with the provisions of the Copyright Licensing Agency, 90 Tottenham Court Road, London W1P 9HE. Users registered with the Copyright Clearance Center, 27 Congress Street, Salem, MA 01970, USA: the item-fee code for this publication is 0305-8719/99/$15.00. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. ISBN 1-86239-060-6 Typeset by Wyvern 21 Ltd, Bristol, UK Printed by The 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]) Australia Australian Mineral Foundation Bookshop 63 Conyngham Street Glenside South Australia 5065 Australia (Orders'. Tel. +61 88 379-0444 Fax +61 88 379-4634 e-mail
[email protected]) India Affiliated East-West Press PVT Ltd G-1/16 Ansari Road, Daryaganj New Delhi 110002 India (Orders: Tel. +91 11 327-9113 Fax +91 11 326-0538) e-mail
[email protected]) Japan Kanda Book Trading Co. 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)
Contents COSGROVE, J. W. Forced folds and fractures: An introduction COSGROVE, J. W. & AMEEN, M. S. A comparison of the geometry, spatial organization and fracture patterns associated with forced folds and buckle folds
1 7
Section 1: Numerical analysis and field study of fractures associated with compactional forced folds
COOKE, M. L., MOLLEMA, P. N., POLLARD, D. D. & AYDIN, A. Interlayer slip and joint localization in East Kaibab Monocline, Utah: field evidence and results from numerical modelling LAUBACH, S. E., SCHULTZ-ELA, D. D. & TYLER, R. Differential compaction of interbedded sandstone and coal COSGROVE, J. W. & HILLIER, R. D. Forced-fold development within Tertiary sediments of the Alba Field, UKCS: evidence of the differential compaction and post-depositional sandstone remobilization
23 51 61
Section 2: Forced folding in extensional environments
MAURIN, J.-C, NIVIERE, B. Extentional forced folding and decollement of the pre-rift series along the Rhine graben and their influence on the geometry of the syn-rift sequences KELLER, J. V. A. & LYNCH, G. Displacement transfer and forced folding in the Maritimes basin of Nova Scotia, eastern Canada TIBALDI, A. & VEZZOLI, L. Late Quaternary monoclinal folding induced by caldera resurgence at Ischia, Italy MANSFIELD, C. S. & CARTWRIGHT, J. A. Stratal fold patterns adjacent to normal faults: observations from the Gulf of Mexico
73 87 102 115
Section 3: Forced folding in compressional and strike-slip environments
COUPLES, G. D. & LEWIS, H. Effects of interlayer slip in model forced folds WICKS, J. L., DEAN, S. L. & KULANDER, B. R. Regional tectonics and fracture patterns in the Fall River Formation (Lower Cretaceous) around the Black Hills foreland uplift, western South Dakota and northeastern Wyoming TEPER, L. Geometry of fold arrays in the Silesian-Cracovian region of southern Poland WATKINSON, A. J. & HOOPER, P. R. Primary and 'forced folds' of the Columbia River basalt province, eastern Washington, USA
129 145 167 181
Section 4: Temporal and spatial relationship between forced folds and buckle folds, crustal-scale folds and fold/fracture relationships
SATTARZADEH, Y., COSGROVE, J. W. & VITA-FINZI, C. The interplay of faulting and folding during the evolution of the Zagros deformation belt STIPSKA, P., SCHULMANN, K. & HOCK, V. Complex metamorphic zonation of the Thaya dome: result of buckling and gravitational collapse of an imbricated nappe sequence LISLE, R. J. Predicting patterns of strain from three-dimensional fold geometries: neutral surface folds and forced folds Index
187 197 213 223
Dedicated to Gilbert Wilson (1899-1987) In appreciation of his inspirational teaching and outstanding contribution to Structural Geology.
Geological Society Special Publications Series Editors A. J. HARTLEY R. E. HOLDSWORTH
A. C. MORTON M. S. STOKER
It is recommended that reference to all or part of this book should be made in one of the following ways: COSGROVE, J. W. & AMEEN, M. S. (eds) 2000. Forced Folds and Fractures. Geological Society, London, Special Publications, 169. SATTARZADEH-GADIM, Y., COSGROVE, J. W. & VITA-FINZI, C. 2000. The interplay of faulting and folding during the evolution of the Zagros deformation belt In: COSGROVE, J. W. & AMEEN, M. S. (eds) Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 187-196.
Forced folds and fractures: An introduction J. W. COSGROVE T. H. Huxley School of the Environment, Earth Sciences & Engineering, Imperial College of Science, Technology & Medicine, Royal School of Mines, Prince Consort Rd., London SW7 2BP Buckle folds and forced folds A considerable body of work exists in the geological literature dealing with the formation of buckle folds, (i.e. folds formed by compression either parallel or at a low angle to the layering or fabric of the rock) and a summary of much of this is presented in Price & Cosgrove 1990. In addition, fractures associated with these folds have been reported and discussed extensively for many decades, (e.g. Stearns 1964). This in part reflects the fact that the formation of folds and their associated fracture patterns frequently plays an important role in controlling the migration and concentration of fluids within the crust and thus has important implications regarding the disposition of water, hydrocarbons and zones of mineralization. However, there are many mechanisms other than buckling operating in the crust which can give rise to folds. One of the most important is that of 'forced folding' defined by Stearns (1978) as 'folding in which the final overall shape and trend [of the fold] are dominated by the shape of some forcing member below' and these folds and their associated fracture patterns have received relatively little attention in the literature. The present volume is an attempt to redress this imbalance. Unlike buckle folds, which are only generated during layer parallel compression, forced folds can be formed in any tectonic environment and are equally common in extensional and compressional regimes. The dominant mechanism operation during forced folding is 'bending', defined as the flexuring of a layer or surface by a compression acting at a high angle to the layering. The two mechanisms of folding mentioned above, i.e. buckling and bending, can be considered as two end members of a complete spectrum. Many folds generated in nature, for example the folds formed in the cover sequence as a result of thrusting in the basement, will involve significant components of compression both parallel to and normal to the layering. Like buckle folds forced folds can control fluid flow and host economically interesting fluid and mineral accumulations and it is therefore important to understand how they form
and the pattern and timing of their associated fractures. Clearly, in order to predict the role that fractures have in controlling fluid movement within and around folds of any type it is important to understand the timing of their formation. Although the intimate relationship between the geometry of folds and their associated fracture patterns strongly suggests that the same stress fields generated both structures, there is considerable uncertainty regarding the timing of fracture formation. Some fractures, for example those filled with vein material, probably formed at the same time as folding. Others however, may have formed much later during the exhumation of the rock, as a result of a decrease in confining pressure and the release of the residual stress locked into the rock at the time of folding. This volume The first paper (Cosgrove & Ameen) is written as an introduction to the volume and deals with the definitions of and the similarities and differences between buckle folds and forced folds. It focuses specifically on the three dimensional geometry, the spatial organization and the fracture patterns that characterize the two types of folds. The aim of this brief study is to establish the criteria that can be used to differentiate between buckle and forced folds and to determine if these features can be used in regions of poor exposure or in areas where it is necessary to rely on seismic data, to indicate the type of folding that has occurred. The study draws on field observations, analogue models and the consideration of conceptual models of folds. The formation of forced folds in three tectonic regimes is considered. These are the regimes of compressional tectonics (where thrusts or early normal faults reactivated as reverse faults, cause folding of the overlying strata), extensional tectonics (where normal fault movement is responsible for folding) and strike-slip tectonics. The association of fractures and folding is examined and a comparison made between fracture patterns associated with buckle folds and those linked to forced folding in both extensional
From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 1-6. 1-86239-060-6/OO/S 15.00 © The Geological Society of London 2000.
2
J. W. COSGROVE
and compressional environments. In addition the larger-scale fractures that form in the cover rocks above normal and reverse dip-slip faults are described and the effect that various amounts of strike-slip motion on these faults would have on the resulting fracture pattern discussed. The remaining 14 papers are grouped into four sections entitled "Numerical analysis & field study of fractures associated with compactional forced folds', 'Forced folding in extensional environments', 'Forced folding in compressional & strike-slip environments', and Temporal & spatial relationships between forced folds and buckle folds, crustal-scale folds & fold/fracture relationships.' The first section contains three papers. The first two describe the use of numerical analyses to investigate the formation of fractures in forced folds, one a monocline in the Navajo formation in Utah and the other a compaction fold formed in Upper Cretaceous coal seams deformed during diagenesis by differential compaction around relatively competent sand lenses. The third paper describes large-scale compaction folding in the Tertiary rocks of the North Sea and shows how the associated fracturing has initiated the development of large, sandstone dykes. The second section contains four papers the first two of which consider the formation of forced folds as a result of normal faulting associated with the formation of the Rhine graben and Maritime Basin of Nova Scotia respectively. The third paper looks at forced folding around a resurgent caldera in Ischia, Italy and the forth examines relatively small-scale folding associated with normal faulting in the Gulf of Mexico. The third section contains four papers related to forced folding in compressional and strikeslip regimes. The first paper describes an experimental study, the remaining three field studies in compressive and strike-slip tectonic environments. The final section contains three papers. The first considers the temporal and spatial relationships between forced folds and buckle folds using the Zagros Mountains as an illustration. The second paper deals with the formation of crustal-scale folds in the Bohemian massive of the Czech Republic and the final paper presents a method of determining areas of maximum strain on a folded surface and argues that these regions are likely to be the most highly fractured. Section 1 The first paper in this group is by Cooke et al. who combine numerical modelling and field
work in order to understand the distribution of fractures in the East Kaibab monocline in Utah, a forced fold in the dune bedded sandstones of the Navajo formation. Two types of joint clusters were documented. One which occurred in the hinge region of the forced fold and which is parallel to the hinge and at right angles to the bedding and the other which occurs on the steep limb of the monocline which is parallel to the hinge but oblique to the bedding. Based on these field observations and numerical modelling the authors conclude that the fold parallel and bedding perpendicular joint clusters form by curvature related stresses within the outer arc of the fold and the fold parallel but bedding oblique fractures formed as a result of interbed slip. The fold curvature and therefore the related joint clusters, relate directly to the shape and amount of displacement of the forcing member which generated the fold and the bedding plane slip and the related joint clusters, relate to the intrinsic mechanical properties (the mechanical anisotropy) of the folded unit. The authors clearly demonstrate that the fracture pattern within a forced fold is not controlled solely by the forcing member but is also sensitive to the material properties of the folded unit. The second paper (Laubach et al.) also combines detailed field observations with numerical modelling in an attempt to account for the distribution of fractures and the variation of fracture type around a forced fold. These folds, which occur in the Upper Cretaceous Mesaverde Group in SW Wyoming, are compaction folds formed in coal seams during burial and diagenesis. The folds form as a result of differential compaction of the coals and the interbedded sand lenses. Like many coals, these coal seams typically contain sub-vertical, open mode fractures (cleat). However, closely spaced normal faults abruptly substitute for open mode fractures in coal beneath some sandstone lenses that have blunt terminations. Finite element modelling of coal deformation shows that shear stress is augmented in coal layers below abruptly tapering edges of sandstone lenses favouring fault development, whereas under gradually tapering lenses shear stresses are not sufficiently enhanced to cause a shift in fracture style. The authors point out that the normal faults formed in the coal have little or no porosity and that the coal that contains them is likely to have low permeability compared to coal having typical, generally porous, open mode fractures. Thus the local change in fracture style may affect both regional and local gas and water flow within the coal. The third paper in this group (Cosgrove & Hillier) also describes the formation of compactional
FORCED FOLDS AND FRACTURES folds during diagenesis. These folds occur in the Eocene of the Outer Moray Firth in the North Sea as a result of differential compaction of mudstones over similar aged sand-rich, deep-marine channel/fan complexes. The study of cores shows that there has been considerable remobilization and redistribution of the sand both within the sand units and out into the surrounding mudstones as small sand dykes. However, in addition to these relatively small injections it is clear from the seismic sections through the structure that sand remobilization has taken place on a larger scale than previously thought. Large dykes almost half a kilometre long and up to eight metres wide emanate from the periphery of the sand lenses and cross-cut the overlying mudsones at an angle of about 60. The authors argue that the positioning of these large-scale dykes was controlled by the stress regime within the flexed overburden which resulted in outer arc fracturing adjacent to the overpressured sand body. These fractures provide ideal sites for sand injection and the proposed process is analogous to that operating during the formation of the peripheral dykes observed at the margin of many igneous intrusions, specifically laccoliths (Pollard & Johnson 1973). Section 2 The second group of papers all relate to the formation of forced folds in extensional settings, i.e. in association with normal faults. The first paper by Maurin & Niviere discusses extensional forced folding associated with the formation of the Rhine Graben. In this example there is an intimate relationship between movement on the basement fault that generated the forced folding and the deposition of some of the cover rocks in which the forced folds occur. The cover sediments lie on the Variscan basement and comprises a prerift sequence of Triassic and Jurassic rocks which contains an important Upper Liasic gypsiferous marl. A major unconformity separates these rocks from the overlying Palaeogene syn-rift sequence. Seismic sections show that the main basement graben-bounding fault to the west is a straight fault dipping 60° to the east. However, within the cover rocks the normal fault links with the sub-horizontal decollement horizon represented by the gypsiferous marls and the resulting geometry is listric. Continued extension on the fault generated a classic roll-over fold in the sediments above the decollement and a typical forced fold in those below. The response of the various sedimentary units to extension was controlled by their rheology. The brittle carbonates
3
of the Dogger extended by the formation of numerous small-scale normal faults whereas the ductile sediments of the Priabonian deformed in a completely ductile manner. The formation of a forced fold in cover rocks above a basement normal fault requires considerable thinning of the resulting monoclinal limb. However, some forced folds formed in extensional settings show no such thinning and in order for such folds to occur it is a geometric requirement that decoupling occurs between the basement and the cover. In the second paper in this group, Keller & Lynch describe an example of extensional forced folding from the Maritimes Basin of Nova Scotia, Eastern Canada, where seismic images and field work indicate that no significant thinning of the limb has occurred. The authors are able to demonstrate that a major detachment horizon has developed in a Visean evaporite sequence near the base of the cover rocks. A variety of kinematic indicators are developed along this horizon including a stretching lineation, a principal schistosity plane and secondary shear planes and intrafolial to upright asymmetric folds. The authors are able to demonstrate that the regionally extensive weak evaporitic layer was remarkably effective in transferring displacement between the normal fault and the decollement horizon in the cover sequence and that the mechanical decoupling of the strata above the detachment can be shown in the Horst block 70km away from the basement fault. In the third paper in this group relating to extensional forced folding, Tibaldi & Vezzoli describe late Quaternary monoclinal folding associated with caldera resurgence on the island of Ischia, Italy. The present level of erosion is such as to enable the various elements of the peripheral monocline (the gently inclined and subvertical limb) to be seen as well as the resurgent block and the peripheral normal faults that define it. They note that the forced folding occurred with the aid of at least one main detachment horizon localized within the pyroclastic succession and argue that the piston-like uplift of a fault bounded block with the generation of forced folds in the overlying volcanoclastic sediments is a viable alternative model of caldera resurgence to that of the classical doming model which is characterized by no peripheral faults, a rounded dome shape in plan view, beds continuous across the dome and the formation of a longitudinal apical graben. The final paper in this section on forced folding in extensional regimes relates to small-scale folding associated with normal growth faults formed as a result of the Mississippi delta tectonics in
4
J. W. COSGROVE
the Gulf of Mexico. Using high-resolution 3D seismic images, Mansfield & Cartwright mapped numerous examples of low amplitude stratal folds in both the footwalls and hangingwalls of these faults. The folds do not generally have the geometry of classical drag folds and the authors explore the possibility that these deflections might occur in regions of fault overlap or linkage. Although their origin remains unclear, they are recognized as a fundamental characteristic of all large growth faults in this part of the Gulf of Mexico. Section 3 The third section of this volume is related to forced folding associated with compressional and strike-slip regimes. In compressional forced folding the strata above the basement faults undergo some layer parallel shortening during the formation of the forced fold and there is therefore the possibility that the resulting structure will have elements of both forced folding and buckling. Indeed there is a complete spectrum between these two end member folds and most natural examples involve both processes. In the first paper in this section Couples & Lewis use rock and rock analogue models to investigate the influence of interlayer slip on the geometry and strain distribution within a forced fold. A variety of 'overburdens' were selected ranging from a simple homogeneous, isotropic single layer to well laminated multilayers and their response to identical basement block movements was recorded. It was observed that in the experiments where interlayer slip was possible the resulting forced fold was more localized than in the experiments with an unlayered overburden. In addition as the number of layers was increased so the fold became progressively more localized. However, as the number of layers increased a point was reached when not all the potential slip planes were activated during folding. The authors comment on this selective amplification of layer parallel slip and plan to investigate the phenomenon in a later publication. Although the results of these proposed studies are not yet available and Couples & Lewis decline to comment on the reason for the activation of some rather than all potential slip planes during folding, it is clear to the present author that the system is one in which two competing mechanisms are operating. If a finely laminated layer is considered there are two possible end member behaviours the layer could adopt during folding. The first is to ignore the layering and to fold as a
homogeneous, isotropic layer. This would maximize the bending stresses and minimize the frictional resistance to interlayer slip. The resulting strain distribution would be that of a tangential longitudinal strain fold (Ramsay 1967). The second type of behaviour is for all the potential slip surfaces to be activated during folding. This would maximize the resistance to folding resulting from the frictional resistance to slip between the laminations but would minimize the resistance resulting from bending stresses. The resulting strain distribution would be that of a flexural flow fold (Ramsay 1967). In practice neither of these two end members occurs and some compromize between the two develops. It is to be expected that more and more layer interfaces will slip as the fold develops in an attempt to reduce the build up of bending stresses. At each stage in the evolution of the fold a balance between the two processes will be achieved where the sum of the bending resistance and resistance to interlayer-slip is a minimum. The relative contribution of the two processes to folding at each stage in fold development will be determined by the mechanical properties of the interface (i.e. the coefficient of sliding friction or some related parameter) and the material properties of the layer (in an elastic model the Young's modulus). In the second paper in this section Wicks et al. examine the jointing in the Lower Cretaceous Fall River Formation, a unit cropping out in South Dakota and Wyoming, and explore the relationship between these fractures and a number of forced folds which occur in the area. These folds are associated with the Larimide compression which was active during the Palaeocene and which reactivated an easterly dipping master thrust in the upper crust. The resulting uplift produced numerous monoclines and anticlines. The authors are able to demonstrate convincingly that the joint sets they examined are in no way related to the regional Larimide compressive stress or the local extensional effects associated with the resulting forced folds. They conclude on the basis of fracture type, orientation and regional distribution that the jointing predated the forced folding and probably formed in the late Early Cretaceous. In the third paper in this section Teper describes the effect of basement faulting on cover rocks when the faulting is predominantly strike-slip. The area studied is the NE margin of the Upper Silesian Coal Basin in southern Poland. Here, Carboniferous molasse was deposited on a pre-Devonian crystalline basement block defined by first-order crustal boundary zones and subdivided into smaller segments by
FORCED FOLDS AND FRACTURES deep seated second order fractures. Many of these faults where reactivated as strike-slip faults during the Variscan compression. Experimental work and field observations (Oliver 1987; Richard 1990; 1991, Richard et al 1991) have demonstrated that pure strike-slip motion on basement faults only produces buckle folds in the cover sequence and that these folds form in en echelon arrays above the faults. Their spatial organization provides an excellent kinematic indicator which declares the sense of motion on the underlying basement fault and the author was able to exploit such folds arrays in the cover rock to determine the first-order movements on the faults during the Variscan deformation. However, because of the effects of releasing and constraining bends along the basement faults, elements of vertical movement occurred along the fault which resulted in the associated cover structures being a combination of both buckle and forced folds. The author shows that the profile geometry of many of the folds in the study area are incompatible with them being pure buckles and points out that pure strike-slip tectonics is just one example in the transtensional-transpressional continuum of tectonic environments (Hartland 1971). Thus, in all environments other than that of pure strike-slip along a perfectly planar fault, varying amounts of vertical motion are to be expected along the faults during their formation and reactivation. He argues therefore that folds with elements of both buckles and forced folds are likely to occur in the cover rocks above a basement fault even when a region is dominated by strike-slip tectonics. This interplay between basement strike-slip faulting, buckle folding and forced folding is further discussed by Cosgrove & Ameen in this volume. The fourth paper in this section tackles the difficult problem of differentiating between buckle and forced folds currently initiating and amplifying in a cover affected by compression. The region studied includes the Yakima fold belt which is made up of a series of asymmetric, E-W trending anticlines separated by broader, open synclines and which formed and are forming in the thick, otherwise horizontal sequence of the Columbia River Basalts Washington State. The authors (Watkinson & Hooper) note that the folds have been growing progressively over the last 17 Ma in response to the regional N-S compression. The study reveals that some of the pre-basalt flow structures in the underlying 'basement', specifically those that lie in an E-W or NW-SE direction, have been reactivated and that as a result a variety of styles of deformation have developed in the basalts including faulting, block uplift and flexure.
5
Despite detailed structural field work which enabled the fold style to be quantified and the fracture and strain distribution around the folds to be determined, the authors did not find the style of fold deformation sufficiently characteristic or distinct enough to be able to distinguish between basement controlled forced folds and buckle folds. The final paper in this section by Sattazadeh et al. considers the possible temporal and spatial relationships between faults, forced folds and buckle folds in a particular tectonic setting. The setting chosen is the Zagros fold/thrust belt situated at the junction of the Saudi Arabian and Central Iranian plate. In the Zagros region the rheological profile of the cover sequence is dominated by the thick basal Hormuz salt, which allows the decoupling of the deformation in the basement and cover, and a second evaporite-rich horizon at the base of the Miocene. The authors conclude that the type of folding is controlled primarily by the rheological profile of the cover, the reactivation of basement faults (wrench faults and the reverse dip-slip reactivation of normal faults) and the generation of new faults (thrusts) in the cover rocks. Dip-slip reactivation of basement normal faults forms forced folds in the overlying Hormuz salt series. The resulting displacements of the more competent units above the salt results in the initiation of important thrusts in the cover. The growth of these thrusts generates large-scale fault-bend folds. Hybrid folds involving elements of both forced folds and buckle folds form above many of the major strike-slip basement fault zones including the Kazarun and Minab lineaments. These are transpressional faults along which considerable horizontal and vertical displacements have occurred. The resulting folds in the cover overlying the Minab fault zone have the en echelon spatial organization of buckle folds formed above a strike-slip basement fault and the characteristic large aspect ratio (hinge length/half wavelength ratio) geometry of forced folds formed over a linear basement scarps. Clearly, these various types of folds, i.e. forced, hybrid and pure buckle folds, can be produced synchronously at different sites along a convergent plate boundary. Section 4 The final section of the volume contains two papers. The first by Stipska et al. considers the formation of the extremely large-scale folds formed in the mid and upper crust and now exposed along the eastern margin of the
6
J. W. COSGROVE
Bohemian massif in the Czech Republic. The authors conclude on the basis of detailed field study and petrological work that the large-scale folds are the result of a complex process which began with the eastward obduction of thrust slices onto the eastern (Brunovistulian) continent and the associated westward underthrusting of the continental margin at a transpressional margin. The authors use PT data in combination with thermal modelling to estimate the rheological evolution of this thrust stack during exhumation and show that when the stack had risen to a depth of about 15 km it encountered an autochthonous granite which acted as a relatively rigid block which inhibited further thrusting. The nappe pile, which represented a mechanical multilayer, continued to deform by folding. By considering the relative rheologies of the different nappe units and their thicknesses the authors conclude that despite the large wavelength (approximately 40 km), folding was by the process of buckling rather than bending. The final paper by Lisle & Robinson focuses specifically on the relationship between folding and fracturing. They propose that the geometries and densities of fractures associated with fold structures can be predicted by assuming that the strains accommodated by fractures mimic the bulk strain induced in the strata during folding. The authors examine, from a theoretical standpoint, the distribution of bedding plane strain expected in folds formed by the various fold mechanisms. The relationship between the state of bedding plane strain (which it is argued will be directly related to fracture density) and fold surface geometry is found to vary according to different fold types, which can be distinguished from each other on the basis of their curvature properties. The first type are developable fold surfaces and these have a Gausian curvature equal to zero. Fold mechanisms which are dominated by the mechanical strength of the layering, such as buckling, produce fold surfaces of this type and it is possible to estimate the bedding plane strains of such folds directly from the geometrical features of the folded layer. The authors illustrate this using flexural slip and neutral surface folds. The other main class of folds has non-developable surfaces, which have non-zero Gaussian curvature. Folded surfaces with this form arise predominantly from mechanisms which involve the passive deflection of the layering in response to displacement gradients originating outside the layer, e.g. forced folds. Although the geometry of these surfaces implies the presence of bedding
plane strain, in contrast to the buckle folds discussed above, this strain cannot be quantified from the geometry alone and requires additional information of the displacement patterns. The papers presented in this volume have one principal aim in common; namely to examine the major similarities and differences between forced folds and buckle folds in order that these differences can be used to recognize the type of folds (and therefore the expected fracture pattern) that are present in regions of poor exposure or where the geologist has to rely on seismic images. It is hoped that a clearer understanding of the differences between the two fold types (their 3D geometry, spatial organization, fracture patterns etc.) and the realization that they represent the two end members of a complete range of fold types will provide a useful predictive tool for Earth scientists concerned with the detailed geometry of fold structures and with assessing their possible role in controlling fluid migration and concentration within the crust. References HARLAND. W. B. 1971. Tectonic transpression in Caledonian Spitzbergen. Geological Magazine. 108, 27-42. OLIVER, D. 1987. The development of structural patterns above reactivated basement faults. University of London. PhD thesis. POLLARD. D. D. & JOHNSON. A. M. 1973. Mechanics of growth of some laccolithic intrusions in the Henry Mountains. Utah. II. Bending and failure of overburden layers and sill formation. Tectonophvsics. 18,311-354. PRICE. N. J. & COSGROVE, J. W. 1990. Analysis of geological structures. Cambridge University Press. RAMSAY. R. G. 1967. Folding and fracturing of rocks. McGraw-Hill, London. RICHARD, P. 1990. Champs de failles audessus d'un decrochement de socle: moderation experimental. University of Rennes. France. PhD thesis. 1991. Experiments on faulting in a two layer cover sequence overlying a reactive basement fault with oblique slip. Journal of Structural Geologv. 13. 459-470. . MOCQUET. B. & COBBOLD. P. R. 1991. Experiments on simultaneous faulting and folding above a basement wrench fault. Tectonophvsics. 188. 133-141. STEARNS, D. W. 1964. Macrofracture patterns on Teton anticline N.W. Montana (abstract) (Eos). Transactions of the American Geophvsical Union. 45. 107. 1978. Faulting and forced folding in the Rocky Mountain foreland. Geological Society of America Memoir, 151. 1-38
A comparison of the geometry, spatial organization and fracture patterns associated with forced folds and buckle folds J. W. COSGROVE1 & M. S. AMEEN 2 1
T. H. Huxley School for Environment, Earth Sciences and Engineering, Imperial College, London SW7 2BP, UK 2 Saudi Aramco, P.O. Box 2817, Rm X-6925, Dhahran 31311, Saudi Arabia
Abstract: In this paper the three-dimensional geometry and spatial organization of folds (both buckle folds and various types of forced folds) are considered, together with their associated fracture patterns, in an attempt to determine if these features can be used in regions of poor exposure or in areas where the geologist must rely on seismic data to indicate the type of folding that has occurred. This study of the relationship between the various fold types and associated fracture patterns draws on theoretical considerations of ideal conceptual models of folds, analogue models and field studies.
The aim of this brief study is to establish criteria that can be used to differentiate between buckle folds and forced folds, defined by Stearns (1978) as 'folds in which the final overall shape and trend are dominated by the shape of some forcing member below'. Three features that are of particular use when attempting to differentiate between these two types of folds are geometry, spatial organization and associated fracture patterns. In this paper these features are used to compare and contrast the two types of folds. Having determined the characteristic geometry, spatial organization and fracture pattern associated with these folds, it may then be possible to predict the role that these fractures play in the migration and concentration of fluids in the upper crust. Fractures generated in association with folding may significantly influence the migration and concentration of fluids within and around these various structures. However, in order to predict the role that these fractures have in controlling fluid movement it is important to understand the timing of their formation. Although the intimate relationship between the geometry of folds and their associated fracture patterns strongly suggests that the same stress fields generated both structures, there is considerable uncertainty regarding the timing of fracture formation. Some fractures, for example those filled with vein material, probably formed at the same time as folding. Others, however, may have formed much later during the exhumation of the rock, as a result of a decrease in confining pressure and the release of the residual stress locked into the rock at the time of folding.
Three-dimensional geometry and spatial organization of buckle folds The three-dimensional geometry and spatial organization of buckle folds have been studied using a variety of techniques including theoretical analyses, field observations and analogue modelling. These latter two studies have shown that buckle folds have a periclinal geometry, i.e. have the form of an elongate dome, basin or saddle. The geometry of periclinal folds is often described by giving the ratio of its half wavelength and hinge length. This is termed the aspect ratio and, although it will increase as the fold amplifies, it is found that the majority of buckle folds in the upper crust have ratios that range between 1:5 and 1:10. Periclinal geometry is characteristic of buckle folds on all scales. For example, the buckle folds of the Zagros mountains of Iran, shown in Fig. 1, have wavelengths in excess of 10km. It is clear from this figure that in plan view the folds are arranged in an en echelon manner. Analogue models (e.g. Dubey & Cobbold, 1977; Blay et al 1977) have shown how this distribution pattern emerges as the folds are initiated and amplify into finite structures. As well as having a limited extent along their hinges, buckle folds often die out rapidly in profile section. A typical profile section through a fold in a multilayer is shown in Fig. 2a. Thus, if the plan and profile data discussed above are combined, we obtain a three-dimensional picture of an isolated fold which exhibits the maximum amount of deformation near the centre of an approximately oblate ellipsoidal space, with the
From'. COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 7-21. 1-86239-060-6/OO/S 15.00 © The Geological Society of London 2000.
8
J. W. COSGROVE & M. S. AMEEN
The random initiation of folds at point irregularities in a multilayer would give rise to the type of fold distribution shown in Fig. 2e, and it is interesting to consider the mechanics behind the en echelon distribution of these structure. It follows from a consideration of the buckling equations (e.g. Ramberg 1960, 1961) that buckle initiation is difficult and generally requires some form of perturbation. However, once it has occurred, folding becomes progressively easier as the fold begins to amplify (see discussion in Price & Cosgrove 1990, pp. 281-282). This strain softening results in the reduction in the compressive stress along the horizon in which the fold has formed and a corresponding increase in stress above, below, in front and behind the fold. This is akin to the stress concentration that occurs around a hole or a body Fig. 1. LANDSAT image of large-scale non-cylindrical of less competent material in a more competent body. The initiation of new folds will, therefore, buckle folds in the Zagros mountains of Iran. be more likely to occur offset from either above, fold progressively 'dying out' away from this below, in front or behind rather than exactly central portion (Fig. 2b & e). Observations of adjacent to the existing fold, and in this way the buckle fold initiation in analogue models show en echelon distribution in both profile and plan that initiation occurs at local, often point, shown schematically in Fig. 2e will arise. irregularities and that the central portion of the structure develops first and the peripheral deformation occurs progressively. In its early Three-dimensional geometry and spatial stages of deformation, the central portion of organization of forced folds the structure passes through the geometrical forms that the outer portions of the fold even- Forced folds associated with basement tually exhibit. That is, the structure exhibits a dip-slip and oblique-slip faults form of 'recapitulation' as the fold spreads both along the fold axis and vertically within In this paper we use Steam's 1978 definition of a the profile section. forced fold, i.e. one in which the final overall
Fig. 2. (a) Typical profile geometry of a fold in a multilayer, (b) Block diagram showing a fold dying out in both profile and plan view, (c) Block diagram and (d) profile geometry of a box fold, (e) The spatial organization of folds within a multilayer.
FORCED FOLD AND BUCKLE FOLD GEOMETRY
Fig. 3. (a) A block diagram of a forced fold model formed over a planar normal basement fault with 3 cm displacement and 60° dip. The fold has been divided into zones according to the level of extension normal (e:) and parallel (ex and ev) to the layer. (Extension is negative and contraction positive.) The values of the strains in the various fields are given in the table, (b) Block diagram of a forced fold model above a planar, reverse basement fault with 5 cm displacement and 60° dip divided into zones according to the level of strain (
shape and trend are dominated by the shape of some forcing member below. One of the most common geological situations where such folds develop is in cover rocks above a fault in a more rigid basement. Such folds can form above pure dip-slip basement faults and above oblique-slip faults. Folds that form in the cover rocks as a result of dip-slip or oblique-slip movement on an underlying basement fault form over the fault scarps that these movements generate, and their geometry is controlled by the geometry of the scarp. As a result they are often long, linear structures and have a much higher aspect ratio than buckle folds. In addition, their spatial distribution is determined by the distribution of the causative basement faults which frequently form part of linear zones of deformation. Their linear plan geometry (i.e. large aspect ratio) and linear spatial organization contrasts sharply
9
with the shorter, more uniformly distributed buckle folds. It is clear from field studies, analogue models (e.g. Ameen 1988, 1992; Richard 1990, 1991; Richard & Krantz 1991) and finite element studies (e.g. Nino et al 1998) that the profile geometry of force folds is controlled by the type of movement (i.e. reverse or normal dip-slip, reverse or normal oblique-slip, etc.) and amount of slip on the basement fault, as indicated in Fig. 3. The evolution of these profile geometries and their associated strain fields and fractures are discussed later in the section on the association of fractures and forced folds. Folding of cover rocks above pure reverse dipslip and reverse oblique-slip faults will involve some layer-parallel shortening of the cover rocks. The forced folds that result will, therefore, have an element of buckling associated with them. However, their aspect ratio will be large, i.e. characteristic of a forced fold, because of the controlling influence of the linear fault scarp in the basement. However, folds can be generated in cover rocks above a basement strike-slip fault and, as discussed in the following section, the resulting folds have aspect ratios and profile geometries that are much closer to buckle folds than forced folds.
Folds associated with basement strike-slip faults Because no fault scarp develops at the basementcover contact during pure strike-slip faulting in the basement, the aspect ratio of the folds that form in the cover is very similar to that of a classical buckle fold. However, they do differ from buckle folds in their spatial organization and in the details of their profile geometry. They generally form in a linear en echelon array above the basement fault, with the folds being consistently offset either to the right or left depending on whether the basement fault is dextral or sinistral, respectively. Thus, the sense of offset can be used as a kinematic indicator to determine the sense of movement along the fault. Many well-documented natural examples exist, including the large-scale en echelon folds formed above a sinistral basement wrench fault in the Darien Basin of eastern Panama (Wilcox et al 1973) and the en echelon folds that formed above the Inglewood fault near Los Angeles, California, USA (Fig. 4). Thus, despite their similarities to classical buckle folds in their aspect ratio and profile geometry, they differ significantly in their spatial organization. As can be
10
J. W. COSGROVE & M. S. AMEEN
Fig. 4. (a) Schematic diagram showing the predicted orientation of fold axes (gently curved lines) in a cover sequence above a dextral strike-slip fault in the basement, (b) En echelon folds formed above the Inglewood fault, near Los Angeles. California, USA. This fault is a dextral strike-slip fault associated with the San Andreas fault zone.
seen from Fig. 4, they are arranged in a linear zone rather than the more uniformly distributed folds associated with buckle folds (Fig. 1). Experimental studies of fold development in cover sequences above basement faults (e.g. Wilcox et al. 1973; Graham 1978; Odonne & Vialon 1983; Oliver 1987; Richard 1990) have enabled the evolution of the profile geometry of these folds to be studied. The profile geometry
shown in Fig. 5a & b is a summary diagram based on a series of analogue experiments performed by Oliver (1987). In these experiments the cover rocks were simulated by thin layers of paraffin wax and the basement by rigid wooden blocks. The profile geometries of the periclines that developed showed a variety of geologically realistic structural forms including upright box folds and overturned folds which may pass laterally into thrusts. When traced from the top to the bottom of the cover sequence, the fold axes tend to rotate towards parallelism to the basement fault. At depths, a major fold sometimes divides into two or more relatively tight folds. The major structures become simpler and more open when traced towards their ends. Oliver also noted that the folds in the cover on each side of the basement fault tend to be asymmetric, and in the upper levels of the cover verge towards the line of the basement fault. Thus, as can be seen from Fig. 5b, the sense of asymmetry of a fold will change as it passes over the basement fault. Complex changes in dip, strike and sense of movement are also observed in faults formed in the cover rocks which straddle the basement fault. The three-dimensional geometry of a fault formed in the cover above a basement strike-slip fault has been reconstructed by Naylor et al. (1986) from horizontal serial sections through a
Fig. 5. (a) Examples of fold structures in layered wax, acting as cover rocks, above a strike-slip fault in the basement, (b) Diagrammatic representation of such a fold. The form and structure are represented by a series of profiles. Note the reversal of symmetry of the fold on opposite sides of the basement fault (from Oliver 1987). Fold profiles formed above basement wrench faults. In models (c)-(e) there has been 8cm of left-lateral displacement on the basement fault, in (f) there has been 20cm (from Richard et al. 1991). (g) Helicoidal geometry of two Riedel shears, reconstructed from horizontal serial sections through a sand body deformed above a pure strike-slip fault in a rigid basement (after Naylor et al. 1986).
FORCED FOLD AND BUCKLE FOLD GEOMETRY model (Fig. 5g). The helicoidal geometry revealed is the result of the en echelon shears at the surface linking to the basement fault at depths. These fractures are further discussed in the section on fractures associated with force folds. The change in fold profile geometry with amount of fault movement is shown in Fig. 5c-f. The experiments illustrated in this figure were carried out using a multilayer cover made up of alternating layers of silicon putty and sand. As there is the possibility of a complete range of movements on basement faults from pure reverse dip-slip, through oblique reverse dipslip motion to pure strike-slip movement, so there will be a corresponding spectrum of folds ranging from buckle folds formed above basement strike-slip faults to the force folds formed above pure reverse dip-slip basement faults. The geometry of other types of forced folds
Fault-bend folds In the previous discussion of forced folding we have attempted to honour Steam's definition of a forced fold, i.e. 4one in which the final overall shape and trend are dominated by the shape of some forcing member below'. It is clear from the field examples and experimental work described above that forced folds are considered to form in cover rocks (usually layered) as a result of fault movement in a relatively rigid basement. Other well-documented relationships between faults and folds are known, one of the most familiar being fault-bend folds (Fig. 6) (Suppe 1983). In this type of forced folding the folding is not the result of the movement of rigid fault blocks in the basement but rather the result of fault movement within the cover rocks. Suppe notes that faults are not perfectly planar surfaces of slip. They generally have gentle undulations and may display substantial curvature or sharp bends. As two fault blocks slip past one another, there must be deformation in at least one fault block because rocks are not strong enough to support large voids. For this reason many major folds in layered rocks exist within the hangingwall fault blocks, formed by bending the fault blocks as they slip over non-planar fault surfaces. This mechanism of folding is called fault-bend folding. The two most widely recognized structural settings in which fault-bend folds form are: (i) in the hangingwall of a listric normal fault; and (ii) in the hangingwall above a ramp in a thrust fault. This folding is not the result of frictional drag but is a result of bending.
11
These structures have been extensively described in the literature and the interested reader is referred to Boyer & Elliot (1982) and to the collection of papers on fault-related folding edited by Anastasio et al (1997). The discussion of the fracture patterns associated with these folds (see Chester 1998) is outside the scope of this paper but, as can be seen from Fig. 6a & b, the fractures formed in association with extensional fault-bend folds, i.e. with the crestal collapse grabens which form in the hinges of roll-over folds, penetrate deep into the fold structure. Clearly, the neutral surface is much lower in these folds than in buckle folds (Fig. 7h). Indeed, if all parts of the roll-over fold undergoes extension, then the neutral surface will lie below rather than within the folded rock unit. Fold fracture association
Fractures associated with buckle folds Fractures formed in association with buckle folding may be the result of the regional stress field or of the local stresses generated as a result of buckling (e.g. extension in the outer arc above the neutral surface and compression of the inner arc below). In addition, because folds form in layered successions and often involve interbed slip, it is clear that the bedding planes cannot sustain a high shear stress. Consequently, the principal stresses are constrained to being either subparallel or subnormal to bedding, and as a result of this stress deflection the fractures also form normal to bedding. This is illustrated in Fig. 7, which shows the predicted orientation of the shear and extensional fractures that would form in response to the regional compression generating the fold (Fig 7a), together with their projection on a stereographic plot (Fig. 7b) and the frequently observed orientation of these fractures on the limbs of the fold which occurs as a result of the principal compressive stress following the layering (Fig. 7c-e). The types and orientations of fractures found in association with buckle folds, which form as a result of both the regional and local stresses, have been summarized by Price & Cosgrove (1990) and are shown in Fig. 7f-h. They note that the relationship between extensional fractures and fold geometry is usually rather simple (Fig. If). Dilational fractures which trend parallel to the axis form perpendicular to bedding, so these fractures will vary in dip through an arc which is determined by the tightness of the fold. The orthogonal set cuts the fold axis at 90° and is also perpendicular to the bedding.
12
J. W. COSGROVE & M. S. AMEEN
Fig. 6. Fault-bend folds formed in (a) and (b) extension and (c)-(f) compression, (a) Hanging-wall deformation above a simple listric fault. A roll-over fold has formed in the pre-rift sediments (the regularly banded dark and light layers). The crestal collapse graben cuts deeply into the fold, (b) A line diagram showing the sequence of fault development, (from McClay, 1990). (c)-(e) Schematic progressive development of a fault-bend fold as a thrust sheet rides over a ramp in the decollement horizon (after Suppe 1983).
Such extensional fractures are sometimes infilled with calcite or quartz, and form at right angles to the minimum principal compression. It can be inferred from the above comments that
the extensional fractures perpendicular and parallel to the fold axis did not develop at the same time and that the orientation of the minimum principal compression was subparallel to bedding.
FORCED FOLD AND BUCKLE FOLD GEOMETRY
13
Fig. 7. (a) Ideal relationship of master joints to a relatively small fold, (b) Stereographic plot of fractures shown in (a), (c) Trend of minor fractures in a folded competent unit, (d) and (e) Stereographic plots of fractures in the two limbs. R and T are shear and extension fractures, respectively (after Price 1966). (f) Typical relationship of dilational fractures to a fold. The orientation of the least principal stress associated with each set (which are of different ages) is also shown, (g) Typical orientation of shear fractures in a thin bedded layer, with associated stress systems, (h) Typical orientation of normal faults and thrusts which may develop in a thick, flexured unit (after Price & Cosgrove 1990).
The relationships which shear fractures exhibit relative to the fold geometry are more complex. The more commonly observed orientations of shear fractures are shown in Fig. 7g & h. It can be seen that these shears include normal, thrust, strike-slip and oblique-slip faults. It should be noted that the orientation and direction of slip on the oblique-slip fractures is determined by the orientation of bedding. Genetically, they are wrench faults and are called oblique-slip only because it is usual to have the horizontal rather than the bedding as the reference coordinate. As with extensional fractures, it is possible to infer the orientation and relative magnitudes of the principal stresses that are associated with the initiation of the various shear fracture systems (Fig. 7g & h). Normal faults tend to be aligned parallel or perpendicular to the fold axis. Those forming parallel to the fold axis form in response to the local extension that occurs in the hinge region of the fold above the
neutral surface. Thrust faults form in the hinge regions below the neutral surface in the region of local compression. It is interesting to note that as these normal and thrust faults form in the hinge region, they influence the effective thickness of the layer which in turn may cause the neutral surface to migrate either up or down. In this way it is possible to find normal faults cutting thrusts or visa versa. However, the most frequently developed shear fractures are strike- or oblique-slip faults, many of which exhibit very little movement and have often been termed joints. The strain distribution (and therefore fracture pattern) within a folded layer is very dependent upon the layer properties. In a homogeneous isotropic layer the strain distribution is likely to be that shown in Fig. 8b, in which a layer-parallel extensional field associated with the outer arc is separated from a layer-parallel compressional field associated with the inner arc by a neutral
14
J. W. COSGROVE & M. S. AMEEN
Fig. 8. (a) Various fractures associated with a pericline (after Stearns 1978). (b) Strain distribution in a tangential-strain fold and (c) a flexural-flow fold (after Ramsay 1967).
surface. This model of strain distribution is known as tangential longitudinal strain folding. In contrast, a homogeneous anisotropic layer may fold by bedding-parallel slip which results in the strain distribution shown in Fig. 8c. This is known as flexural flow if the shear strain parallel to the layer boundary is uniformly distributed across the layer or flexural slip if it is concentrated along distinct bedding planes. These two models can be considered to be end members of a spectrum. Many geological folds develop in multilayers and, as the folds amplify, various bedding planes may become active slip planes. Thus, although initially the multilayer may behave as a single unit with a single neutral surface, if at some point in the fold development slip occurs on a bedding plane the fold is divided into two units, each of which will have its own neutral surface. As more and more slip surfaces become activated, a proliferation of neutral surfaces will occur together with a complex superposition of different strain fields and associated fracture patterns. Returning to the folding of a single layer, it must be emphasized that not all the individual
fracture sets illustrated in Fig. 7 are likely to develop in one fold. In addition, it has already been noted that the three-dimensional geometry of buckle folds, particularly in the upper levels of the crust, is periclinal. Thus, the idealized orientation of the fractures relative to the fold axis shown in Fig. 7 only holds for the central, approximately cylindrical, portion of the fold. Field studies indicate that the orientation of the fracture patterns tends to be related to the slip direction along the bedding. The idealized relationship between a pericline and its associated fractures is shown in Fig. 8a. Fractures associated \\ithforcedfoldsformed above dip-slip basement faults In this section the results of some experimental work (Ameen 1988) in which analogue materials were used to study the initiation and growth of forced folds above dip-slip basement faults are briefly described. The models consist of two basic parts, the 'basement' which was simulated by wooden blocks and the 'cover' which was
FORCED FOLD AND BUCKLE FOLD GEOMETRY simulated by layered wax. Variations in competence between different layers within a sedimentary cover were modelled using waxes of different rheologies. The layers were imprinted with strain marker grids both in the profile section of the fold and on the surface of the layers. These experiments enabled the progressive development of the forced folds to be studied. In particular, they allowed the strain field and associated fracture patterns to be determined at various stages during fold amplification and, thus, gave a clear indication of how these strain fields and fracture patterns changed as the folds grew. Analysis of the results showed that the cover accommodated itself to the faulting in the basement by two main mechanisms, namely rigid-body rotation and internal deformation. Not surprisingly, the deformation was highly concentrated in the vicinity of the basement fault and died out away from these areas. Having determined the extension normal and parallel to the layering (in directions normal and parallel to the strike of the basement fault) using the strain marker grids, these values were then contoured and their evolution tracked during the course of the experiment. Two examples of strain distribution around force folds are shown in Fig. 3. Figure 3a shows a fold formed over a normal dip-slip fault and Fig. 3b shows a fold formed over a reverse normal fault. The models have been divided into different strain fields on the basis of the values of the three principal extensions. These fields are numbered and the magnitudes of the three principal strains in each field given in the accompanying tables. It is clear from these experiments that the strain fields and associated fractures around forced folds formed in cover rocks as a result of dip-slip movement on basement faults are very sensitive to the sense of movement, i.e. either normal or reverse. In extensional forced folding (i.e. where the causative basement faults are normal dip-slip faults) the strain patterns generated in the folded layers are characterized by layer-perpendicular shortening and layer-parallel extension in the direction normal to the strike of the fault, regardless of the amount of movement on the fault (Fig. 3a). Thus, an extensional strain field forms throughout the fold and, although the fracture density is predicted to increase with increased movement on the fault, the fracture type will remain the same. This is in marked contrast to the fracture patterns generated during the forced folding of a cover sequence above a basement dip-slip fault reactivated by reverse movement. Whereas basement normal faults cause strain patterns in the
15
Fig. 9. Schematic block diagrams showing potential geometry and sense of movement of macrofaults in experimentally produced forced folds above (A), reverse and (B) normal basement faults (after Ameen 1990).
cover that are homogeneous in both time and space, the strain pattern and associated fracture pattern formed in the cover above a reverse dip-slip basement fault is found to vary both in time and space (see Fig. 3b). The result is that one set of fractures, characteristic of a particular strain field, may subsequently be overprinted by another set characteristic of a different strain field. In this way extensional fractures can be cut by compressional fractures and visa versa. Field studies of compressional forced folds show this superposition of different fracture patterns very clearly (Ameen 1992). Schematic block diagrams, illustrating the geometry and the potential orientation of macrofaults in experimentally produced forced folds formed above basement dip-slip faults, are shown in Fig. 9. The overlapping of different strain fields during fold amplification is a characteristic feature of forced folds formed over reverse dipslip basement faults. As mentioned in the previous section, a migration of the boundary between two strain fields may also occur during buckling when, as a result of fracturing at the layer boundaries, the neutral surface, which separates extensional and compressional strain fields, migrates through the layer. However,
16
J. W. COSGROVE & M. S. AMEEN
Fig 10. Plan and profile view of folding and fracturing of cover sediments above a basement normal fault. (A)-(D) shows four experiments with 2.6cm of pure, normal dip-slip movement on the basement fault and no strike-slip movement. (E)-(H) shows four experiments with the same normal oblique-slip (4cm) and the same strike-slip dip-slip ratio (ss/ds= 1). The line of the sections shown in the right-hand column are marked on the left-hand columns (after Richard 1990).
FORCED FOLD AND BUCKLE FOLD GEOMETRY this migration is much less in buckle folds than in forced folds, and in many buckle folds, does not occur at all.
Fractures associated with forced folds formed above oblique-slip basement faults In the previous section the discussion relates to forced folds formed in the cover above dip-slip basement faults, either normal or reverse. Forced folds also form above oblique-slip basement faults. The geometry of these structures and their associated fracture patterns has been studied experimentally by Richard (e.g. Richard 1990, 1991). He investigated the effects of the relative amount of dip-slip vs strike-slip movement on the basement fault on the geometry and orientation of the resulting folds and fractures generated in the cover rocks. Three different basement faults were used in the experiments: a normal fault dipping at 45°, a vertical fault and a reverse fault dipping at 54°; and he investigated the effect of pure dip-slip movement (normal and reverse) and oblique-slip movement on the cover units. Richard also considered the effect on the cover structures of having a weak, incompetent layer between the rigid basement and the less rigid cover. Figure 10 shows the different structures that form over identical basement faults dipping at 45° when the movement is 100% normal dipslip and when the motion on the fault is oblique-slip normal faulting with a dip-slip/ strike-slip ratio of 1 (ds/ss=l). It can be seen that the introduction of an element of strikeslip motion to the basement fault has a firstorder effect on the orientation of the macrofaults that develop in the cover. As can be seen from Fig. 10A-D, when the basement fault is a pure dip-slip fault, the strike of the faults in the cover is parallel to the basement fault. When the movement on the basement fault is obliqueslip, the strike of the faults in the cover are oblique to the trend of the basement fault (Fig. 10E-H). This is confirmed by the experiments shown in Fig. 11, which illustrate the influence of a buffer layer of weak material between the basement and cover on the distribution of strain in the cover rocks. As the buffer layer gets thicker so the fault domain in the cover widens. In the reverse wrench fault experiments with no buffer layer, discrete reverse Riedel shears develop in the cover immediately above the basement fault. As the thickness of the buffer layer is increased the reverse faults become less important and less localized, and
17
eventually cease to form. In such cases the cover accommodates the movement on the basement fault by forming a monoclinal flexure. Inspection of Fig. 11 shows how the thickness of the buffer layer controls the partitioning of the strain induced in the cover by movement on the basement fault between brittle and ductile deformation. The formation of folds in cover rocks above a basement strike-slip fault has been discussed earlier. Fold curvature and fracture density In addition to the type and orientation of fractures that form around the different types of folds, geologists interested in rock strength and fracture permeability are also concerned with fracture density. It has been suggested (Lisle 1992, 1994) that the fracture density around a fold may be directly related to the curvature of the fold and that by plotting out the variation in curvature around a fold the likely variation in fracture density can be determined. The essential idea behind this analysis is summarized by Lisle (1992). He notes that the folding of a sheet without stretching of lines within it is referred to, in the language of differential geometry, as isometric bending. As can be inferred by flexing sheets of paper, the no-stretch condition imposes important constrictions on the curvature changes which points on the sheet can undergo during folding. These constraints are embodied in Gauss's Theorema Egregium which state that 'the total curvature (equal to the product of the two principal curvatures) at any point remains invariant under isometric bendings'. It follows from this theorem that there is a limited range of fold geometries that an initially planar non-stretching sheet can adopt. These developable surfaces, which include cylindrical and conical fold (but not periclinal folds), have the property that points of equal dip and strike of the surface are arranged in straight lines. This property allows a simple check to be made of the validity of the constant bed-length assumption in the case of natural fold structures. For any fold represented by structure contours, points of equal strike on the structure are linked by isotrend lines that will be straight if the structure is developable. Curved isotrend lines indicate that the structure has a geometry incompatible with the constant bed-length model. The patterns of isotrend lines constructed for a fold help to indicate the parts of the structure
18
J. W. COSGROVE & M. S. AMEEN
Fig 11. Plan and profile views of folding and fracturing of cover sediments above a 45 normal basement fault, a 90: basement fault and a 54(reverse basement fault. All faults had the same strike-slip dip-slip ratio (ss ds = 1). The results of four experiments are shown for each fault type. The four experiments differ from each other only in the thickness of the weak layer at the boundary between the basement and the cover. The line of section of the profiles is shown on the corresponding plan (after Richard 1990).
FORCED FOLD AND BUCKLE FOLD GEOMETRY
Fig. 12. (a) Isofracture map of Goose Egg dome, Casper, Wyoming, USA (after Harris et al. 1960). (b) Map of absolute values of Gaussian curvature. The darkness of the shading indicates the absolute magnitude of curvature. Dark areas are those parts of the structure that possess a geometry incompatible with the isometric folding model and thus requires bedding plane strain. It is argued that these areas will be the most intensely fractured (after Lisle 1994).
where layer stretching, and therefore fracturing, is likely. This curvature analysis can be applied to any type of fold including buckle folds and forced folds, and an attempt has been made to examine the correlation between fracture density and curvature by the careful study of well-exposed folds such as the Goose Egg dome structure in Wyoming, USA (Fig. 12). A good correlation is found, and the reader interested in this approach to determining fracture density is referred to the papers by Lisle cited here and to his paper in this volume. Lisle's analysis presented above relates to the folding of a surface. It is argued that any surface involving a double curvature (e.g. a dome, basin or saddle structure) cannot be without some stretching or contraction of the surface. The amount of straining at any point on the surface relates directly to the magnitude of the Gaussian curvature and this, Lisle argues, determines the fracture density. In this example it is possible to relate geometry to strain and thus, fracture distribution. However, in general, when folds are being considered it is the straining and fracturing of a layer (rather than a surface) that is important and it is generally not possible to determine the strain within a folded layer from its profile
19
geometry. This is convincingly demonstrated by considering the strain distribution within a tangential longitudinal-strain fold (the type of fold that might develop in an isotropic homogeneous layer) and a flexural-slip fold (which might form in an anisotropic homogeneous layer) (Fig. 8b & c). Both folds have the same profile geometry, i.e. that of a parallel fold, but they have remarkably different strain distributions and therefore different predicted fracture patterns. Thus, it is important to note that geometric models of fold formation such as the fault-bend fold shown in Fig. 6c, which is constructed by building in certain assumptions for example constant bed length during fold formation and the development of a particular strain pattern (e.g. simple shear), cannot be used to predict the strain distribution and related fracture patterns Timing of fracture formation We have looked at the geometry, spatial organization and probable fracture patterns associated with forced folds and buckle folds in an attempt to determine if these features can be used in regions of poor exposure (see e.g. Mitra & Mount 1998) or in areas where the geologist must rely on seismic data to indicate the type of deformation that has occurred. It is also important to consider when the fractures formed. Fractures which relate directly to the fold geometry are probably either syn- or post-folding, and it is generally difficult to say which. The relative age of the fractures can sometimes be established using abutting relationships but, as is discussed in the following paragraph, the absolute age is more difficult to ascertain. It is interesting to note that the density of fractures in the Earth's crust decreases with depth. The upper 0.5-1 km is much more densely fractured than the rocks below. This is thought to be the result of the reduction of the overburden and confining stresses combined with weathering associated with exhumation. The fractures form as a result of the release of residual stresses which may have been locked in the rock since the time of folding (Price 1964). Thus, if the residual stresses were locked into the rock at the time of folding then the orientation of the fractures that form when they are released will be symmetrically oriented with respect to the folds. Thus, although the same stress field was responsible for the formation of both the folds and fractures, the time of formation of the two structures may be separated by several hundred million years. Clearly, an
20
J. W. COSGROVE & M. S. AMEEN
understanding of the time of formation of the fractures is crucial to geologists concerned with the migration and concentration of fluids in and around folds. In this brief paper we have attempted to compare and contrast the three-dimensional geometry, spatial organization and fracture patterns of buckle folds and various types of forced folds, with the aim of using this understanding in regions of poor exposure or in areas where the geologist must rely on seismic data to indicate the type of folding that has occurred. Although the temporal and spatial association of forced fold and buckle folds is outside the scope of this study, it is clear that at many convergent plate margins both buckle folds and forced folds occur together. An impressive example of this can be found in the Zagros orogenic belt, and the interested reader is referred to the paper by Sattarzadeh et al. (this volume).
References AMEEN, M. S. 1988. Folding of Layered Cover Due to Dip-slip Basement Faulting. PhD Thesis, University of London. 1990. Macrofaulting in the Purbeck-Isle of Wight monocline. Proceedings of the Geologists' Association. 101, 31-46. 1992. Strain pattern in the Purbeck-Isle of Wight monocline, a case study of folding due to dip-slip fault in the basement. In: BARTHOLOMEW, M. J.. HYNDMAN, D. W.. MOCK, D. W. & MASON. R. (eds) Characterisation and Comparison of Ancient (Precambrian-Mesoioic) Continental Margins. Proceedings of the 8th International Conference on Basement Tectonics at Butte Montana, USA. Kluvver. Dordrecht, 559-578. ANASTASIO, D. J., ERSLEV, E. A., FISCHER, D. M. & EVANS, J. P. (eds) 1997. Fault-related folding. (Special Issue.) Journal of Structural Geologv, 19. 243-602. BLAY, P. K.. COSGROVE, J. W. & SUMMERS. J. M. 1977. An experimental investigation of the development of structures in multilayers under the influence of gravity. Journal of the Geological Societv, London. 133. 329-342. BOYER, S. E. & ELLIOT, D. 1982. Thrust systems. AAPG Bulletin, 66, 1196-1230. CHESTER, J. S. 1988. Geometry and fracture distribution in fault-propagation folds in nature and experiments. Abstracts of the American Association of Petroleum Geologists, 72, 171. DUBEY, A. K. & COBBOLD, P. R. 1977. Non-cylindrical flexural slip folds in nature and experiment. Tectonophvsics. 38. 223-239. EVANS, J. P. (ed.) 1997. Fault-related folding. (Special Issue.) Journal of Structural Geologv, 19, 243602.
GRAHAM. R. H. 1978. Wrench faults, arcuate fold patterns and deformation in the southern French Alps. Proceedings of the Geologists' Association. 89, 129-142. HARRIS, J. F.. TATLOR. G. L. & WALPER. J. L. 1960. Relation of deformational structures in sedimentary rocks to regional and local structures. AAPG Bulletin. 44. 1853 1873. LISLE, R. S. 1992. Constant bed-length folding: threedimensional geometrical implications. Journal of Structural Geology, 14. 245-252. 1994. Detection of zones of abnormal strains in structures using Gaussian curvature analysis. AAPG Bulletin. 78, 1811-1819. 1999. Predicting patterns of strain from threedimensional fold geometries: neutral surface folds and forced folds. This volume. McCLAY. K. R. 1990. Extensional fault systems in sedimentary basins: a review of analogue model studies. Marine and Petroleum Geologv. 7. 206233. MITRA. S. & MOUNT. VAN S. 1998. Foreland basementinvolved structures. AAPG Bulletin. 82, 70-109. NAYLOR, M. A.. MANDL. G. & SIJPESTEIJN. C. H. K. 1986. Fault geometries in basement-induced wrench faulting under different initial stress states. Journal of Structural Geologv. 8, 737-752. NINO, F.. PHILIPS, H. & CHERY. J. 1998. The role of bedding parallel slip in the formation of blind thrust faults. Journal of Structural Geologv. 13. 503 516. ODONNE. F. & VIALON. P. 1983. Analogue models of folds above a wrench fault. Tectonophvsics. 99. 31-46. OLIVER. D. 1987. The Development of Structural Patterns Above Reactivated Basement Faults. PhD Thesis. University of London. PRICE. N. J. 1964. A study of time-strain behaviour of coal-measure rocks. International Journal of Rock Mechanics & Mining Science. 1. 277-303. 1966. Fault and Joint Development in Brittle and Semi-brittle Rock. Pergamon. Oxford. & COSGROVE. J. W. 1990. Analysis of Geological Structures. Cambridge University Press. Cambridge. RAMBERG. H. 1960. Relationships between length of arc and thickness of ptygmatically folded veins. American Journal of Science, 258. 36-46. 1961. Contact strain and fold instability of a multilayered body under compression. Geologische Rundschau. 51, 405-439 RAMSAY, R. G. 1967. Folding and Fracturing of Rocks. McGraw-Hill. London. RICHARD, P. 1990. Champs de Failles Audessus d'un Decrochement de Socle: Modelisation Experimentale. Ph.D. Thesis. University of Rennes. France. 1991. Experiments on faulting in a two layer cover sequence overlying a reactive basement fault with oblique slip. Journal of Structural Geology, 13. 459-470. & KRANTZ, R. W. 1991. Experiments on fault reactivation in strike-slip mode. Tectonophvsics. 188,117-131.
FORCED FOLD AND BUCKLE FOLD GEOMETRY , MOCQUET, B. & COBBOLD, P. R. 1991. Experiments on simultaneous faulting and folding above a basement wrench fault. Tectonophysics, 188, 133-141. SATTARZADEH, Y., COSGROVE, J. W. & VITA-FINZI, C. 1999. The interplay of faulting and folding during the evolution of the Zagras deformation belt. This volume.
21
STEARNS, D. W. 1978. Faulting And forced folding in the Rocky Mountain foreland. Geological Society of America Memoir, 151, 1-38 SUPPE,J. 1983. Geometry and Kinematics of fault-bend folding. American Journal of Science, 283, 684721. WILCOX, R. E., HARDING, T. P. & SEELY, D. R. 1973. Basic wrench tectonics. AAPG Bulletin, 57, 74-96.
This page intentionally left blank
Interlayer slip and joint localization in the East Kaibab Monocline, Utah: field evidence and results from numerical modelling M. L. COOKE1, P. N. MOLLEMA2, D. D. POLLARD & A. AYDIN Rock Fracture Project, Stanford University, Stanford, CA 94305-2115, USA ( l Present address: Geosciences Department, University of Massachusetts, Amherst, MA 01003-5820, USA; 2 Present address: Western Atlas Logging Services, 10205 Westheimer OB6, Houston, TX 77042, USA) Abstract: The mechanics of deformation in multilayer flexures is analysed by comparing field observations of joint clusters from the East Kaibab Monocline, Utah with fracture patterns produced in analytical and numerical experiments. Dune boundaries (bedding planes) were mapped through the thickness of the aeolian Navajo Formation, and the occurrence of joints related to dune boundary slip and fold curvature was documented. Slip along dune boundaries, as evidenced by joint clusters oblique to bedding, occurs along the steep limb of the fold and in the middle of the Navajo. Joints perpendicular to bedding and parallel to the fold axis occur near the synclinal hinge. Numerical experiments examine a layer flexed to match the Navajo at Hackberry Canyon with both uniform and observed distribution of dune boundaries. Within the numerical experiments, horizontal frictional interfaces slip within the centre of the layers where dips are steepest, and opening-mode fractures related to curvature form within the anticlinal and synclinal hinges of the fold. Thus, the first-order numerical results match field observations. This study illustrates the important roles of mechanical stratigraphy and interlayer slip in multilayered folding and the contribution of bedding-plane faults and fold curvature in the production of joint clusters.
The deformation of layered sedimentary strata upon flexure is, in part, controlled by the nature of the contacts between the layers. If the contacts are bonded, the strata deform like a single beam; if the layer contacts are weak in shear, they may slip past one another like the pages of a flexed telephone directory. In addition to the nature of layer contacts, deformation of flexed multilayers is also controlled by the thickness and material properties of each layer. Together, these three attributes describe the mechanical stratigraphy of sedimentary strata and govern flexural deformation. Joints grow in response to stresses in the rock; the presence of joints indicates that, at some time, the stresses were great enough within that portion of the rock to initiate fracture. In this manner, joints are indicators of the palaeostresses within rock and can be used to unravel the geological history of a structure. We postulate that the clustering of joints within folds may be a result of both variation in fold curvature and slip along layer contacts (beddingplane faults). Local increases in curvature raise the fibre stress in the outer arc of mechanical layers and thereby promote joint growth. As geological layers fold, slip along bedding planes alters the local stress state, potentially concentrating stresses and promoting joint growth; thus, joint clusters may be used as evidence for
slip along bedding planes. Although beddingplane faults do not often exhibit offset markers, they may produce joint clusters, slickenlines and/or other fault textures. Other mechanisms, such as lithological contrasts, both among and between layers, and relative position with respect to an inferred underlying basement fault, may also contribute to joint localization but are not investigated in this study. In this paper we compare field observations of interlayer slip and joint localization in the East Kaibab Monocline, Utah (USA) with patterns produced in numerical experiments. Numerical models of fracture localization using the boundary element method (BEM) explore the character and extent of interlayer faulting and joint localization. Previous investigations of bedding-plane faulting and joint localization
Curvature-related joints As a layer is flexed, the outer arc stretches promoting the development of opening-mode fractures (e.g. Price & Cosgrove 1990). The region of highest tension occurs where the fold curvature is greatest. In linear elastic materials, the layer-parallel normal stress is proportional
From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 23-49. 1-86239-060-6/OO/S 15.00 © The Geological Society of London 2000.
24
M. L. COOKE ET AL.
to layer curvature and increases linearly from the centreline to the top and bottom surfaces of the layer (e.g. Timoshenko & Goodier 1934). The maximum stress at the layer surfaces is proportional to the cube of the layer thickness (e.g. Timoshenko & Goodier 1934); the thicker the layer, the greater the tension in the outer arc of the fold. Thus, in nature, curvature-related joints are expected to form parallel to the fold axis, perpendicular to bedding and within fold hinges. Correlations between fracture density and fold curvature have been documented both by surface mapping (e.g. Gorham et al. 1979) and by inference of fracture density from increased well productivity (e.g. Murray 1968; Gorham et al. 1979).
Bedding-plane faults Slip along bedding planes contributes to, and is evidence for, folding. As elastic layers flex, the shear stress on initially horizontal planes increases with increasing dip of the planes (e.g. Timoshenko & Goodier 1934). When the shear stress exceeds some critical strength of the material, a slip patch which will develop between the layers and reduce the shear stress. This shear 'decoupling' of layers allows a fold to accommodate a greater flexure than if the stack deformed as a single layer. Mechanical models show that if layers slip along interfaces when flexed, folds will have greater amplitude and sharper hinges than folds with bonded layer contacts (Chappie & Spang 1974; Freund 1979; Koch el al 1981; Roth et al. 1982). Evidence for bedding-plane slip is often limited to slickenlines and offset markers. As potential offset markers are rare in sedimentary strata, bedding-plane slip may frequently occur without prominent evidence. Nevertheless, geologists have observed beddingplane slip within drape folds over laccoliths (e.g. Jackson & Pollard 1990) and faults (e.g. Lewis & Couples 1993; Becker 1994).
1991; Mollema & Aydin 1997) and splay cracks (Martel et al. 1988). Within this paper we will refer to these features as splay cracks to describe the oblique trace of the fractures with respect to the associated fault. The presence of joints at the tips of faults has motivated investigations into the fault tip stress field as a causative agent for fracturing (Segall & Pollard 1983; Cruikshank 1991; Cruikshank & Aydin 1994). Using linear elastic fracture mechanics, researchers have shown that, within brittle materials, the concentration of stresses near tips of faults promotes the initiation of an obliquely oriented opening-mode fracture (e.g. Lawn 1993). These faults can be analysed as Mode II fractures when viewed along a cross-section parallel to slip direction. Theoretical prediction of the propagation path of single splaycracks propagating from the tips of Mode II fractures has been verified with both laboratory (Brace & Bombolakis 1963; Erdogan & Sih 1963; Nemat-Nasser & Horii 1982) and numerical experiments (Thomas & Pollard 1993). Standard linear-elastic fracture mechanics does not predict the formation of joint clusters away from the tips of faults, such as along bedding planes. At least three mechanisms have been postulated for the growth of splay cracks away from fault tips: (1) splay cracks mark the locations of previous slip-patch terminations (Martel et al. 1988); (2) deviations from a purely planar slip surface could provide regions of local stress concentration for the initiation of multiple splay cracks (Cruikshank 1991; Cruikshank & Aydin 1994); and (3) lateral variations in fault strength may produce broad zones of elevated stresses within which splay cracks may localize (Cooke 1997). Changes in frictional strength along bedding planes may be produced by lateral changes in thickness, grain size, composition or texture in overlying and/or underlying layers.
Fault-related join ts
Geological setting of the East Kaibab Monocline
Joint clusters that are highly localized and restricted in length are associated with faulting of clastic sediments (Granier 1985; Cruikshank & Aydin 1994; Mollema & Aydin 1997), carbonates (Rispoli 1981; Petit & Mattauer 1995), and granite (Segall & Pollard 1983; Granier 1985; Martel et al. 1988). Secondary fractures associated with fault terminations are referred to variously as horsetail fractures (Granier 1985; Cruikshank 1991), tail cracks (Cruikshank
The Kaibab Monocline is an approximately IMS-trending fold in the eastern portion of the Colorado Plateau (Fig. 1). One of many uplifted blocks within the Colorado Plateau tectonic province, the Kaibab Monocline is believed to be a forced fold, i.e. formed by vertical movement on high-angle faults underlying the folded strata (Hodgson 1965; Davis 1978; Huntoon 1993). Layer-parallel compression may have contributed to folding either during or after
NUMERICAL MODELLING OF JOINT LOCALIZATION
25
Fig. 1. Geological map of part of the East Kaibab Monocline and its location within the Colorado Plateau (in inset). Three sections of the fold, outlined with boxes on figure, were mapped by Mollema (1994) and Mollema & Aydin (1997). This study focuses on the northern most section, Hackberry Canyon. (Taken from Mollema and Aydin, 1997.)
26
M. L. COOKE ET AL.
Fig. 2. Cross-sectional profile across the East Kaibab Monocline at Hackberry Canyon (located on Fig. 1). The fold has a broad anticlinal hinge and a sharper synclinal hinge. A normal fault with 30-50m throw crosses this section near the steep limb of the fold. Four structural domains, characterized by deformational style, are expressed within this section. Domain 4 of fold-parallel and bed-perpendicular joints is exposed in the Cretaceous Straight Cliffs Formation (Ks) within the synclinal hinge of the monocline. Bedding-plane faults (domain 2) are observed in the Jurassic Navajo Formation (Jn. denoted in light grey), which is exposed within the anticlinal hinge and left portion of the monocline's steep limb. The exposed region of the Navajo is highlighted in dark grey (Adapted from Mollema and Aydin, 1997).
faulting (Reches & Johnson 1978a; Himtoon 1993; Mollema 1994). The Kaibab Monocline is about 240km long and 5 km wide with the west side up relative to the east side. Although the fold axis diverges to the south, the northern portion of the fold (East Kaibab Monocline) contains a single sublinear NNE-trending fold axis. The study area lies in the northern portion of the structure where the monocline trends NNE (020°) and has relatively few major faults. The strata in this area have a regional dip of 05° to the east with a maximum dip within the fold of about 60° to the east. In the study region, the monocline is asymmetric with a relatively tight syncline and broad anticline. Rocks of Permian-upper Cretaceous age are exposed in the area and the total thickness of the stratigraphic section is about 4600 m. Mollema (1994) and Mollema & Aydin (1997) characterized four structural domains across the East Kaibab Monocline in this area through detailed mapping of three sections (e.g. Hackberry Canyon cross-section, Fig. 2). The structural domains were delineated on the basis of recurrent fracture assemblages. The westernmost
domain contains a systematic joint set, striking obliquely to the monocline, accompanied by a non-systematic set of joints at high angle to the systematic set. Eastward of the domain of systematic joints is a domain with dips up to 45° and containing interlayer faults and associated splay cracks. To the east of the domain of bedding-plane faults is a domain with the greatest bed dips and steeply dipping faults striking parallel to the fold axis accompanied by deformation bands, small faults with cataclastic gouge. The easternmost domain has a systematic joint set striking parallel to the monocline along with joints at high angle to the systematic set. Generally, interlayer faults and associated joint clusters are interpreted to have formed earlier in the development of the monocline than deformation bands, fold-parallel faults and fold-parallel joints (Mollema 1994; Mollema & Aydin 1997). For the purpose of this study, we compare the field observations of fracture localization and interlayer slip to the results of numerical experiments. As we do not investigate the development of steeply dipping fold-parallel faults and associated fractures, we focus our study on the northernmost and least faulted of
NUMERICAL MODELLING OF JOINT LOCALIZATION
27
Fig. 3. Large joints perpendicular to bedding and parallel to fold axis within domain 4 of the Buskin Gulch section of the East Kaibab Monocline (located on Fig. 1) . The subvertical joints in the aeolian sandstone weather to form fin-like erosional structures. The region within the shadows at the centre of the image is composed of a closely spaced cluster of joints. The geometry of the large joints suggests that they formed due to high layer-parallel stresses as the layers were flexed and the outer arcs of the Navajo stretched. The non-linear trace of some finer joints within the zone at the centre of the photo suggest that these joints were subjected to relatively complex stress fields; these finer joints may have developed after the larger curvature-related joints.
the three cross-sections investigated by Mollema (1994): the Hackberry Canyon section (Figs 2 and 3). All four structural domains are well represented within this section. We focus our investigation on the domains of interlayer slip and fold-parallel joints. Field evidence for curvature-related joints Joints that grow in response to bending of a layer are expected to develop parallel to the fold axis and perpendicular to bedding along the outer arc of the fold where fold curvature is highest. At Hackberry Canyon, joints perpendicular to bedding and parallel to the fold axis are exposed within the Straight Cliffs Formation along the synclinal hinge of the fold (Fig. 2). The geometry and location of these joints suggests that they formed under layer-parallel tension due to increased fold curvature. Observation of foldparallel, bed-perpendicular joints is not limited to the Straight Cliffs Formation; the Navajo Formation exhibits such potentially curvaturerelated joints in the Buckskin Gulch section (Figs 1 and 3). In Fig. 3, concentrated erosion
enhances the expression of the large joints and joint clusters. Generally, fold-parallel and bedperpendicular joints are observed in the synclinal hinge and not the anticlinal hinge of the East Kaibab Monocline. The relative curvatures of the hinges, which may account for this observation, will be discussed later in this paper. Field evidence for interlayer slip At Hackberry Canyon interlayer slip and associated splay cracks are observed within the Jurassic Navajo Formation (Fig. 2). The domain of interlayer slip is exposed within the anticlinal hinge and upper portion of the steeply dipping limb of the monocline. Dune boundaries and inter dune deposits The Navajo Formation is a massive cross-bedded aeolian sandstone containing fine-grained interdune deposits (Stokes 1968; Middleton & Blakey 1980). Within aeolian sandstones, the bounding surfaces of dunes act as bedding
28
M. L. COOKE ET AL.
planes and interfaces for potential frictional slip. Bounding surfaces cut across cross-strata and other structures within the underlying rock unit (Stokes 1968; Brookfield 1977). These discontinuities represent a pause in deposition with possible erosion (Brookfield 1977). There may be a change in depositional environment associated with bounding surfaces, as evidenced by overlying interdune deposits (Kocurek 1981). These deposits may range from a few centimetres to a few metres in scale (Kocurek 1981) and contain lacustrine, carbonate or intratidal deposits (Middleton & Blakey 1980; Fryberger et al 1990). Interdune strata may be either fine grained and easily eroded or highly cemented and resistant to erosion (Fryberger et al 1990). First-order bounding surfaces are planar and extensive surfaces that cross-cut higher-order (smaller-scale) bounding surfaces (Brookfield 1977) and may be overlain by interdune strata (Kocurek 1981). Higher-order bounding surfaces dip relative to the first-order bounding surface and may be non-planar (Brookfield 1977). For the purposes of this study, we are interested in first-order bounding surfaces that are overlain by interdune deposits, because these surfaces may localize deformation. Within this paper first-order bounding surfaces underlying interdune strata are referred to as dune boundaries. At least two processes have been proposed for the formation of planar first-order bounding
surfaces and associated fine-grained sediments. An elevation in water-table may erode dune sand to water level and deposit finer-grained material in extensive planar and horizontal layers (Stokes 1968). Alternatively, planar bounding surfaces may form within a migrating interdune basin (Brookfield 1977; Kocurek 1981). In this case bounding surfaces climb as the dunes and interdune basins migrate. The angle of climb for migrating bedforms that form a first-order bounding surface is calculated to be up to 2° for fast (4 cm year"1) bedforms as large as 20 m (Brookfield 1977). For the purposes of this study we will assume first-order bounding surfaces to have developed horizontally. In southern Utah, the Navajo Formation contains somewhat regularly spaced, horizontal first-order bounding surfaces and overlying interdune deposits of carbonate and lacustrine sediments (Stokes 1968; Middleton & Blakey 1980). Interdune deposits at Hackberry Canyon are most easily recognized by differential weathering. For example, interdune material often erodes more easily than the dune material and the erosion results in a groove in the surface of the canyon wall (Figs 4 and 5). The expression of interdune layers as grooves aids the identification of dune boundaries from the canyon floor. For example, the highest and lowest visible dune boundaries in Fig. 4 are about 30 m apart and are easily detected as grooves in the canyon
Fig. 4. Dune boundary 16, as numbered up-section from the Kayenta-Navajo contact. The fine-grained interdune material has eroded making this dune boundary recognizable from a distance. The interdune material thins and disappears to the left of the tree in the foreground. The oblique splay cracks facilitate dune boundary identification and in this view indicate a top-to-the-right (up-dip) sense of slip. Some of these splay cracks are 40m long and may be significant conduits for fluid flow in the subsurface.
NUMERICAL MODELLING OF JOINT LOCALIZATION
29
Fig. 5. The field geologist is pointing out dune boundary 5. The red-stained and fine-grained interdune material varies in thickness and erodes easily. This dune boundary can also be recognized by the oblique splay cracks indicating top-to-the-right (up-dip and to the northwest) sense of slip along the dune boundary.
wall. The nature of erosion may change along dune boundaries, according to the thickness and composition of the interdune material. The lowest dune boundary in Fig. 4 appears to terminate to the left of the tree in the foreground where the groove in the canyon wall ends. While interdune material thins at this point, producing a less evident groove, the dune boundary continues to the left and downward as the
exposure curves towards the camera. Similarly, the interdune layer is most eroded at the right edge of Fig. 5 where it is thickest. Within the Navajo Formation, there are a few intensely cemented interdune layers which are more resistant to erosion than the dune material (Fig. 6). Interdune deposits within the Navajo Formation at Hackberry Canyon contain silt-sized grains, finer than the overlying and underlying
Fig. 6. Dune boundary 9. The camera lens cap rests on a small ledge produced by the relative resistance of the interdune layer to erosion. The interdune deposit is thin and contains layers of pale and dark fine-grained silts. The dark colour suggests high organic content of the interdune layer. Cross-bedding is evident in the sandstone above the interdune layer.
30
M. L. COOKE ET AL.
aeolian sandstone. The fine-grained nature of these layers may contribute to increased erosion, as observed in Figs 4 and 5. In contrast, the thin interdune deposit of Fig. 6 is fine grained but highly cemented, lending it more resistant to erosion. Most of the interdune layers at Hackberry Canyon are fine grained and relatively uncemented. Many interdune deposits within Navajo Formation of Hackberry Canyon are stained red. In the black and white photographs of Figs 4 and 5, red-stained dune boundaries appear darker grey than the surrounding rock. This staining aids in the identification of dune boundaries. Some portions of dune boundaries contain white material, which resembles a bleached gouge layer associated with faulting. After locating potential dune boundaries, cross-cutting relationships within underlying strata were examined to verify that the surfaces are firstorder bounding surfaces. Joints associated with dune boundary slip Localized splay cracks draw attention to dune boundaries (Figs 4 and 5). These oblique fractures do not cross the dune boundaries and are interpreted to be indicative of slip along the dune boundary. Although the majority of the exposed splay cracks developed just above the dune boundaries, some splay cracks were observed below the dune boundaries (Fig. 8). Joint spacing is non-uniform along dune boundaries. Splay crack density increases near irregularities in dune boundaries, such as changes in interdune layer thickness (Fig. 4) and bends in slip surface (Fig. 7). Deformation along dune boundary 7 is unusual in that the throughgoing slip surface develops along three segments of cross-bedding rather than along the more planar dune boundary. The result is a non-planar slip surface along which fractures are highly localized near steps between segments (Fig. 7a). The angle between slip surface and splay crack, the crack initiation angle, is oblique to bedding and is non-uniform. Along dune boundaries 7 and 11 the crack initiation angle ranges from about 20° to 85°; the average crack initiation angle is 52° + 9° for boundary 7 and 49° + 16° for boundary 11 (Fig. 7). The splay cracks along boundary 7 have more consistent orientations than those along boundary 11. Splay cracks observed in the bedding-plane slip domain of Hackberry Canyon showed consistent sense of relative slip (e.g. Figs 5, 7 and 8). Throughout the Hackberry Canyon exposure of the Navajo Formation, dune boundaries dip
Fig. 7. Map of splay cracks along dune boundary 7. Equal-area stereonets (lower hemisphere) of splay crack orientation along dune boundaries 7 and 11. Whereas dune boundaries dip gently to the southeast, splay cracks along these boundaries are more steeply dipping to the northwest. The layer above the dune boundary slips to the northwest, along the dune
shallowly to the southeast (Table 1, Figs 5, 7 and 8), whereas splay cracks (both above and below the dune boundaries) dip approximately to the northwest (Figs 5, 7 and 8). This evidence indicates that the top layer slipped to the northwest relative to the bottom layer; the top layer slipped up-dip relative to the bottom layer. Slip indicators within the fault material, such as slickenlines, were not observed. Location and extent of dune boundary slip Within the Navajo Formation, the distribution of interdune slip is non-uniform; there are regions of no joints (Fig. 6) and regions of many joints related to interlayer slip (Figs 4 and 5). In order to document the location of interdune slip, dune boundaries were mapped from the Kayenta-Navajo contact to the Navajo-Carmel contact. Twenty-three dune boundaries were delineated within the exposed Navajo Formation along Hackberry Canyon. Table 1 presents a brief description of fractures associated with slip along dune boundaries. Dune boundaries were mapped by recording the dune boundary slip and the distance between successive dune layers. The thicknesses of dune layers are part of the mechanical stratigraphy of the Navajo which will aid in modelling flexure of the East Kaibab Monocline. The dune boundaries are best expressed on canyon walls where direct measurement of the distance between bounding surfaces is difficult; the canyon walls are up to 200m high and nearly vertical in most places. The safest (i.e.
Fig. 8. Dune boundaries 7 and 8. The dune boundaries 20m apart can be recognized from a distance as they form characteristic 'ledges' or 'breaks' along canyon wall. Cross-beds (medium-weight lines in sketch) are distinguished from dune boundaries (thick lines in sketch) by their steeper dips. Splay cracks, sketched in thin lines, indicate top up-dip (to the northwest) sense of slip.
M. L. COOKE ET AL.
32
Table 1. Compilation of field data: dime boundary orientation, dune thickness and description of splay cracks. Dune boundary
Strike dip ( / )
Thickness (m)
Jk-Jn contact 1
3 4
320 12 020 12 030 19 017/20 055/21
5 5 15 20 25
5 6
-
20 35
7
-
22.5
8
022 22
22.5
9 10
000 30 -
45 30
11
10
12
022 15:030 180 28/18 -
10
13
028 18
15
14 15 16 17 18 19
038 22 048 24 032/33
10 10 23.3 11.7 15 35
20 21 22 23
028:31 028/32 025 35 -
30 35 25 40 -
9
Description of splay cracks [length in m]
no splay cracks no splay cracks no splay cracks no splay cracks few scattered splay cracks along cross-bed boundaries between 4 and 5, slight localization [<3m] few splay cracks with some localization along 5 [<10m] (Fig. 5) few splay cracks along cross-bed boundaries between 5 and 6 [c. 5m] many splay cracks with some localization between 7 and 8. cracks along cross bedding, cross-bed and dune boundaries, spots of high localization along 7 [<15m] (Fig. 8) many splay cracks with some localization between 8 and 9. cracks along cross-bedding, cross-bed and dune boundaries [<15m] (Fig. 8) no splay cracks some localized splay cracks along 10 [8 m]: few splay cracks along cross-bed boundaries below 11 [<5m] many splay cracks with high localization in spots [20-25 m]: some splay cracks along cross bedding between 1 1 and 12 [10m] many splay cracks [10m]; some splay cracks along cross-bedding between 12 and 13 [10m] many splay cracks with some localization [20 m]: some splay cracks along cross-bed boundaries between 13 and 14 [<10m] some splay cracks and some localization [20m] few splay cracks some splay cracks and some-high localization [25-30 m] (Fig. 4) few splay cracks along cross-beds between 17 and 18 [<2m] few splay cracks along cross-beds between 18 and 19 [<2m] few scattered splay cracks on cross-bed boundaries and crossbeds near and above boundary 19 [<1 m] few splay cracks [<8 m] some splay cracks [c. 8 m], deformation bands no splay cracks, deformation bands 35 = 5 no splay cracks, many deformation bands, some cracks related to brecciated fault zone
Relative number of splay cracks: few < 5; 5 < some < 25; many < 25 Degree of localization: slight = 5 cracks in 0.5 m; some = 10 cracks in 0.5 m; high = greater than 10 cracks in 0.5m.
most efficient) method to determine the dune thicknesses is to map the location of dune boundaries along the canyon floor and to find the true thickness using geometric projections. The quality of dune boundary expression varies along canyon surfaces. Where dune boundaries were difficult to distinguish from cross-bed boundaries, their attitudes could be used to discriminate between the different boundary types. For example, Fig. 8 shows a section of the canyon near dune boundaries 7 and 8. The cross-bedding is nearly indistinguishable from the dune boundaries based on staining or degree of erosion. However, the cross-bed boundaries can be identified because they dip more steeply relative to the dune boundaries. Although exposures along the canyon floor
may not provide enough evidence to discriminate between dune and cross-bedding boundaries, once a dune boundary is identified along the canyon wall its location along the canyon floor can be found by extrapolation. Dune boundaries were numbered sequentially up-section from the base of Navajo Formation (Kayenta-Navajo contact). Stations were established within the middle of the canyon where each dune boundary approximately crosses the canyon bottom. Station numbers correspond to numbers of the dune boundaries. Each station was located using the bearing and distance from the previous station starting at the Kayenta-Navajo contact. Dune boundaries 8, 12, 15 and 17 were not discernible along the canyon floor but their locations were estimated
NUMERICAL MODELLING OF JOINT LOCALIZATION
between the adjacent boundaries. If the strike and dip of the dune boundaries are known, then the layer thicknesses can be calculated from horizontal distances between stations along a section perpendicular to dune boundary strikes; the distances between stations along canyon floor are projected into the dip-direction. The distance between boundaries is the sine of the dip (from 12° to 30°, Table 1) multiplied by the projected distance in the horizontal plane. The estimated layer thicknesses were rounded to the nearest 5 m in recognition that variations in attitude and dune thickness incorporate imprecision into the thickness calculations (Table 1). Changes in dip of 5° can produce changes in bed thickness of up to 8%. The bedperpendicular distance from the base of the Navajo to the top, as calculated from this method, is 515m. However, between station 23 and the top of the Navajo is a normal fault observed to have 30-50 m of offset (Fig. 2) which increases apparent thickness. The bedperpendicular distance between station 23 and the Navajo-Carmel contact is 40 m and provides an upper limit on the amount of normal fault offset. As the thickness of the Navajo should not include the repeated horizontal length of the exposed Navajo due to slip on the normal fault, we remove 35m from this section. The summed total thickness of the Navajo is approximately 480m. The Navajo Formation is not exposed through the entire Hackberry Canyon cross-section but, rather, in a small exposure window illustrated in the cross-section (Fig. 2). We hypothesize that our observations of dune boundary spacing made within this window are representative of the spacing across the 5-km width of the monocline. Observed parallel to ancient wind direction, first-order bounding surfaces are very extensive (Kocurek 1981). Because wind direction for the Navajo Formation in southern Utah is generally southeast (Middleton & Blakey 1980) and subparallel to the Hackberry cross-section, we can confidently extrapolate from the observed dune boundary distribution. The distribution of dune boundaries through the Navajo Formation is presented in Fig. 9 with regions of splay cracks indicated with hatched patterns. Near the Kayenta-Navajo contact the dune layers are thin (about 5m) and generally increase in thickness up-section to the middle of the formation. The thickest dune layers (approximately 45m) occur just below the middle of the layer. At the middle of the Navajo Formation the dune thickness dramatically decreases from 45 to 10m. From the middle of the formation to the Navajo-Carmel
33
contact, the dune thicknesses generally increase. Splay cracks are most prevalent among the thin layers near the centre of the Navajo, and are not found near the base and top of the formation. The thickest dune layer just below the centre of the layer does not exhibit significant splay cracks. The region of high splay crack density occurs along the steep limb of the monocline and part of the anticlinal hinge. As splay cracks are evidence of interlayer slip, dune boundaries at Hackberry Canyon slip primarily among the thin layers near the centre of the Navajo and along part of the anticlinal hinge and steep limb of the monocline. Numerical experiments of Navajo Formation flexure We modelled the flexure of the East Kaibab Monocline in order to understand the nature of multilayer deformation upon flexure. To this end, we compared the character and extent of interlayer slip and curvature-related joints of the model results with field observations. In order to model the flexure of the East Kaibab Monocline, the mechanical stratigraphy had to be documented. Rather than modelling flexure of the entire 4.6km of strata within the East Kaibab Monocline, we simplified the task by only considering the flexure of the Navajo Formation. Because the Navajo displays both interdune slip and curvature-related joints, we were able to compare the model results to the field observations in a straightforward fashion. Our field observations contained the data required to describe the mechanical stratigraphy of the Navajo Formation: the material properties of the dune layers were assumed homogeneous and the thicknesses of dune layers are documented in Table 1. We performed boundary element method (BEM) numerical experiments to examine the slip along interdune strata and associated opening-mode fracture growth, as well as the development of curvature-related joints during flexure. The frictional slip along horizontal layers was examined, as well as the stresses that initiate opening-mode fractures.
Boundary element method The BEM is a numerical technique for solving the governing equations of continuum mechanics, including heat and mass transport and solid deformation. In principle, this method can be used to describe the deformation
Fig. 9. The distribution of dune boundaries through the Navajo Formation in Haekberry Canyon. The observations of dune boundaries distribution are extrapolated from the window of exposure throughout the Navajo within this fold. Dune layer thiekness increases from the base of the Navajo to the middle of the formation. At the middle of the layer, the dune thickness decreases sharply while, above the middle, layer thicknesses increase to the top of the Navajo Formation. Regions and intensity of splay cracks are indicated in shades of dark grey. Interdune slip and associated splay cracks occur primarily within two regions: among moderately t h i n dune layers below the thickest layers at the middle of the Formation, and above the thick layers where the dune layers are t h i n . There is no evidence of inlerdune slip near the base and top of the Navajo Formation, nor is there evidence for slip w i t h i n the thickest layer near the middle of the formation.
NUMERICAL MODELLING OF JOINT LOCALIZATION
of any elastic body if either the tractions or displacements are prescribed along the internal and external boundaries of that body. The advantage of BEM compared to other techniques, such as finite element method (FEM) or the finite difference methods (FDM), is that only the internal and external boundaries and interfaces need to be discretized into boundary elements (Crouch & Starfield 1990; Becker 1992). The stress, strain and displacement field throughout the body are uniquely determined by these boundary conditions. Another advantage of the BEM is that it is particularly adept at solving problems with linear (2D) or planar (3D) discontinuities in the displacement field which are characteristic of faults and fractures (Crouch & Starfield 1990). The numerical experiments described in this paper use a computer code which is based upon the displacement discontinuity formulation of the BEM in two dimensions (Crouch & Starfield 1990). The boundaries of the model are discretized into linear elements, each associated with a normal and shear displacement discontinuity. Whether tractions or displacements are prescribed on the elements, the displacement discontinuity is constant along the length of each element. Dune boundaries are modelled with special elements that use the Coulomb friction criterion to determine if interface elements slip (Crouch 1979; Crouch & Starfield 1990). These frictional elements have been used to investigate slip on bedding-planes (Cooke & Pollard 1994; Cooke & Pollard 1997; Roering et al 1997) and faults (Schultz & Aydin 1990; Cooke 1997). Slip along elements is inelastic in the sense that it is not necessarily recovered during unloading, but this slip can be reversed by applying the appropriate combination of shear and normal stresses. The loading of the boundaries of the model is applied monotonically in small increments to minimize path dependency of inelastic slip. As slip along one element may influence the shear tractions on nearby elements, the problem is iterated until the solution converges, as defined by a user-prescribed tolerance (Cooke & Pollard 1997). After the solution converges the model loading can be incremented to the next step. An important constraint on the development of flexural slip is the depth of burial during folding. As the slip on a frictional interface is related to the normal stresses across it, the weight of sedimentary overburden is a significant deterrent to development of bedding-plane faults at depth. The weight of sedimentary overburden is modelled by superposing a linearly increasing compressional normal load with depth.
35
Frictional interfaces may act to concentrate stresses and promote joint development. Opening-mode fractures (joints) are predicted to initiate where the tensile stresses exceed the tensile strength of the rock. After the convergence of all boundary elements, the maximum tensile stress is determined from the normal (crvv), shear (crVY) and tangential (<J YY ) stresses at numerous points above and below each frictional interface. While the shear and normal stresses must be continuous across the interface, the tangential stress may be discontinuous. The angle of fracture growth is determined from the orientation of the maximum tensile stress at the initiation point according to the relationship which assumes that openingmode fractures grow perpendicular to the direction of maximum tension (Jaeger & Cook 1979; Lawn 1993). Three models analyse the flexure of the Navajo Formation within the East Kaibab Monocline. The first model examines joint development from the outer arcs of the fold and contains no horizontal frictional interfaces. This model investigates the potential for curvature-related joints to develop. The second model has eight evenly spaced frictional interfaces, whereas the third model duplicates the observed distribution of dune boundaries in Hackberry Canyon. Our intention is to examine the first-order effects of interdune slip and on jointing with the evenly spaced model. The third model will examine the second-order effects of mechanical stratigraphy, in this case variations in layer thickness, on distribution of interdune slip.
Numerical experiment set-up While the exact mechanism that produced the East Kaibab Monocline is unknown and a topic of ongoing discussion and research (Hodgson 1965; Davis 1978; Reches & Johnson 19786; Huntoon 1993), the deflection profile can be estimated from the exposed structure. Rather than model a specific driving mechanism to produce the fold observed at Hackberry Canyon we use the available data, the final fold shape. However, the regional dip, 05° to the east, does not contribute to folding deformation and is removed from final fold shape data. Within the numerical experiment an initially horizontal layer is deflected into the shape of East Kaibab Monocline at Hackberry Canyon, minus the regional dip and with a much-reduced amplitude. The BEM code considers only elastic behaviour and infinitesimal strains. For a 4.5-km wide monocline, a total throw of 45m will produce shear strains of around 1% which is
Th
M. L. COOKE ET AL
approximately the limit for infinitesimal strain approximations. The models of Navajo flexure are deflected up to 30 m with fold shapes similar to the deflection profile at Hackberry Canyon where the actual monocline throw is 1300m. We assume that the development of the East Kaibab Monocline has been self-similar through time. This assumption may not accurately describe the actual development of the East Kaibab Monocline but provides a first-order approximation with which to start our numerical investigation. Because we deflect the structure less than 45 m we may neglect topographic effects on gravitational overburden loading. In general the stresses imposed by topographic variations will only be felt within a region of radius equal to the length scale of these variations (McTigue &Mei 1981). Analytical functions of the constructed deflection profile are needed to deflect the top and bottom boundaries of the numerical models. The fold profile of Hackberry Canyon (Fig. 3) was sampled at 100-m intervals and the point data used to develop an analytical function which simulates the fold shape (Fig. 10). Before developing a function the regional dip of 05° to the east was subtracted from the data; uniform regional dip rotates all of the rock strata and does not contribute to the deformation within the Navajo Formation. A transcendental function was used to capture the overall monoclinal shape of the fold as follows:
where z is the deflection profile and x is the horizontal distance in metres from the western edge of the cross-section (Fig. 10). The symmetric function forms a curve similar in shape to the monocline, but does not capture the minor asymmetry of Hackberry Canyon cross-section. The anticlinal portion of the Hackberry Canyon cross-section is broader and the synclinal portion tighter than the hyperbolic tangent function (Fig. 10). As the transcendental function captures the first-order fold shape, we will use this function for the analytical and numerical analyses of the East Kaibab Monocline. Further considerations of the fold asymmetry are expressed in the Discussion of this paper. The upper and lower boundaries of the model are vertically displaced according to equation (1) in monotonic increments (Fig. 11). The prescribed shear tractions on the upper and lower boundaries are zero, so the Navajo Formation is decoupled from the rest of the strata within the fold. The geological formations above and below the Navajo contain abundant shale-siltstone sequences, weak in shear, which will slip and reduce the horizontal shear stress. We recognize that the actual shear tractions on the top and bottom surfaces of the Navajo are non-zero and describe the implications of our assumption in the Discussion of this paper. The left side of the model is fixed (neither horizontal nor vertical displacements allowed), while the right side displaces downward incrementally accordingly to the monocline amplitude described in equation (1). The right side is not allowed to displace
Fig. 10. Sampled deflection profile from cross-section of Hackberry Canyon (Fig. 3) with an approximate analytical function. The transcendental function captures the overall shape of the monocline but neglects its asymmetry. The anticlinal portion of the fold is broader and the synclinal portion tighter than the hyperbolic tangent function.
NUMERICAL MODELLING OF JOINT LOCALIZATION
37
Fig. 11. Boundary conditions to describe the deformation of the Navajo Formation within the East Kaibab Monocline at Hackberry Canyon. The left side is held fixed while the top and bottom surfaces are incrementally displaced according to equation (1). There are no shear tractions on the upper and lower surfaces of the model. The right side displaces downward in monotonic increments. Lithostatic loading equivalent to 1.5km of sedimentary overburden is superposed onto the stress state as a result of flexure.
horizontally. There is evidence at the East Kaibab Monocline for overall horizontal contraction across the structure (Reches & Johnson 19780; Mollema & Aydin 1997); however, contraction will have only a secondary effect on flexural slip and we neglect this component of deformation within this study. The displacement profiles on the upper and lower boundaries of the fold are the same, resulting in a similar fold (i.e. the layer becomes thinner along the fold limbs than at the hinges). The exposure within the East Kaibab Monocline does not give adequate evidence to determine if this fold is parallel, similar or of some hybrid fold class. Because the Navajo Formation is stiff relative to the adjacent strata, a parallel fold is expected to develop (Ramsay & Huber 1987). Within this study we limit fold amplitudes to less 30m which produces a maximum layer dip of 1°. The maximum layer thinning is estimated at 0.015% (Ramsay & Huber 1987); the difference between parallel and similar fold shape is negligible in this study. The exact depth of burial of the Navajo Formation during folding is unknown. Currently, there are approximately 1500 m of folded Jurassic and Cretaceous strata above the top of the Navajo. This is the minimum depth as strata may have been eroded since the uplift of the
Colorado Plateau. For this study we apply a lithostatic (crYV = a rv ) compression equivalent to 1500m of sedimentary rocks (density = 2600 kg m~ 3 ) above the Navajo Formation (Fig. 11). The material properties assigned to the layers are equivalent to coarse sandstone; Young's modulus of 30 GPa and Poisson's ratio of 0.25. In comparison with the dune layers, the interdune layers deform like frictional interfaces and are assigned constitutive properties corresponding to a moderately weak contact; cohesion is zero and friction coefficient is 0.4. Pore pressures are not considered in this model, however, elevated pore pressures at 1.5km burial are likely to contribute to the frictional weakness of the interfaces.
Development of curvature-related joints As the layer curvature increases, the layerparallel normal stresses that initiate fracture increase. The layer-parallel stresses are most tensile at the outer arcs of the flexed layer. Therefore joints related to fold curvature are most likely to initiate at the top and bottom surfaces of the fold. Irregularities along bedding may act as flaws which concentrate stresses and provide initiation points for fractures (Pollard & Aydin
38
M. L. COOKE ET AL.
Fig. 12. The maximum principal stresses along the top and bottom of the layer flexed from 12- to 22-m fold amplitude. The stresses are greater at the crest of the anticline than the trough of the syncline due to increased lithostatic compression. For rock tensile strengths of 5-25 MPa curvature-related joints are expected to form perpendicular to bedding at 14-20m of fold amplitude.
1988). In general, joints develop where the maximum principal stress exceeds the tensile strength of the rock (Jaeger & Cook 1979). Thus, we examine the maximum principal stress along the top and bottom surfaces of the fold to assess the potential for joint development. The maximum principal stress increases everywhere in the outer arc of the Navajo with fold amplitude (Fig. 12). Along the top surface of the fold the highest stresses occur at the crest of the anticline, while along the fold's bottom surface the trough of the syncline contains the highest maximum principal stresses. Opening-mode fractures grow perpendicular to bedding at these locations. The maximum principal stress is greater in the anticlinal hinge than in the synclinal hinge. The larger compressive lithostatic loads along the bottom surface of the fold decrease the maximum principal stress in the trough of the syncline relative to the crest of the anticline. While the fold shape is symmetric the potential for joint growth is not. The tensile strength of rock is measured to be between 5 and 25 MPa (Bieniawski 1984). For this range of values, joints initiate within the fold at amplitudes of 14-20m. The greater the tensile strength, the greater the fold amplitude at fracture initiation.
Frictional slip with uniform dune boundary distribution The influence of interdune layers within the Navajo on joint development is investigated by incorporating frictional horizontal interfaces into the numerical BEM model. Eight interfaces are evenly spaced 60 m apart through the 480-m thick layer. The layer is monotonically deflected in 2m amplitude increments to a maximum amplitude of 30 m. At 20-m fold amplitude, the shear stresses along the interfaces exceed the friction coefficient multiplied by the compressive stress across the interfaces, and the frictional layers begin to slip. Frictional slip initiates within the middle of the multilayer and near the fold centre (Fig. 13). As the fold amplitude increases so does the length of slip patches. The region of interface slip occurs along the steeply dipping limb of the fold. The greatest magnitude of slip (approximately 25cm) occurs along the middle layer, located at a depth of 1740m (Fig. 14a). The horizontal shear stresses within a flexed layer are greatest at the middle of the layer and decrease towards the top and base (Timoshenko & Goodier 1934). Although, the interfaces at depths of 1800 and 1680m are an equal distance
NUMERICAL MODELLING OF JOINT LOCALIZATION
39
Fig. 13. Locations of interface slip for uniform distribution of layers and fold amplitudes of 20-30 m at 2m increments, (a) Slip first develops at 20-m fold amplitude along the frictional interfaces at depths of 1740 and 1680m. With increasing fold amplitude the length of slip patches increases along each layer. At 22-m fold amplitude the 1800-m deep interface slips. The slipped interfaces lie along the steep limb of the fold and near the middle of the layer. Slip patches along the interface at 1680m depth are longer than those along the 1800-m deep interface; increased lithostatic loading on deeper interfaces inhibits frictional slip, (b) The region of frictional slip within a cross-section of the monocline at various fold amplitudes.
from the middle of the layer, the shallower interface has greater slip and a longer slip patch. Owing to increasing lithostatic compression with depth, frictional slip is inhibited on deeper interfaces, producing shorter slip patches (Fig. 13) and smaller offsets (Fig. 14a) on interfaces below the middle of the layer. The slip gradient is highest for the interface at a depth of 1740m and lowest for the interface at a depth of 1800m. Longer slip patches have greater slip magnitudes and higher slip gradients (Fig. 14a and b). The magnitude of slip increases nearly linearly with fold amplitude (Figs 14b and 15a). The distribution of slip is roughly symmetric along each of the slip patches (Fig. 14a and b). The point of greatest offset along the slip patches is located at about 2400m and lies within the region of greatest dip along the monocline (Fig. 14a and b). We analyse the potential development of joints associated with interdune slip from the slip distribution along the sliding frictional interfaces. Regions of high slip gradient are most likely to develop opening-mode fractures.
The local normal stress, acting tangent to the interface, aXX9 is calculated from the slip gradient, ddjdx, from the following formula (Crouch & Starfield 1990):
where 4 is the interface slip, avv the interface normal stress, p is rock density, g the gravitation constant, y the interface location (-depth), and E and v are the material properties (Young's modulus and Poisson's ratio, respectively). The first part of equation (2) is the contribution of slip gradient to the tangential stress and the second is the lithostatic contribution. The interface shear and normal stresses from the BEM results and the interface tangential stress from equation (2) can be used to determined the maximum principal stress (Timoshenko & Goodier 1934). The maximum principal stresses at 30-m fold amplitude are plotted in Fig. 16. There are two curves for each interface corresponding to the maximum principal stresses above and below
40
M. L. COOKE ET AL.
Fig. 14. (a) The slip distribution along uniformly distributed frictional interfaces at 30-m fold amplitude. The interface at 1740m depth has the greatest amount of top-to-the-left slip (approximately 25cm). The interface at 1680m depth has greater frictional slip than the 1800-m deep interface because of reduced lithostatic compression at the shallower depth. The point of greatest offset lies near the centre of each slip patch, producing an approximately symmetric slip distribution. The points of greatest offset along all the slip patches (at approximately 2400m) are located within the region of steepest dip along the monocline. The slip gradient for the interface at 1740m depth is greater than that at 1680 depth which in turn is greater than 1800m depth, (b) The slip along the interface at 1740m depth for fold amplitudes of 20-30 m in 2-m increments. The slip patch increases in length and magnitude with increased fold amplitude. The points of greatest offset along the slip patches (near 2400m) for all fold amplitudes are located within the steeply dipping region of the monocline.
the slip patch. The stresses above and below the interface are the same for regions of the interface which do not slip. The interfaces at depths 1680 and 1800 m have asymmetric stress curves along the layer as a result of flexure. Interfaces above the middle of the fold (e.g. 1680m depth) have greater maximum principal stresses in the anticlinal portion of the fold where the interfaces are within the stretched outer arc of the fold. Interfaces below the middle (e.g. 1800m depth) have greater maximum principal stress in the synclinal portion of the fold. The local stress perturbations along the interfaces, due to top-to-the-left sense of slip, increase the stresses within the bottom left and top right portions of the slip patch (Fig. 16). Stresses are most tensile along the bottom left portion of interfaces located above the middle of the layer; joints are most likely to
develop within these regions. Increasing slip gradient produces increasing concentration of stresses; the interface with the largest slip gradient (1740m depth) has the largest stress perturbation (approximately lOMPa) (Fig. 16). The stress perturbations are less than the magnitude of the lithostatic load (approximately 35 MPa) and do not result in the tensile maximum principal stress required for joint initiation. Therefore, joints will not develop along frictionally slipping interfaces under the modelled conditions. Frictional slip with sampled dune boundary distribution Numerical modelling of the observed dune boundary distribution requires a large number
NUMERICAL MODELLING OF JOINT LOCALIZATION
41
Fig. 15. Maximum offset along each interface for fold amplitudes 20-30 m in 2m increments, (a) The maximum slip for the uniform distribution of interfaces increases with fold amplitude. At each fold amplitude, the middle interface has the greatest slip and the slip on the deepest interface is less than that on the shallowest interface, (b) The maximum slip for the non-uniform distribution of interfaces increases with fold amplitude. At fold amplitudes of less than 28m, the interfaces from depths of 1645 to 1765m slip and the interface at 1725-m deep has the greatest offset. At fold amplitudes of 28 and 30m the interface at a depth of 1832.5m slips and the 1765-m deep interface has the greatest offset.
of boundary elements. The minimum dune thickness, 5 m (Fig. 9), controls the maximum element size. Simulating the observed mechanical stratigraphy at 5-m element lengths will require greater computing power than available at this time. In order to work within the limits of available computing power, the observed distribution is sampled to produce a similar distribution with
fewer dune boundaries. The thicknesses of successive dune boundaries were added producing a distribution with half the number of interfaces. The resulting distribution resembles the observed distribution with greater layer thicknesses (Fig. 17). The greater thickness increases the magnitude of curvature-related stresses in each layer but does not change the spatial pattern of
42
M. L. COOKE ET AL.
Fig. 16. The maximum principal stress along uniformly distributed frictionally slipping interfaces at 30-m fold amplitude. Solid and dashed curves represent the stresses along the bottom and top of the interface, respectively. The curves are coincident along the interfacial regions that have not slipped. Within the slip patches, the maximum principal stress is discontinuous, resulting in local stress perturbations. Along the slipped regions (top-to-the-left sense of slip), the local stresses increase along the bottom left and top right of the slip patches. The interface with the greatest slip gradient, 1740m depth, has the largest stress perturbation (approximately lOMPa). Increasing the slip gradient increases the stress concentration. These local stress increases are not great enough to produce tensile stresses or initiate splay cracks.
stresses. From the base of the Navajo up-section the dune thickness increases. There is then a sharp decrease and then a more gradual increase in dune thickness.
Fig. 17. (a) The observed and sampled distributions of dune boundaries through the thickness of the Navajo formation. The computation power required to model the observed distribution is greater than that available at this time. In order to reduce the number of elements necessary to model the Navajo Formation we sampled the distribution with half as many interdunes. (b) Schematic stratigraphic columns show that the sampled distribution has thicker layers than the observed distribution.
In the numerical experiment the monocline is monotonically flexed in 2-m fold amplitude increments to 30m total throw. Slip initiates at 20-m fold amplitude along the frictional interfaces at depths of 1765, 1725, 1700 and 1680m. These interfaces are located at the centre of the fold, where layers are most steeply dipping, and above the middle of the layer (Fig. 18). Within increasing deflection the length of each slip patch increases. At 24-m fold amplitude, the interface at 1645m depth slips, and at 28-m fold amplitude the 1832.5-m deep interface slips. At less than 28 m of monoclinal deflection, all of the frictional slip occurs among the thin layers between depths of 1645 and 1765m; contacts between thicker layers above and below this region do not slip at these amplitudes. Similar to the uniformly distributed interface model, the region of greatest offset along each of the interfaces (approximately 2400m) lies within the area of largest dip along the monocline (Fig. 19). However, the variation in layer thickness in this numerical experiment influences deformation and the slip patch length does not correlate simply with slip magnitude. For example, at 30-m fold amplitude the interface at 1725m depth has the longest slip patch, and the interface at 1765m depth has the greatest
NUMERICAL MODELLING OF JOINT LOCALIZATION
43
Fig. 18. Locations of interface slip for non-uniform interface distribution and fold amplitudes of 20-30 m at 2m increments, (a) Slip first develops at 20-m fold amplitude along the frictional interfaces at depths of 1765, 1725, 1700 and 1680m. Within increasing fold amplitude the length of slip patches increases along each layer. At 24-m fold amplitude, the interface at 1645m depth starts to slip, and at 28-m fold amplitude the 1832.5-m deep interface slips. The slipped interfaces lie along the steep limb of the fold and near the middle of the layer. The interfaces with the longest slip patches are those among the thin layers between two areas of thicker layers, (b) The region of frictional slip within a cross-section of the monocline at various fold amplitudes. offset (Fig. 19). Whereas the maximum offset along interfaces shallower than 1725 m decreases gradually with decreasing depth, there is a large change in slip magnitude between the interfaces
at depths of 1832.5 and 1765m (Figs 15b and 19). These results differ from those for the model with uniform layer thicknesses and indicate that layer thickness variations influence
Fig. 19. The slip distribution along non-uniformly distributed frictional interfaces at a 30-m fold amplitude. The interface at a depth of 1725m has the longest slip patch, while the 1765m deep interface has the greatest offset (approximately 15cm). The slip gradient along the interface at a depth of 1765m is greater than the other interfaces. The centre of the slip patches (approximately 2400m) lies within the region of steepest dip along the monoclines.
44
M. L. COOKE ET AL.
the slip pattern; mechanical stratigraphy controls distribution of strain. The results from the model with uniform interface distribution suggest that slip magnitude increases with proximity to the middle of the layer and decreasing depth; this does not strictly hold for the model with non-uniform layer thicknesses. At fold amplitudes of less than 28m the interface at 1725m has greater slip than the 1765-m deep interface; however, the difference between the two decreases as the fold grows (Fig. 15b). For fold amplitudes of 28 and 30m the 1765-m deep interface has the largest slip, even though it lies 10m farther from the middle of the layer and has greater lithostatic compression acting across it than the 1725-m deep interface. It is remarkable that the distribution of shear strain along interfaces is not similar as the fold grows; the location of maximum slip changes with fold amplitude. This shift in the maximum slip location is not expressed in the model with uniform distribution of interfaces and, therefore, must be related to the variation of layer thicknesses. The maximum principal stress along the top and bottom of the interface was examined to evaluate the potential for splay crack initiation. The results show that increased slip gradient produces greater local stress perturbations. Similar to the case of uniformly distributed interfaces, the stress perturbations are not great enough to overcome the overall compressive stress field due to lithostatic loading. Joints do not develop along the frictional interfaces under the modelled conditions. Discussion Fold asymmetry The East Kaibab Monocline at Hackberry Canyon is asymmetric with a broader anticline and tighter syncline than the approximated shape used in this study (Fig. 11). The increase in curvature at the trough of the syncline will promote the development of curvature-related joints in that region. During the incremental growth of the fold, we expect curvature-related joints to initiate first within the synclinal portion of the fold where the curvature is greatest. The influence of increased curvature is greater than the suppression of joint growth due to the increased lithostatic load. In the Buckskin Gulch section of the East Kaibab Monocline, the Navajo is exposed in the synclinal portion of the fold and there contains extensive jointing (Fig. 3) as well as contractional deformation features (Mollema 1994).
Fold asymmetry may be indicative of folding mechanism because there is asymmetry within the stress field around steeply dipping faults. Thrust faults result in greater vertical tension above the footwall than the hangingwall while, for normal faults, the relationship is reversed; in both cases, the synclinal hinge has greater vertical tension than the anticlinal hinge of the fold (Cooke & Pollard 1997). If a steeply dipping fault underlies and drove the folding of the East Kaibab Monocline, we would expect that the synclinal portion of the fold would have greater tension induced on horizontal planes and would have greater slip along dune boundaries. Slip along bedding planes acts to tighten fold hinges (Cooke & Pollard 1994). Thus, the tighter syncline observed at Hackberry Canyon may be due to asymmetry of the stress field around an underlying fault which increased interlayer slip within that region. Alternatively, the asymmetry of Hackberry Canyon may have developed in the later stages of folding and may not necessarily have been present during the initial folding stages investigated in this paper. As the fold amplitude increases, the anticlinal region of the fold is raised higher than the synclinal region. If there is erosion off the top of the anticline, the two regions of the fold will be subjected to different levels of lithostatic stress and may deform differently. Influence of non-uniform distribution of inter dune layers on slip pattern The frictional interfaces slip in similar locations, for both uniform and non-uniform layer thicknesses; slip occurs where limb dips are steepest and near the middle of the layer. For both models, most of the frictional slip occurs above the middle of the layer where lithostatic compression is reduced. The magnitude of slip along frictional interfaces differs between the models with uniformly and non-uniformly distributed interfaces. The model with uniform distribution has fewer slipping interfaces (three vs six) and greater maximum slip (25 vs 15cm) than the model with non-uniform distribution; increasing the number of available slip planes reduces the maximum slip. We may extrapolate from these results to the observed distribution of dune boundaries in the East Kaibab Monocline. Because this structure has more interfaces available for slip than the model, it would yield less slip on each of the sliding planes for a given fold amplitude. The results from the uniform interface distribution model suggest that slip magnitude
NUMERICAL MODELLING OF JOINT LOCALIZATION
increases with proximity to the middle of the layer and decreasing depth. These two factors act together to produce the observed pattern of slip in the uniformly distributed interface model. However, within the unevenly distributed interface model, variations in layer thickness may also contribute to the slip pattern. At the largest applied fold amplitude four interfaces accommodate shear strain in the form of inelastic frictional slip within the 100m above the middle of the layer (1640-1740m depth), whereas within the 100m below the middle of the layer (17401840m depth) only two frictional interfaces slip. Consider two sets of strata with relatively strong layers and weak interfaces, one has a few thick layers and one has many thinner layers. If the same amount of shear strain is applied to both sets, the strata with thicker layers will have greater interfacial slip than the strata with many thinner layers because the latter contains more interfaces along which to distribute shear strain. However, within the numerical experiments, the four interfaces above the middle of the layer have greater slip than the two interfaces below the middle of the layer. This may indicate that greater shear strain was applied to interfaces in the upper half of the layer and/or that increased lithostatic compression inhibited slip within the lower half of the flexed layer. We infer that the distribution of frictional slip among layers of the East Kaibab Monocline depends on variation in shear strain with depth, lithostatic compression and variation in the number of interfaces available for frictional slip (mechanical stratigraphy). These variables are interrelated since mechanical stratigraphy may control the distribution of shear strain. Within the numerical experiments, this effect was expressed as a change in slip pattern with fold amplitude. For fold amplitudes of less than 28m the maximum slip occurs above the middle of the fold; for fold amplitudes larger than 28m, the maximum slip occurs below the middle of the fold. The change in slip pattern reflects a change in the distribution of shear strain. For fold amplitudes of less than 28m, the region below 1765 m does not slip and the shear stresses are not reduced within the region. The thick layer between 1765 and 1832.5m depth stores shear strain energy by deforming elastically through its thickness. At 28-m fold amplitude, the interface underlying this stiff layer slips and the shear stresses and strains within the layer are reduced. This strain is transferred from the layer as frictional slip, which localizes shear strain along the interfaces. The change in shear strain distribution at fold amplitude higher
45
than 28m influences the pattern of slip and brings the location of maximum slip below the middle of the layer. Thus, the slip pattern may change during fold evolution and the mechanical stratigraphy may produce strain partitioning.
Nature of contact between layers Within this study we consider all the dune boundaries to have the same frictional properties. Field observations suggest that while most dune boundaries erode easily (Figs 4 and 5) and are weak in shear, some dune boundaries are highly cemented. In addition, interdune material varies in thickness (Figs 4 and 6). These factors may lead to frictional and stiffness material differences that would alter the resulting pattern of interface slip. For example, the map of dune boundary and splay crack locations within the Navajo (Fig. 9) shows a region without many splay cracks near the middle of the formation. Dune Boundary 9, which lies within this region, is thinner and more highly cemented than other interdune deposits in the Navajo (Fig. 6, Table 1). While the numerical results with homogenous layer contact properties show frictional slip throughout the middle of the flexed layer, it is unlikely that this highly cemented dune boundary easily slipped. The secondary pattern of dune boundary slip location and extent may be influenced by the heterogeneous nature of layer contacts.
Development of splay cracks The numerical experiments did not produce local stress concentrations near tips of frictional slip patches large enough to initiate fracturing. An increase in fold amplitude would increase the local stress perturbation and a decrease in depth would decrease the perturbation necessary to produce joints. Alternatively, other mechanisms may act to concentrate stresses, such as changes in properties of interdune material along the dune boundary. For the purposes of these numerical experiments we have assumed a constant friction coefficient along the length of the dune boundaries. This assumption does not consider variations in dune boundary planarity and interdune layer thickness that were observed at the East Kaibab Monocline. Irregularities along sliding dune boundaries may produce locally high slip gradients which concentrate stresses and promote joint initiation. A change in interdune layer thickness is observed along dune
46
M. L. COOKE ET AL.
boundary 16 in Fig. 4 As the canyon surface curves towards the viewer, the interdune thickness is reduced dramatically from that on the right side of the image. At the point of interdune thickness change there is a large joint cluster. As a material weak in shear thins, its effective friction increases (e.g. Scholz 1990). Such an increase in friction along a sliding interface will locally increase the slip gradient and concentrate stresses, which in turn promote joint initiation (Cooke 1997).
Influence of lithology Within the East Kaibab Monocline, fractures associated with interlayer slip are only observed in the exposed Navajo Formation. The Kayenta Formation exposed to the south of Hackberry Canyon has the same dip as the Navajo within the domain of bedding-plane slip at Hackberry Canyon yet displays no opening-mode fractures oblique to layers. The Kayenta contains relatively thick layers of fine-grained material. As the dip of layers is proportional to the shear stress on bedding, we would expect similar deformation within formations of similar dip. It is possible that the thick fine-grained layers between the sandstone layers of the Kayenta prevent the development of high slip gradients which promote development of fracture clusters. The thick finer-grained layers are able to deform internally and reduce stress concentration at irregularities along sandstone bedding. Within this study we have assumed that the lithologies above and below the Navajo Formation slip freely, producing negligible shear stresses. The Kayenta and Carmel Formations below and above the Navajo cannot reduce the shear tractions to zero because they are not frictionless materials. Non-zero shear tractions immediately along the upper and lower boundaries of the fold will increase the size of the region of interlayer slip. For the numerical experiments of this paper, the slip is limited to the centre of the layer and the outer interfaces do not slip because the shear stresses are reduced near the boundaries. If the shear tractions along the top and bottom surfaces of the layer are non-zero then the outer bedding-planes away from the centre of the layer are more likely to slip than presented in this study.
Implications of age relationships The numerical experiments suggest that curvature-related joints develop prior to frictional
slip along interdunes and associated fractures. Within the study region of the East Kaibab Monocline there are no outcrops that expose cross-cutting relationships between interdune faults of domain 2 and fold-parallel fractures of domain 4. The fold-parallel and bedding-perpendicular joints within the synclinal hinge of the East Kaibab Monocline have been inferred to be younger than bedding-plane faults because within domain 4 at Buckskin Gulch the joints are younger than deformation bands and at the edge of domains 2 and 3 at Hackberry Canyon some bedding-plane faults are older than deformation bands (Mollema 1994). These observations may not be conclusive evidence for age relations as the deformation bands at Hackberry Canyon may have formed at a different time to the deformation bands of domain 4 at Buckskin Gulch. The deformation bands between domains 2 and 3 at Hackberry Canyon occur near the exposed normal fault and may have formed during a later stage of folding related to the normal fault development. Both field evidence and numerical experiment support the following deformation evolution during fold development: deformation bands within domain 4; later curvature-related joints within domain 4; then interdune slip and associated jointing within domain 2; and, lastly, deformation bands in domain 2. This paper does not address the mechanical conditions of deformation band development. In order to fully understand the evolution of the East Kaibab Monocline, deformation band development should be studied in numerical experiments similar to those described in this paper. Conclusions We characterized and mapped the occurrence of joint clusters and bedding-plane slip within the Navajo Formation of the East Kaibab Monocline, Utah. Two types of joint clusters were documented: those perpendicular to bedding and parallel to the fold axis; and those parallel to the fold axis and oblique to bedding. The first occur predominantly within the synclinal hinge of the monocline, while the second occur along the steeply dipping limb of the fold and within the middle of the formation (Fig. 20). The fold-parallel and bedding-perpendicular joints are inferred to have formed by curvaturerelated stresses within the outer arc of the fold. Based on the numerical model, curvature-related joints are most likely to form in the anticlinal hinge of the symmetric fold where the lithostatic
NUMERICAL MODELLING OF JOINT
LOCALIZATION
47
and pattern of slip among layer interfaces. Knowing the mechanical stratigraphy of rock strata one can evaluate the contribution of both fold curvature and bedding-plane slip to joint localization. This knowledge will help in the interpretation of potential fracture clustering mechanisms and will lead to the development of a method for predicting subsurface fracture distributions associated with forced folding. This work was supported by DOE grant FG0394ER14462 and the Rock Fracture Project, an industrial affiliates program at Stanford University. Field assistants Juliet Crider and Gavin Bell assisted greatly in the mapping efforts. The manuscript was improved by comments of Drs Byron Kulander and Peter Cobbold. Fig. 20. Comparison of field observation and numerical results: curvature-related joints develop in the hinge of the syncline, whereas interlayer slip and associated joints occur within the steep limb of the monocline.
compression is least. However, within the East Kaibab Monocline the synclinal hinge is tighter than the anticlinal hinges, and we expect curvature-related joints to form within the trough of the syncline (Fig. 20). The field evidence and inferences drawn from the numerical results concur. Fold-parallel and bedding-oblique joints (splay cracks) are inferred to be evidence of interdune slip. For both evenly distributed frictional interfaces and the distribution based on the mapped dune boundaries, the region of interface slip occurs along the steeply dipping limb of the fold and in the middle of the layer (Fig. 20). These numerical results match the field evidence, which exhibit splay cracks in the middle of the Navajo and along the steep limb and part of the anticlinal hinge of the monocline. Investigation of the stress around slipped frictional interfaces shows that the concentration of stresses along uniform-friction slip patches is not great enough to produce joints at this depth. However, there will be greater slip concentration near heterogeneities along the layer contacts. For example, irregularities in dune boundary planarity and interdune thickness can both increase the friction coefficient locally and concentrate stresses during slip. This study illustrates the important role of bedding-plane slip during the flexure of multilayers. Slip along interlayers effectively decouples layers, allowing greater folding amplitude, and may produce clusters of joints. Variations in dune layer thicknesses influence the sequence
References BECKER, A. A. 1992. The Boundary Element Method in Engineering. McGraw-Hill, Cambridge. 1994. Bedding-plane slip over a pre-existing fault, an example: the Ramon Fault, Israel. Tectonophv5/^,230,91-104. BIENIAWSKI, Z. T. 1984. Rock Mechanics Design in Mining and Tunneling. Balkema, Rotterdam. BRACE, W. F. & BOMBOLAKIS, E. G. 1963. A note on brittle crack growth in compression. Journal of Geophysical Research, 68, 3709-3713. BROOKFIELD, M. E. 1977. The origin of bounding surfaces in ancient aeolian sandstones. Sedimentology, 24, 303-332. CHAPPLE, W. M. & SPANG, J. H. 1974. Significance of layer-parallel slip during folding of layered sedimentary rocks. Geological Society of America Bulletin, 85, 1523-1534. COOKE, M. 1997. Fracture localization along faults with spatially varying friction. Journal of Geophysical Research, 102, 22,425-22,434. & POLLARD, D. D. 1994. Development of bedding plane faults and fracture localization in a flexed multilyer: A numerical model. In: LAUBACH, N. A. (ed.) First North American Rock Mechanics Symposium, Balkema, Austin, Texas. & 1997. Bedding plane slip in initial stages of fault-related folding. Journal of Structural Geology, 19, 567-581. CROUCH, S. L. 1979. Computer simulation of mining in faulted ground. Journal of the South African Institute of Mining and Metallurgy, 79, 159-173. & STARFIELD, A. M. 1990. Boundary Element Methods in Solid Mechanics. Unwin Hyman, Boston. CRUIKSHANK, K. M. 1991. Analysis of minor fractures associated with joints and faulted joints. Journal of Structural Geology, 13, 865-886. & AYDIN, A. 1994. Role of fracture localization in arch formation, Arches national Park, Utah. Geological Society of America Bulletin, 106, 879-891.
48
M. L. COOKE ET AL.
DAVIS, G. H. 1978. Monocline fold pattern of the Colorado Plateau. Geological Societv of America Memoir 151, 215-233. ERDOGAN, F. & SIH, G. C. 1963. On the crack extension in plates under plane loading and transverse shear. ASME, Journal of Basic Engineering, 85, 519-527. FREUND, R. 1979. Progressive strain in beds of monoclinal flexures. Geology, 7, 269-271. FRYBERGER, S. G., KRYSTINIK, L. F. & SCHENK, C. J. 1990. Modern and Ancient Eolian Deposits; Petroleum Exploration and Production, SEPM, Denver, Co. GORHAM. F. D., JR, WOODWARD, L. A., CALLENDER, J. F. & GREER, A. R. 1979. Fractures in Cretaceous rocks from selected area of the San Juan basin, New Mexico - exploration implications. AAPG Bulletin 63, 598-607. GRANIER, T. 1985. Origin, damping, and pattern of development of faults in granite. Tectonics, 4, 721-737. HODGSON, R. A. 1965. Genetic and geometric relations between structures in basement and overlying sedimentary rocks, with examples from Colorado Plateau and Wyoming. AAPG Bulletin, 49, 935-949. HUNTOON. P. W. 1993. Influence of inherited precambrian basement structure on the localization and form of Laramide monoclines. Grand Canyon, Arizona. In: Schmidt, C. J., Chase, R. B. & Erslev, E. A. (ed.) Laramide Basement Deformation in the Rocky mountain Foreland of the Western United States, Geological Society of America, Special Paper. 280, 243-256. JACKSON, M. D. & POLLARD, D. D. 1990. Flexure and faulting of sedimentary host rocks during growth of igneous domes, Henry Mountains. Utah. Journal of Structural Geology, 12, 185-206. JAEGER, J. C. & COOK, N. G. W. 1979. Fundamentals of rock mechanics. Science Paperbacks, 18, 18. KOCH, F. G., JOHNSON, A. M. & POLLARD, D. D. 1981. Monoclinal bending of strata over laccolith intrusions. Tectonophysics, 74. T21-T31. KOCUREK, G. 1981. Significance of interdune deposits and bounding surfaces in aeolian dune sands. Sediment olog\\ 28, 753-780. LAWN, B. 1993. Fracture of Brittle Solids (2nd edn). Cambridge University Press, New York. LEWIS, H. & COUPLES, G. D. 1993. Production evidence for geological heterogeneities in the Anschutz Ranch East Field. Western USA. In: NORTH, C. P. & PROSSER, D. J. (eds) Characterization of Fluvial and Aeolian Reservoirs, Geological Society Special Publication, 73, 321-338. MARTEL, S. J., POLLARD, D. D. & SEGALL, P. 1988. Development of simple strike-slip fault zones, Mount Abbot quadrangle, Sierra Nevada, California. Geological Societv of America Bulletin, 100, 1451-1465. McTiGUE. D. F. & MEI. C. C. 1981. Gravity-induced stresses near topographies of small slope. Journal of Geophysical Research, 86, 9268-9278. MIDDLETON. L. T. & BLAKEY, R. C. 1980. Processes and controls on the intertonguing of the Kayenta and Navajo formations, northern Arizona; eolianfluvial interactions. In: BROOKFIELD, M. E. &
AHLBRANDT. T. S. (eds) llth International Association of Sedimentologists Congress. Eolian Sediments and Processes. Developments in Sedimentology, 38, 613-634. MOLLEMA, P. 1994. The Influence of Structural Position and Lithology on the Fracture Distribution in the East Kaibab Monocline, SE Utah: Implications for Fluid Flow Properties. MSc Thesis. Stanford University. & AYDIN. A. 1997. Fracture patterns and fault architecture in the East Kaibab Monocline. /;/: CLOSE. J. & CASEY. T. A. (eds) Natural Fracture Systems in the Southern Rockies, Four Corners Geological Society, 63-75. MURRAY, G. H., JR. 1968. Quantitative fracture study - Spanish pool. McKenzie County. North Dakota. AAPG Bulletin, 52, 57-65. NEMAT-NASSER, S. & HORII, H. 1982. Compressioninduced nonplanar crack extension with application to splitting, exfoliation, and rockburst. Journal of Geophysical Research, 87. 6805-6821. PETIT. J.-P. & MATTAUER. M. 1995. Palaeostress superposition deduced from mesoscale structures in limestone: the Matalles exposure. Languedoc. France. Journal of Structural Geologv, 17, 245256. POLLARD, D. D. & AYDIN, A. 1988. Progress in understanding jointing over the past century. Geological Society of America Bulletin, 100, 1181-1204. PRICE, N. J. & COSGROVE. J. W. 1990. Analysis of Geological Structures. Cambridge University Press. Cambridge. RAMSAY. J. G. & HUBER. M. I. 1987. The Techniques of Modern Structural Geology. Academic Press. London. RECHES, Z. & JOHNSON, A. 1978
NUMERICAL MODELLING OF JOINT LOCALIZATION 49 SEGALL, P. & POLLARD, D. D. 1983. Nucleation and growth of strike slip faults in granite. Journal of Geophysical Research, 88, 555-568. STOKES, W. L. 1968. Multiple parallel-truncation bedding planes - A feature of wind-deposited sandstone formations. Journal of Sedimentary Petrology,^ 510-515.
49
THOMAS, A. L. & POLLARD, D. D. 1993. The geometry of echelon fractures in rock: implications from laboratory and numerical experiments. Journal of Structural Geology, 15, 323-334. TIMOSHENKO, S. P. & GOODIER, J. N. 1934. Theory of Elasticity. McGraw-Hill, New York.
This page intentionally left blank
Differential compaction of interbedded sandstone and coal STEPHEN E. LAUBACH, DANIEL D. SCHULTZ-ELA & ROGER TYLER Bureau of Economic Geology, University of Texas at Austin, University Station Box X, Austin, TX 78713-7508, USA Abstract: Cretaceous and Tertiary coal beds in the western United States typically contain subvertical opening-mode fractures (cleat). However, closely spaced normal faults abruptly substitute for opening-mode fractures in coal beneath some sandstone lenses having blunt terminations. Differential forced-fold compaction of coal beds around and beneath lensshaped sandstone bodies accounts for such shifts in fracture style. Finite element modelling of coal deformation shows that shear stress is augmented in coal layers below abruptly tapering edges of sandstones lenses, favouring fault development, whereas under gradually tapering lenses shear stresses are not sufficiently enhanced to cause shifts in fracture style. Upper Cretaceous Mesaverde Group coal beds in southwest Wyoming have significant variations in fracture style over distances of a few to tens of metres. Because these faults have little or no porosity, the coal that contains them is likely to have low permeability compared to coal having typical (generally porous) opening-mode fractures. Thus, shifting fracture style may affect regional and local gas and water flow in coal beds.
This paper illustrates how fracture style in coal varies around subtle folds caused by differential compaction. Results have implications for the hydrology of coal beds, which commonly are important aquifers in which fluid is transmitted through fracture (cleat) networks. Fracture attributes in coal that are relevant to fluid flow may vary on regional and local scales to a greater extent than is commonly assumed. Syndepositional and early burial compaction is caused by sediment loading. In flood-basin settings, compaction is greatest in peat and mud, and least in channel belt sandstones. Sandstone deposition in any given vertical sequence may be concentrated over compacting flood basin and/or thin channel margin deposits, producing stacked, lenticular, sandstone-coal sequences. Shapes of sandstone lenses, or 'coal splits', vary from gradually to abruptly tapering, depending on depositional setting and history. During subsequent burial loading of interbedded sandstone and coal in late-stage compaction, relatively rigid sandstone lenses can distort adjacent coal beds, creating forced folds. Because most sandstones tend to taper gradually, compacted coal forms generally subtle, open folds. Most differential compaction occurs by ductile flow before consolidation. Accordingly, fractures develop in coal beds during late-stage compaction of a few per cent which may occur shortly after burial to depths of about 1000m or more. Observations of several coal-bearing sequences in the western United States show that gradually tapering sandstone lenses have little effect on coal fracture patterns. In contrast, blunt, rapidly tapering sandstone lenses having
abrupt lateral changes in proportions of rigid sandstone to coal are, in some cases, associated with a transition from opening-mode fractures to arrays of closely spaced normal faults. Mapping of fractures in coal beds shows that exposures of Upper Cretaceous Rock Springs Formation (Mesaverde Group) in the western United States (Fig. 1) represent an end-member of fracture patterns associated with coal seams that diverge around blunt lens-shaped sandstones (Tyler et al. 1991). In this paper we use finite element modelling to show how contrasts in fracture style - abrupt shifts from openingmode fractures to normal faults - can be accounted for by effects of late-stage compaction beneath lens-shaped sandstone bodies. In a reservoir setting the observed variations in fracture attributes might mean the difference between a conductive and a non-conductive fracture system. Geological setting The Upper Cretaceous Mesaverde Group is exposed around the Rock Springs Uplift, a Ntrending anticline within the Green River Basin of southwestern Wyoming and northwestern Colorado. The uplift is a broad early Tertiary (Laramide) structure that formed as a result of basement-involved deformation in a foreland setting (see references in Schmidt & Perry 1988). Mesaverde Group rocks on the uplift probably experienced maximum depths of burial during Late Cretaceous of less than -3000m.
From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 51-60. 1-86239-060-6/OO/S 15.00 :r The Geological Society of London 2000.
52
S. E. LAUBACH ET AL.
have one or more dominant sets. These are opening-mode fractures oriented normal to bedding, having spacing that ranges from less than 0.02cm apart to more than 50cm, depending on coal rank, lithotype and bed thickness. Calcite, and locally other minerals, partly fill fractures, but typically fracture porosity is visible in outcrop and core. Fracture age relative to that of other fractures in the same outcrop can commonly be determined from abutting or crossing relations, where younger fractures abut or cross older fractures. In Rock Springs Formation coal beds, the oldest and commonly the most prominent systematic fractures (face cleats) strike northeastward to east-northeastward (060-085°) (Laubach et al. 1993), but subsidiary fractures with more northerly strikes (045°) are common. Other systematic and non-systematic fracture sets are also present. Fractures in a forced fold Fig. 1. Rock Springs Uplift, Green River Basin, Wyoming, USA. Locality discussed in text is located at T19N, R104W in Rock Springs.
Differential compaction of sandstone and coal creates structures that can be interpreted as low-amplitude, open-drape folds. Relationships between sandstone and coal geometry and fracCampanian Rock Springs Formation is an ture style is particularly clear at an exposure in important coal-bearing stratigraphic unit in the Rock Springs, the Roosevelt School locality. Mesaverde Group. It was deposited in a series of Here a lens-shaped upper delta-plain fluvial wave-dominated deltas formed in an embayment channel or crevasse-splay deposit is interbedded of the western shoreline of a Cretaceous seaway. with several thin (less than 3m) coal beds. In The formation consists of interbedded sandstone, the coal bed directly beneath the sandstone, fracsiltstone, shale and coal, and is about 600 m thick. ture style shifts from closely spaced planar extenDepositional environments range from marine sion fractures to closely spaced curviplanar prodelta to marginal marine delta front, and normal faults over a distance of about 9m strandplain and non-marine lower delta plain. (Figs 2 and 3). Although areally (or volumetriCoal beds are interbedded with delta-plain and cally) such fault arrays are exceptional, other distributary-mouth-bar sandstones (Levey 1985). exposures in Rock Springs coals near sandstone Coal beds range from less than 1 m to more lenses show that these features are widespread. than 6m thick, but typically are less than 4.5m thick. In the subsurface, coal rank is high volatile A bituminous (A. R. Scott, pers. comm. 1992), Sandstone-coalbed geometry but rank in outcrops we studied is sub-bituminous (B. E. Law, pers. comm. 1992). The thickest coals Coal beds and sandstones at the Roosevelt were preserved landward and parallel to ancient School exposure are associated with upper shorelines. Although these coals are the most Rock Springs Formation coal seam No. 7 of laterally continuous deposits of the delta-plain Levey (1985). Bed dips are 9-21° to the southassemblage, they contain numerous lenticular west. Exposures within N- and NE-striking inacsandstone to sandy shale 'splits' that reflect tive excavation highwalls contain a lens-shaped deposition by channels and overbank splays. point-bar or crevasse-splay sandstone (Fig. 3). The sandstone lens is exposed for 137m and is about 12m at its maximum development, thinning to less than 1.5m on the northern end of Regional fracture patterns the exposure. Sediment transport was to the In the Green River Basin, coal beds commonly southeast as the shoreline and delta plain contain closely spaced fractures (cleats) that advanced southeastward and seaward (Levey
FOLDING OF COAL BY DIFFERENTIAL COMPACTION
Fig. 2. Cross-section, north limb of sandstone lens, in Rock Springs, Wyoming, USA, showing forced-fold anticline in coal bed, underlying coal and faults. Fold hinge is indicated.
1985). Therefore regional and local trends of distributary and alluvial channel-belt deposits are northwest-southeast in the Rock Springs Formation. The axis of the channel and local palaeoflow within the channel was southeastward, but the lens-shaped sandstone deposit tapers to the northwest, northeast and southwest; the thickest sandstone accumulation is on the southeastern side of the exposure. The lower part of the sandstone lens is composed of a massive medium sandstone that is as much as 2m thick and locally has load casts at its base. This unit is overlain by medium to fine sandstone, grey siltstone and, in its lower section, thin (less than 25cm) discontinuous coal and carbonaceous shale stringers that are overlain and partly truncated by a fine to medium sandstone of variable thickness, having prominent SW-dipping lateral accretion surfaces. The upper part of the sandstone lens has a planar contact with overlying fine sandstone and siltstone. The lens splits the seam into an upper and lower coal bed.
Fig. 3. Distribution of contrasting fracture style (opening-mode fractures vs faults) relative to channel trend and area of thick sandstone; attitude of one representative small fault.
53
The surface defined by the upper contact of the sandstone lens with the overlying coal bed is relatively planar, but over a distance of about 64 m the contact between the lower coal bed and the sandstone lens defines a curved surface having an abrupt convex-downward inflection where the sandstone lens thickens. This area of maximum inflection ('drape fold') is about 30m from the thickest sandstone accumulation and is about 26 m from the north end of the outcrop where thin sandstone overlies the coal bed (Fig. 2). As a result, the lower coal bed diverges from parallelism with the upper contact by 10-15°. Dip of the lower coal bed where folded under the channel is 21° southwestward. Discordance between upper sandstone stratification and the lower coal bed is greatest on the flanks of the sandstone lens, and is least beneath the thickest and thinnest parts of the sandstone lens. The termination of the sandstone lens is a relatively blunt one in the spectrum of lenticular sandstone shapes at coal splits. For Mesaverde coal 'splits' the height-to-length ratio of 0.3 represents a moderately to strongly divergent contact between channel sandstones and coal beds. The height-to-length ratio is the maximum sandstone thickness value, or thickness where sandstone is tabular with parallel-sided contacts, divided by the distance from this location to the position where the coal splits. It measures the shape of the sandstone termination rather than overall shape of the sandstone body, and our height-to-length ratio is commonly less than the overall aspect ratio of sandstone lenses depicted on regional cross-sections. We found no evidence that sandstone is incised into underlying coal. The coal bed below the sandstone is about 2 m thick where fully exposed under the northeast limb. However, the entire thickness of the coal bed is not exposed underneath the thickest part of the lens, so we could not document any decrease in its thickness. Decreases in coal bed thickness under deltaplain channel sandstones in the Rock Springs Formation were ascribed to compaction by Levey (1985). Throughout, the coal contains a strong foliation defined by flattened coal macerals; such fabrics are widespread in coal and are typically attributed to bedding-parallel flattening caused by compaction. We detected no changes in intensity of compaction fabric across the outcrop. The attitude of the compaction fabric conforms to the shape of overlying sandstone. The upper 0.6m of the coal bed is relatively bright, whereas the lower 1.2m is medium to dull and grades downward into dark carbonaceous shale, but along-strike the character of the coal appears constant. Coal
54
S. E. LAUBACH ET AL.
overlies a grey carbonaceous shale that is more than 2m thick along a planar contact. The inflection in the contact between lower coal and overlying sandstone lens does not extend into the underlying shale.
Fracture patterns In areas of parallel sandstone and coal contacts, cleats are orthogonal to bedding and compaction fabrics, and there is no evidence of slip parallel to fracture walls, consistent with opening-mode development. Fractures rarely cross coal-noncoal boundaries. The oldest fracture set (face cleat) based on cross-cutting relations, and a set of younger (butt) cleats have strikes of between 060° and 070°, and between 335° and 340°, respectively. A third, intermediate-age set of fractures striking between 020° and 040° abuts and is younger than the face cleat, but is older than the butt cleat. Fractures have vuggy, porous calcite mineral lining and locally preserved fracture porosity. There are no large map-scale faults that cut both sandstone and coal in this outcrop, except at the western end of the highwall where a subvertical NE-striking normal fault with about 1 m of displacement cuts across a stratigraphically higher sandstone. Where this fault cuts higher coal stringers there are no unusual coal fracture patterns associated with it. In the coal bed under the sandstone lens, opening-mode fractures of the face-cleat set are locally replaced by an array of closely spaced small normal faults with identical orientation (065°) and spacing (2.5-15 cm). Although opening-mode fractures of the face-cleat set are not present in areas where this fault array is present, intermediate and butt cleats have identical patterns in both areas (Figs 4 and 5). Faults have curviplanar slip surfaces that are mineral filled, with calcite having a growth-fibre microstructure striated in a down-dip direction. In contrast to calcite in cleats, calcite on fault surfaces has no visible porosity. The amount of displacement on faults ranges from indiscernible to 0.3 m, with most faults having displacement of less than 1.2cm. Closely spaced faults have the same cross-cutting and abutting relations to other fracture (cleat) sets, as do face cleats in other parts of the coal bed. Intermediate and butt cleat sets end against either face cleats or faults, depending on whether face cleats or faults are present. Because fault slip is small, it is clear that these are abutting relations rather than offsets of pre-existing fractures. The fault array extends laterally outward from beneath the thickest part of the sandstone lens
Fig. 4. Cross-section illustrating transition in fracture type. Lower-hemisphere equal-area projections show: (a) poles to opening-mode fractures (cleat) at a station north of sandstone lens, outside of faulted area; and (b) poles to faults at a station beneath lens. Note that cleats and faults have nearly identical strike.
for at least 64m. Beyond this an opening-mode fracture array exists; the transition occurs within an interval that is about 9m wide and more than 30m from the hinge of the forced fold in the coal bed (Fig. 4). In cross section, fault surfaces are moderately to highly curved. Many faults have sigmoidal shapes that have steep dips near upper and lower coal-non-coal contacts, and have shallow dips (as low as 45° to bedding) in the centre of the coal bed. Toward the upper contact of the coal bed, some faults have highly variable dips, ranging from vertical to 60° northwest. Although faults are curved in profile, in plan view fault traces are not notably curved. Under the northeast flank of the sandstone lens, faults dip predominantly to the southeast,
Fig 5. Faults in a coal bed north of the sandstone channel axis. Faults face south towards the thickest part of the sandstone lens, are strongly curved (see highlighted fault) and cut the subhorizontal compaction fabric at an angle less than 90°. Scale bar is 30cm.
FOLDING OF COAL BY DIFFERENTIAL COMPACTION
toward the thickest part of the overlying sandstone lens. The pattern is one of unidirectional fault facing. There also is a subsidiary, rare set of faults with dips to the northwest, but these are typically much smaller than SE-dipping faults. Faults underneath the thickest exposed part of the sandstone dip southeastward and northwestward, and have criss-crossing patterns. Faults in the coal bed under the southwest flank of the lens have both southeast and northeast dips. Most faults end at the coal-non-coal contact, but about 20% of the faults perceptibly offset the sandstone, and a few offset the contact by as much as 15cm. These faults die out in the sandstone within a few centimetres of the coal-sandstone contact. Fault dips in sandstone are commonly gentler than in coal. Where faults cross into sandstones, they are diffuse zones as much as 5 cm wide; there is no discrete fault surface or evidence of grain breakage. Such faults could have formed when the sandstone was relatively unconsolidated. These normal faults are unlikely to be cleats that have experienced later slip. The angle, in cross-section, at which faults intersect bedding or compaction fabric is the cutoff angle. These range from 40° to 90°, but are typically 60-85°. Adjacent faults and different parts of the same fault intersect the compaction fabric at different angles. Opening-mode fractures (face cleats), on the other hand, have cutoff angles that are consistently 90°. If faults were reactivated openingmode fractures, we would expect to see the 90° cutoff angle preserved, as is the case elsewhere in the Rock Springs Formation where later faults cut cleated coal and slip on pre-existing fractures has occurred. Abutting and crossing relations among fracture sets suggest that when earliest fractures (face cleats) were developing as opening-mode fractures, closely spaced faults were also forming nearby. Both fracture types formed after compaction fabrics but pre-date other cleat sets. The transition from face cleats to normal faults under sandstone lenses and the unidirectional dip of faults toward the thickest part of some sandstones suggests that sandstone shape influenced fracture style. Deformation beneath sandstone lenses In a typical coal seam, opening-mode fractures (cleats) are nearly perpendicular to bed boundaries and compaction fabric. They may be open or mineralized, but show negligible offset parallel to fracture walls. Ubiquity of cleats in coal
55
suggests that a process common to coal seams, such as extreme compaction and desiccation of peat (coalification) during burial relative to these processes in other sediments (for example, Ting 1977), is responsible for the formation of cleats and their close spacing. Depending on the time of formation of structures in coal seams, the large degree of compaction in coal (compaction ratios of as much as 10 or more (e.g. Ryer & Langer 1980; Elliott 1985; White 1986)) will affect how these fractures are modified by subsequent compaction. Thus, early formed structures such as clastic dykes (initially normal to bedding) are compressed into ptygmatic folds (Fearnsides 1916; Raistrick & Marshall 1939), and early steeply dipping faults are passively rotated to low dips (Elliott 1985). However, analysis of geometrical evolution of structures during compaction shows that final compaction fabrics will always appear nearly parallel to layer boundaries, and that normal faults could not have initiated as subvertical extension fractures that subsequently rotated into their present orientation. Sediment loading causes all or most early compaction; assuming uniaxial strain, volume loss accommodates all vertical shortening. If a sediment layer is horizontal during compaction, geometry of deformation will be pure shear, and material lines (e.g. clastic dykes) that are precisely vertical will remain so. However, even a small deviation from vertical orientation or pure shear results in a large passive rotation with progressive compaction. For example, for a compaction ratio of 10, a line originally dipping 85° rotates to a final dip of 48°. Material lines with moderate dips rotate to nearly horizontal. Compaction effects will be prominent even if compaction ratios are as low as 4 or 5. Deformation of clastic dykes into ptygmatic folds indicates that they formed early and near vertical, but opening-mode fractures in coal are nearly always planar and normal to bedding and are not visibly rotated or deformed, so they must have formed relatively late in compaction history. Compaction of a dipping layer, such as a coal bed beneath a sandstone lens, can be analysed in terms of transpression (Sanderson & Marchini 1984), where deformation is by a combination of movement parallel and perpendicular to its boundaries (simple shear and shortening). If shortening is accommodated by loss of volume rather than movement of material out of the plane, the system is non-isochoric plane strain. Kinematics of vertical compaction of a dipping layer impose a unique relationship between simple shear and shortening components of
56
S. E. LAUBACH ET AL.
Fig. 6. Diagram illustrating orientation of strain axes during compaction-induced transpression in a dipping layer. Strain during compaction can be resolved into pure shear and simple shear components, determined by layer dip (0) and compaction ratio (R); t, earlystage thickness; t/R, late-stage thickness. Ellipses show strain at an early and late stage of compaction. Maximum elongation orients slightly counterclockwise from horizontal initially, rotating toward parallelism with layer boundaries as compaction increases. Deformation in the left half of the body would be symmetrically related about a vertical centre line.
transpression (Fig. 6). The two-dimensional deformation tensor for plane strain is completely determined by bed dip and compaction ratio. This tensor can be used to determine deformation of any material line or to compute various measures of finite strain. Relations of use here are for change in orientation of a line and orientations of principal strain axes. In a horizontal reference frame with dips to the left positive, with R compaction ratio and 0 dip of compacting layer, deformed orientation of a line of initial orientation (p is thus
(modified from Sanderson & Marchini 1984, equation 4). Orientation of maximum principal strain axis (/?) relative to horizontal is
For large compaction ratios characteristic of coal and for different plausible origins, compaction fabric is always nearly parallel to layer boundaries. For example, fabric in a layer compacting 10-fold under a sandstone lens with a dip of 10° will be oriented less than 1° from layer boundaries. Therefore, regardless of reasonable variations of initial orientations of compaction fabric, final orientation will always nearly parallel the layer boundary. Thus, parallelism of fabric and layer boundaries does not constrain the origin of the fabric, but rather indicates the intense compaction of the sediment.
This type of kinematic analysis cannot explain the change from vertical extensional cleats in the horizontal coal layer to normal faults beneath the sandstone lens. The planar, vertical configuration of face cleats indicates that they formed fairly late in compaction history and have not been significantly modified by that deformation, although a fracture that was initially oriented precisely vertically would remain so after deformation. Is it possible that Mode I fractures formed slightly away from vertical, then rotated to the orientation of the observed normal faults? An extensional fracture in a homogeneous medium will form parallel to the maximum shortening axis for infinitesimal strain. Using a small value for compaction ratio R, such as 1.1, in equation (2) gives the orientation of the extension axis as 0.9°, so the shortening axis would dip about -89°. This direction is opposite to the dip of the bed; that is, away from the centre of the sandstone body. Compaction would only rotate it further in the direction opposite that observed for the normal faults. In general, the maximum shortening axis rotates from vertical to a maximum of 90° — 0 as R increases from 1 to infinity. If compaction is invoked to rotate initial fractures to the orientation of observed normal faults, fractures must originate with inclinations in the opposite direction from the maximum shortening axis, which is mechanically unlikely, and extreme compaction after fracture formation would be required. Inward-dipping lines rotate much more slowly away from near vertical than outward-dipping lines. A fracture initiating at 85° dip toward the lens centre requires a subsequent R value of about 8 to rotate it to a fault dip of 60°. Such a large compaction is unlikely if fractures form relatively late in compaction history. Predominant dip of faults towards the centre of the sandstone body indicates that most of them originated as faults with an initial dip in that direction.
Generation of normal faults Although kinematic analysis constrains possible rotation of fractures in various plausible orientations, it does not describe why normal faults replace cleats beneath the sandstone lens. Dynamic analysis explains fracture orientation and why normal faults form beneath the sandstone lens rather than typical openingmode fractures. Data discussed above show faults formed in the same deformation that created the cleats and compaction fabric; that is, late-stage compaction deformation.
FOLDING OF COAL BY DIFFERENTIAL COMPACTION
57
Fig. 7. Mesh of 2583 elements used for finite element simulation of sandstone lens compacting a uniform thickness coal layer. Coal layer, outlined by heavier lines, is sandwiched between stiffer sandstone layers. The mesh is most refined in the coal layer, where nearly all deformation occurs. Stress increasing with time is applied to the top of the model to simulate progressive compaction under an increasing sedimentary load. Geometry of the model is simplified from outcrop observations.
We have used a finite element model incorporating realistic rheologies to simulate stresses and strains in a compacting layer. The software GEOSIM-2D can simulate large strains for a variety of rheologies, including elastic and elasto-plastic with frictional yield, which is appropriate for rock types and conditions analysed here. Figure 7 shows the geometry of the model. Here we assume that structures in the coal and sandstone lens are symmetrical about a centre point. Therefore, the model only needs to simulate half of the body. The left half has been shown in Figs 2 and 4, but the right half is modelled numerically for convenience to make positive the distance from the centre point. A simulated coal layer somewhat thicker than the actual one allows better resolution of yielding zones. Coal splits with a variety of shapes are known, so we experimented in the simulation with several shapes for the sandstone termination. Results for the area beneath most of the sandstone lens are not sensitive to the exact shape of the sandstone termination. Elastic and frictional properties in the numerical model are within reported ranges for sandstone: Young's modulus 20 GPa, Poisson's ratio 0.25, density 2500 kg irT3 (Turcotte & Schubert 1982). Because the exact state of coal at the time of fracturing and faulting is unknown, the parameters used mainly reflect average values. To simulate compactibility of the coal, shear and Young's moduli are reduced to one tenth that of sandstone. A bulk modulus was chosen to make Poisson's ratio -0.1. This unusual negative ratio for coal simulates loss of volume during compaction. These elastic parameters are too stiff to represent the entire compaction process, but they allow compaction of 3% (R = 1.03) for a load representing about 1 km of overlying sediment. This range of shortening seems appropriate for simulating initiation of fractures if they formed late in the compaction process. Variation of rheological parameters in both
coal and sandstone had little effect on stress and strain patterns developed. For example, changing sandstone to an effectively rigid elastic body produced the same results. The only critical relationship is that the coal layer is elastically weaker than the sandstone. A downward-directed stress on the top surface elements of the model drives deformation. Analysis of the stress field during elastic deformation of coal suggests orientations in which shear fractures would form. Figure 8a shows principal stress axes at element centres at the edge of the simulated sandstone lens for an overburden load of about 1000m of sediment. Maximum compressive (longest) stress axis is close to vertical in the horizontal part of the layer, which combined with the small tensional horizontal stress would cause vertical Mode I fractures as observed in outcrop. The thickening sandstone lens reorients these vertical and horizontal stress axes, rotating them counterclockwise. Assuming a common friction angle of 30°, Mohr-Coulomb failure criteria predicts that normal faults should form at 30° from maximum compression axes. Deformation kinematics determine which facing direction is favoured for a finite deformation. Below the right half of the sandstone lens, transpression imposes a left-lateral shear. Because the left-dipping predicted normal faults have the same sense of shear, they should be favoured for continued movement. Conjugate faults dipping opposite to the layer dip would shear in a sense opposite to that of the transpression, so significant offset across them should be inhibited. Figure 8a shows two favoured failure trajectories for stresses beneath the sandstone lens. The orientation of the trajectories and steepening curvature near the layer boundaries accord with normal faults observed in outcrop. Opening-mode fractures forming parallel to the right-dipping compression axis would obviously be in the wrong orientation to evolve
58
S. E. LAUBACH ET AL.
Fig. 8. Finite element model results after compaction due to 1 km of overburden. Only the coal layer is shown, (a) Rotation of principal stress axes beneath the edge of the lens during elastic deformation. Short axes (shown heavier) are tensional. Predicted fault trajectories shown as grey lines, (b) Shear stress is high everywhere beneath the thick sandstone. Shear stress increases near the edge of the lens, (c) Equivalent plastic strain (strain due to non-recoverable plastic yield). Only areas that have reached the strain-weakening stage are shown. Leftdipping contours are consistent with failure on left-dipping faults.
into the left-dipping normal faults observed in outcrop. Principal strain axes (not shown) are oriented slightly counterclockwise from stress axes, reflecting non-coaxial strain accumulation in the zone of transpression. Strain axes are vertical in the horizontal part of the layer, consistent with formation of observed vertical cleats. Figure 8b suggests why normal faults instead of opening-mode fractures develop in the dipping part of the layer. Maximum shear stress is relatively constant in the horizontal layer, but increases substantially over a short distance at the edge of the sandstone lens and remains elevated beneath the entire lens. This increased shear stress favours failure on faults. To simulate strain weakening and localization as faults form, initial (plastic) failure of coal is followed by weakening from a friction angle of 31-23 as effective plastic strain increases from 0.01 to 0.02. Over the same strain interval,
cohesion drops from the initial 0.25kPa to 0. simulating destruction of cohesion as faults slip. Figure 8c displays contours of effective plastic strain in simulated coal, highlighting weakening failure zones growing at the boundary of the sandstone lens. These zones of shear failure dip at 60°, closely resembling failure on normal faults observed in outcrop. Zones shown are the only areas of plastic failure in the model to reach the weakening stage. The width of the localized shear zones in the model depends on the mesh size; decreasing the mesh size would correspondingly decrease the shear zone width until the shear zones would appear as planar discontinuities (faults) at this scale. Plastic strain in the simulated sandstone body is negligible at this stage. Failure zones in the coal do not propagate into the sandstone except in a small zone at the layer boundary. This behaviour is similar to that in outcrop.
FOLDING OF COAL BY DIFFERENTIAL COMPACTION where normal faults die out in sandstone within a few centimetres from the coal contact. For a model with double the load of Fig. 8a-c, dominant contour orientation of effective plastic strain is left-dipping, in the same direction as the dip of the layer. Zones of plastic failure would form in that direction, as observed in outcrop. Individual normal faults in outcrop are not seen in the model because of the lack of boundary perturbations and material inhomogeneities to localize failure zones, and the coarse mesh resolution. Plastic strain contours and failure direction in the horizontal part of the simulated coal layer are generally vertical.
Implications for fluid movement in fractures The coal exposures described in this paper illustrate that significant variations in coal-bed fracture style exist over short distances of tens of metres where typical opening-mode fracture (cleat) patterns are replaced by arrays of closely spaced faults. Because cleats and faults likely have contrasting properties as fluid conduits, transitions from cleats to faults can affect subsurface fluid flow in coal, where matrix permeability is low. For example, one study of enhanced fluid flow in cleats at simulated reservoir conditions shows permeability parallel to fractures at 1.7mD but only O.OOVmD through unfractured coal matrix (Gash 1992). In contrast, dosely spaced faults in Rock Springs coal beds have little or no primary porosity and have dense, syn-kinematic, calcite fill that is likely to impede fluid flow. Thus, shifting fracture style may affect regional and local gas and water flow in coal beds.
59
faults to a dip in the opposite direction to that observed. Overall, the system is that of a relatively rigid indenter impinging on a weaker plastic body. Because these faults are mineral filled and have little or no porosity, the coal that contains them will have low permeability compared to coal having mineral lined, but porous, cleats. Such shifts fracture style could affect fluid flow in coal beds. How widespread are shifts in fracture style from cleat to fault arrays? Most accounts of cleats in coal seams describe opening-mode fractures, but closely spaced fault arrays are also known (Kendall & Briggs 1933; Ammosov & Eremin 1963; McCulloch et al. 1974). There are many possible origins for these fault arrays. Compaction faults are associated with Tertiary sandstone lenses in the Powder River Basin of Wyoming (Law 1976; Law et al. 1991), and similar faults are locally exposed in Cretaceous coal beds in Colorado (Tremain et al. 1991). However, in coal beds under many lenticular sandstones in the western United States there is no difference in the style or type of fracture pattern. Coal beds interbedded with several broadly lens-shaped littoral Cretaceous sandstones have fracture patterns that have little or no contrast in orientation (dip) or style from those where beds are tabular, although size and frequency of fractures in these settings may vary (Tyler et al. 1991). Abrupt shifts in fracture type may only occur under lens-shaped sandstones when and where certain conditions are met, and the shape of the sandstone lens may be a decisive factor. Shear stresses may only be sufficiently augmented to produce normal faults where the sandstone lens has an abrupt, blunt, termination.
Conclusions Brittle failure in a coal seam changes from vertical opening-mode fractures, where sandstone and coal bed layers have parallel boundaries, to normal faults dipping in the same direction as the inclined base of the sandstone lens, a change that can be explained by kinematics and dynamics of coal differential compaction beneath a lens-shaped sandstone body. Shear stress increases in dipping parts of the coal layer, favouring development of normal faults instead of opening-mode fractures seen elsewhere. Normal faults could not have originated as opening-mode fractures because faults intersect the compaction fabric at a range of angles mostly less than 90°. Moreover, deformation beneath the indenting sandstone lens would have rotated
Acknowledgements We appreciate reviews of the paper by Stuart Dean, Terry Engelder and W. R. Kaiser. This study was funded by the Gas Research Institute under contract No. 5087-214-1544. GEOSIM2D was written and supported by EUROSIM Sari and is jointly owned by EUROSIM and Total C.F.P. References AMMOSOV, 1.1. & EREMIN, I. V. 1963. Fracturing in Coal (translated from Russian). IZDAT, Moscow (available from Office of Technical Services, Washington, DC).
60
S. E. LAUBACH ET AL.
ELLIOTT, R. E. 1985. Quantification of peat to coal compaction stages, based especially on phenomena in the East Pennine Coalfield, England. Mercian Geologist, 11, 63-172. FEARNSIDES, W. G. 1916. Some effects of earth-movement on the Coal Measures of the Sheffield district (South Yorkshire and the neighbouring parts of West Yorkshire, Derbyshire, and Nottinghamshire): Part 1. Transactions of the Institute of Mining Engineering, 50, 109-125 & 573-624. GASH, B. W. 1992. The effects of cleat orientation and confining pressure on cleat porosity, permeability, and relative permeability in coal. Log Analyst, 33(3), 176-177 (Abstract). KENDALL, P. F. & BRIGGS, H. 1933. The formation of rock joints and the cleat of coal. Proceedings of the Royal Society of Edinburgh, 53, 164-187. LAUBACH, S. E., SCHULTZ-ELA, D. D. & TYLER, R. 1993. Analysis of compaction effects on coal fracture patterns, Upper Cretaceous Rock Springs Formation, southwestern Wyoming. Mountain Geologist, 30, 95-110. LAW, B. E. 1976. Large-scale compaction structures in the coal-bearing Fort Union and Wasatch Formations, northeast Powder River basin, Wyoming. In: LANDON, R. B., CURRY, W. H., In, & RUNGE, J. S. (eds) Geology and Energy Resources of the Powder River. Wyoming Geological Association, Annual Field Conference Guidebook, 28, 221-229. , RICE, D. D. & FLORES, R. M. 1991. Coalbed gas accumulations in the Paleocene Fort Union Formation, Powder River Basin, Wyoming. In: SCHWOCHOW, S. (ed) Coalbed Methane of Western North America. Rocky Mountain Association of Geologists, 179-190. LEVEY, R. A. 1985. Depositional model for understanding geometry of Cretaceous coals: Major coal seams, Rock Springs Formation, Green River Basin, Wyoming. AAPG Bulletin, 69, 1359-1380. McCuLLOCH, C. M., DUEL, M. & JERAN, P. W. 1974. Cleat in Bituminous Coalbeds. US Bureau of Mines Report of Investigations, 7910.
RAISTRICK, A. & MARSHALL, C. E. 1939. The Nature and Origin of Coal and Coal Seams. English Universities Press, London. RYER, T. A. & LANGER, A. W. 1980. Thickness change involved in the peat-to-coal transformation for a bituminous coal of Cretaceous age in central Utah. Journal of Sedimentary Petrology, 50, 987992. SANDERSON, D. J. & MARCHINI, W. R. D. 1984. Transpression. Journal of Structural Geology, 6, 449458. SCHMIDT, C. J. & PERRY, W. J. (eds) 1988. Interaction of the Rocky Mountain Foreland and the Cordilleran Thrust Belt. Geological Society of America, Memoir, 171. TING, F. T. C. 1977. Origin and spacing of cleats in coal beds. Journal of Pressure Vessel Technology, 99, 624-626. TREMAIN, C. M., LAUBACH, S. E. & WHITEHEAD, N. H. 1991. Coal fracture (cleat) patterns in Upper Cretaceous Fruitland Formation, San Juan Basin, Colorado and New Mexico: implications for exploration and development. In: SCHWOCHOW, S. (ed.) Coalbed Methane of Western North America. Rocky Mountain Association of Geologists, 4959. TURCOTTE, D. L. & SCHUBERT, G. 1982. Geodynamics: Applications of Continuum Physics to Geological Problems. Wiley, New York. TYLER, R., LAUBACH, S. E. & AMBROSE, W. A. 1991. Effects of compaction on cleat characteristics: Preliminary observations. In AYERS, W. B., JR, KAISER, W. & LAUBACH, S. E. (eds) Geologic and Hydrologic Controls on the Occurrence and Producibility of Coalbed Methane, Fruitland Formation, San Juan Basin. University of Texas at Austin, Bureau of Economic Geology Topical Report prepared for Gas Research Institute (GRI-91/0072), 141-151. WHITE, J. M. 1986. Compaction of Wyodak coal, Powder River Basin, Wyoming, U.S.A. International Journal of Coal Geology, 6, 139-147.
Forced-fold development within Tertiary sediments of the Alba Field, UKCS: evidence of differential compaction and post-depositional sandstone remobilization JOHN W. COSGROVE1 & ROB. D. HILLIER2 Department of Geology, Imperial College of Science , Technology & Medicine, Prince Consort Road, London SW7 2BP, UK. (
[email protected]) 2 Amoco (UK) Exploration Company, Amoco House, West Gate, London W5 1XL, UK 1
Abstract: It is argued that the present trapping geometry within the Eocene Alba Field of the Outer Moray Firth, UKCS, has combined structural-stratigraphic elements. The structural component formed by significant topographic inversion of the deep-marine channel sandstones that constitute the field's reservoir. A study of core and three-dimensional (3D) seismic demonstrates that topographic inversion took place in response to two processes. The first was differential compaction of mudstone over the channel resulting in a forced fold. This structure was further enhanced by post-depositional sandstone fluidization and remobilization, the latter being focused toward the structural crest of the sandstone body which formed the core of the forced fold. The process of post-depositional remobilization caused significant sand injection into the mudstone overburden and the formation of seismic-scale dykes at the margins of the channel fairway. The growth of the forced fold (by sand remobilization) and the position of the peripheral dykes is analogous to that of igneous intrusions, specifically laccoliths.
The exploration for hydrocarbon traps within Tertiary sediments of the North Sea has proved to be a highly successful venture (Bain 1993). Many of the discovered fields have their trapping mechanism attributed to dip-closure resulting from drape, or differential compaction over deeper structural features such as Mesozoic horst blocks or salt domes (Parsley 1990). Traps in the Tertiary have also formed by differential compaction of mudstones around similar aged sandstone-rich deep-marine channel-fan complexes, for example the Gannet East Field which is reservoired within the Eocene Tay Sandstone Member (Armstrong 1987). Another example where differential compaction is thought to have been instrumental in the generation of the present-day trapping configuration is the Eocene Alba Field, block 16/26 of the Outer Moray Firth (Fig. 1). In this paper we present and interpret seismic data over the down-dip aquifer portion of the Alba Field, together with observations from core material that indicates that, in addition to differential compaction, another mechanism, namely post-depositional sand remobilization, may, at least in part, be responsible for Alba's trap configuration. Existing geological models for the Alba Field trapping geometry The mid to late Eocene Brodie Sandstone was deposited in a predominantly mudstone-prone
environment as a series of narrow, discontinuous 'shoestring' channels within bathyal water depths (Mattingly & Bretthauer 1992; Newton & Flanagan 1993). One such channel comprises the reservoir at the Alba Field. Reservoir quality is exceptional, with an average porosity of 35%, permeability approaching 5 darcys (D), and a high net to gross ratio which approaches unity in certain areas (Fig. 2). Two main facies have been identified by Newton & Flanagan (1993), the dominant being the Bedded Sandstone Facies. It exhibits a 'boxcar' log motif and the sandstone contains sedimentary structures such as dish and pillar structures with rare tractional features. This facies is thought to have been deposited by high-density turbidites. The second facies are the Injected Sandstones, comprising sedimentary dykes and sills thought to have been sourced from the main Bedded Sandstone Facies (Fig. 2). The Alba Field was initially described as a submarine channel system bounded by mudstone levees forming a topographic high on the sea-floor, as portrayed in Fig. 3a (Harding et al 1990; Mattingly & Bretthauer 1992). Following subsequent appraisal drilling and the acquisition of 3D seismic data, this model was challenged by Newton & Flanagan (1993), who proposed a more viable alternative. They suggest that the Alba reservoir was deposited within a pre-existing erosional scour, this representing a significant topographic low during the time of deposition (Fig. 3b). This
From'. COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 61-71. 1-86239-060-6/OO/S 15.00 © The Geological Society of London 2000.
62
J. W. COSGROVE & R.. D. HILLIER
Fig. 1. Location map showing the Alba Field sandstone channel outline, with well and seismic lines referred to in the text.
FOLDING BY DIFFERENTIAL COMPACTION
63
Fig. 2. Late Eocene Brodie Sandstone Reservoir, in the Alba Field well 16/26-8, with core coverage and interpreted sedimentary facies. Note the high net to gross ratio of the Bedded Sandstone Facies, and thin bedded nature of the Injected Sandstone Facies. Depth in metres below Kelly bush.
scour, or channel, constrained the deposition of the sand-prone flows to such a degree that very little 'over-bank' sandstone has been discovered by appraisal drilling off the main channel axis. The incised nature of the channel sandstone within the surrounding Horda Formation mudstone supplies the lateral stratigraphic trapping
element to the Alba Field. Whatever the exact depositional architectural of the Alba channel was, all previous authors accept that the present-day structural component to trapping formed as a result of differential compaction of the surrounding Horda Formation mudstone over the NW-SE-trending sandstone channel
64
J. W. COSGROVE & R.. D. HILLIER
Fig. 3. Differing models explaining the deposition and post-depositional origin of the Alba Field, (a) The channel-levee complex as described initially by Harding et al. (1990) and Mattingly & Bretthauer (1992). In this model, the Alba channel was a high on the depositional sea-floor, (b) The erosional scour model of Newton and Flanagan (1993). Here the Alba channel is perceived to have been constrained within a pre-existing erosional gully, (c) The presentday geometry of Alba, believed by previous authors to be the result of significant differential compaction of the overlying mudstone during burial.
axis (Fig. 3c). If the incised channel model of Newton & Flanagan (1993) is indeed correct, inversion of the depositional topography must have occurred, as the depositional low channel axis is presently a structural high feature. This topographic inversion has resulted in a southeasterly plunging, near symmetrical, anticlinal nose along the axis of the sandstone channel. Seismic expression of the Alba Field topographic inversion The seismic expression of this topographic inversion is illustrated in Fig. 4, a series of east-west lines taken from a recently acquired 3D data set over the down-dip, aquifer portion of the Alba channel. Within Fig. 4a, the dark blue horizon represents a regional, well-tied maximum flooding surface termed the 'Blue Marker' (Harding et al. 1990, see Fig. 2). Above this surface is a convex-up zone of high seismic amplitude that corresponds to 51m of sandstone-bearing interval in the nearby well
22/1 b-6. The interpreted well-tied top sandstone surface is the pink horizon, and the base of the sandstone rests close to the Blue Marker. This sandstone unit represents the main Alba Field channel system. The red horizon marks the top of the Eocene section. The near symmetrical, convex-up, mounded nature of the sand body is common across the Alba Field channel. Above the channel sandstone it is evident that the overburden has been deformed in a variety of ways. First, the convex-up nature of the top sandstone surface is mirrored at the Top Eocene reflector, suggesting that bending of the overburden has taken place over the channel sandstone. Secondly, it is apparent that the whole Tertiary section has been deformed by an array of normal faults which often form in conjugate sets. As is discussed later in the paper, there is considerable evidence for the occurrence of hydraulic fracturing during the evolution of the Alba Field, and these faults are thought to have been caused by such fracturing related to catastrophic, probably transient pore-fluid escape, as suggested elsewhere in the Central North Sea Tertiary by Cartwright (19940,6). It is evident from isopach analysis that mudstone thickness above the main channel axis is approximately 85% that of the equivalent age section off-channel, implying that the latter has undergone less compaction. Therefore some differential compaction over the main sandstone channel must have taken place. It is interesting to note, however, that isopach variances also occur in zones thought to be sandstone free. Such zones commonly exhibit a structural high at the Top Eocene reflector and the structures defined by these isopach variations were drilled by some explorers in the belief that they were caused by differential compaction over channel sandstones. Only after numerous exploration failures, together with the commonplace acquisition of 3D seismic for exploration, was this flawed philosophy exposed by Cartwright (19946) who argued that the structure could be related to faulting and fault-block rotation.
Evidence of post-depositional sandstone remobilization Another mechanism that might generate or contribute towards the generation of the convex-up, mounded geometry of the Alba channel is that of large-scale post-depositional remobilization of the conventionally deposited sand body. This process has been advocated to explain the
FOLDING BY DIFFERENTIAL COMPACTION
65
Fig. 4. Seismic examples of the forced-fold development over the Alba channel. For location see Fig. 1; red horizon - Top Eocene; pink horizon - top channel sandstone; blue horizon - Blue Marker, (a) Note welldeveloped, convex-up nature to the channel sandstone, mirrored at the Top Eocene reflector. Note also the highamplitude, high-angle reflectors emanating from the channel margins along interpreted fault planes B-B', and the reflector offset by later faulting at C-C'. (b). Similar folded overburden to that in (a). Note the isopach thickness between the Top Eocene and the Blue Marker beneath point X, interpreted as the result of fault-block rotation.
66
J. W. COSGROVE & R.. D. HILLIER
Fig. 5. Conceptual model for the post-depositional formation of mounded sand bodies, as advocated by Brooke et al. (1995). (a). Gas charging of unconsolidated sands with minor topographic irregularities on their upper surface, (b). Sand remobilization and focused transportation toward the growing topographic high.
generation of Quaternary, gas-charged mounded sandstones from the Outer Moray Firth by Brooke et al. (1995). Within this conceptual model, the mounded sand body is formed by liquefaction and subsequent entrainment of unconsolidated sands by high-pressure gas charging. It is argued that sand entrainment and thus remobilization would be focused toward initial topographic highs, being driven by significant buoyancy effects (Fig. 5). This process would accentuate any pre-liquefaction topographic highs to the sand body, possibly developed through early differential compaction. As noted by Brooke et al. (1995), this process may be driven by other pore fluids, for example highpressure brines. Such post-depositional modification should be recognizable in core, and with this aim in mind we will now focus attention on core observations from the Alba Field. Cores from the main channel sandstone (the Bedded Sandstone facies) are either structureless or dominated by an array of dish and pillar structures. Evidence of tractional deposits are rare. The dishes are thin (2-4 mm thick) dark coloured, subhorizontal, flat to concave-upward
laminations delineated by fine streaks of argillaceous material. Beneath each argillaceous horizon is a zone of clean sandstone 3-5 mm thick (Fig. 6). Dish length varies with morphology: strongly concave-upward dishes (almost convolute laminations) are typically 3-4 cm long, whereas flatter, less deformed dishes often exceed the observed core width in length. The flatter dishes are seen to cut across and truncate the strongly concave dishes, suggesting several periods of dish development. In close association with the dish structures are small sandstone pillars. These originate from discrete clean sandstone horizons, and are vertical to subvertical discordant bodies of clean sandstone that clearly cross-cut preexisting dish structures and laminations. The pillars are up to 5cm in length and up to 1.5cm wide. Larger-scale, subvertical 'bedding' discordant pillars are also present, being several feet in length and up to 8 cm in width. Again, the latter clearly cross-cut earlier dishes, and appears to emanate from discrete source beds (Fig. 6). Although dish structures have long been attributed to post-depositional modification of sandstone bodies (Wentworth 1967), it was Lowe & Lopiccolo (1974) who first recognized the importance of pore-water escape to their formation. In essence, these authors argued that pore water followed horizontal flow paths beneath semi-permeable laminations. The clean sandstone zones beneath the dish structures were seen as localized horizontal, high-permeability pathways from which argillaceous material was removed or elutriated, to be ultimately deposited at the periphery of the fluidization fronts, or at localized low-permeability heterogeneities that formed barriers to fluid flow. Later, Lowe (1975) described these clean sandstone zones as 'horizontal fluidization channels'. Allen (1984) questioned this interpretation, suggesting an alternative process that relied on sediment stoping. In this process, horizontal cavities formed within the sand after localized fluidization of the bed. The cavities would be water filled, having a suspension of fine-grained sediment entrained by the upward water flow through the underlying sand layer. It was argued that such a system is temporally unstable and that roof failure would result in a cohesive mass of sand sinking through the cavity. This sinking mass would 'filter' fines from solution, resulting in a concentration of fines forming at the base of the collapsing roof. Fluid drag would be greatest at the margins of the mass, resulting in an upturning at the sides of the sinking unit. This process explains the formation of key features exhibited by dish structures, notably the textural grading and the upward concavity of the dish.
FOLDING BY DIFFERENTIAL COMPACTION
67
Fig. 6. Oil-stained core from 16/26-16 (1991.8-1994.6m MD, top is to the left). Note the well-developed flat 'bedded' dish structures (A), being strongly concave-up (B) in association with large discordant sandstone pillars (C).
Tsuji & Miyata (1987) verified this process experimentally, expanding our understanding of the development of dish morphology. In their experiments, natural sand was fluidized by a mixture of water and clay suspension, with void growth and subsequent collapse producing clay-rich dishes. The close association of dish and pillar structures was also explained in these experiments when it was observed that discrete pillars were formed as fluidized sediment was injected into fracture dilations associated with roof collapse. Pillars are seen as vertical to subvertical flow pathways that may feed successive dish developments, a phenomena earlier described by Lowe & Lopiccolo (1974). It is clear that multiple phases of fluid flow would lead to the obliteration of any primary stratification, a feature characteristic of many channel sands containing dish and pillar structures.
Lowe (1975) indicates that fine-grained sands tend to have a greater susceptibility to fluidization due to their lack of cohesion, the extremely low mass of individual grains and a low frictional resistance to fluid drag. It is probable that fluid escape through the sand body was not one single event, but occurred as discrete pulses of fluid movement. Evidence for this comes from the discordant relationships of both dish structures and pillars, with highly distorted concave-upward dishes clearly predating later cross-cutting large-scale pillars. Mattingly & Bretthauer (1992) interpret two main phases of fluidization and deformation. The first occurred close to the sediment-water interface and was associated with the rapid deposition of sediment from later high-density gravity flows. These early formed dishes are associated with small-scale pillars. Later, a
68
J. W. COSGROVE & R.. D. HILLIER
Fig. 7. Reconstruction seismic line along the axis of the Alba Sandstone channel (see Fig. 1 for location). Note the high-amplitude reflector associated with fault A-A7 at the abrupt down-dip channel termination, and almost monoclinal bending of the overburden at this point. Red horizon - Top Eocene; pink horizon - top channel sandstone; blue horizon - Blue Marker.
FOLDING BY DIFFERENTIAL COMPACTION
second, more violent, phase of sediment deformation formed the large-scale pipes, and it is during this second phase that the present authors argue that large-scale sand remobilization occurs which has the effect of increasing the amplitude of the sand body and contributing to the formation of the present-day mounded trap geometry. It is probable that this latter phase is also responsible for the bulk of the small-scale sandstone dykes and sills (the Injected Sandstone facies, Fig. 2) hosted within surrounding mudstones. Of particular relevance to this suggestion that significant remobilization of sand occurs during the burial and compaction of the sand body are the two high seismic amplitude, high-angle, conjugate reflectors dipping at approximately 65° and emanating from the margins of the sandstone channel (Fig. 4a and b). These reflectors are seen to be structurally discordant to the well-bedded off-channel mudstone stratigraphy, and dim at approximately 2.05s TWT. When traced further up-section, these zones of planar, stratigraphic offset, continue suggesting that the reflections are related to faulting at the margins of sandstone channel. A similar high-amplitude feature is seen to emanate from the abrupt down-dip termination of the Alba Sandstone channel (Fig. 7). Note that the high amplitude reflectors are only present within the fault zones that are marginal to the sand body, and interpreted fault planes away from the sand body are typically dim (Figs 4 and 7). We propose that the high-amplitude reflectors represent significant volumes of sand that have been remobilized and injected into penecontemporaneous fault sets. The high seismic amplitude reflectors probably represent a seismic tuning effect resulting from the thin nature of the sandstone present. Depth conversion for the area suggests that the dyke lengths are approximately 440m, with true maximum dyke widths of 8 m. This evidence for the extensive remobilization of sand during burial and compaction supports the suggestion made earlier that the amplitude of the sand mound has been increased by the processes of fluidization. The positioning of the dykes at the margins of the sandstone channel may be analogous to peripheral dykes associated with some igneous intrusions, specifically laccoliths (Fig. 8a). Such peripheral dykes owe their origin to differences in longitudinal strain within the bending overburden above the laccolithic igneous intrusion (Pollard & Johnson 1973). Within such a forced fold the strains due to bending along the upper contact of an intrusion are contractional at the crest of the structure and extensional at the periphery (Fig. 8b). The formation of dykes
69
Fig. 8. (a) Analogous peripheral dykes at the margins of laccolithic intrusions (Price & Cosgrove 1990). (b) Differences in longitudinal strain within a flexed overburden above a laccolithic intrusion (Pollard & Johnson 1973). The strains are represented by the length of the arrows, and vary from being tensional at the periphery (X) to compressional at the crest (Y). Strains vary from being zero at the neutral surface to a maxima at the layer boundaries.
would be favoured at sites of maximum extensional strain, namely the zones of maximum bending in direct contact with the source magma. Such zones are encountered at the periphery of the laccolith or, by analogy, at the margins of the Alba Sandstone channel. The theoretical maximum intrusion of the dyke (assuming passive intrusion) would be to the point within the overburden termed the neutral surface (Fig. 8b). Above this point, the tensional stresses necessary for passive dyke emplacement change to that of compression, making intrusion difficult. If, indeed, the dykes do terminate at the neutral surface, and the dyke length of 440 m is taken as typical within the bending overburden, then the theoretical minimum amount of overburden at the time of injection would therefore be double the dyke length, that is 880m. It is uncertain whether the sandstone injection along the fault planes occurred during or after the large-scale hydraulic fracturing of the lowpermeability mudstones described by Cartwright (19940, b). Certainly, the high-cohesion, lowpermeability mudstones would deform in a different way during this basin-wide pore fluid expulsion compared to the high-permeability, low-cohesion sand body (Dott 1966). Whereas the mudstones, as a result of their high cohesion, would release overpressured pore fluids along
70
J. W. COSGROVE & R.. D. HILLIER
discrete hydraulic fractures, the sand body, because of its low cohesion, would fluidize in respond to high fluid pressures (Cosgrove 1995). This fluidization of the sand would permit injection of the entrained sand into the surrounding mudstone, with large-scale injection taking place at the margins of the sand body for reasons described above. Conclusions Similar observations of sand remobilization and clastic injection have been made in other Palaeogene deep-marine sediments of the North Sea, most notably in the Forth Field (Alexander et al. 1992; Dixon et al 1995), the Balder Field (Jenssen et al. 1993) and the Gryphon Field (Jaffri 1993; Newman et al 1993). These observations suggest that fluidization of low-cohesion sand bodies within the mudstone-dominated basin fill is common within the Tertiary burial history of the North Sea basin. It has been argued by previous authors that the presentday trapping geometries of many Paleogene fields is primarily the result of differential compaction during burial. The recognition of seismic-scale sandstone injections within the overburden above Alba Field suggests that sand remobilization has taken place on a larger scale than previously thought. The present study suggests that the trapping geometry in the Alba Field was topographically accentuated by the process of sand remobilization. Within this process, sand remobilization was focused toward the crest of the Alba channel, resulting in a convex-up, symmetrical forced fold within the overburden. The positioning of large-scale dykes was controlled by the stress regime within the flexed overburden, ie analogous to the peripheral dykes observed at the margins of many igneous intrusions, specifically laccoliths. Thus, the geometry of the forced folds that form over the sand channels within the Alba field is the combined result of compaction folding over original channel sands and the increase in amplitude of the sand body brought about by sand remobilization during burial. Acknowledgements The authors thank the P.653 partner group (Amoco, BG, Conoco, Idemitsu and Sands) for granting permission to publish this paper; and Chevron for the kind use of their core store, and granting permission to publish the core photograph from the Alba Field. We acknowledge
Anthony Fogg and Keith Martin for many fruitful discussions, and Tony Game for drafting the figures. Thanks also to two anonymous referees for their helpful suggestions and constructive criticism. References ALEXANDER, R. W. S.. SCHOFIELD, K. & WILLIAMS. M. C. 1992. Understanding the Eocene Reservoirs of the Forth Field, UKCS Block 9 23b. In: SPENCER. A. M. (ed.) Generation, Accumulation and Production of Europe's Hydrocarbons III. European Association of Petroleum Geoscientists. Special Publication, 3, 3-15. ALLEN, J. R. L. 1984. Sedimentary Structures, Their Character and Physical Basis. Unabridged Onevolume Edition. Developments in Sedimentology 30, Elsevier. Amsterdam. ARMSTRONG, L. A. 1987. The geology of the Gannet Fields, Central North Sea, UK sector. In: BROOKS. J. & GLENNIE, K. (eds) The Petroleum Geology of North West Europe. Graham and Trotman, London. 533-548. BAIN, J. S. 1993. Historical overview of exploration of Tertiary plays in the UK North Sea. In: PARKER. J. R. (ed.) Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference. Geological Society, London, 5-13. BROOKE, C. M., TRIMBLE. T. J. & MACKAY, T. A. 1995. Mounded shallow gas sands from the Quaternary of the North Sea: analogues for the formation of sand mounds in deep water Tertiary sediments. In: HARTLEY, A. J. & PROSSER, D. J. (eds) Characterization of Deep Marine Clastic Systems. Geological Society, London. Special Publications. 94. 95-101. CARTWRIGHT, J. A. 19940. Episodic basin-wide fluid expulsion from geopressured shale sequences in the North Sea Basin. Geology, 22, 447-450. 19946. Episodic basin-wide hydrofracturing of overpressured Early Cenozoic mudrock sequences in the North Sea Basin. Marine and Petroleum Geology, 11, 587-607. COSGROVE, J. W. 1995. The expression of hydraulic fracturing in rocks and sediments. In: AMEEN. M. S. (ed.) Eractography: Fracture Topography as a Tool in Fracture Mechanics and Stress Analysis. Geological Society, London, Special Publications, 92, 187-196. DIXON, R. J., SCHOFIELD, K., ANDERTON, R., REYNOLDS. A. D., ALEXANDER, R. W. S., WILLIAMS, M. C. & DAVIES. K. G. 1995. Sandstone diapirism and clastic intrusion in the Tertiary Submarine fans of the Bruce-Beryl Embayment. Quadrant 9, UKCS. In: HARTLEY, A. J., & PROSSER, D. J. (eds) Characterization of Deep Marine Clastic Systems. Geological Society, London. Special Publications. 94, 77-94. DOTT. R. H. JR. 1966. Cohesion and flow phenomena in clastic intrusions. AAPG Bulletin, 50. 610-611 (Abstract).
FOLDING BY DIFFERENTIAL COMPACTION HARDING, A. W., HUMPHREY, T. J., LATHAM, A., LUNSFORD, M. K. & STRIDER, M. H. 1990. Controls on Eocene submarine fan deposition in the Witch Ground Graben. In: HARDMAN, R. F. P. & BROOKS, J. (eds) Tectonic Events Responsible for Britain's Oil and Gas Reserves. Geological Society, London, Special Publications, 55, 353-367. JAFFRI, F. 1993. Cross-cutting Sandbodies of the Tertiary, Beryl Embayment, North Sea. PhD Thesis, University College, London. JENSSEN, A. I., BERGSLIEN, D., RYE-LARSEN, M. & LINDHOLM, R. M. 1993. Origin of complex mound geometry of Paleocene submarine-fan sandstone reservoirs Balder Field, Norway. In: PARKER, J. R. (ed.) Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference. Geological Society, London, 135-143. LOWE, D. R. 1975. Water escape structures in coarsegrained sediments. Sedimentology, 22, 157-204. & LOPICCOLO, R. D. 1974. The characteristics and origins of dish and pillar structures. Journal of Sedimentary Petrology, 44, 485-501. MATTINGLY, G. A. & BRETTHAUER, H. B. 1992. The Alba Field - a Middle Eocene deepwater channel system in the U.K. North Sea. In: HALBOUTY, M. T. (ed.) Giant Oil and Gas Fields of the Decade 19781988. American Association of Petroleum Geologists, Special Publication, 54, 297-305. NEWMAN, M. ST. J., REEDER, M. L., WOODRUFF, A. H. W. & HATTON, I. R. 1993. The geology of the
71
Gryphon Oil Field. In: PARKER, J. R. (ed.) Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference. Geological Society, London, 123-133. NEWTON, S. K. & FLANAGAN, K. P. 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. PARSLEY, A. J. 1990. North Sea hydrocarbon plays. In: Glennie, K. W. (ed.) Introduction to the Petroleum Geology of the North Sea. Blackwell, Oxford, 362388. POLLARD, D. D. & JOHNSON, A. M. 1973. Mechanics of growth of some laccolithic intrusions in the Henry Mountains, Utah, II. Bending and failure of overburden layers and sill formation. Tectonophvsics, 18,311-354. PRICE, N. J. & COSGROVE, J. W. 1990. Analysis of Geological Structures. Cambridge University Press, Cambridge. TSUJI, T. & MIYATA, Y. 1987. Fluidisation and liquefaction of sand beds - experimental study and examples from the Nichinan Group. Journal of the Geological Society of Japan, 93, 781-808. WENTWORTH, C. M., JR. 1967. Dish structure, a primary sedimentary structure in coarse turbidites. Bulletin American Association of Petroleum Geologists, 51, 485 (abstract).
This page intentionally left blank
Extensional forced folding and decollement of the pre-rift series along the Rhine graben and their influence on the geometry of the syn-rift sequences JEAN-CHRISTOPHE MAURIN 1 & BERTRAND NIVIERE 2 1
Centre Littoral de Geophysique, 17042 La Rochelle Universite, France 2 BRGM, BP 6009, 45060 Orleans, France (e-mail:
[email protected] [email protected])
Abstract: On the western margin of the Rhine graben, forced folding and decollement of competent Dogger strata occurred during the latest Eocene-Early Oligocene (Priabonian) as a result of normal faulting in the basement. The Dogger series are represented by NNEstriking thick layers of oolitic limestones strongly disrupted by extensional structures and incipient boudinage. To the east, the overlying Priabonian (syn-rift) sequences exhibit divergent onlaps, steeply dipping intraformational unconformities and a general drag syncline geometry. These syn-depositional structures attest to a close control by the progressive tilting and normal dip-slip sliding of the Dogger strata. Interpretation of a shallow seismic reflection profile reveals that this tilting is related to an underlying master normal fault which dips 60-70° to the east. This fault offsets the Variscan crystalline basement and the gently dipping Triassic cover by about 2 km. To be balanced, the structural arrangement requires the presence of a decollement layer located between the top of the Triassic series and the Dogger series, corresponding to the Late Triassic incompetent gypsiferous clays and marls. The response of the competent Dogger to the imposed forced folding is through the development of a large-scale extensional syncline in the hangingwall of the master detachment, compensated by the development of a piggy-back half-graben in the footwall. In the light of this new interpretation, a mechanical hypothesis is proposed including a two-stage evolution model of extension and subsidence for the Rhine graben.
Natural examples and field evidence of forced the Rhine graben by Maurin (1995). The folding within extensional regimes, are quite author identified folding of the pre-rift Mesozoic rare compared to the long-recognized cases in cover in the hangingwall of the master detachcompressional regimes (e.g. Heim 1922; 111 & ment fault of the graben, on the ECORS Work 1978; Palmquist 1978; Stearns & Stearns (Etude de la Croute Continentale et Oceanique 1978). For instance, such thin-skinned tectonics par Reflexion Sismique) depth-migrated seismic have seldom been described from any of the profile presented by Brun et al. (1991, 1992). great intracontinental rift systems. Recently, a The author also focused on the influence of foldvery good example was illustrated by Sharp ing on the geometry of the syn-rift sedimentary et al. (2000) along the Tertiary Rift of Suez, infill. He suggested that such deformations north of the Red Sea Rift. Another example could be accommodated by decollement and by from the Rhine graben, part of the Tertiary stretching within the pre-rift series in the footwall West European Rift System, is described in this of the master detachment fault, paper. In fact, this rift system provides several The aim of the present paper is to examine and examples of basins where such thin-skinned tec- balance all the thin-skin tectonic features affecttonics have previously been described, in the ing both pre-rift and syn-rift successions, related French Southeastern and Bresse basins (Bergerat to drag forced folding within both the hanginget al. 1990; Roure et al. 1992). Although these wall and the footwall of the master detachment authors described folds related to decollement fault of the Rhine graben, on the basis of the layers within the sedimentary pile, such as drag interpretation of a shallow seismic section and synclines, they did not develop any conceptual field observations, models to explain their origin and the relationship between tectonics and sedimentation. In the Rhine graben, drape folding over a basement Geology and tectonics of the Rhine graben: a fault has been suggested across the eastern brief review shoulder of the graben (Black Forest) by Laubscher (1982). Recently, drape folding was The Rhine graben is 300km long and 30-40 km also recognized along the western shoulder of wide. It trends approximately north-south in From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 73-86. 1-86239-060-6/OO/S 15.00 © The Geological Society of London 2000.
74
J.-C. MAURIN & B. NIVIERE
its northern part, N30 E in the central section and N10 E at its southern end. It is bounded to the east and west by the Black Forest and the Vosges massifs, respectively (Fig. 1). These features represent the shoulders of the rift where Variscan (Hercynian) basement is exposed. The rift cuts obliquely across a Variscan structure where two domains are separated by the major Variscan Lalaye-Lubine shear zone (Fig. 1). Soutn of the shear zone, the Variscan domain is formed mostly of high-grade polymetamorphic rocks which are intruded by Carboniferous granitoids. North of the shear zone, the Variscan domain is made up of low-grade metamorphic rocks and volcano-sedimentary sequences. The pre-rift Mesozoic sedimentary cover which overlies the Variscan basement, is more or less continuous from the Triassic to the Late Jurassic (Fig. 2). Mid-Triassic carbonate sediments overlie Lower Triassic sandstones. Gypsiferous marls end the Triassic series. The Lias is represented by a mainly marly sequence and carbonate facies developed during the Jurassic. The lack of Cretaceous and Early Cenozoic sediments is the result of a large-scale uplift of the area and the associated erosion, due to an early Alpine compressive event (lilies 1975; Ziegler 1992). The erosional unconformity is emphasized by the development of an Eocene karstic network within the Dogger limestones (Wittmann 1955). The first syn-rift sediments consist of both Late Eocene-Early Oligocene (Priabonian) alluvial and fluviatile sediments near the rift shoulders, and of lacustrine sediments elsewhere (Duringer 1988). During Middle-Late Oligocene (Rupelian), a shallow-marine transgression homogenized the sedimentation all over the Rhine graben area (Sittler 1969). Subsidence in the graben has remained active from the Late Eocene to the present in the northern part of the graben (Ziegler 1992), whereas it stopped in the south after a 1500m shoulder uplift of the Vosges and Black Forest which occurred from the Middle Miocene to the Pliocene (Villemin el al. 1986). Contemporaneously, the Kaiserstuhl volcanism occurred 17 Ma ago (Sittler 1969). Within the rift, the main structuration consists of a succession of four depocentres (Fig. 1) separated by transverse highs. These depocentres are known as the Mulhouse. Selestat. Strasbourg and Karlsruhe sub-basins, respectively. Field investigations and the shallow seismic interpretation presented in this paper concern mainly the western and northwestern edge of the Mulhouse sub-basin. A deep seismic reflection line (ECORS), shot across the central part of the Rhine graben (Fig. 1), shows a marked asymmetry of the rift
(Brun et al. 1991, 1992). The depth-migrated section of this seismic profile reveals the occurrence of an E-dipping master detachment fault along the western border of the Rhine graben (Fig. 3). The syn-rift Oligocene sedimentary infill wras controlled by approximately 3 km of vertical throw along this fault. The rift detachment persists to the base of the crust w-here a ductile shear zone seems to affect the crust-mantle boundary. The section shows that the depth of the Moho discontinuity decreases from the western Lorraine basin to the Rhine graben. an observation comparable with the conclusion of Edel et al. (1975). The estimation of the upper crust extension ranges from 5 to 7 km (Villemin et al. 1986). The subsurface interpretation of the ECORS deep seismic profile also reveals that block-tilting is directly related to this master fault, with dips reaching 10-15 . However, new accurate field investigations by Maurin (1995) show steeply tilted (up to 80 ) pre-rift and syn-rift strata within the hangingwall of the master fault. These shallow deformation patterns were not discernible on the ECORS profile. Their mechanical significance is examined here in the light of a shallow industry generated seismic profile and from several examples taken in the field.
Deformation in the hangingwall of the master detachment Interpretation of a seismic reflection profile The seismic profile shown in Fig. 4 was recorded a few kilometres southwest of the city of Colmar by Clyde Petroleum for ESSO. but its precise location remains confidential (Duringer 1988). The master basement normal fault is dipping about 60 to the east (Fig. 4). The hangingwall comprises: • syn-sedimentary syncline. filled by the Priabonian series, characterized by progressive unconformities, with a roll-over type deformation in the Rupelian series; • an incipient boudinage of the Dogger competent series which developed simultaneously with faulting in the basal Triassic series and in the basement; • below the Dogger series, the deformation is accommodated by thinning and thickening of the incompetent Lias marls series; • below the Lias series, faulting of the Triassic. and the Variscan basement, developed independently of the faults affecting the Lias-Dogger series, with the former exhibiting classical compensation faults of the master detachment.
EXTENSIONAL FORCED FOLDING IN THE RHINE GRABEN
Fig. 1. Structural map of the Rhine graben and its Variscan shoulders.
75
76
J.-C. MAURIN & B. NIVIERE
Fig. 2. Detailed lithostratigraphic column of the pre-rift Mesozoic series of the Rhine graben region.
These interpretations suggest that a progressive tilting of the Jurassic (Dogger-Lias) layers along the master fault occurred (in a reversedrag configuration) during the deposition of the Priabonian sediments. Indeed, this progressive tilting is marked out by progressive syn-sedimentary unconformities in the Priabonian infilling and by a characteristic drag-syncline geometry
of these deposits. We also observe a similar drag-syncline geometry of the lowermost Rupelian sediments (Fig. 4) suggesting a similar structural control during the earliest Rupelian times. Beneath the whole Tertiary series, the dragsyncline geometry is closely controlled by drag folding of both the Dogger and the Lias series.
EXTENSIONAL FORCED FOLDING IN THE RHINE GRABEN
77
Fig. 3. Interpreted section of the ECORS deep seismic reflection pr ile, after Brim et al. (1992) (see Fig.l for location).
This folding is accommodated by interna deformation (thinning and thickening) of the incompetent Lias marl sequences, while the Dogger competent limestones exhibit a brittle behaviour which produces incipient boudinage. This structural arrangement strongly suggests sliding and drag folding of the Jurassic series over the basement master detachment fault. Forced folding over the basement fault is also attested by the half-anticline that can be observed along the fault (Fig. 4). This half-anticline, affecting the Dogger, the Priabonian and the base of the Rupelian, is interpreted as a portion of an earlier forced anticline formed over the top of the master fault and which was offset later on by normal displacement along the master fault. The presence of a roll-over within the younger Rupelian sequences argues against it being the result of inversion tectonics. On the other hand, the roll-over type deformation of the Rupelian infilling cannot be accounted for by the same kind of structural control operating during the formation of the earlier drape folds. We suggest this is more likely to be related to the deflection of the master detachment fault along a decollement. Then, while the Priabonian and the base of the Rupelian sediments are associated with growth on an approximately planar fault, the younger Rupelian sediments are related to growth on a listric fault which produced the roll-over fold. In order to balance the fault profile according to the roll-over structure of the Rupelian sequences, we applied the construction method of listric normal fault proposed by Faure & Chermette (1989). Following this method (Fig. 5), for a horizontal reference surface with known vertical throw (/?) of the fault, one can construct the decollement-rooted normal fault profile at depth. If one assumes that the constant
displacement H (corresponding to the horizontal throw) of all points of the block base opened a potential void, the accommodation of this theoretical void implies that the point on the hangingwall reference surface moves along 60° dipping segments (a, b, c, d, e, f) which correspond to simple-shear slip lines. Then in the example illustrated in Fig. 5, the listric normal fault profile is rooted roughly at the depth of the very bottom of the Lias series or at the very top of the Triassic. This analysis also suggests a decollement of the Jurassic series along the master fault during the deposition of the Rupelian series.
Field interpretations Two-dimensional sections (Figs 7-9) summarize our main field observations made, respectively, a few kilometres west and west-northwest of the city of Colmar (Fig. 6). The three main lithostructural units comprise the Variscan granitic basement, its Mesozoic cover (Triassic, Lias, Dogger) and the overlying syn-rift sedimentary infill (Priabonian-Rupelian). A major N-trending normal fault dips at 60° towards the east and separates gently tilted Triassic series dipping 10° west from steeply tilted Lias and Dogger strata, dipping at about 60°, in the east. The competent oolitic Dogger limestones are strongly disrupted by small-scale normal faults and outcrops of these strata exhibit numerous slickensides which indicate top-to-the-east dipslip sliding of these units. These observations are consistent with incipient boudinage of the competent Dogger series observed on the shallow seismic profile. At the Florimont Hill (Fig. 7) Lias and Dogger layers are forced-folded above the top of the major N-trending basement fault. The layers
78
J.-C. MAURIN & B. NIVIERE
Fig. 4. (A) Shallow seismic reflection profile west of the Mulhouse basin and (B) structural interpretation.
EXTENSIONAL FORCED FOLDING IN THE RHINE GRABEN
79
Fig. 5. Reconstruction of the normal fault profile controlling the roll-over structure identified on the shallow seismic reflection profile (see text for explanation).
are flexed into a monoclinal shape, with both the gentle (30°) and the steep (80°) dipping limbs preserved. Poles of Jurassic bedding planes plotted on a lower-hemisphere Schmidt stereonet (Fig. 7) define a horizontal north-south fold axis whose trend is roughly parallel to the master fault. Poles of bedding planes of the Priabonian deposits (Fig. 7) also define a similar rotation axis which indicates that both the Jurassic and Priabonian series were folded and tilted as a consequence of motion on the master fault. Along the Letzenberg cliff (Fig. 8) the syn-rift Priabonian sequence exhibits divergent onlaps and steeply dipping intraformational unconformities, which attest to a close control of their sedimentation and orientation by the progressive tilting and normal-slip sliding of the underlying Dogger strata. As for the Florimont exposure, both the Jurassic and Priabonian bedding-plane poles (Fig. 8) indicate a progressive tilting with a N-S-trending rotation axis parallel to the master fault. Sedimentary sequence and flow analysis of the Priabonian infill was performed by Duringer (198§) and his results are summarized in Fig. 8. The sequence comprises an alluvial fan conglomerate containing 95% Dogger oolitic limestones pebbles. The sedimentary flow direction as indicated by pebble orientations, is trending to the east and the northeast towards the Rhine basin. This shows that the syn-rift sedimentary infill is closely controlled by the palaeomorphotectonic environment created by the master fault scarp along which the Dogger oolitic limestones were draped and tilted towards the basin. Thus, the field observations made in the Florimont and the Letzenberg areas strongly
suggest forced folding of the Jurassic layers as a result of normal faulting in the basement along the E-dipping master fault, concurrent with the deposition of Priabonian sediments. These observations corroborate the interpretations made for the shallow seismic profile. The problem now is to determine how stretching is accommodated between the competent Dogger limestones and the faulted basement. An insight into this problem can be obtained by looking at the footwall of the master detachment. Deformation in the footwall of the master detachment The structural section (Fig. 9) summarizes field observations made a few kilometres northeast of Bergheim city (Fig. 6). One observes, from west-northwest to east-southeast (Fig. 9), that Late Triassic gypsiferous marls constitute a N20°E-trending structural salient whose limbs dip 20° to the WNW and ESE. Syn-tectonic gypsiferous fibres provide numerous ductile shear kinematics indicators which are discussed later. The overlying Lias marls have been cut and gently tilted (20°) eastward by W-dipping normal faults. The relationship between the overlying Priabonian alluvial conglomerates, the Liassic marls and the Dogger oolitic limestone can be seen in Fig. 9. In the east, a steep W-dipping N20°E trending normal fault zone brings the Priabonian conglomerate into contact with Dogger oolitic limestones. The limestone, incised by conglomerate channels, is disrupted by the N20°E-trending normal faults network. The dip of the layers increases towards the east
80
J.-C. M A U R I N & B. N I V I E R E
Fig. 6. Detailed lithostructural map of the study area which is west and northwest of the city of Colmar.
EXTENSIONAL FORCED FOLDING IN THE RHINE GRABEN
81
Fig. 7. Structural profile constructed from field data across the Florimont area, west of Colmar (see Fig. 6 for location).
(from 10° to 25°) and then suddenly becomes horizontal. To the west the structure is dominated by gently E-dipping normal faults with an important horizontal throw compared to the conjugate steeply dipping normal fault network to the east (described above), which defines a domino system. This geometry suggests that the E-dipping normal faults are part of a listric fault system, while the conjugate domino system faults correspond to antithetic compensating faults, the whole leading to the formation of a classical asymmetrical half-graben structure. As
a consequence of the listric geometry of the fault, tilting of the layers increases westward (Fig. 9). This leads to the structural denudation of the Lias marls during which erosion removed most of the overlying Dogger oolitic limestones. Associated with the development of this faulting is the generation and the remobilization of the Priabonian alluvial fan system on a Lias substratum, and the build-up of sediments in the depocentre located at the west of the structure. An accurate sedimentary sequence and flow analyses of the section in Fig. 9 has been carried out by Duringer (1988). Here, as at Letzenberg,
82
J.-C. MAURIN & B. NIVIERE
Fig. 8. Structural profile constructed from field data across the Letzenberg area, west of Colmar (see Fig. 6 for location). Sedimentary sequence (a) and flow analysis (b) of the Priabonian infill after Duringer (1988). The sequence comprises an alluvial fan conglomerate containing 95% Dogger oolitic limestones pebbles.
EXTENSIONAL FORCED FOLDING IN THE RHINE GRABEN
83
Fig. 9. Structural profile constructed from field data west of Bergheim (see Fig. 6 for location). Sedimentary sequence (a) and flow analysis (b) of the Priabonian infill, after Duringer (1988).
the fan conglomerate contains more than 90% Dogger oolitic limestone pebbles. The sedimentary flow, as indicated by pebble orientation (see Fig. 9), trends to the west or to the northwest. These relationships suggest that this local Priabonian alluvial fan system filled a halfgraben developed in the footwall of the master Rhine graben fault and, unlike the fans described above from the Florimont and Letzenberg areas, was not connected to the main basin in the hangingwall. The authors have also been able to show that stretching of the cover, and the development of the asymmetrical half-graben in the footwall, is compensated by a decollement layer within the Late Triassic gypsiferous marls. As already mentioned, these rocks exhibit two kinds of kinematic indicators which are related to a ductile shear zone. Gypsiferous fibres, ductilely strained, show ductile shear bands. The stereographic net of the stretched mineral lineations (Fig. 9) shows a main east-west shear trend.
The shear intensity is maximum for a trend from N90° to N100°, i.e. where the fibre dip is minimum. Shear sense indicates top-to-the-east movement. Associated with these ductile shear bands are 'en echelon' gypsum-bearing tension gashes which present the same kinematics. The heterogeneous strain with a strong non-coaxial component observed at the top of the Trias is replaced near the surface by the gently E-dipping fault system which breaks up the competent Dogger layers. This confirms that the faults are listric and become horizontal at depth near the evaporite layer. The creep, which is observed in the gypsiferous marls, is associated with the listric faults and leads to! the gypsiferous marls structural salient (Fig. 9). The tectonic interpretation presented here shows an eastward decollement movement of Lias and Dogger series over a Late Triassic gypsiferous weak layer. It is assumed that such a decollement is accommodated by stretching of the Jurassic series as they slide eastward and
84
J.-C. MAURIN & B. NIVIERE
Fig. 10. Schematic section showing the structural pattern across both the hangingwall and the footwall of the western master detachment fault of the Rhine graben (see Fig. 6 for location).
become draped over the master detachment. Simultaneously, stretching and the decollement of the Jurassic series result in the formation of an a'symmetrical graben whose Priabonian sedimentary infill is not connected with the main depocentre, which developed to the east within the hangingwall of the master detachment. This structural arrangement is summarized in the schematic diagram shown in Fig. 10 which incorporates data from the Ostheim well (Blanalt 1972). Summary and conclusion Based on an interpretation of a shallow seismic reflection profile (Fig. 4) and field observations made in the Florimont and in the Letzenberg areas (Figs 7 and 8), forced folding is clearly evidenced in the hangingwall of the west master detachment fault of the Rhine graben. Folding affects both the Dogger and the Lias series. If is accommodated by internal deformation (thinning and thickening) of the incompetent Lias marl sequences which is quite obvious on the seismic line (Fig. 4), although it is not clearly identified on a mesoscopic scale in the field. In contrast, the competent Dogger limestones exhibit brittle behaviour on all scales as a result of folding. This is indicated by incipient boudinage on the seismic line (Fig. 4) and by numerous small-scale normal faults observed in the field (Figs 7 and 8). Folding does not affect the
Triassic series, which is faulted with the Variscan basement. From field observations made in the footwall of the master detachment around Bergheim (Fig. 9), the authors show that a decollement of the Jurassic series occurred on top of the Triassic. The decollement horizon corresponds to a weak ductile layer composed of Keuper gypsiferous marls. Non-coaxial internal deformation affects this layer and the resulting structures act as kinematic indicators indicating top-to-the-east simple shear. This non-coaxial deformation at depth is accommodated by block faulting of the overlying Jurassic series, which is strongly disrupted by an asymmetrical graben structure. The development of this extensional thin-skin structure was caused by the decollement and the eastward sliding of the Jurassic series as a result of forced folding on top of the master detachment fault. From a conceptual point of view, this thin-skin half-graben could be considered to be an extensional 'piggy-back' structure. According to the syn-sedimentary response to thin-skinned tectonics described above, a new reconstruction of the tectono-sedimentary evolution of the western border of the Rhine graben is proposed. This evolution comprises two stages (Fig. 11): during Late Eocene and earliest Oligocene times, subsidence in the hangingwall of the E-dipping master detachment fault is accommodated by antithetic faulting in the basement and by forced folding of the Jurassic cover. The syn-rift Priabonian and earliest Rupelian
EXTENSIONAL FORCED FOLDING IN THE RHINE GRABEN
85
Fig. 11. The two stages in the suggested reconstruction of the tectono-sedimentary evolution along the western master detachment fault of the Rhine graben (no scale).
sedimentary fill are progressively deformed with a syncline geometry which is controlled by progressive forced folding of the underlying Jurassic series. At the same time, a 'piggy-back' asymmetrical sub-basin is formed independently on the footwall of the master detachment fault by decollement of the Jurassic series on top of the Triassic; during the Early Oligocene, faulting propagates towards the surface and controls the growth of the Rupelian deposits which are affected by a roll-over type deformation. This
roll-over structure also attests that the decollement of the Jurassic series along the master fault is still active during the deposition of the Rupelian series. This new model of the fault mechanics along the western border of the Rhine graben should be tested along its eastern border where forced folding was also suggested in the area around Basel (see Fig. 1 for location) by Laubscher (1982). However, the problem in this area is a strong interference between two deformations,
86
J.-C. MAURIN & B. NIVIERE
the Oligocene normal faulting and the MioPliocene inversion associated with the Alpine thrust-front propagation from the Jura. In addition, lack of exposure makes it difficult to observe the relationships between the syn-rift sedimentation and the structures in the field. The only solution is to obtain information from industry derived shallow seismic profiles, which remain confidential. The present authors studied the eastern side of the Rhine graben immediately to the east of the area of Colmar described in this paper (Fig. 1). No field evidence for forced folding was found. Study of the ECORS deep seismic profile (Fig. 3) by Brun el al. (1992) shows that this latter area is affected by compensation faults of the master detachment, with small vertical throws and gentle antithetic block faulting. The difference between the two sides of the graben poses several questions. For example, does the difference indicate that decollement and forced folding of the Jurassic series were restricted to the zone of maximum subsidence, i.e. along the master detachment? Could gravity sliding of the pre-rift cover be the essential prerequisite for forced folding along a master normal fault? These problems will be the focus of further work. This is a contribution of the CNRS UMR 7516, Institut de Physique du Globe de Strasbourg. We wish to thank P. Duringer for fruitful scientific discussions. Comments from J. W. Cosgrove, G. D. Couples and an anonymous reviewer resulted in significant improvement of the manuscript.
References BERGERAT, F., MUGNIER, J. L.. GUELLEC. S.. TRUFFERT. C. CAZES, M., DAMOTTE, B. & ROURE, F. 1990. Extensional tectonics and subsidence of the Bresse basin: an interpretation from ECORS data. ///: ROURE, F., HEITZMANN, P. & POLINO, R. (eds) Deep Structure of the Alps, Memoires de la Societe Geologique de la France. 156, 145-156. BLANALT. J. G. 1972. Carte Geologique 1150000 Colmar-Anolsheim* BRGM edition, XXXVII-18. BRUN. J. P.. WENZEL, F. & ECORS-DEKORP team 1991. Crustal scale structure of the southern Rhinegraben from ECORS-DEKORP seismic reflection data. Geology. 19. 758-762. . GUTSCHER. M. A. & DEKORP-ECORS team 1992. Deep crustal structure of the Rhinegraben from DEKORP-ECORS seismic reflection data: a summary. Tectonophysics, 208. 139-147. DURINGER, P. 1988. Les conglomerate de la Bordure du Rift Cenoioique Rhenan. Dynamique Sedimentaire et Controle Climatique. These d'etat. ULP Strasbourg.
EDEL. J. B., FUCHS, K.. GELBKE, C. & PRODEHL. C. 1975. Deep structure of the Rhinegraben area from seismic refraction investigations. Journal of Geophysics, 41, 333-356. FAURE, J. L. & CHERMETTE. J. C. 1989. Deformation of tilted blocks, consequences on block geometry and extension measurements. Bulletin de la Societe Geologique de la France. 3. 461-476. HEIM. A. 1922. Geologic der Sclnvei:. Band 11.2 parts. Tauschnitz. Leipzig. ILL, M. V. & WORK. D. F. 1978. Laramide folding associated with basement block faulting along the northeastern flank of Front Range. Colorado. Memoirs of (he Geological Society of America. 151. 101-24. ILLIES. J. H. 1975. Recent and paleo-intraplate tectonics in stable Europe and the Rhinegraben rift system. Tectonophysics. 29. 251-264. LAUBSCHER, H. P. 1982. Die Sudostecke des Rheingrabens, ein kinematisches und dynamisches Problem. Eclogae Geologicae Helvetiae. 75. 101-116. MAURIN. J. C. 1995. Drapage et decollement des series Jurassiques sur la faille de detachement majeure du rift rhenan sud: implications sur la geometrie des depots syn-rifts Oligocenes. Comptes Rendus de I'Academic des Sciences. Paris. 321. serie 11. 1025-1032. PALMQUIST, J. C. 1978. Laramide structures and basement block faulting: two examples from the Big Horn Mountains, Wyoming. Memoirs of the Geological Society of America. 151. 125-138. ROURE. F.. BRUN. J. P.. COLLETTA. B. & VAN DEN DRIESSCFLE, J. 1992. Geometry and kinematics of extensional structures in the Alpine foreland of southeastern France. Journal of Structural Geology. 14, 503-519. SHARP. I. R.. UNDERHILL. J. R. & GAWTHORPE. R. L. 2000. The synsedimentary response to forced folding: Oligo-Miocene rift sequences. Gulf of Suez. Sinai, Egypt. Journal of the Geological Society. London. SITTLER. C. 1969. Le fosse rhenan en Alsace: aspect structural et histoire geologique. Revue de Geographic Phvsique et de Geologic Dvnamique. Paris. 11. 465-494. STEARNS. M. T. & STEARNS, D. W. 1978. Geometric analysis of multiple drape folds along the northwest Big Horn Mountains front. Wyoming. Memoirs of the Geological Society of America. 151, 139-156. VILLEMIN. T.. ALVAREZ. F. & ANGELIER. J. 1986. The Rhinegraben: extension, subsidence and shoulder uplift. Tectonophysics. 128. 47-59. WITTMANN. 1955. Bohnerz und praeozane Landoberflache im Markgraflerland. Jahrbuch Geologisches Landesamt. Baden Wur.ttemberg* \. 267-299. ZIEGLER, P. A. 1992. Geodynamics of rifting and implications for hydrocarbon habitat. Tectonophvsics. 215.221-253.
Displacement transfer and forced folding in the Maritimes basin of Nova Scotia, eastern Canada J. V. A. KELLER 1 & G. LYNCH2 Geological Survey of Canada, Centre Geoscientifique de Quebec, C.P. 7500, Sainte-Foy, Quebec, Canada Gl V 4C7 1 Present address: Chevron Petroleum Technology Company, San Ramon, CA 94583, USA 2Present address: Shell Canada Ltd, P.O. Box 100, Station M, Calgary, Alberta, Canada T2P 2H5 Abstract: Field and seismic data in northern Nova Scotia, eastern Canada, demonstrate that displacement transfer from steep basement faulting to bedding-parallel detachment is necessary in the development of forced folds. Lateral translation of the strata above horsted and down-dropped blocks generates a monoclinal structure which, as the faults are kinematically linked, evolves in a manner similar to fault-bend folds in thrust-and-fold belts. In the case of partial transfer of displacement a breached drape syncline is developed. The breached syncline is characterized by steep upturned beds against the fault that truncates them. In the Nova Scotia example the detachment horizon is located at the base of a Visean evaporitic sequence, and is exposed in the study area showing shearing structures within the evaporites. Brittle fault rock types (gouge and cataclasite) and meso- to microstructures were formed, including stretching lineation, principal schistosity plane and secondary shear planes, as well as intrafolial to upright asymmetrical folds. The regionally extensive weak evaporitic layer was remarkably effective in transferring displacement between the two faults, with mechanical decoupling of the strata above the evaporitic detachment being observed in the horsted block 70km away from the steep basement fault. Moreover, displacement was also transferred as much as 40km onto the down-dropped block at the frontal part of the system.
Forced (drape) folds develop by flexure of cover rocks (usually sediments) above an underlying dip-slip fault, with the cover rocks being deformed into a broad monoclinal structure. Classic examples are known from the Rocky Mountains foreland in the USA (Stearns 1971, 1978), the Rhine graben (Laubscher 1982) and the Gulf of Suez (Robson 1971). Extensional forced folds, formed by normal dip-slip faulting in the basement, are preferentially developed when weak or more ductile units are present in the cover sequences, forming a layered anisotropy and allowing the decoupling and lateral movement of the detached and folded strata (Stearns 1978; Withjack et al 1988, 1989). Although the best characterized examples present good evidence linking forced folds to steep basement faults, associated detachments are usually unexposed and inferred for material balance or in order to accommodate specific geometric models. Moreover, a definite kinematic link has not been clearly demonstrated. In this paper we present the geometry of extensional forced folds (or drape folds) developed in the Maritimes basin of Nova Scotia, eastern Canada. We also attribute these folds to the presence of an associated shallow decollement horizon which transfers upward displacement
from the underlying basement fault onto a layer-parallel detachment in the cover sequence. Regional setting and stratigraphy The Maritimes basin of eastern Atlantic Canada, located in the Gulf of St. Lawrence and surrounding provinces (Fig. 1), is a late- to post-tectonic extensional basin that developed following the Acadian (Late Silurian-Early Devonian) orogeny in the northern Appalachian belt. The basin is approximately 900km long by 400km wide (Fig. 1), and is up to 12km thick (Marillier et al. 1989). Five groups (Horton, Windsor, Mabou, Cumberland and Pictou), ranging from Late Devonian to Lower Permian age, define the basin stratigraphy (Fig. 2). The sedimentary succession is characterized by a relatively undeformed flat-lying sequence dominated by coarse- to fine-grained continental siliciclastics and subordinate marine carbonates, elastics and evaporites (Poole et al. 1970). The Horton, Cumberland and Pictou groups form the main clastic cycles, whereas the Windsor and Mabou groups contain marine carbonates and evaporites, as well as siltstones and shales, marking a general transgression in Visean-Namurian time
From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 87-101. 1-86239-060-6/OO/S 15.00 (r The Geological Society of London 2000.
88
J. V. A. KELLER & G. LYNCH
Fig. 1. Location of the Maritimes basin in eastern Canada; the extent of the basin is shown by the stippled pattern. Location of the main structural features discussed in the text is indicated with the cross-sections represented by thick straight grey lines.
(Giles 1981). At the base of the Windsor Group a distinctive and regionally extensive 10m thick laminated, intraclastic limestone unit (the Macumber Formation) occurs (Giles 1981) (Fig. 2), blanketing the full extent of the of the Maritimes basin (Howie & Barss 1975). The Macumber Formation passes abruptly upwards into a thick evaporite succession (up to 2km; Giles 1981) consisting of gypsum, anhydrite and halite (Boehner 1986; Howie 1988).
Fig. 2. Simplified stratigraphic column for the Maritimes basin in the area of the Gulf of St. Lawrence. AD - Ainslie detachment. Hatch pattern above the Ainslie detachment represents the salt.
The early evolution of the basin is marked by intense rifting, subsidence and coarse faultcontrolled clastic sedimentation (Horton Group). Late Devonian crustal-scale thinning of the Appalachian orogenic edifice has been proposed for the early evolution of the basin (Marillier & Verhoef 1989), in association with low- and high-angle extensional faults (Lynch & Tremblay 1994). Two phases of extension are recognized within the basin, both involving low-angle normal faults. The first phase, in the Late Devonian, mainly affected basement rocks. A low-angle ductile shear zone (Margaree shear zone; Lynch & Tremblay 1994) accommodated crustal thinning and led to the denudation of high-grade metamorphic rocks in the central portion of the basin; at present the Margaree shear zone is exposed onshore in Cape Breton island (Fig. 3). A younger low-angle extensional fault, the Ainslie detachment (Lynch & Giles 1995), was described as an upper crustal carapace fault (Fig. 3). It was active in the Late Carboniferous and developed during the second period of extension affecting predominantly the marine carbonates and evaporites of the Windsor Group. The Ainslie detachment is exposed onshore, both in Cape Breton island and mainland Nova Scotia, where it created regionally extensive stratigraphic gaps (Figs 1 and 3). The detachment is rooted offshore in a highangle basement master fault (Lynch 1995), which
DISPLACEMENT TRANSFER AND FORCED FOLDING
89
Fig. 3. Regional cross-section exhibiting the geometry of the main structural elements discussed in the text. Note the relationship between the Hollow fault and the Ainslie detachment, as well as the role of the Ainslie detachment in the transfer of displacement from the horsted block (zone of extension) to the down-dropped one (zone of compression). The cross-section joins an offshore segment, constructed from seismic data (segment A), with an onshore segment in Cape Breton island, constructed from field data (segment B).
runs along the western coastline of Cape Breton island (the Hollow fault of Benson 1970; Sanford & Grant 1990) (Fig. 1). Seismic imaging indicates up to 10-12 km of normal displacement of Carboniferous units along the Hollow fault (Durling & Marillier 1993,1994; Langdon & Hall 1994). However, sedimentological (Yeo & Ruixiang 1987) and kinematic analysis (St. Jean et al. 1993) on nearby onshore exposures of the Hollow fault in Nova Scotia appear to indicate that strike-slip movements have also occurred in the southern part of fault. The Ainslie detachment is exposed for more than 250km along strike (Fig. 1), from Cape Breton island in the northeast to mainland Nova Scotia in the southwest. Calc-mylonites and breccias, that occur at the top of the Macumber Formation, are frequently in contact with sheared Visean Mabou Group rocks due to the excision of 1500-2000 m of Visean Windsor Group stratigraphy (Giles & Lynch 1994) (Fig. 3). In addition, seismic lines from the Gulf of St. Lawrence indicate that offshore the Ainslie detachment is associated with a broad diapir field and pillows of Windsor salt that formed as buckle folds above the detachment (Lynch & Keller 1998) (Fig. 3). Folding was enhanced by salt diapirism, developing into large structures with wavelengths on the order of 18km and amplitudes of approximately 3.5km. A minimum of 12km of shortening has been measured in the detached Windsor and Mabou Group sequences immediately above the non-folded Macumber carbonates (Lynch & Keller 1998) (Fig. 3). Moreover, along the western coast of Cape Breton island, Windsor and Mabou
Groups are seen to be sharply upturned against the Hollow fault forming a well developed drape syncline (Fig. 3). The geometry imaged suggests that the syncline developed as an extensional forced fold by displacement transfer between the steep Hollow fault and the bedding-parallel Ainslie detachment. Extensional forced folds The geometry of the forced fold adjacent to the Hollow fault can be studied in seismic lines from the Gulf of St. Lawrence, and from fieldwork in mainland Nova Scotia where the Hollow fault runs onshore (Merigomish area). Furthermore, a smaller forced fold associated to normal dip-slip and bedding-parallel movement occurs in SW Cape Breton island associated to a subsidiary of the Hollow fault (Creignish fault) and the Ainslie detachment. The geometry of these forced folds is presented and discussed below from a seismic line immediately offshore Cape Breton island, and from structural cross-sections and maps of the two onshore areas. Bedding dips are illustrated in the cross-sections with data adjacent to the sections being incorporated using down-plunge projection. Offshore west Cape Breton island A northwest-southeast seismic reflection profile from offshore west Cape Breton island shows a largely undeformed sedimentary sequence, of
90
J. V. A. KELLER & G. LYNCH
Fig. 4. Line drawing interpretation of a seismic profile in the Gulf of St. Lawrence showing the flat-lying (relatively undisturbed) attitude of the sediments west of Cape Breton island, and the local upturning of the sequences against the Hollow fault.
Devonian-Lower Permian age, that encompasses the Horton-Pictou groups (Burling & Marillier 1993) (Fig. 4). At the southeastern end of the line Windsor-Cumberland Group strata are upturned against the Hollow fault forming a drape syncline (Fig. 4). The strata also appear truncated at a low angle to the fault. Conversely, Horton Group rocks do not show evidence of draping and are transected by the fault at a high angle to bedding (Fig. 4). This geometry suggests that the sediments overlying the Horton Group must have been detached from the underlying strata, with displacement on the Hollow fault being transferred along the interface between the Horton and the Windsor groups (the Ainslie detachment). Creignish A steep subsidiary fault to the main Hollow fault crops out along the western side of the Creignish hills, north of Port Hastings, SW Cape Breton island (Figs 5a and 6a). The Creignish fault is a basement fault that acted as a growth fault during Tournasian Horton Group sedimentation. Substantial thickness variations are observed in Horton Group sediments across the fault (approximately 500-2000 m from the footwall to the hangingwall). Later reactivation of the Creignish fault occurred (after Namurian Mabou Group deposition) when the Ainslie detachment became active. The Macumber Formation, as well as Mabou and Horton Group rocks, are exposed in the hangingwall to the west, where along the coast Mabou Group shales and fine sandstones and Horton conglomerates show moderate dips (35-45°) (Figs 5b and
c). Dips become subvertical and nearly fault parallel close to the fault surface (Fig. 6b), with the strata being upturned and forming a drape syncline against the fault (Fig. 5b and c). Horton Group strata, however, exhibit shallow to moderate dips and high-angle cut-offs with the fault, even when near to the fault surface (Fig. 5b). This indicates that the zone of displacement transfer is located above the Horton Group sediments. In the footwall Horton Group strata exhibit low to moderate dips (10-40°) towards the east to southeast. The monocline developed above the fault has a length of approximately 1.5km and an amplitude of 2.0km (Fig. 5b). The fold profile shown in Fig. 5b has been reconstructed using a kink geometry and a thicknessconstant assumption for the draped sedimentary sequences. The Ainslie detachment is well exposed along the top of the Macumber Formation in both the horsted and down-dropped block on opposite sides of the Creignish fault. Calc-mylonite and fault breccia are readily observed in the field. Moreover, the trace of the detachment is enhanced by the stratigraphic gap whereby the entire Windsor Group above the basal Macumber limestone is missing. Mabou Group rocks have been technically down-ramped onto the detachment, and the Windsor Group has been transported out of the section towards the west. A direct kinematic link between the Ainslie detachment and the Creignish fault is indicated by the fact that bedding in the Mabou Group sediments is parallel to both faults. As such, the cut-off angle with bedding (i.e. zero) was transferred from the detachment to the steep basement normal fault. In this kinematic model the upper tip of the Creignish fault is constrained
DISPLACEMENT TRANSFER AND FORCED FOLDING
91
Fig. 5. (a) Geological map of the Creignish study area. The location of cross-section D in (b) is indicated. Map is based on Weeks & Ferguson (1949) and the author's mapping. 1500-2000 m of sediments from the Windsor Group are missing along the Ainslie detachment between the Macumber limestone and the Mabou Group, (b) Structural cross-section of the Creignish monocline with no vertical exaggeration. Bars show apparent bedding dips in the plane of cross-section from down-plunge projection, (c) Lower-hemisphere stereographic projection of the plunging Creignish drape syncline showing bedding planes and poles (fault plane is also plotted).
to terminate at the Ainslie detachment, generating a forced fold with geometric features similar to those of a fault-bend fold (e.g. Suppe 1983).
Merigomish In northern fault runs scarp (Figs over 40km
Nova Scotia the basement Hollow onshore and forms a prominent 7a and 8a) that can be traced for along strike. In the study area
(Fig. 7 a), the fault is defined by the juxtaposition of Mabou Group sediments in the hangingwall against crystalline basement in the footwall. Although not exposed in this locality, the Ainslie detachment is seen to crop out in the horsted block, to the southeast of the Merigomish area, where isoclinally folded calc-mylonites of Macumber limestone, as well as a fault breccia, occur. A cross-section was constructed parallel to the East French river, from the town of Merigomish in the north across the Hollow fault in
92
J. V. A. KELLER & G. LYNCH
Fig. 6. (a) Photograph of the high-angle W-dipping Creignish fault as seen in outcrop. Footwall rocks are mainly basement gneisses and granitic intrusions, (b) Upturned beds of Mabou Group sediments 300m to the west of the previous photograph. the south (Fig. 7a and b). Along the section Mabou Group rocks are observed to be moderately dipping (25-45°) towards the north-northwest away from the Hollow fault (Fig. 7c), but dips become subvertical to vertical close to it (Fig. 8b) forming an open syncline. A monocline
has been reconstructed above the basement fault (Fig. 7b), having a length of approximately 1.0km and amplitude of 1.0-1.5 km. As indicated by the upturning of Mabou strata against the basement Hollow fault, displacement must have been transferred between the Hollow fault
Fig. 7. (a) Geological map of the Merigomish study area. The location of cross-section E in (b) is indicated. Map is based on Benson (1973) and the author's mapping, (b) Structural cross-section of the Merigomish monocline with no vertical exaggeration. Bars show apparent bedding dips in the plane of cross-section from down-plunge projection, (c) Lower-hemisphere stereographic projection of the Merigomish drape syncline showing bedding planes and poles (the fault plane is also plotted)
DISPLACEMENT TRANSFER AND FORCED FOLDING
Fig. 8. (a) Photograph of the exposed scarp of the Ndipping Hollow fault as seen in the field. The scarp can be followed for approximately 40 km along strike. Footwall rocks are basement gneisses and volcanic rocks, (b) Verticalized beds of Mabou Group rocks 300-400 m to the north of the previous photograph.
and the Ainslie detachment, with the monocline and the drape syncline forming above the region where the two faults intersect. Deformation aspects of the evaporitic extensional detachment As described above, the Ainslie detachment is a regional-scale extensional fault which occurs at the interface between a thick evaporite succession and the limestones from the underlying Macumber Formation. Shear structures, as well as calc-mylonite and breccia, have developed at the top of the Macumber limestones where locally stratigraphic gaps have been generated with the excision of most of the Windsor Group. Here, Mabou Group rocks have been down-ramped and are in tectonic contact with sheared Macumber limestones (Fig. 3). Less commonly, detached rafts or 'boudins' of evaporitic rocks (mainly gypsum and anhydrite)
93
are also observed. Numerous seismic surveys from other regions have shown that evaporitic horizons are frequently the sites where basinscale extensional detachments sole out (Gibbs 1984; Balkwill & Legall 1989; Mohriak et al. 1989; Tankard et al 1989; Withjack et al 1989). However, field descriptions of regional-scale extensional evaporitic detachments are rare. Conversely, in compressional settings the mechanical role of salt horizons in transferring horizontal displacements is not only well documented (Davis & Engelder 1985), but field examples of salt-based compressional decollement horizons are more common than their extensional counterpart (Davis & Engelder 1985 and references therein; Marcoux et al 1987). In this section we describe the shear structures developed at the detachment level, and within the detached evaporites, from preserved sections that crop out onshore Cape Breton island; textures and microstructures in the calcareous fault rocks have been presented elsewhere (Lynch & Giles 1995; Lynch etaL 1998). In outcrop the gypsum and anhydrite display well-developed mesostructures, including secondary shear planes and folds, in contrast to the rocks above and below the evaporites which display no evidence of shear. This indicates that shear is concentrated along the limestone-evaporite interface at the top of the Macumber limestone. The development of the regional-scale Ainslie detachment at the base (Fig. 3), instead than at the top, of the evaporitic sequence has been attributed to a pronounced layered anisotropy (i.e. a low-strength-low-viscosity horizon) and abnormally high fluid pressures beneath the evaporites (Lynch et al 1998). In the locality of Mabou Mines, sheared evaporites affected by the Ainslie detachment crop out along the coast (Fig. 9a and b). Brittle fault rocks are preferentially developed and exposed in the detachment, including a fine calcareous gouge along the principal displacement zone and a cataclasite (Fig. 9a). Synthetic secondary shear planes indicate a top-to-the-west direction of transport (Fig. 9a). In the hangingwall of the detachment the bedded evaporites have been deformed into an outcrop-scale roll-over anticline (Fig. 9b). The geometry of the roll-over is in good agreement with the westward direction of displacement on the detachment (Fig. 9b).
Secondary shear planes and lineation Shear structures attributed to slip along the Ainslie detachment formed within the detachment and in the detached evaporites. These
94
J. V. A. KELLER & G. LYNCH
DISPLACEMENT TRANSFER AND FORCED FOLDING
include a principal shear fabric, secondary shear planes, a stretching lineation and intrafolial to upright asymmetrical folds. The principal displacement zone of the Ainslie detachment is characterized by a 25-35-cm thick zone of calcareous gouge that occurs immediately below the thick cataclasites in Fig. 9a. This calcareous gouge is mainly cohesionless and exhibits a distinct and penetrative schistosity that is oriented along the Y direction (i.e. parallel to the shear zone boundary; Logan et al. 1979) (Fig. 9c). Secondary shear planes are also observed parallel to the R\ (or Riedel) and P directions (Logan et al. 1979). Within the gouge 0.5-1-cm thick zones of lighter coloured evaporite (gypsum) smears are found, producing a layered internal structure (Fig. 9c). These sheared gypsum planes are usually oriented along the Y direction, but smears can also be observed parallel to the R\ (or Riedel) and P directions (Fig. 9c). Typically, deformation within the sheared gypsum is characterized by distinct planar structures (Fig. 9d and e). The most common planar structure consists of a pervasive, intense schistosity developed parallel to bedding (Fig. 9d and h), which coincides with the Y direction. A stretching lineation is observed on the main schistosity planes (F-shears), defined mainly by elongated grains and nodules. Gypsum slickenfibres can also be developed on the 7-shears, which indicates bedding-parallel slip within the evaporites. The main schistosity frequently exhibits an anastomosing geometry with less deformed centimetric pods, or foliation fish, being preserved (Fig. 9d). A second set of shear planes oriented parallel to the R\ (or Riedel) direction occurs (Fig. 9d). These shears can truncate or curve asymptotically into the 7-shears, or into a further set of planar structures oriented in the P direction (Fig. 9d). As a result many R\ shears exhibit a sigmoidal geometry. The Poriented planes locally accommodate displacements on the order of millimetres, and therefore
95
appear to be P-shears. Alternatively, they could correspond to planes developed parallel to one of the finite flattening planes. R\ shear planes are usually more strongly developed than the P-shears, and in some cases measure up to 50 to 60cm in length (Fig. 9e). Folds Several fold types occur in the deformed evaporites (Fig. 9f and g). The first type, asymmetrical folds (Fig. 9f), commonly occur in trains and have steeply inclined axial planes oriented at high angles to bedding. Hinge lines are oriented at high angles to the stretching lineation direction. Most folds have wavelengths that can reach a few metres (wavelengths of < 1 m are the most common), whereas amplitudes do not exceed tens of centimetres in size. Fold limbs may exhibit pronounced thickness variations (Fig. 9f) with fold profiles that mainly conform to class 2 (or similar), but with classes 3 and la also present. Fold asymmetry (vergence) is usually in agreement with the bulk sense of shear towards the west. A second fold type correspond to tight intrafolial to isoclinal folds which typically show axial planes that are parallel to bedding, as well as the main schistosity planes (Fig. 9g). Fold hinges are also parallel to the main schistosity (Fig. 9g), but at high angle to perpendicular to the stretching lineation direction. Wavelengths are on the order of a few centimetres and amplitudes can reach up to 25-35 cm. Fold profiles usually correspond to class Ib (or parallel), with rare class 2 (or similar) profiles. Although most of the folds described above have hinge lines oriented at high angles to the stretching lineation direction, some intrafolial folds appear to have hinge lines parallel to the lineation direction. Furthermore, a number of these folds have eye-like closures typical of sheath folds.
Fig. 9. Photographs of the Ainslie detachment and related fault rocks and mesostructures. (a) Exposure of the Ainslie detachment in Mabou Mines showing brittle fault rocks, especially a fine calcareous gouge along the principal displacement zone and coarser-grained cataclasites towards the hangingwall. Metric-scale synthetic Riedel (R\} shears occur at the base of the cataclasite and sole out along the principal displacement zone, indicating a top-to-the-west direction of transport. Note the person standing at the base of the cliff, (b) Detached evaporites located above the Ainslie detachment deformed into an outcrop-scale roll-over anticline, (c) Detail of the principal displacement zone in (a) showing the structures in the calcareous gouge. Note the light-coloured planar zones of gypsum smears within the gouge, (d) Mesostructures formed at the base of the detached evaporites during shearing, including a principal shear fabric (7-shears) and secondary shear planes (R\- and P-shears). Sense of displacement as indicated by the R\-shears is towards the west, (e) Pervasive Yshears in intensely deformed gypsum-anhydrite showing the development of longer R{ -shears than in the previous zone of deformation, (f) Upright W-vergent asymmetrical folds within the banded evaporites. (g) Centimetre-scale complex intrafolial isoclinal folds, (h) E-dipping gypsum veins oriented at high angles to the main bedding/schistosity. Veins show centimetric disharmonic buckle folding and up to 50% of shortening.
96
J. V. A. KELLER & G. LYNCH
A series of small-scale (up to 75cm) E-dipping gypsum veins cross-cut bedding the main schistosity at Mabou Mines (Fig. 9h). These veins are oriented at high angles to the main schistosity and show centimetric disharmonic buckle folding (Fig. 9h). Such structures are interpreted as the result of syn-kinematic dilation and fluid flow, followed by flattening across the shear zone. Flattening, as measured by the amount of shortening in the veins, was on the order of 50%.
Comparisons with other regions and laboratory models A brief comparison with forced folds from other field areas and from laboratory physical models is helpful in placing the example presented here from Nova Scotia into context. A classic example of a forced fold is the Rattlesnake Mountain anticline (Stearns 1971), in the Rocky Mountains foreland, Wyoming, USA. Figure lOa shows a diagrammatic cross-section of Rattlesnake Mountain illustrating the geometry of the monoclinal structure. Similar to the observed drape fold over the Hollow fault, Rattlesnake Mountain exhibits an open anticlinal-synclinal pair formed above an upthrown basementinvolved steep fault (Fig. lOa). Moreover, Stearns (1971, 1978) pointed out that during the forced folding the cover sequence must have become detached, which implies significant layer-parallel slip. Indeed, Couples et al. (1994) have documented bedding-parallel slip within Cambrian shales along the base of the cover sequence. Accordingly, Rattlesnake Mountain shares first-order geometrical similarities and provides a close analogy to the Ainslie-Hollow fault system presented here. Other examples of forced folds further demonstrate that layerparallel detachments play an important role in generating this kind of folding (Haltenbanken, offshore Norway, Withjack et al. 1988, 1989; Gulf of Suez, Robson 1971; Coffield & Schamel 1987, 1989). Detachments and bedding-parallel slip have also been shown to significantly affect the geometry of deformation associated with extensional forced folding in physical laboratory models (Vendeville 1987, 1988; Withjack et al. 1990) (Fig. lOb). Displacement transfer was observed between the steep fault and the layer-parallel detachment in multilayer models containing a lower ductile decoupling horizon or internal detachment surfaces. This produced open monoclinal structures above the steep basement fault and shallower extensional structures (such as
Fig. 10. Forced fold geometries similar to the ones observed in Cape Breton island from field and physical modelling studies, (a) Cross-section of the monoclinal structure of Rattlesnake Mountain, northwest Wyoming. USA (modified from Stearns 1971. 1978). (b) Summary diagram illustrating the structures developed during clay modelling of extensional forced folds (redrawn from Withjack et al. 1990).
normal faults, grabens and roll-over anticlines) above and away from the basement fault in the horsted block (Fig. lOb). Moreover, the models also showed that a strong layered anisotropy is required for the generation of extensional forced folds. Discussion Extensional forced folds form when cover rocks flex as a result of normal dip-slip faulting in the underlying basement. As described above, lateral transport is required within the sedimentary cover rocks for the draping to occur without appreciable thinning of the layers. This results in the partition of the overall displacement between the main basement fault and a bedding-parallel detachment, with mechanical decoupling of the cover rocks. However, layers below the detachment are truncated by the basement fault and display high-angle cut-offs to the fault. Decoupling, and displacement transfer, can be remarkably effective, especially when
Fig. 11. Retrodeformable kinematic model for the sequential development of forced-fold geometries according to the degree of displacement transfer between a steep basement fault and a layer-parallel detachment. See text for further discussion.
98
J. V. A. KELLER & G. LYNCH
associated with evaporites. In Cape Breton, the zone of deformation in the hangingwall of the Ainslie detachment extends up to 70km inland from the location of the basement Hollow fault (Fig. 3). Layer-parallel slip, with mechanical decoupling of the strata above the detachment from the Horton Group and basement, generated displaced rafts, roll-over anticlines and listric normal faults in the Windsor Group and overlying sediments. Extensive stratigraphic gaps due to tectonic incision formed above the Ainslie detachment in the horsted block, as observed in eastern Cape Breton (Fig. 3). Westward displacement of the decoupled strata also caused the detachment folding and salt upwelling imaged offshore in the Gulf of St. Lawrence (Fig. 3). Here, displacement was transferred as far away as 40km to the west of the Hollow fault above the flat basal Windsor sediments. Folding is the result of less effective decoupling and pinning at the front end of the allochtonous strata. Shortening was estimated to be on the order of 12-15 km along this section (Fig. 3), but is likely to be much greater in the central portion of the basin to the north where diapirs are more extensive (Fig. 1). The effects of detachment faulting in extensional forced folding are illustrated by restorable kinematic models presented in Fig. 11. In these models we compare the geometry of deformation resulting from basement normal faulting with no displacement transfer (Fig. 11 a) to that of basement normal faulting with complete displacement transfer to an overlying detachment (Fig. lie). In the complete displacement transfer model, slip along the bedding-parallel detachment laterally transfers extension from the basement normal fault directly to a zone of extension above the horsted block (zone of extension in Fig. lie). This results in the formation of, for example, shallow listric normal faults and stratigraphic gaps in that zone. For material balance, a zone of compression is required at the front of the system (zone of compression in Fig. lie). Buckle or detachment folds, reverse faults, as well as salt pillows and diapirs develop to accommodate the shortening. Most of these deformation features are observed within the region of allochtonous salt in the Maritimes basin. A more general geometric model (a partial displacement system), which incorporates components of the two end-members, is shown in Fig. lib. The general model (Fig. lib) shows the fold draping occurring associated with displacement transfer to a shallower detachment at the earliest stage of basement faulting. During the displacement transfer a zone of extension above the horsted block and a zone
Fig. 12. Schematic model for the progressive evolution of the Hollow fault forced fold, especially the layer-parallel transport of the allochtonous sequence above the Ainslie detachment and the consequent development of salt pillows at the front of the system.
of compression at the front of the system are formed for material balance, as in the complete transfer model (Fig. lib). However, further movements along the basement fault breach the previously formed drape fold as displacement along the detachment is locked (e.g. inefficient transfer between the two systems; Fig. lib). The upturned beds against the basement fault retain their steeper attitude even after the upward breaching of the drape fold, but they are no longer parallel to the basement fault and show non-zero cut-offs (Fig. lib). Therefore, a comparison of the attitudes of the basement
DISPLACEMENT TRANSFER AND FORCED FOLDING
fault and the upturned beds constitutes a practical way to ascertain if displacement was complete or partial between the two systems of faults. Such a model is most consistent with the seismic features illustrated in Fig. 4. The evolution of the forced fold over the Hollow fault can be summarized with a schematic model as shown in Fig. 12. Extension above the Ainslie detachment in the horsted block of Cape Breton island was transferred to the steeply dipping Hollow fault (Fig. 12a). This created a monoclinal structure located at the upper tip of the normal fault (Fig. 12a), whereas the layers below the detachment were truncated by the Hollow fault at high-angles (Fig. 12a and b). The Windsor evaporites constitute a regionally extensive low-strength-lowviscosity layer that, together with high fluid pressures immediately beneath them, control the localization of the Ainslie detachment. Further displacement on the Hollow fault and the Ainslie detachment enlarged the drape syncline at the margin of the basin, and salt diapirs started to develop to the west (Fig. 12c). The detached evaporites formed salt-cored detachment folds probably in relation to the pinning of the Ainslie detachment in the western part of the basin (Fig. 12c). Conclusions Field and seismic data in northern Nova Scotia demonstrate that steep basement faulting is kinematically linked to detachment faulting in the development of forced folds. Displacement transfer from the basement fault to the shallower detachment fault usually occurs along a weak decollement horizon in the cover sequence. Lateral translation of the strata above horsted and down-dropped blocks, as well as over the steep basement fault, results in a monoclinal structure where the two faults intersect. As the faults are linked, forced folds in the overlying strata evolve in a manner similar to fault bend folds. In the case that transfer of displacement is not complete between the two faults, or the decollement horizon has become inefficient, a breached drape syncline is developed. The breached syncline is characterized by a steep upturned limb which is no longer parallel to the steep fault, but rather is truncated by it. In the study area an extensional evaporitic basal decollement is exposed onshore showing the effects of the shearing in the Windsor evaporites. Brittle fault rock types and meso- to microstructures were formed, including a principal schistosity plane in the Y direction and secondary
99
shear planes, a stretching lineation and intrafolial to upright asymmetrical folds. The weak layer represented by the Windsor evaporites, together with underlying high fluid pressures, created a low-strength-low-viscosity horizon that was remarkably effective in transferring displacement. Mechanical decoupling of the strata above the detachment is observed onshore in Cape Breton island, 70 km away from the location of the basement fault. Roll-over anticlines and listric normal faults in the Windsor, and younger groups, formed to accommodate layer-parallel slip on the Ainslie detachment. Moreover, at the frontal part of the system to the west, salt-cored detachment folds (pillows and diapirs) indicate layerparallel shortening above the detachment related to less effective decoupling and a stick point. Nevertheless, displacement was transferred as much as 40km offshore from the Hollow fault. The authors wish to thank Leopold Nadeau for discussions, comments and suggestions on an earlier version of this manuscript Geological Survey of Canada contribution no. 1996255.
References BALKWILL, H. R. & LEGALL, F. D. 1989. Whale Basin, offshore Newfoundland: extension and salt diapirism. In: Tankard, A. J. & Balkwill, H. R. (eds) Extensional Tectonics and Stratigraphy of the North Atlantic Margins. American Association of Petroleum Geologists, Memoir, 46, 233-245. BENSON, D. G. 1970. Notes to Accompany Maps of the Geology of Merigomish and Malignant Cove Map Areas, Nova Scotia. Geological Survey of Canada, Paper 70-9. 1973. Merigomish and Malignant Cove, Nova Scotia. Geological Survey of Canada, Map 1361 A, scale 1:50000. BOEHNER, R. C. 1986. Salt and potash resources in Nova Scotia. Nova Scotia Department of Mines and Energy Bulletin, 5, 346 pp. COFFIELD, D. Q. & SCHAMEL, S. 1987. Kinematics of drape fold evolution over normal faults in the Gulf of Suez, Egypt. Geological Society of America Abstracts with Programs, 19, 622. & 1989. Surface expression of an accommodation zone within the Gulf of Suez, Egypt. Geology, 17, 76-79. COUPLES, G. D., STEARNS, D. W. & HANDIN, J. W. 1994. Kinematics of experimental forced folds and their relevance to cross-section balancing. Tectonophysics, 233, 193-213. DAVIS, D. M. & ENGELDER, T. 1985. The role of salt in fold-and-thrust belts. Tectonophysics, 119, 67-88. DURLING, P. & MARILLIER, F. 1993. Structural elements of the Magdalen basin, Gulf of St. Lawrence, from seismic reflection data. In: Current Research Part D. Geological Survey of Canada, Paper 93-1D, 147-154.
100
J. V. A. KELLER & G. LYNCH
& 1994. Tectonic setting of Middle Devonian to Lower Carboniferous rocks in the Magdalen basin. Atlantic Geology, 29, 199-217. GIBBS, A. D. 1984. Clyde field growth fault secondary detachment above basement faults in North Sea. AAPG Bulletin, 68. 1029-1039. GILES, P. S. 1981. Major Transgressive-Regressive Cycle in Middle to Late Visean Rocks of Nova Scotia. Nova Scotia Department of Mines and Energy, Paper 81-2. GILES, P. S. & LYNCH. G. 1994. Stratigraphic omissions across the Ainslie Detachment in east-central Nova Scotia. In: Current Research Part D. Geological Survey of Canada, Paper 94-1D, 89-94. HOWIE. R. D. 1988. Upper Paleozoic Evaporites of Southeastern Canada. Geological Survey of Canada. Bulletin 380. & BARSS. M. S. 1975. Upper Paleozoic Rocks of the Atlantic Provinces, Gulf of St. Lawrence, and Adjacent Continental Shelf. Geological Survey of Canada, Paper 74-30, 35-50. LANGDON. G. S. & HALL, J. 1994. Devonian-Carboniferous tectonics and basin deformation in the Cabot strait area, eastern Canada. AAPG Bulletin, 78, 1748-1774. LAUBSCHER, H. P. 1982. Die Sudostecke des Rheingrabens - ein kinematisches und dynamisches problem. Eclogae Geologicae Helve tiae. 75, 101-116. LOGAN. J. M.. FRIEDMAN. M., HIGGS, N. G., DENGO C. & SHIMAMOTO, T. 1979. Experimental studies of simulated gouge and their application to studies of natural fault zones. In: Proc. Conf. VIII, Analysis of Actual Fault Zones in Bedrock. United States Geological Survey, Open-File Report 79-1239, 305-343. LYNCH, G. 1995. Salt diapirism in relation to fault displacement transfer from a basement normal fault to an extensional detachment in the Carboniferous Maritimes basin. Nova Scotia. Canada. Geological Society of America Abstracts with Programs, 27. 385. & GILES. P. S. 1995. The Ainslie Detachment: a regional flat-lying extensional fault in the Carboniferous evaporitic Maritimes Basin of Nova Scotia, Canada. Canadian Journal of Earth Sciences, 33. 169-181. & KELLER, J. V. A. 1998. Association between detachment faulting and salt diapirs in the Devonian-Carboniferous Maritimes Basin, Atlantic Canada. Canadian Petroleum Geology Bulletin, 46, 189-209. & TREMBLAY, C. 1994. Late Devonian-Carboniferous detachment faulting and extensional tectonics in western Cape Breton Island. Nova Scotia, Canada. Tectonophysics, 238, 55-69. , KELLER, J. V. A. & GILES, P. S. 1998. The tectonic setting of the Maritimes Basin and influence of the Ainslie Detachment on mineralization in the Windsor Group of northern Nova Scotia, Canada. Economic Geology, 93, 703-718. MARCOUX, J., BRUN, J.-P., BURG, J.-P. & Ricou, L. E. 1987. Shear structures in anhydrite at the base of thrust sheets (Antalya, Southern Turkey). Journal of Structural Geology, 9, 555-561.
MARILLIER, F. & VERHOEF, J. 1989. Crustal thickness under the Gulf of St. Lawrence, northern Appalachians, from gravity and deep seismic data. Canadian Journal of Earth Sciences. 26. 1517 1532. MARILLIER. F.. KEEN. C. E.. STOCKMAL. G. S.. QUINLAN. G., WILLIAMS, H.. COLMAN-SADD. S. P. & O'BRIEN. S. J. 1989. Crustal structure and surface zonation of the Canadian Appalachians: implications of deep seismic data. Canadian Journal of Earth Sciences. 26. 305 321. MOHRIAK. W. U.. MELLO, M. R.. KARNER. G. D.. DEWEY. J. F. & MAXWELL. J. R. 1989. h and Stratigraphic evolution of the Campos Basin, offshore Brazil. In: TANKARD. A. J. & BALKWILL. H. R. (eds) Extensional Tectonics and Stratigraphy of the North Atlantic Margins. American Association of Petroleum Geologists Memoir. 46. 577-598. POOLE, W. H.. SANFORD. B. V.. WILLIAMS. H. & KELLEY. D. G. 1970. Geology of southeastern Canada. In: DOUGLAS. R. J. (ed.) Geology and Economic Minerals of Canada. Geological Survey of Canada Economic Geology Report. 1. 229-304. ROBSON, D. A. 1971. The structure of the Gulf of Suez (Clysmic) rift, with special reference to the eastern side. Journal of the Geological Society. London. 127. 247-276. ' SANFORD. B. V. & GRANT. A. C. 1990. Bedrock geological mapping and basin studies in the Gulf of St. Lawrence. In: Current Research Part B. Geological Survey of Canada. Paper 90-1B. 33-42. STEARNS, D. W. 1971. Mechanisms of drape folding in the Wyoming province. In: 23rd Annual Field Conference Guidebook. Wyoming Geological Association, 125-144. STEARNS, D. W. 1978. Faulting and forced folding in the Rocky Mountains foreland. ///: MATTHEWS. V. (ed. ) Laramide Folding Associated with Basement Block Faulting in the Western United States. Geological Society of America. Memoir. 151. 137. ST. JEAN. J. A. R.. NANCE, R. D. & MURPHY. J. B. 1993. Tectonic significance of Late Paleozoic deformation in the Cape George Peninsula. Antigonish Highlands, Nova Scotia. Atlantic Geology. 29. 27-42. SUPPE, J. 1983. Geometry and kinematics of fault-bend folding. American Journal of Science. 283. 684721. TANKARD. A. J., WELSINK. H. J. & JENKINS. W. A. M. 1989. Structural styles and stratigraphy of the Jeanne d'Arc Basin. Grand Banks of Newfoundland. In: TANKARD, A. J. & BALKWILL, H. R. (eds) Extensional Tectonics and Stratigraphy of the North Atlantic Margins. American Association of Petroleum Geologists. Memoir. 46. 265-282. VENDEVILLE. B. 1987. Champs de Failles et Tectonique en Extension: Modeli~ation Experimental'. PhD Thesis, Universite de Rennes, France. 1988. Modeles experimentaux de fracturation de la couverture controlee par des failles normales dans le socle. Comptes Rendus de I'Academic des Sciences. Paris, 307. 1013-1019.
DISPLACEMENT TRANSFER AND FORCED FOLDING WEEKS, L. J. & FERGUSON, S. A. 1949. Mulgrave, Nova Scotia. Geological Survey of Canada, Map 995A, scale 1:63360. WITHJACK, M. O., MEISLING, K. E. & RUSSEL, L. R. 1988. Forced folding and basement-detached normal faulting in the Haltenbaken area, offshore Norway. AAPG Bulletin, 72, 259. , & 1989. Forced folding and basementdetached normal faulting in the Haltenbaken area, offshore Norway. In: TANKARD, A. J. & BALKWILL, H. R. (eds) Extensional Tectonics and Stratigraphy of the North Atlantic Margins.
101
American Association of Petroleum Geologists, Memoir, 46, 567-575. , OLSON, J. & PETERSON, E. 1990. Experimental models of extensional forced folds. AAPG Bulletin, 74, 1038-1054. YEO, G. M. & RUIXIANG, G. 1987. Stellarton Graben: an Upper Carboniferous pull-apart basin in Northern Nova Scotia. In: BEAUMONT, C. & TANKARD, A. J. (eds) Sedimentary Basins and BasinForming Mechanisms. Canadian Society of Petroleum Geologists, Memoir, 12, 299-309.
This page intentionally left blank
Late Quaternary monoclinal folding induced by caldera resurgence at Ischia, Italy ALESSANDRO TIBALDI1 & LUIGINA VEZZOLI2 Department of Geological Science and Geotechnologies, University of Milan-Bicocca, Italy 2 Department of Chemistry, Physics and Mathematics, University of Insubria, Italy Abstract: The main structure of the active resurgent caldera of Ischia is uncommonly well exposed. The very young age of the resurgence (<33 ka BP) has enabled the uppermost part of the uplifted block to be preserved, whereas the severe erosion on the steep flanks, made up of low-resistant pyroclastic deposits, has resulted in the rest of the resurgent block structure cropping out. The uppermost part is characterized by strata gently dipping at 1-5° in a general south-southeast direction. Along the southern flank of the resurgent block, the strata gradually form a complex double monocline fold with ENE-trending hinge lines. Some tens of metres below the topographic surface, the anticline part of each monocline fold gradually turns into a vertical fault. The middle limb strata are vertical to subvertical. This draping occurred with the aid of at least one main detachment horizon localized inside the pyroclastic succession. Total resurgence here is about 350m. The opposite northern flank of the resurgent horst, which has experienced maximum uplift of the order of about 900m, is characterized by a few vertical to subvertical normal faults with draping of strata. These data suggest that the outward-dipping strata commonly found in Quaternary resurgent calderas could represent; (i) the surface expression of forced folding induced by the pistonlike uplift of a fault-bounded concealed block; or (ii) the classical doming induced by radial and vertical growth. The fault-bounded horsts sometimes recognized in old, deeply eroded, resurgent calderas could be the forcing block of non-preserved drape folds. These monocline folds could also be disrupted by upward propagation of the bounding faults after an incremental offset beyond the limit of folding of a given volcanic succession.
The island of Ischia is located in the Southern Tyrrhenian Sea to the west of the Bay of Naples (Fig. 1 A). Lava and pyroclastic eruptions of alkalitrachytic and trachytic composition built up the island of Ischia from the late Pleistocene at least. The oldest rocks (>147-130kaep) are the remnants of a volcanic field which collapsed in the form of summit caldera (Gillot et al. 1982; Vezzoli 1988) (Fig. IB and C). The area has undergone several phases of deformation mainly as a result of magma inflation and deflation (Poli et al. 1989) but also from tectonic regional stresses (Fusi et al. 1990; Tibaldi & Vezzoli 1997). Resurgence started approximately 33kaBP (Gillot et al. 1982), uplifting the western-central part of the caldera floor from sea level to the present altitude of 787m, creating the Mount Epomeo horst (Rittmann 1930), and triggering several landslides and earthquakes. The resurgent block is approximately square shaped in plan view, and is mostly bounded by vertical to outward-dipping faults (Fig. IB and C). The escarpment on the southern flank has previously been interpreted as a normal fault (Vezzoli 1988; Orsi et al. 1991), whereas here, on the basis of new structural field data, we propose that it represents a flexure linked to the piston-like uplift of a fault-bounded concealed
block. From this case, we move to examine similar structural features of other resurgent calderas. The Ischia caldera The current Ischia caldera formed around 55 ka BP during eruption of alkalitrachytic ash and pumice flows constituting the Mount Epomeo Green Tuff (MEGT) (Vezzoli 1988) (Fig. 2). The caldera collapse occurred throughout a piston-like subsidence of the entire MEGT intracaldera block. This Ischia caldera probably reactivated an older volcanic collapse structure defined by a discontinuous ring of lava domes of 130-73 ka BP. Collapse and deposition of the MEGT erased further evidence of the previous caldera history. The Ischia caldera is elliptical in shape, about 1 0 x 7 km in diameter, with the major axis oriented in the east-west direction (Fig. 1). The depth of the caldera is between 200 and 400 m in the western flank, and 350 m in the southern flank. This caldera depression is filled by the MEGT with a thickness of between 140 and 400m. After the MEGT eruption, the caldera floor was covered by the sea, resulting in the
From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 103-113. 1-86239-060-6/OO/S 15.00 © The Geological Society of London 2000.
104
A. TIBALDI & L. VEZZOLI
LATE QUATERNARY MONOCLINAL FOLDING
105
Fig. 1. (A) Location map of the island of Ischia in the Bay of Naples, Italy. Bathymetry in metres. (B) Location of the geological cross-sections (lines) and geothermal drillings (solid dots) of Figs 3 and 4. The box is the enlarged area in Fig. 5. Contour lines in metres above sea level (a.s.l.) show the morphology of the resurgent block. (C) Geological sketch map of the island of Ischia showing the deformation features developed during the caldera resurgence.
deposition of shallow-marine sediments, while explosive hydromagmatic centres were active along the external side of the caldera rim (Citara Tuff) (Fig. 2). During this stage (5533kaBp) about 200m of sediments and distal
pyroclastic deposits accumulated and buried the caldera floor (Colle Jetto Formation) (Fig. 2). Ring faults of the caldera offset the volcanic rocks older than 73kaep and are mantled by pyroclastic deposits of the Citara Tuff. Outflows of the MEGT are 5-60 m thick and cover the external side of the caldera rim. General structure of the Ischia resurgence
Fig. 2. Stratigraphy of the units involved in the Ischia caldera resurgence. Not to scale.
The resurgent block of Mount Epomeo has a diameter of 4.5km. Pyroclastic rocks of the Citara Tuff (43-33 ka BP) and marine sediments of the caldera fill are deformed, tilted and uplifted up to 650m of altitude along the Mount Epomeo flanks. The uplift at Ischia has been large enough to expose caldera floor rocks (lavas dated at 133kaBp, Fig. 2) along the western flank of the resurgent block (Fig. 3A—A'). Syn-resurgence volcanism is located along the caldera rim and in the eastern side of the caldera floor peripheral to the resurgent block. The only volcanic activity that occurred in the deformed area is represented by very viscous trachytic lavas repeatedly emplaced after 10 ka BP as endogenous and exogenous domes. Monolithological talus breccias and debris flow derived from the
Fig. 3. Geological cross sections of the western and southern H a n k s of the M o u n t Hpomco resurgence structure. Core locations and traces of sections in Fig. I B .
LATE QUATERNARY MONOCLINAL FOLDING
107
Fig. 4. General view towards the southeast and geological cross section (E-E7) of the northwestern flank of the resurgent block of the Ischia caldera. The topographic escarpment marks the drape fold induced by the resurgence. On the right of the frame it is possible to see NNW-striking normal faults (little arrows). On the background the Mount Epomeo summit graben is shown (large arrow).
108
A. TIBALDI & L. VEZZOLI
resurgent block intertongue with these lavas (Fig. 2). An abrupt change from undeformed deposits to 100m high fault escarpments marks the transition from the caldera floor to the resurgent block. This geometry can be appreciated in the geological cross sections of Fig. 3, which have been reconstructed using drilling stratigraphic data (Penta & Conforto 1951; Ippolito el al. 1973) and field work by the authors. A narrow transition zone (50-200 m wide) is characterized by a variation in bedding inclination around the base of the fault escarpments. In the caldera floor, bedding attitude is regular with horizontal or subhorizontal strata (inclination <2°). In the transition zone the strata gradually bend to reach an inclination mostly of 15-50°. A fault zone connects the transition zone with the horst block along the northern, eastern and western flanks. In the fault-bounded blocks of these fault zones, sedimentary and pyroclastic strata dip outward with an inclination of 30-45° in
the western flank and 50-70° in the northern flank (Figs 3 and 4). The fault zones are made of two or three main parallel faults that have guided the entire uplift and align various fumarole fields of the caldera. Percolative hot fluids would have favoured the concentration of deformation on only a few preferential planes, which repeatedly acted as weakness zones. The resurgent core of the Ischia caldera is mainly represented by an almost unbroken block of MEGT dipping homoclinally to the south-southeast and culminating at the Mount Epomeo summit. A detailed description of the structural characteristics of each flank of this block follows. The northern flank is bounded by parallel normal faults striking, in decreasing frequency, east-northeast, north-northeast and west-southwest. These have a dip of 65-80° towards the north (outward with respect to Mount Epomeo). Along this northern flank, a few smaller inwarddipping normal faults have created a localized
Fig. 5. Detailed geological and structural map of the double monocline fold in the southern flank of the Mount Epomeo resurgent block. Results of structural analysis are shown in Schmidt stereographic projections: lines indicate fault planes, small arrows indicate the sense of the hangingwall block tectonic transport, circles indicate poles to bedding of monocline, great arrows outside the stereoplot indicate direction of
Fig. 6. Draping of strata along the double monocline fold of the southern flank of the resurgent block of the Ischia caldera. (A) Eastern edge of the drape fold. Arrow points to the location of B. (B) Anticlinal hinge of the drape fold. (C) Western edge of the double drape fold. Arrow points to the location of D. (D) Synclinal hinge between upper and lower drape folds.
no
A. TIBALDI & L. VEZZOLI
horst and graben structure at the Mount Epomeo summit zone (Fig. 4). These smaller faults strike northwest and north-northwest, and are the oldest in the area, being displaced by the other fault sets. All the faults described here show normal displacements and belong to the post33 kasp main resurgence phase, as they offset deposits dated to be 33 ka BP and younger. The western and eastern flanks are limited by subvertical to vertical normal faults dipping southwest to west and east-northeast, respectively. These faults show incremental offset from south to north and displace deposits young as 33 ka BP The southern flank is marked by a monocline which will be described in more detail below. The Ischia drape fold
Fig. 7. Sketch of the anticline core of the northern (A) and southern (B) drape zones of the double monocline fold of the southern flank of the resurgent block of the Ischia caldera.
The monocline of the southern flank of the Mount Epomeo resurgent block affects pyroclastic deposits dated at 33 ka BP in the western part of the fold and at 38.4kaBP in the eastern part (Vezzoli 1988). The monocline faces toward south-southeast with horizontal hinge lines that trend around N67° (Figs 1C, 3D-D' and 5). The upper strata are subhorizontal (<2°) and locally dip towards the caldera margin (southward). These strata define the uppermost part of the resurgent block and gradually pass first to a dip of <5° towards south-southeast (outward) and then to a steeper inclination (20-30°) in a 20-30 m wide zone. The middle limb strata dips 80° towards southsoutheast. The zone between the low-inclined upper strata and the middle limb strata defines the anticlinal part of the monocline (Fig. 6A and B). Farther to the south-southeast, the strata of the middle limb abruptly change to a dip of approximately 15° and, a few metres farther, to a subhorizontal attitude (<5°) that defines the synclinal part of the monocline (Fig. 6C and D). About 100m farther to the south-southeast, the strata change once again to a steeper inclination (40°) and then to a vertical dip in a 20-m wide zone (Fig. 5). More to the south-southeast, the strata dip 10° towards the south-southeast. These dip changes define another monocline with hinge lines approximately parallel to the previous ones (Fig. 6C). The present topography reflects the complex structure of the entire double monocline (Figs 3D-D' and 6C). A planar surface gently dipping towards the south-southeast represents the upper limb of the northern (and upper) monocline. A strong increase of the topographic gradient,
given by an escarpment approximately 100m high and inclined about 50% coincides with the middle limb of this monocline. South of the upper synclinal hinge, the altitude slowly decreases and then another, but smaller, escarpment marks the second, lower and southern, middle limb. A third escarpment along the coast has been produced by sea erosion. The presence of deep gullies cut into the pyroclastic deposits found in this zone of the island has enabled us to study the structure of the double monocline at a deeper level. At the northern and upper anticline hinge the folded strata are affected by brittle structures in the form of systematic ENE-striking faults (Fig. 7). These faults are described below starting from the uppermost exposed strata. The first 10-20m consist of draped strata without any brittle deformations. About 20m down in the stratigraphic succession, a conjugate system of faults is present. Faults dip towards north-northwest and south-southeast with dips ranging from 50° to 80C forming a splay with a downward apex (Fig. 7A). Displacements are in the order of decimetres. Approximately 10m farther down in the stratigraphic succession, the faults converge in a single main vertical plane. This fault can be followed for about 60-80 m down where it shows a displacement of several metres with a relative downthrow of the southern block. The strata of the northern, uplifted block are horizontal to subhorizontal (<6 C ), whereas those of the southern block are gently dipping at 5-10° to the south-southeast. The strata of the lower (i.e. southern) monocline defines anticline and syncline zones with flank angles of approximately 90 . A few
LATE QUATERNARY MONOCLINAL FOLDING
metres stratigraphically below the anticline zone, an ENE-striking vertical fault puts the northern uplifted block into contact with the vertical middle limb strata (Fig. 7B). This fault can be followed up to a detachment horizon located inside the subhorizontal pyroclastic stratigraphic succession. This horizon acted as a surface of transfer displacement between the folding zone and the uppermost part of the uplifting horst. Thinning of strata in the middle limbs or shortening in the synclinal hinge zones of the monocline fold were not observed. However, it is necessary to note that the pyroclastic deposits involved in the flexure are not characterized by a constant depositional thickness, which makes it difficult to document changes during deformation events.
Other structures of the drape-fold zone Some other normal faults crop out close to the hinge lines of the drape fold on the southern flank of the resurgent block (Figs 1C and 5). These faults also affect landslide deposits that cover 38.4-33 kasp old pyroclastic rocks deformed by the drape fold. These landslides have been attributed to the Holocene by Vezzoli (1988), and thus their deformation indicates that the process of resurgence of Mount Epomeo is still active. The normal faults on the southern flank of the horst have triangular facets, offset river streams and strike parallel to the monocline hinge lines. Fault planes are vertical to subvertical, with dips towards the south-southeast. The tectonic blocks immediately to the south of the fault planes are always relatively downthrown. Possible draping of strata along these faults cannot be ascertained because outcropping landslide deposits are mostly made up of chaotic breccia without visible markers.
111
Along the southern flank of the horst, two main faults can be seen in the lower part of the rock succession inside deep river gullies. These faults strike east-northeast with a vertical dip. At a higher level of the succession, some tens of metres below the uppermost topographic surface of the resurgent block, the strata are clearly draped and depict a monocline facing southsoutheast. The low offset values measured along the faults, in the order of metres at maximum, cannot account for the Mount Epomeo uplift, which is in the order of at least 350m along the southern flank. This can be explained by: (i) the fact that the exposed structures are at a shallow level, where deformation is mainly represented by the flexure of the strata, and (ii) the possible existence of other hidden faults. All these data indicate the presence of a complete drape fold dominating in the upper part of the rock succession, and brittle structures in the lower part. Draping has been favoured by at least one main detachment horizon localized inside the pyroclastic succession. Folding is quite rare in volcanic rocks at a surface level. At Ischia, this can be explained by the lithological composition, and thus rheological behaviour, of the volcanic succession which is made up primarily of pyroclastic deposits interspersed with occasional lava flows. This succession characterizes the whole island (Vezzoli 1988) and, thus, cannot explain the presence of a monocline exclusively along the southern flank of the resurgent horst. An explanation can instead be provided by an analysis of the differential absolute uplift along the various flanks of the horst. The uplift is maximum, about 900 m, along the northern flank, while it gradually
Discussion The draping of the volcanic succession of Ischia The Mount Epomeo resurgence occurred in a zone of highly concentrated deformation, which rapidly merges with the practically undeformed surrounding caldera floor. This deformation zone is marked by a few vertical to subvertical outward-dipping normal faults which bound the horst. Close to the faults, the strata is flexured Fig. 8. Block diagram of the southwestern flank of with an outward dip of about 15-50°, which the double monocline fold of the Ischia resurgent caldera. locally reaches a value of 75°.
112
A. TIBALDI & L. VEZZOLI
decreases southward to about 350m. Even if detachment horizons favoured draping, we suggest that uplift along the southern flank did not reach a threshold beyond which this volcanic succession would have completely behaved in a brittle manner, while in the other flanks the threshold was bypassed and subvertical faults propagated upward to the surface (Fig. 8). Is drape folding a new model of caldera resurgence? The mechanics of caldera resurgence has been considered as a (i) dome-like or (ii) piston-like uplift. Both cases can be symmetric or asymmetric. Dome-like resurgence (Fig. 9A and C) is usually interpreted as a result of arching of the caldera floor (Smith & Bailey 1968; Marsh 1984) and is the case of several calderas with resurgence of Neogene-Quaternary age, such as the Toba caldera (Chesner & Rose 1991), Valles caldera and other several calderas of western North America (Steven & Lipman 1976; Lipman 1984).
Fig. 9. Possible models and stages of caldera resurgence. Piston-like and dome-like resurgence mechanisms are schematized during incremental stages tl,t2 and t3 for symmetric and asymmetric geometries. At early stages, the surface geometry of doming by the two models may be similar. The Ischia case is characterized by an asymmetric uplift with the resurgent block bounded by a preserved double drape fold at one flank and faulted drape folds along the other flanks.
Dome-like resurgence is characterized by: (a) no peripheral faults; (b) a rounded dome shape in plan view; (c) beds continuous across the dome; and (d) radially outward-dipping strata of the caldera fill with increase of dip from the flat top to the margins. In the dome-like resurgence the extension over the dome is accommodated by a longitudinal apical graben. Piston-like resurgence (Fig. 9B and D) is represented by the uplift of a well-defined cylindrical or subcylindrical block bounded by vertical to subvertical faults. This resurgence is characterized by (a) small deformation of the overburden and (b) large-scale shear fractures at the periphery. Symmetric piston-like resurgence is recognized for (a) flat top and steep flanks and (b) peripheral faults with the same displacement along the entire core flanks. Asymmetric piston-like resurgence can be (a) uplift along one flank only with a fixed fulcrum in the opposite flank and (b) different amounts of uplift along all flanks. Both types are recognized for the regularly inclined top. The first type of asymmetric piston-like resurgence has rotation of dip of the fault plane along maximum uplift. The second type is recognized for the presence of peripheral faults along the side with maximum uplift and folded beds along side with minimum uplift. Examples of purely piston-like resurgence are rare. The Platoro caldera (Lipman 1984) is an example of a typical asymmetric hinged resurgence. Many resurgent calderas are more complex that these simple models. The data presented here on Ischia suggest that there can be some linking between the dome-like and piston-like resurgence at certain stages of development of the uplift. The outward-dipping strata, commonly found in Quaternary resurgent calderas, could represent the surface expression of forced folding induced by the piston-like uplift of a fault-bounded concealed block (tl in Fig. 9B). In the same way, the fault-bounded horsts recognized in old, deeply eroded, resurgent calderas could represent the forcing block of non-preserved drape folds. Apart from erosion, the non-preservation of these monocline folds could also be a result of the absence or locking of a weak horizon of detachment and the relationships between the amount of uplift and the capacity of a given volcanic succession to absorb the deformation by folding. In the absence of transfer of displacement along horizontal or subhorizontal detachment surfaces, if the incremental vertical offset increases beyond the limit of folding of the rock succession, the caldera resurgence drape folds could be disrupted by upward propagation of the bounding faults (t3 in Fig. 9B and D).
LATE QUATERNARY MONOCLINAL FOLDING
Outward-dipping strata can be found also when resurgence occurs with vertical and radial growth induced by diffuse arching above a developing shallow magma body (tl in Fig. 9A and C). In this case, incremental deformation commences with the formation of an apical graben (t3 in Fig. 9A and C). The graben that affects the summit part of the Mount Epomeo horst of Ischia is the earliest brittle structure of the resurgence phase, as shown by the cross-cutting relationships between the faults. As a consequence, it can be argued that resurgence at Ischia first begun by asymmetric dome-like uplift with formation of an apical graben (tl of Fig. 9D), and subsequently developed by wide flexuring along the southern flank and faulting up to the surface along the other flanks induced by asymmetric uplift of a rigid indenter (t2 and t3 of Fig. 9D). This reconstruction denotes that caldera resurgence can also occur throughout successive steps with different models of fold draping, which can be attributed to the mobility in time and space of the subsurface magma bodies. Acknowledgements We acknowledge many useful suggestions from Joao Keller. The authors have benefitted of funds from the Ministero delFUniversita e della Ricerca Scientifica e Tecnologica (Italy).
References CHESNER, A. & ROSE, W. I. 1991. Stratigraphy of the Toba Tuffs and the evolution of the Toba Caldera Complex, Sumatra, Indonesia. Bulletin of Volcanology, 53, 343-356. Fusi, N., f IBALDI, A. & VEZZOLI, L. 1990. Vulcanismo, risorgenza calderica e relazioni con la tettonica
113
regionale nell'Isola di Ischia. Memorie Societa Geologica Italians 45, 971-980. GILLOT, P. Y., CHIESA, S., PASQUARE, G. & VEZZOLI, L. 1982. <33.000yr K-Ar dating of the volcanotectonic horst of the Isle of Ischia, Gulf of Naple. Nature, 229, 242-244. IPPOLITO, F., ORTOLANI, F. & Russo, M. 1973. Struttura marginale tirrenica delFAppennino campano: reinterpretazione di dati di antiche ricerche di idrocarburi. Memorie Societa Geologica Italiana, 12, 227-239. LIPMAN, P. W. 1984. The roots of ash-flow calderas in Western North America: windows into the tops of granitic batholits. Journal of Geophvsical Research, 89, 8801-8841. MARSH, B. D. 1984. On the mechanics of caldera resurgence. Journal of Geophysical Research, 89, 8245-8251. ORSI, G., GALLO, G. & ZANCHI, A. 1991. Simple-shearing block resurgence in caldera depressions: A model from Pantelleria and Ischia. Journal of Volcanological and Geothermal Research, 47, 1-11. PENTA, F. & CONFORTO, B. 1951. Risultati di sondaggi e di ricerche geominerarie nell'isola dTschia dal 1939 al 1943 nel campo del vapore, delle acque termali e delle forze endogene in generale. Annali di Geofisica, 4, 1-23. POLI, S., CHIESA, S., GILLOT, P. Y., GUICHARD, F. & VEZZOLI, L. 1989. Time dimension in the geochemical approach and hazard estimates of a volcanic area: the isle of Ischia case (Italy). Journal of Volcanological and Geothermal Research, 36, 327-335. RITTMANN, A. 1930. Geologie der Insel Ischia. Z. f. Vulkanol. Erganzungsband, 6, 1-265. SMITH, R. L. & BAILEY, R. A. 1968. Resurgent cauldrons. Memoir Geological Societv of America, 116,613-667. STEVEN, T. A. & LIPMAN, P. W. 1976. Calderas of the San Juan Volcanic Field, Southwestern Colorado. United States Geological Survey, Professional Paper, 958. TIBALDI, A. & VEZZOLI, L. 1997. The space problem of caldera resurgence: A lesson from Ischia. Geologische Rundschau, 87, 53-66. VEZZOLI, L. (ed.). 1988. Island of Ischia. Quaderni de La Ricerca Scientifica, 114, 10. CNR, Rome.
This page intentionally left blank
Stratal fold patterns adjacent to normal faults: observations from the Gulf of Mexico C. S. MANSFIELD1 & J. A. CARTWRIGHT2 Nederlandse Aardolie Maatschappij B.V., Business Unit Offshore, Grote Hout of Koningsweg 49, 1950 AA, Velsen-Noord, The Netherlands 2 Department of Earth Sciences, Cardiff University, PO Box 914, Cardiff CF 10 3 YE, Wales 1
Abstract: This paper describes the study of a subsurface array of young growth faults from the Gulf of Mexico, using a high-resolution three-dimensional (3D) seismic data set. The seismic data have resolved all of the faults within the survey area in detail, enabling the full three-dimensional structure and displacement patterns of the faults to be accurately mapped. In this paper, special attention is paid to the magnitude and orientation of small 'drag' folds that occur in the wall-rock of these faults, ubiquitous throughout the data set and well resolved by the seismic data. The results indicate that in both the footwalls and the hangingwalls the folds occur with variable amplitude, and also that they are small in relation to the dimensions of the respective faults. That is, they are spatially impersistent, often exhibiting rapid changes from normal drag to reverse drag over distances as short as 100m. Furthermore, strong discordance with the recorded patterns of throw (the vertical component of displacement) suggests that no simple apparent correlation exists with the respective displacement distributions. By comparison with a number of field and theoretical examples, various mechanisms are discussed to explain the probable origin and variability of these folds. In general, however, evidence for the cause of the folds is limited and remains largely inconclusive. It is suggested that one potential explanation is stratal deformation in the vicinity of points of overlap and linkage between faults in the dip direction (dip linkage). In an earlier study of the fault throw distributions it is has been argued that this is the preferential mechanism by which faults in this part of the Gulf of Mexico enlarged their surface area during growth.
In the vicinity of dip-slip and strike-slip faults the volumetric strains associated with movement are often apparent as combinations of smaller-scale stratal folding and brittle failure adjacent to the slip plane. Such patterns are widely recognized, most frequently described in association with normal faults exposed along strike (e.g. Hamblin 1965). However, the often extensive and clear exposure of normal faults in plan view at the Earth's surface contrasts with their typically poor and limited exposure in cross-section. Far greater attention has, therefore, been paid in previous studies to studying the structure of such faults along strike than in the dip direction, although good cross-sectional exposures have been documented in a small number of cases (e.g. Muraoka & Kamata 1983; Peacock & Zhang 1994). In either case, however, the inability to qualify the behaviour along strike of those faults exposed in cross-section, and vice versa, has necessarily precluded complete description of fault structural character in all three dimensions. With the relatively recent availability of large, good-quality 3D seismic data sets, it is now possible to interpret faults in three dimensions,
Using this technique, examples have been described in which small-scale deformation adjacent to fault slip surfaces has also been imaged (e.g. Jones & Knipe 1996; Hesthammer & Fossen 1997; Hesthammer 1998). Constrained, however, by the quality of the seismic data and the subsequent resolution limitations, these descriptions have largely been qualitative and limited in scope. The aim of this paper is to describe a detailed study of exceptionally well-resolved stratal folds found adjacent to a series of normal faults that are imaged in a good-quality, high-resolution 3D seismic data set from the Gulf of Mexico. Careful mapping throughout the volume of the available seismic data has enabled the position, the orientation and the magnitude of the folds to be determined precisely. In this way, a pattern of folding was established for the footwall and hangingwall stratal terminations over extensive areas of the slip plane of each fault. The results are presented and discussed here with reference to a number of representative examples. They show that the structure and magnitude of the folds is characterized by considerable spatial variability in both
From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 115-128. 1-86239-060-6/00/S15.00 C The Geological Society of London 2000.
116
C. S. MANSFIELD & J. A. CARTWRIGHT
Fig. 1. Two-way travel time contour map of an intermediary level horizon interpreted from the seismic data. After Mansfield & Cartwright (1996). Survey dimensions are approximately 5 x 10km and the contour interval is 6ms. The labelled faults, 1-3, have been selected for discussion of their throw and wall-rock fold patterns: see text. Inset shows a location map of the survey area.
FOLD PATTERNS ADJACENT TO NORMAL FAULTS
the foot wall and the hangingwall of the faults. Evidence is also presented which shows that no direct correlation is readily apparent with the distribution of throw on the faults. Neither do they exhibit any apparent preferred orientation or extensive alignment. Based on these results, and in the light of other published examples, the paper concludes with a discussion of possible explanations for the origin of these folds and the causes of their spatial variability. The observations presented here represent a very recent discovery for which a singular rationale is not readily apparent. The intention is not, therefore, to argue for any particular mechanism. Rather, we wish to examine a range of plausible causes that may in future be used as a starting point to define where and under what conditions such folds occur and, also, as a basis to interrogate fault-related fold structures mapped elsewhere. Geological setting The faults discussed in this paper are those for which mapped distributions of throw were presented in a separate, earlier study by Mansfield & Cartwright (1996). The reader is referred to this paper for a more complete description of the geological setting of the faults. Only the most salient points with respect to the study of the fault folds are reiterated here. The data set used in this study is part of a highresolution, 3D seismic survey provided by Schlumberger Geco-Prakla. Located offshore Louisiana, in shallow water (Fig. 1), the survey area covers a part of the continental margin dominated by a series of interbedded deltaic sands, silts and clays of predominantly Late Cenozoic age (Murray 1951; Burgess 1976; Jackson & Galloway 1984; Galloway 1986). In the seismic data these interbeds appear as a highly reflective sequence with good lateral continuity. As a result, individual faults are exceptionally well defined (Fig. 2). All of the imaged faults in the data set are normal offset, i.e. there are no reverse faults or other compressional features evident within the survey area. A thickening of strata from footwall to hangingwall, recording fault growth contemporaneous with sediment deposition, is indicated by the seismic data at most intervals across the planes of all of the faults in the survey area. The largest expansion ratios (Thorsen 1963) measured across the faults average about 1.1, with a maximum value of 2.7 recorded over a limited stratigraphic interval. All of the faults in the survey area appear marginally listric in the deepest parts of
117
the data set, becoming subvertical near welldefined upper tips. They also exhibit considerable variations in maximum throw from one to another, the largest recorded being 160 milliseconds (ms) one-way travel time (OWTT) on Fault 2 (Fig. 1). Applying an average velocity of 3500ms"1 over the entire depth range of the fault this corresponds to a throw of 560 m. All of the mapped fault displacement fields were found to be modified by a gentle folding of the wall-rock immediately adjacent to each of the fault planes. Most commonly these folds are developed in both the footwalls and the hangingwalls of the faults (Fig. 2). The quality of the seismic data is such that these features can be reliably interpreted throughout the survey area and across the entire imaged vertical and lateral extents of each fault. Confident interpretation is possible even in the deepest parts of the data set at 4000ms two-way travel time (TWTT). Each of the folds represents a local deformation of the strata where they are truncated by the fault plane, and in general are seen on seismic time slices to have evolved with their long axes subparallel to fault strike. Generally they represent only a small proportion of the fault displacement at any location. Nevertheless, they are highly variable both in magnitude and in style, accounting for up to 11 ms deflection of bedding across zones stretching up to 500 m from some of the largest fault planes (Fig. 3). Fold mapping Interpretation of the faults and adjacent folds in this data set was carried out on a workstation using Landmark seismic interpretation software. Following established convention, the mapped folds were defined either as 'normal' or 'reverse', depending on their orientation with respect to the dip-slip displacement direction of the faults (Fig. 4). Normal folds were recorded in those places where strata in the immediate vicinity of the fault plane are deflected in a direction opposite to that of the displacement direction (Fig. 4a). Reverse folds were then defined in those places where the strata are deflected in the same sense as the displacement direction on the fault (Fig. 4b). The magnitude and spatial distribution of the folds was established by taking a series of strike-normal profiles across each of the faults, approximately 50-100m apart. Along the fault slice on each profile, the fold amplitudes were then recorded at closely spaced intervals of 2550ms, as a deflection of bedding (d) in ms (Fig. 5). Where reverse folding occurred a
118
C. S. MANSFIELD & J. A. CARTWRIGHT
Fig. 2. Seismic cross-section showing fault structure in the southern half of the survey area. After Mansfield & Cartwright (1996). Inset section shows the typical detail of folding adjacent to the fault planes apparent in the data. Assuming an average velocity of 3500ms" 1 over the entire interval, the vertical exaggeration in this figure is approximately 1.5.
FOLD PATTERNS ADJACENT TO NORMAL FAULTS
Fig. 3. Three-dimensional perspective view showing the typical variability of stratal folding adjacent to faults in the survey area. The example presented shows an horizon offset across Fault 1, represented as a wireframe surface based directly on the interpreted data. The vertical scale is in milliseconds (ms), two-way travel time (TWTT).
Fig. 4. Schematic diagram illustrating the classification of (a) normal folding, and (b) reverse folding adjacent to a normal fault.
119
Fig. 5. Schematic illustration showing how the amplitudes of folds adjacent to the faults in this data set were recorded. Reverse folds were assigned a negative amplitude (d\), as the deflection of strata increases the apparent throw on the fault at these sample locations. Similarly, normal fold amplitudes were assigned positive values (d2) as the deflected strata reduce the apparent throw.
negative amplitude was recorded, as the stratal deflection acts to increase the net value of displacement measured across the fault. Conversely, at those sample locations where normal folding occurred a positive value was reported. To minimize spatial distortion of the distributions resulting from curvature of the fault planes along strike these data were then projected onto strike-parallel vertical planes for contouring (Fig. 6). The distributions are bounded in each case laterally by the edges of the data set, by a branch line with a neighbouring fault or by a lateral tip-line. At the base they are bounded by a lower limit of measurement and at the top by an upper tip-line. Tip-lines represent zero displacement contours defined by the limit of displacement that can be resolved from the seismic data, estimated in this case to be no more than about 15m. The only anticipated error in the recorded fold amplitudes arises from the accuracy with which travel times could be measured from the peaks or zero crossings of the seismic reflections, at each of the sample locations on the fault planes. This is estimated to be approximately ± l m s OWTT, i.e. about 4m. Additional error associated with the depth of each of the folds arises because of uncertainty in the positions of the stratal terminations adjacent to the faults as represented by the seismic wavelets. Because expansion ratios across the faults are, in general, relatively small, it is assumed that seismic wavelets in the footwalls and hangingwalls, representing correlative sedimentary units across the faults, are identical. Consequently, the error is simply a function of the frequency content of
120
C. S. MANSFIELD & J. A. CARTWRIGHT
the seismic data. In this data set it ranges between approximately 45 and 80 Hz. Assuming an average velocity of 3500 m s"1 over the entire interval of interest, this corresponds to positioning errors of between 11 and 20m, the average being approximately equivalent to ±15m. These positioning errors do not introduce any additional uncertainties into the recorded fold amplitudes as they are expected to be identical at corresponding positions in the footwalls and hangingwalls. Fold description The stratal folds recognized in this data set have been mapped across most of the faults in the survey area. Despite the complex and diverse patterns found in this way, each of the mapped distributions shares a number of fundamental similarities. These similarities are discussed below with reference to three of the longest faults that transect the survey area. They are the E-W-striking fault in the southernmost part of the survey area (Fault 1), the arcuate fault in the centre of the survey area (Fault 2) and the NE-SW-striking fault in the northeast-
ern part of the survey area (Fault 3), as indicated in Fig. 1.
Spatial distribution of stratal folding ('drag folding') Generally referred to as 'drag folding*, minor stratal folds have been well resolved by the seismic data adjacent to the slip planes of all the faults in the survey area. When mapped in detail it is immediately apparent that these folds have evolved with considerable spatial variability (Fig. 6). There are numerous but distinct changes in fold style along both the dip and strike directions of the faults. Patterns clearly alternate from reverse folding to normal folding, to areas of the fault planes adjacent to which no perceptible folding is developed at all. Furthermore, it is clear from the many small isolated closures on the fault plane projections, evident for example at point A on Fault 1 (Fig. 6a) and at point B on Fault 3 (Fig. 6b), that most of these changes occur rapidly, sometimes over distances as short as 100m. Within the resolution of the seismic data, the closures that define the spatial extent of each
FOLD PATTERNS ADJACENT TO NORMAL FAULTS
121
Fig. 6. Vertical strike projections showing the distribution of (i) footwall and (ii) hangingwall stratal folding adjacent to three example faults, 1-3. See Fig. 1 for location. The faults are all plotted at the same scale, each with a horizontal exaggeration of approximately 2.25. Labels A-H are referenced in the main text.
122
C. S. MANSFIELD & J. A. CARTWRIGHT
Fig. 7. Vertical strike projection of Fault 1 showing the variability of the cumulative fold amplitudes, i.e. footwall and hangingwall fold amplitudes summed at each sample location. The complexity of this distribution is typical of those mapped on all the faults in the survey area. The summing method assumes positive values for normal folds and negative values for reverse folds; see Fig. 5 and text for details. Contours are in milliseconds (ms), one-way travel time (OWTT). The fault is plotted with a horizontal exaggeration of approximately 1.25.
fold type exhibit very irregular patterns. Systematic or repeated shapes are nowhere apparent within any of the mapped distributions, and no correlations can be established between either the footwall or the hangingwall patterns of any given fault or between those of different faults. Particularly, there appears to be no correlation in depth at which the different folds occur. For example, the large area of normal folding at point C in the footwall of Fault 3 (Fig. 6c) corresponds to a region of short and discontinuous folds, of both normal and reverse polarity, at the equivalent position in the hangingwall. The correlation is equally poor with the fold patterns recorded at the same depths on Faults 1 and 2. The predominantly dark coloration of Fault 2 indicates that normal folding dominates the near-field deformation in both the footwall and hangingwall of this fault. In addition, there are large areas on the fault where no folding could be resolved at all. Very little reverse folding was found anywhere. In general, however, there is no clear bias in the development of the folds on any given fault. There is no systematic pattern that can be established with the folding distribution in relation to the shape, the orientation or the positions of the faults within the survey area. For example, the relative absence of resolved reverse folding adjacent to Fault 2 is repeated in the footwall of Fault 3. This fault has an entirely different shape, trend and throw
direction to Fault 2. Furthermore, reverse folding is strongly present in the hangingwall of this fault, whereas it is almost entirely absent in that of Fault 2. The complexity evident in the fold patterns is, in part, caused by many offset sedimentary intervals having contrasting fold styles at their respective footwall and hangingwall stratal terminations. A further consequence of this is that the summation of correlative fold amplitudes at each of the sample locations on the faults demonstrates equally non-uniform behaviour. An example of this is presented in Fig. 7, where the cumulative contribution of the fold geometries to the measured throw at each sample location on the plane of Fault 1 is expressed in contours of one-way travel time, in ms. Positive values are recorded at locations where the dominant folding style is normal. Conversely, where a reverse folding style dominates the contour values are negative. The poor correlation between the spatial distributions of footwall and hangingwall fold geometries is reflected in the appreciable complexity that characterizes this cumulative distribution. In particular, the pattern is dominated by numerous local positive and negative contour closures that exhibit no apparent spatial clustering, alignment or other related geometrical ordering. Similar complex cumulative distributions were recorded adjacent to the slip planes of all the mapped faults in the survey area.
FOLD PATTERNS ADJACENT TO NORMAL FAULTS
123
Fig. 8. Vertical strike projections showing the distribution of throw on the three example faults, 1-3. Throw is contoured in milliseconds (ms), one-way travel time (OWTT), and labels I-M refer to points in the main text. The faults are all plotted at the same scale, each with a horizontal exaggeration of approximately 2.25.
The complex fold distributions described above appear to be genuine. Although many of the folds are small and have low amplitudes, the high sampling density and clarity of the seismic data have ensured reliable definition of their true spatial extent. Nevertheless, despite
the resolution achieved, few systematic patterns are apparent in the arrangement of the mapped folds. Close inspection does show, however, that sometimes the different fold styles occur as subhorizontally aligned, elongated but discontinuous regions adjacent to the fault planes. For
124
C. S. MANSFIELD & J. A. CARTWRIGHT
example, point D in the footwall of Fault 1 indicates a subhorizontal zone of normal folding at approximately 1700ms TWTT. This aligns with other regions of normal throw on the fault at the same depth. Similarly at point E on Fault 2 and at point F on Fault 3. Numerous other examples are also apparent in the distribution of reverse folding on each of the faults. Nevertheless, the strongly contrasting spatial distributions of the different fold styles exhibited by each of the faults determines that such mapped alignments rarely correspond in the footwall and hangingwall of the same fault. Neither is any relationship apparent between the positions of the various aligned folds on the different faults. For example, the strong alignment of normal folding at 2000ms in the hangingwall of Fault 2 (point G) corresponds to regions of discontinuous, mixed folding styles at equivalent positions in the hangingwalls of Faults 1 and 3.
Comparison with throw patterns The respective distribution of throw for each fault (Fig. 8) was established using a technique similar to that employed to map the fault folds described above. The complete details of this procedure are discussed by Mansfield & Cartwright (1996). For comparative purposes, the results are presented here as strike-parallel vertical projections for the three example faults shown in Fig. 6. In each case the patterns of throw distribution are expressed as contours of one-way travel time, in ms. The contours of the throw projections define complex distributions that are typical of all the faults in the survey area. In each case they show that the faults are characterized by a general increase of throw with depth, but in addition they indicate that the mapping has also resolved a series of numerous local, higher frequency throw variations. Each of the distributions are modified by these frequently steep changes in the throw gradient, most of which are believed to be genuine and which represent local perturbations of the fault displacements (Mansfield & Cartwright 1996). Comparing the recorded throw distributions with the patterns established for the respective footwall and hangingwall stratal folds indicates only very poor apparent spatial correlations. Although subhorizontal alignments are often apparent in both, a definitive correspondence between the two patterns can be identified only in a small number of cases. For example, a clear change from normal to reverse folding at
point H in the footwall and hangingwall of Fault 2 (Fig. 6b) corresponds to a localized increase in throw (point I in Fig. 8). In this instance, the shape of the throw contours closely matches that of the boundary defining the change in the style of folding. In most cases, however, such recognized correspondence between the two patterns at one point on a fault plane is usually contradicted elsewhere on the same fault. This is largely as a consequence of the strongly contrasting fold patterns that have developed in the footwalls and hangingwalls of the faults. For example, the local increase in throw at point J on Fault 1 (Fig. 8a), defined by closure of the 16ms and the 18ms (OWTT) contours, is not mirrored by any similar patterns in the mapped folds in either the footwall or the hangingwall (Fig. 6a). Likewise, the partial throw contour closures recorded along the deepest mapped parts of Fault 1, at points K. L and M, occur where there is predominantly reverse folding in the hangingwall but discontinuous mixed folding styles in the footwall. Discordance between the two sets of distributions is further apparent along the upper tip-lines of all the mapped faults. Here the throw contours are subhorizontal and are broadly parallel. However, these parts of all the faults are also typified by significant variations in the folding style along strike. Discussion Folded wall-rock strata have been recognized adjacent to the planes of all the faults in this data set. Detailed mapping has established that the folds occur with different amplitudes and with complex spatial distributions where they are truncated against the fault planes. The folds also occur with differing orientations, that can be classified as 'normal' or ^reverse' with respect to the normal offset displacement direction of the faults (see Fig. 4). The normal folding is probably synonymous with normal drag, that is frequently described in association with field exposures and seismic sections of normal faults in cross-section (Hobbs et al. 1976). Different mechanisms have been proposed to explain the occurrence of normal drag. These include the shearing of a monocline that grows ahead of a radially expanding fault tipline (Chamberlain & Salisbury 1912; Walsh & Watterson 1987) and systematic variations of compaction in the hangingwall with distance from a fault plane, as a function of sedimentary thickness (Coward 1992). The most widely held belief, however, is that its development is a
FOLD PATTERNS ADJACENT TO NORMAL FAULTS
mechanical response by wall-rock lithologies to the frictional resistance to sliding on a fault plane (Ramsay & Huber 1987). Potentially, all three of these mechanisms may have contributed to the normal folding of strata adjacent to the faults in this data set. Reverse drag is regarded to be the geometrically necessary, ductile deformation which maintains continuity in the volume surrounding a fault (Barnett et al. 1987) whose cumulative sum adjacent to the fault is therefore equivalent to displacement on the fault surface (Hamblin 1965; Barnett et al 1987; E. Willemse pers. comm.). Of various explanations for the origins of reverse drag reviewed by Hamblin (1965), the most likely was considered to be a response to listric fault geometries at depth. In this case the hangingwall subsides to fill a void that develops along the upper reaches of a fault, where the slip vector of the deeper detachment surface is not parallel to the fault plane. This idea is integral to a number of models of extensional faulting (e.g. Gibbs 1983, 1984). However, recent theoretical studies (Gibson et al. 1989; Ma & Kusznir 1993) and modelling of crustal-scale faulting (King et al. 1988; Stein et al. 1988), in which faults are treated as dislocations in an elastic half-space, have shown that listric fault plane geometries are not an essential prerequisite for reverse drag to develop. In agreement with earthquake data they confirm that reverse drag can also be characteristic of planar faults (Stein & Barrientos 1985; King et al. 1988; Stein et al. 1988). Subsidence in the hangingwall is explained as being largely an elastic response to coseismic slip (Roberts & Yielding 1994), whilst uplift in the footwall it is thought to be an expression of isostatic rebound (e.g. Jackson & McKenzie 1983; King et al. 1988; Stein et al. 1988). It is argued by Barnett et al. (1987) that nearfield stratal geometries similar to those discussed above should also be characteristic of synsedimentary normal faults, of the type mapped in the seismic data set described here. Clearly, however, the elastic properties of poorly or partially compacted sediments cannot be compared with those of lithified rock (Lambe & Whitman 1979). Consequently, reverse folding recorded adjacent to the faults in this data set should probably not be directly equated with reverse drag found elsewhere. In this data set, the recorded fold geometries are spatially impersistent and where present the amplitudes are nowhere large enough to account for measured offsets across the faults. That is, there is no systematic relationship between the magnitude of the reverse folding and the fault displacements.
125
It therefore represents only a small proportion of the volumetric strains that must exist in the immediate vicinity of the fault plane surfaces. The reverse folding does, however, share a number of characteristics in common with those previously described for reverse drag. In particular, in the hangingwall it is generally better developed on the lowermost parts of a fault plane, whilst in the footwall it is usually most pronounced on the uppermost parts (Barnett et al. 1987). Where present, reverse folding in the footwalls is commonly much more weakly developed than in the hangingwalls, i.e. it has much smaller amplitudes (Hamblin 1965, Gibson et al. 1989). In general, it is also much less persistent along strike in the footwalls of the faults than in their hangingwalls, such that the spatial distributions of each appear completely independent (Hamblin 1965). To account for the observed spatial variability of the different folding styles adjacent to the planes of faults in this data set a number of additional factors need also to be considered: (1) the contrasting mechanical responses of different lithologies to movement on the faults; (2) variable spatial and temporal patterns of friction on the fault planes; and (3) the influence of fault plane topology. All are potentially important elements governing near-field deformation patterns during consecutive slip events, and at certain localities may even be interdependent. For example, rheological contrasts in multilayer lithologies have been shown to have important consequences for shaping fault plane geometries (Peacock & Sanderson 1992). Equally, variations in shear strength between different lithologies can dictate the preferential development of fault clustering and fault zone thickening, which subsequently can be expected to influence the extent to which the wall-rocks develop strain hardening or strain softening during fault growth (Sibson 1993). This is likely to constrain the spatial distribution and the magnitude of folding, but its influence on the preferential development of one particular folding style is, however, less apparent. The same is true also of the influence of fault plane friction, the magnitude and spatial extent of which will largely be a function of the juxtaposition of different lithologies, and the shear and re-healing of fault plane asperities (Scholz 1990). Irregularities in fault plane topology could, however, govern both the magnitude and the style of folding (Fig. 9). Based on a detailed study of irregularities in the fault throw distributions (Mansfield & Cartwright 1996) it has previously been argued that linkage in the dip direction, or 'dip linkage', between individually
126
C. S. MANSFIELD & J. A. CARTWRIGHT
Fig. 9. Schematic illustration showing potential mechanisms for generating different drag fold geometries at (a) right-stepping (extensional) and (b) left-stepping (compressional) fault overlaps. The folds are generated where strata are displaced across the fault plane topological irregularities represented by the overlap zones; see text for details.
propagating precursor segments was prevalent during the evolution of the faults in this area. In growth fault settings such as the US Gulf Coast and the Niger Delta, this style of segmentation in the dip direction is likely to be further perpetuated in those instances where small segments propagate and overlap ahead of the upward expanding tip-lines of growing faults. Providing that sediment supply continues to outpace movement on an established fault, then each slip event will be preceded by a period of time in which a previously faulted horizon is blanketed in a thick succession of poorly consolidated material. Propagation of the upper tip-line into this material is likely to be accommodated first by monoclinal flexuring, or 'forced folding' (Stearns 1978; Price & Cosgrove 1990) in the region ahead of the tip, analogous to a 'ductile bead' surrounding the leading edge of blind, propagating thrust faults (Elliott 1976;
Williams & Chapman 1983). Within the region of flexure, small discrete slip surfaces nucleate at positions of weakness within the material, enabling enlargement of the main fault slip plane to be achieved principally by coalescence at the upward propagating tip. This can be expected to be a characteristic of growth throughout fault activity provided that sediment supply continues to outpace movement on the fault. As neighbouring fault segments are rarely coplanar, overlap and linkage introduces a local change in fault dip. That is, at the relay structure the slip vector is no longer subparallel to the fault plane. Where strata are displaced over such topological irregularities, additional footwall and hangingwall strains will be imposed. Subsequent faulting and folding should be localized, although it can be expected that different fold styles develop at extensional and compressional overlaps (Peacock & Zhang 1994) (Fig. 9). The subhorizontal alignment recognized in some of the patterns of folding are consistent with this idea, as it has previously been argued that the dip-linkage structures are also subhorizontal over extensive areas of the fault planes (Mansfield & Cartwright 1996). Implicit in this mechanism is a correlation between the mapped regions of dip-linkage and the development of the different fold styles. As previously noted, however, such a correlation is not apparent in the data. One possible explanation for this disparity is illustrated in Fig. 10. At dip-linkage relay structures the fault plane is locally no longer parallel to the fault slip vector. Therefore, in those instances where the displacement of strata across a fault exceeds the dip-dimension of a given relay structure, a reconfiguration of the folds could occur in the presence of additional stresses. As shown in Fig. 10, voids can be expected to develop at extensional overlaps as juxtaposed strata are pulled apart at the relay during continued slip on the fault. Collapse of the hangingwall into this void where the stresses are locally relieved may appear at the scale of seismic data as downward folding in the dip direction. Conversely, a step in the fault plane at compressional overlaps will represent a residual asperity, causing local faulting and folding where the strata are forced to slide across it. It is envisaged that these same arguments can be extended to deformation of footwall strata at extensional and compressional overlaps in those cases where there is significant footwall uplift. The observed stratal folding in the wall-rocks of the faults described in this data set is expected to be essentially a ductile behaviour that can be expected to be dependent largely on ambient
FOLD PATTERNS ADJACENT TO NORMAL FAULTS
Fig. 10. Schematic illustration depicting a change in stratal folding at an extensional fault overlap. In this case the relay has breached and normal folding is preserved at the stratal termination in the footwall (cf. Fig 9a). However, dilational strains in the hangingwall, at the void where the fault plane pulls apart, dictate a progressive change from normal folding to reverse folding as displacement on the fault increases.
strain rate and the presence of fluids. Under these circumstances additional heterogeneities should be expected. Ruptures of different magnitude and periods of stable sliding may also produce contrasting fold geometries. Where these overlap, interference can be expected to produce complex patterns. The superimposed effects of lithological contrasts and variabilities in fault plane friction are likely to dictate that over most parts of fault plane surfaces there is unlikely to be direct, simple correlations describing the spatial distributions of the different fold styles. Modification by continued slip and variations in the topological expression of relay structures at different stages of linkage will also introduce additional heterogeneities into any otherwise simple spatial relationship. Conclusions • Using high-resolution 3D seismic data, numerous low-amplitude stratal folds have been mapped in the footwalls and hangingwalls of an array of growth faults in the Gulf of Mexico. On the basis of their orientation with respect to the displacement direction of the faults, these folds have been classified either as normal or reverse. • Detailed mapping has revealed that both types of fold are characterized by complex spatial distributions. High-frequency changes in folding style are common in both the dip and strike directions of the faults, defining domains in which folds of one particular type are confined to small areas of a fault plane. Rarely, some of
127
the folds occur in subhorizontal, discontinuous bands but nowhere else is any alignment or other geometrical ordering apparent in the data. Despite the complexity of the mapped distributions, each of the patterns is believed to be genuine and to reflect the true spatial definition of the folds, as a consistently high sampling frequency was maintained throughout the mapping. • The complex fold geometries have been found to exhibit no simple spatial relationship with the respective distributions of throw. Neither do they coincide with suspected regions of dip-linkage on the faults. The absence of any apparent correlative relationship can probably be explained by the multiplicity of potential physical factors that govern the development and interaction of the different fold styles. • Comparable fold geometries adjacent to all of the faults in the survey area lead to the conclusion that these complex patterns are probably a fundamental characteristic of the evolution of all large growth faults in this part of the Gulf of Mexico. We propose a mechanism for fault development in which growth is achieved primarily by coalescence at the upward propagating tip-line between coplanar, precursor fault segments. These small fault segments develop in response to the monoclinal flexuring of each successive sedimentary pile deposited between individual slip events on the main fault. Folding of the adjacent wall rock is then expected to be governed largely by variations in the topological expression of compressional and extensional relay structures at different stages of fault segment linkage. Thanks are due to Schlumberger Geco-Prakla for kindly providing the seismic data set upon which this work is based and for their permission to publish the results. Shell International Petroleum Company Ltd (C. S. Mansfield), Fina UK Ltd and the Nuffield Foundation (J. A. Cartwright) are thanked for their financial support of this research. We gratefully acknowledge the helpful review of the original manuscript by an anonymous referee.
References BARNETT, J. A. M., MORTIMER, J., RIPPON, J. H., WALSH, J. J. & WATTERSON, J. 1987. Displacement geometry in the volume containing a single normal fault. AAPG Bulletin, 71, 925-937. BURGESS, W. J. 1976. Geologic evolution of the midcontinent and Gulf Coast areas a plate tectonics view. Transactions of the Gulf Coast Association of Geological Societies, 26, 132-143.
128
C. S. MANSFIELD & J. A. CARTWRIGHT
CHAMBERLAIN & SALISBURY. 1912. ///: GEIKIE. A. Structural and Field Geology. Oliver and Boyd. Edinburgh. COWARD, M. P. 1992. Structural Interpretation With Emphasis on Extensional Tectonics. JAPEC Course Notes 124. Geological Society, London. ELLIOT, D. 1976. Energy balance and deformation mechanisms of thrust sheets. Philosophical Transactions of the Roval Society, London, A283. 289-312. GALLOWAY. W. E. 1986. Growth faults and faultrelated structures of prograding terrigenous clastic continental margins. Transactions of the Gulf Coast Association of Geological Societies, 36. 121-128. GIBBS, A. D. 1983. Balanced cross-section construction from seismic sections in areas of extensional tectonics. Journal of Structural Geology, 5, 153-160. 1984. Structural evolution of extensional basin margins. Journal of the Geological Society, London, 141, 609-620. GIBSON. J. R., WALSH, J. J. & WATTERSON. J. 1989. Modelling of bed contours and cross-sections adjacent to planar normal faults. Journal of Structural Geology, 11, 317-328. HAMBLIN, W. K. 1965. Origin of'reverse drag" on the downthrown side of normal faults. Bulletin of the Geological Society of America, 76, 1145-1164. HESTHAMMER. J. 1998. Evaluation of the timedip. correlation and coherence maps for structural interpretation of seismic data. First Break, 16, 151-167. & FOSSEN, H. 1997. Seismic attribute analysis in structural interpretation of the Gullfaks field, northern North Sea. Petroleum Geoscience, 3, 13-26. HOBBS. B. E.. WINTHROP. D. M. & WILLIAMS. P. F. 1976. An Outline of Structural Geology. Wiley. New York. JACKSON. J. & MCKENZIE. D. 1983. The geometrical evolution of normal fault systems. Journal of Structural Geology, 5. 471-482. JACKSON, M. P. A. & GALLOWAY, W. E. 1984. Structural and depositional styles of Gulf Coast Tertiary continental margins: Application to hydrocarbon exploration. Continuing Education Course Note Series 25. American Association of Petroleum Geologists. JONES. G. & KNIPE, R. J. 1996. Seismic attribute maps; application to structural interpretation and fault seal analysis in the North Sea Basin. First Break, 14,449-461. KING. G. C. P.. STEIN, R. S. & RUNDLE, J. B. 1988. The growth of geological structures by repeated earthquakes. 1. Conceptual framework. Journal of Geophysical Research, 93, 13,307-13,318. LAMBE, P. C. & WHITMAN, R. V. 1979. Soil Mechanics. McGraw-Hill, London. MA, X. Q. & KUSZNIR, N. J. 1993. Modelling of nearfield subsurface displacements for generalized faults and fault arrays. Journal of Structural Geology, 15, 1471-1484.
MANSFIELD, C. S. & CARTWRIGHT, J. A. 1996. High resolution fault displacement mapping from three-dimensional seismic data: evidence for diplinkage during fault growth. Journal of Structural Geology, 18. 249-263. MURAOKA, H. & KAMATA. H. 1983. Displacement distribution along minor fault traces. Journal of Structural Geology. 5. 483-495. MURRAY. G. E. 1951. Sedimentary volumes in Gulf Coastal plain of the United States and Mexico. Part 3: Volume of Mesozoic and Cenozoic sediments in central Gulf Coastal plain of the United States. Bulletin of the Geological Society of America, 63. 1177 1192. PEACOCK. D. C. P. & SANDERSON. D. J. 1992. Effects of layering and anisotropy on fault geometry. Journal of the Geological Society. London. 149. 793-802.' & ZHANG, X. 1994. Field examples and numerical modelling of oversteps and bends along normal faults in cross-section. Tectonophysics. 234, 147167. PRICE. N. J. & COSGROVE. J. W. 1990. Analysis of Geological Structures. Cambridge University Press. Cambridge. RAMSAY. J. G?& HUBER. M. I. 1987. The Techniques of Modern Structural Geology. Academic Press Inc.. London. ROBERTS. A. & YIELDING. G. 1994. Continental extensional tectonics. In: HANCOCK, P. L. (ed.). Continental Deformation. Pereamon Press. Oxford. 223-250. SCHOLZ, C. H. 1990. Mechanics of Faulting and Earthquakes. Cambridge University Press. Cambridge. SIBSON. R. H. 1993. Load-strengthening versus loadweakening faulting. Journal of Structural Geology. 15. 123-128. STEARNS, D. W. 1978. Faulting and forced folding in the Rocky Mountains foreland. Bulletin of the Geological Society of America, 151. 1-37. STEIN. R. S. & BARRIENTOS, S. 1985. Planar high-angle faulting in the Basin and Range Geodetic analysis of the Borah Peak. Idaho earthquake. Journal of Geophysical Research. 90. 11.355-11.366. .KING. G. C. P. & RUNDLE. J. B. 1988. The growth of geological structures by repeated earthquakes. 2. Field examples of continental dip-slip faults. Journal of Geophysical Research. 93. 13.31913.331. THORSEN. C. E. 1963. Age of growth faulting in southeast Louisiana. Transactions of the Gulf Coast Association of Geological Societies. 13. 103-110. WALSH. J. J. & WATTERSON. J. 1987. Distribution of cumulative displacement and seismic slipon a single normal fault surface. Journal of Structural Geology. 9. 1039-1046. WILLIAMS. G. & CHAPMAN. T. 1983. Strains developed in the hanging walls of thrusts due to their slip propagation rate: a dislocation model. Journal of Structural Geology. 5. 563-571.
Effects of interlayer slip in model forced folds G. D. COUPLES & H. LEWIS Department of Petroleum Engineering, Heriot-Watt University, Edinburgh EH 14 4AS, UK. (e-mail:
[email protected]/
[email protected]) Abstract: Forced-fold models, constructed from multiple sheets of rock layers and other materials, are deformed, under confining pressure, by the translations and rotations of a steel-block forcing assembly. The first-order response of these models is like that observed in earlier sets of models: an asymmetric fold pair develops over the uplifted edge of the rigid, primary forcing block; the layers translate into the folding area without being pushed; and bending deformation imparts spatially variable strains onto the layered package. The second-order response of the new experiments is quite different from earlier work - the new multilayered models develop temporally variable bending strains caused by the progressive development of interlayer slip on some of the layer-layer interfaces. The presence of the layer-parallel slip surfaces reduces the bending resistance, and it causes the resulting folds to be more localized and to have different shapes than in similar experiments where layering is less effective. In the multilayer models, faulting in the layered package is much reduced compared to similar models without multiple layering, and material strains are everywhere smaller in the layers. Layer contractions and elongations of the layering are produced in these forced-fold models as a consequence of the flexure process, without regard to farfield causative loads. The models reveal a progression of deformation consequences that is related to the mechanical effectiveness of the layering. If the understanding of process gained from these models is extended to natural folds, it becomes possible to explain both the observed differences of fold shape and some of the variations in subsidiary deformations.
Our purpose is to investigate how interlayer slip affects the development of forced-fold models, and to explore how this knowledge impacts the interpretation of forced folds in nature. We use a modelling approach developed by Weinberg (1979) and Couples et al (1994), which was derived from earlier investigations (e.g. Handin et al 1972, 1976; Friedman et al 19760). In our experiments, packages of layers are forced folded under confining pressure by the translations and rotations of an assembly of machined steel blocks (Fig. 1). The layered packages are mechanically isolated from the remainder of the experimental apparatus (in particular, they are isolated from the pistons), so the folds that are created can be attributed completely to the process of flexure - that is, they are ideal forced folds in the sense of Stearns (1978). The layered packages that comprise our models were made of a variety of materials, including machined layers of rock and sheets of lead; we also used paper card-stock, gypsum crystals and potter's clay. Some layer interfaces were lubricated to enhance the potential for slip. Our motivation for this study derives from a wider investigation of the flexural-slip process. In natural folds, discrete displacements commonly occur on only some bedding planes (Tanner 1989; Lewis & Couples 1993; Cooke et al this volume), leading to the ordered and hierarchical partitioning of both stress and
strain (Couples et al 1998; and references therein). The selective activation of layer-parallel slip is also observed in numerical models of the flexural-slip process (Couples et al 1996; Cooke & Pollard 1994, 1997; Cooke 1997; Cooke et al this volume). The focus of this study is on determining the effects resulting from the availability and activation of layer-parallel slip surfaces in forced-fold models. By comparing our results with the responses of otherwise-similar models (but without significant layering), we can deduce that the observed differences in response are attributable to the enhanced degree of layering, and hence to flexural slip. In common with other model studies, our results lead to greater understanding of fold processes, and this knowledge can be used to develop better hypotheses concerning the development of natural forced folds. Experimental design
Previous work Physical models of forced folds using rock materials, which are deformed under confining pressure, have been the subjects of a series of earlier contributions. Friedman et al. (\916a, b, 1980) and Gangi et al. (1977) describe a set of experiments in which thick rock layers (and
From'. COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 129-144. 1-86239-060-6/OO/S 15.00 O The Geological Society of London 2000.
130
G. D. COUPLES & H. LEWIS
Couples et al. (1994) developed a mechanical interpretation of the structural events in the forced folding of their models. This development can be summarized as: Fig. 1. Schematic drawing of pistons and deformed model (all located inside the pressure vessel for experiment). The top layer package is drawn with three layers (like the configuration described by Couples et al. 1994). The bottom package is wholly schematic. Each of the four machined steel blocks are shaded differently. Note especially the plasticine plugs between the layers and the pistons. Model + plasticine is enclosed by lead side-jackets and plastic heat-shrink tubing.
packages of layers) are deformed by high-angle reverse motion on a 'basement' fault that is pre-cut into a strong sandstone block (which acts as a forcing member when shortened by the advance of pistons). These models produce deformation that is dominated by faulting in the layer(s), although some of the later models (Friedman et al. 1980) produce a greater degree of folding (instead of faulting) by adding a ductile basal layer and/or by enhancing the slip of the layers relative to the forcing blocks. Folds that look much more like the shapes observed in classic forced folds - such as at Rattlesnake Mountain, where the forcing member (basement) is a rigid, rotated block (Stearns 1971) - are produced by the experimental configuration that is reported by Stearns & Weinberg (1975), Weinberg (1978, 1979), Couples et al. (1994) and Couples & Lewis (1998). These sets of models use a steel-block forcing assembly that produces both localized uplift and rotational displacements (Fig. 1). These studies reinforce the importance of a ductile basal layer in creating natural-looking forced folds (see also Chester et al. 1988; Haneberg 1993). The structural development of such rock-layer models is revealed through the interpretation of information derived from the individual experiments that comprise a suite. Typically, such information consists of: the final geometries; the distortions of inscribed grids; the generation of inter- and intra-granular deformations; and externally recorded force-displacement records. In practice, a suite of experiments is run using initially identical models deformed to different magnitudes of structural relief. The individual results represent "snapshots' of the deformation history, from which a progression of structural changes (a deformation path) is inferred. From such experimental information, together with field observations and numerical modelling,
• the rock layers develop bending stresses (Hafner 1951; Couples 1977) represented by finite strains; • the basal ductile unit develops a pattern of differential pressure (Jamison 1979; Patton & Fletcher 1995) that causes it to flow from the uplifted block to the downthrown block; • consequently, the layered package is transported laterally into the folds; • the shapes of the forced folds are controlled primarily by the shapes of the forcing members. The first-order response of the models described here is essentially identical with these previous results, but we distinguish higher-order mechanical effects associated with increases in layering. It is important to emphasize that the layered packages in the physical models of Weinberg (1979), Couples et al. (1994), Couples & Lewis (1998) and those reported here are shorter than the steel-block forcing assembly, so that the layers do not touch the loading pistons (Fig. 1). The space between the ends of the layers and the pistons is filled with plasticine, which, under the experimental conditions (jacketed, 50MPa confining pressure, room temperature, dry), behaves like a very weak (but non-invasive) fluid, transmitting only confining pressure to the ends of the layered package. Therefore, the observed longitudinal movements of the layered package (relative to the adjacent steel blocks) must be a self-induced mechanical response of the folding/faulting process.
Composition of the new models The models described by Weinberg (1979) and Couples et al. (1994), and particularly the mechanical interpretations drawn from them, serve as a reference against which we compare the kinematic and dynamic evolution of the new forced-fold models described here. The primary difference between the new experiments, and the reference models of Weinberg (1979) and Couples et al. (1994), is that most of the models described here are constructed in such a way as to 'replace' the single, thick layers used in those previous studies by a stack of multiple, thinner layers that yields the same total thickness. Therefore, any differences between the new models and those reported earlier can be attributed to the greater degree of layering. Additional models, which exploit the potential for
INTERLAYER SLIP IN MODEL FORCED FOLDS
131
Fig. 2. Model set 1. Line drawings of the slabbed medial surfaces of deformed models 40, 42, 43 and 44, with additional detail added from thin sections. Solid black areas are void spaces that appeared during epoxy impregnation of samples. Lines at the base of each model show the top of the steel-block forcing assembly (see Fig. 1). Short lines within or cutting the layers represent small fractures or microfaults. L - limestone; D dolostone; P - lead.
enhancing the degree of layering through the use of novel materials, are also reported. In the following paragraphs we describe five sets of models that allow us to address the aims noted above. In one set of three experiments (Fig. 2, Table 1) the layered package is patterned after the configuration used by Couples et al. (1994) in its use of a basal, ductile layer of lead (very ductile) and an overlying layer of dolostone (very stiff). However, this set (models 42, 43 and 44) differs from the earlier work (e.g. model 40) in that the single layer of limestone (moderately ductile), located above the dolostone in the previous model series, is replaced with four layers of limestone. Taken together, these four layers total the same thickness as the single limestone layer in the previous model configuration. Different patterns of interlayer lubrication (with MoS2, a highpressure lubricant), and amount of imposed uplift, distinguish the three members of this set. These models are designed to: investigate the development of interlayer slip in the forced folds; assess the resulting shape of the folds; and determine the bending strength of the layered package. A second set of experiments (Fig. 3, Table 1) has a similar configuration to that described above, except that the layer of dolostone is replaced by another layer of limestone, so that these models have a package of five limestone layers overlying the basal lead unit (i.e. no stiff layer). The two models of this set are different
in that one model has lubrication on all of the layer interfaces (model 1428), while the other has no interlayer lubrication (model 1427). These experiments are designed to provide a comparison against the response of an existing model where a single, thick layer of limestone (without a dolostone layer) overlies a basal lead layer (Weinberg 1979; his specimen 7). In the next set of experiments (Fig. 4, Table 1), the perfect (010) cleavage of a gypsum single crystal is used to assess the effects of a finely divided, multilayered package. In one model, a thick 'layer' of single-crystal gypsum is used, overlying a basal layer of lead (model 47). This configuration might represent a thick, monotonously bedded succession of moderately ductile rocks (e.g. sandstones) lying above a basal ductile unit. In the other two experiments (models 49 and 52), thinner 'layers' of single-crystal gypsum are sandwiched between layers of machined, leather-hard potter's clay (above) and lead (beneath). The latter two experiments differ in that model 49 is completely unlubricated, whereas model 52 had its layer interfaces enhanced by lubricant. These two models might represent the situation where a succession of sandstones is encased by very ductile rocks. The layered packages in the remaining experiments (Fig. 5, Table 1) are dominated by sheets of lead. These models are intended to investigate if the effects that we observe, in models with multilayers of slightly ductile or of brittle rock, also occur in multilayers that are dominated by
132
G. D. COUPLES & H. LEWIS
Table 1. Configuration of models
42
Cpt OCl 11
L L L L D P
0.170 0.200 0.195 0.190 0.165 0.220
1427
Set 2
L L L L L P
Set 4
G P
L L L L D P
L L L L D P
0.190 0.185 0.195 0.195 0.160 0.220
40
0.195 0.190 0.190 0.195 0.160 0.220
L D P
0.190 0.190 0.190 0.190 0.190 0.220
L L L L L P
0.190 0.190 0.190 0.190 0.190 0.220 52
49 0.850 0.170
C G P
1.1250 0.3850 0.3500
C G P
0.9800 0.2750 0.3350
50
51
53
54
P P P
P P P
P P P P P P S S (17 total)
__ P
(10 total) P
0.730 0.155 0.230
1428
47
Set 3
44
43
1.1850 (total)
(10 total)
1.1900 (total)
P
s s p
0.6850 (total)
S P 0.370 (total) 0.325
S
1.4350 (total)
P
p
For each model, left column shows composition and lubrication; right column shows layer thickesses (in cm). L limestone layer (Indiana Limestone); D - Dolostone layer (Blair Dolomite); P - lead layer; G - gypsum crystal (layering' parallel to [001] cleavage); C - potter's clay (leather hard); S - paper card. Underlining denotes lubrication at the base of the layer: this is in the form of MoS:, except for paper-paper contacts, which use powdered graphite. In model 54, there was a total of 10 sheets of lead and 18 sheets of paper. Model 40 from Couples etal. (1994).
very ductile materials. One set of two models has a layered package consisting only of lead sheets. One of the models has lubrication on all interfaces (model 51), while the layer interfaces in the other model's package are unlubricated (model 50). If unlubricated lead sheets become welded together by the confining pressure, this latter model should deform like similar all-lead configurations reported by Weinberg (1979; his specimens 3, 4 and 5). Another set of experiments uses sheets of paper card-stock (cut from old computer data cards). Model 54 has a layered
package in which single sheets of lead are sandwiched between double-sheets of paper, and where the interfaces between the paper sheets are lubricated with graphite dust (ground up from pencil leads). Model 53 also uses the paper card-stock, but in this case the paper is assembled into a package of 17 sheets, with each interface being lubricated with graphite dust. This paper package is encased in packages of lead sheets (both above and below) that have no interfacial lubrication (i.e. are welded together, as in model 50).
INTERLAYER SLIP IN MODEL FORCED FOLDS
133
Fig. 3. Model set 2. Line drawings of the slabbed medial surfaces of deformed models 1427 and 1428, with additional detail added from thin sections. Solid black areas are void spaces that appeared during epoxy impregnation of samples. The top layer in the forelimb of model 1428 was lost during examination (see text). Lines at the base of each model show the top of the steel-block forcing assembly (see Fig. 1). Short lines within or cutting the layers represent small fractures or microfaults. L - limestone; P - lead.
Results
Fold and fault geometries All of the models described in this paper produce asymmetric forced folds (antiform and synform pairs situated above and adjacent to the central, major uplifted block; see Fig. 1) whose overall shapes are a response to the displacements (rotation and translation) of the forcing assembly. However, the shapes of these experimental forced folds (Figs 2-5) differ somewhat from the geometries produced in previous models that have less layering (compare models 42 and 40, Fig. 2). In all of the new experiments, except for model 50 (all lead, but unlubricated, so no effective layering), the forelimb (i.e. the common limb between the antiform and the synform) at the top of the package dips more gently than does the forelimb in the less-layered experimental forced folds described by Couples et al. (1994) or in those models described by Weinberg (1979). Another distinct difference is that the prominent graben, which disrupts the crestal region of the antiform in the older models (those of Couples et al. 1994, refer to model 40, Fig. 2, this paper; and Weinberg 1979), is either missing or greatly subdued in this new series. Couples et al. (1994) show that two sharp, angular hinges form in the dolostone layer
Fig. 4. Model set 3. Line drawings of the slabbed medial surfaces of deformed models 47, 49 and 52. Solid black areas are void spaces that appeared during epoxy impregnation of samples. Lines at the base of each model show the top of the steel-block forcing assembly (see Fig. 1). C - leather-hard potter's clay; G - gypsum single crystal; P - lead.
early in the folding, defining a long, straight limb (the 'interlimb') that is subsequently disrupted by further hinge formation at higher structural relief (model 40, Fig. 2). The three models in the present study that contain a dolostone layer (models 42, 43 and 44, Fig. 2) also develop such an interlimb, but the distance between the initial hinges is only about half that observed in the previous models where the overlying limestone was composed of a single layer. The other new models reported here lack the stiff layer of dolostone, but they nevertheless exhibit distinct upper and lower hinge areas that bound a zone of deformation comparable in size to that in models 42, 43 and 44, which have the dolostone layer. The distance between the antiform and synform is universally smaller in the present suite of models than it is in the lesswell-layered models of Weinberg (1979) and Couples et al. (1994).
134
G. D. COUPLES & H. LEWIS
Fig. 5. Model set 4. Line drawings of the slabbed medial surfaces of deformed models 50, 51. 53 and 54. Solid black areas are void spaces that appeared during epoxy impregnation of samples. Lines at the base of each model show the top of the steel-block forcing assembly (see Fig. 1). P - lead; S - paper sheet(s).
There are no throughgoing faults in the new rock-layer models (e.g. models 42, 43 and 44, Fig. 2; models 1427 and 1428, Fig. 3; models 47, 49 and 52, Fig. 4), in contrast to previous results as noted above. At most, small faults offset one layer-layer contact (see models 44 and 1427). In the lead-only experiments (models 50 and 51, Fig. 5) there also is no throughgoing faulting, although it might be argued that one or two layers of lead are faulted or greatly attenuated immediately above the uplifted forcing block (the lead-lead contacts are indistinct due to the welding effect of the confining pressure, so this point cannot be fully addressed). In the model that has a paper package encased in lead (model 53, Fig. 5), there is no faulting at all (except for movement along 'bedding planes'). However, the model with the lead-paper interleave (model 54) develops an unusual curved fault (Fig. 6); it is a curved reverse fault near its base, and a curved normal fault where it intersects the top of the model. This fault everywhere lengthens the paper layers that it cuts, as these are steeply dipping in the forelimb of the fold where the fault is of reverse type. One final observation on model geometries is warranted. Models 51 and 54 (Fig. 5) are distinctly layered: in model 51 this is due to the lubrication between the lead sheets, and in model 54 it is due to the presence of the distributed paper
sheets. Both of these models are dominated by quite ductile material (lead), and both develop a comparable overall shape at their top surface
Fig. 6. Photograph of slabbed medial surface of model 54 showing curved reverse normal fault. The pre-experiment thickness of package is 1.43cm.
INTERLAYER SLIP IN MODEL FORCED FOLDS
135
Fig. 7. Plot of layer-parallel translations (lateral motion) vs magnitude of uplift for the present series of models and those of Couples et al. (1994). Solid line and circles are for the downthrown area, dashed lines and squares are for the upthrown region. Long dashes represent simple fit of data, while short dashes indicate a possible fit patterned after the kinematic analyses of Weinberg (1978). CW - clockwise shear; CCW counter-clockwise shear (see Couples et al. 1994 for details of sign conventions).
(allowing for the extra 'wiggle' in model 54 associated with the curved fault described in the previous paragraph). This shape is noticeably different to the shape developed in model 50, which is also quite ductile (all lead, but unlubricated). In this latter instance, the forelimb is quite steep and there is a clear thinning of the layered package (see also Weinberg 1979; his specimens 3, 4 and 5). Enhancing the layering (with lubrication) clearly affects the outcome.
Layer-parallel motions Significant layer-parallel translations occur along the steel-layered package interface in all models of both the present and previous model series (Fig. 7). Early in the folding, whole-package layer-parallel translation is away from the uplift (on both upthrown and downthrown blocks), and later layer-parallel translation is toward the uplift. These lateral motions of the entire layered package are passive responses of the folding, being the result of the flow of the basal ductile unit (Couples et al. 1994), and possibly related to kinematic requirements associated with offset hinges (Weinberg 1978; Weinberg & Stearns 1978). Variations in layerparallel motion, and their interpretation in terms of laterally varying strains, are considered by Couples & Lewis (1998).
At a smaller scale of observation, the present series of models provides direct evidence of layer-parallel slip within the layered package, as exemplified by the forelimb of model 1428 (Fig. 8). Parts of the uppermost limestone layer of this model were removed following the experiment, allowing examination of the layer-layer contact beneath. Prominent striations cover the steep forelimb portion of this lubricated interface (the offset grids prove major layer-parallel slip on it), but the gentle limb on the upthrown block does not show these striations. (During the examination, the removed forelimb piece of the top limestone layer disintegrated, but the gentle-limb piece was restored prior to epoxy impregnation; the image in Fig. 3 reflects this loss of material.)
Fig. 8. Photograph of external portion of model 1428 before epoxy impregnation showing offset of inked grids in the forelimb area. The pre-experiment thickness of the total package is 1.17cm.
136
G. D. COUPLES & H. LEWIS
Fig. 10. Line drawing of the exterior portion of model 53 showing segmentation of the paper-card package into distinct, multilayer packages, and slip between the paper package and the overlying lead package. Interlayer contacts within the paper package are not shown for clarity. The pre-experiment thickness of the total package was 1.38cm. Fig. 9. Line drawings of the exterior portions of models 50 and 51 showing displacements of scribed lines. The apparent 'rounded* corner of uplift is due to the flow of lead along the fold axis and over the edge of steel forcing block. Note the segmentation of the lead package by interlayer slip on some interfaces in model 51. but the absence of slip in model 50. The pre-experiment thickness of the total package in model 50 was 1.185cm.
While the lubricated layer-layer contacts of model 1428 experience slip in the forelimb, the unlubricated contacts of model 1427 do not show measureable slip in this position. The evidence from similar ink grids on others of the rock-layer models is comparable: lubrication is associated with distinct slip, while unlubricated interfaces do not show such visible offsets. There are other examples of layer-parallel motions. In the lead-only, and lead-plus-paper, models straight, layer-normal marker lines were inscribed onto the edges of the lead and/or paper sheets before folding. Where intra-package layer-parallel slip has occurred, these markers have been offset (Fig. 9). In model 50 (all lead layers, unlubricated) there is little suggestion of such movements, but in model 51 (all lead layers, all lubricated) layer-parallel translations are obvious, with a sense of motion such that the upper layers move towards the crestal region of the antiform. Note, however, that layer-parallel offsets are not uniformly distributed across the layer-layer contacts; instead, they segment the package of lead sheets in model 51 into discrete packages, consisting of two-to-four sheets each, that are delimited by active slip surfaces. A comparable segmentation of the layered package is revealed within the paper-card stack of model 53, where offsets of inscribed grid lines illustrate discrete, localized layer-parallel slip in the forelimb (Fig. 10). Within the multilayer package of 17
paper sheets, four slip-bounded packages are developed during the folding, and there is an additional prominent layer-parallel slip surface at the top of the paper package where it is in contact with the overlying package of lead sheets. Within the slip-bounded packages of paper sheets, the other interfaces do not experience discernable slip. In the experiments that use the layers of gypsum (models 47, 49 and 52, Fig. 4) there is further evidence of intra-package slippage, albeit at a very fine scale. In the model that has a single, thick layer of gypsum (model 47), the forced-fold flexures are accomplished by an array of kink bands which transect the entire layer of gypsum (Fig. 11), suggesting a wholesale activation of slip on the (010) cleavage. In the two models that have thinner gypsum layers (models 49 and 52), a different form of response is observed (Fig. 12). Here, the gypsum also has deformed by developing multiple slip surfaces on the (010) cleavage planes, but, as well as kink bands, the slip has allowed a series of asymmetric folds and fault-fold systems to develop
Fig. 11. Photomicrograph illustrating the development of kink bands in model 47 with a 'layer' composed of a gypsum single crystal (0.85cm thick). Crossed nicols. Note that the entire forced fold is not a kink, although kink-folding is a prominent mechanism.
INTERLAYER SLIP IN MODEL FORCED FOLDS
Fig. 12. Photomicrograph of small-scale folds and thrusts at the base of a thin layer1 of gypsum in model 52. Note also the kink bands transecting the remainder of the layer. The thickness of the gypsum is 0.275 cm.
associated with the detachment of thin packages of cleavage-bounded sheets. In the lower forelimb area, and in the synform hinge region, the asymmetry of these small folds and thrusts within the gypsum indicate a sense of movement like that suggested for all the other movement indicators: higher layers move towards the crest of the antiform.
Bending resistance Force-displacement records show that the presence of layering reduces the external force required to deform a model that is otherwise identical to a model without layering (Fig. 13). We record axial force and axial displacement externally to the pressure vessel. The measured displacement accurately reflects the closure of the two end steel blocks (Fig. 1), and this motion is kinematically linked to the translations
Fig. 13. Axial force (normalized to peak force in model 46) and axial displacement (cm) for representative experiments. Note the steep rise in axial force, and then the constant or decreasing force level as the experiment proceeds. Letters refer to the composition of the layered packages: P - lead; D dolostone; L - limestone; G - gypsum; C - potter's clay.
137
and rotations of the central two steel blocks (at the join of which are located the forced folds that we investigate in this study). However, the externally measured total axial force includes the effects of: the friction between the pistons and the O-ring pressure seals (Handin et al. 1972); the slippage between lubricated steelblock surfaces; and the bending resistance of the layered package. In our experiments, the 'apparatus' elements are the same for each model, and the differences that we observe can be related to changes in the overall flexural resistance of the layered package. All of the models in the study reported by Couples et al. (1994) follow a consistent forcedisplacement path (Fig. 13), as illustrated by model 46 (which has the highest magnitude of uplift from that series). Model 43 of the present study, which has the same configuration as model 46 (except for a four-layer package replacing the single thick layer of limestone), shows the strength-reducing effect of multiple layers. The force required to deform model 43 is initially as large as that needed for model 46, but the force necessary to continue the folding drops off to a smaller level than that needed for deforming model 46. Because the other conditions are identical (e.g. piston and O-ring friction, and steel-block lubrication) this difference must represent a smaller bending resistance in the multilayer model. A slight further reduction in bending resistance is seen for model 1428 (which lacks the dolostone layer), suggesting that the absence of the single, thin layer of dolostone is reducing both the initiation force and the force needed to continue the bending. The models with gypsum single crystals show two different resistances: model 47 (thick gypsum layer) is substantially stronger than model 49 (thin gypsum layer). The lead-dominated models have even less bending resistance than those with rocks or thick gypsum crystals, but we cannot distinguish meaningful variations within this latter set of configurations (here represented by model 51). These results taken together demonstrate a consistent decrease in the force needed to initiate and continue folding as the degree of layering and lubrication are increased, and as the original strength of the rocks is reduced. Interpretation of fracture fabrics In this section, we focus on the models that contain rock layers, allowing us to use the empirical and theoretical knowledge of deformation mechanisms that operate in rocks (e.g. petrofabric
138
G. D. COUPLES & H. LEWIS
techniques; see Friedman & Sowers 1970) to interpret the state of stress-strain that has developed in the rock layers as a result of the folding process. Here, we describe the stress fields that can be deduced by such techniques (e.g. Friedman et al. \916a, b) when they are applied to our models. The interpreted stress fields are spatially very heterogeneous (see following paragraphs); the domains of 'homogeneous' stress state are typically only a few millimetres in size (about the thickness of the layers). This scale limitation precludes our use of calcite twinning to quantify the state of strain. That technique requires that a considerable number of crystals be used, but, because the grain size of the limestone in our models (> 1 mm) is nearly as large as the identified 'homogeneous' domains of stress, it is impossible to find enough crystals within a homogeneous strain domain to make the technique valid. In addition, the grains comprising the limestone consist of a wide range of bioclastic types, many of which are optically 'dirty' (pellets/microcrystalline). Therefore, our comments on the state of strain rely on an inference that the stress state and strain state are coaxial, and so we emphasize the orientation of strain axes without knowing the precise magnitude of the strain components. Friedman et al. (\916b) argue that most of the microfractures in their experiments are Mode I 'tensile' cracks, and that the alignment of observed microfractures reflects the local orientation of the maximum compressive principal stress (
Fig. 14. Stress states inferred for limestone portion of three models that have limestone layer(s): each also has dolostone layer overlying a basal lead layer (dolostone and lead combined in this figure). Lines represent o\ (maximum principal stress) trajectories inferred from observed deformation fabrics (see Fig. 2). (A) Model 40 (from Couples et al. 1994). Note the single, thick-beam stress state. (B) Model 44. Note two beams. (C) Model 43. Note four beams.
• an inverted, mirror image of this pattern in the synform region; • an intervening region where the trajectories are inclined to the layering. The fracture fabrics of our present suite of models show something different. Let us compare three models that have similar structural relief (Fig. 14): one from the previous series (having a single limestone layer), one multilayer model that has only the middle interface lubricated (in its stack of four limestone layers), and one multilayer model that has lubricant on all of its limestone-limestone interfaces. Microfracture fabrics indicate that a single bending stress state develops in the single-layer model (Fig. 14A), but that two. two-layer thick bending states develop in model 44 (only its middle interface is lubricated; Fig. 14B). while there are four one-layer 'beams' developed in model 43 (all of its layer-layer contacts are lubricated; Fig. 14C). The deformation fabric observed in
INTERLAYER SLIP IN MODEL FORCED FOLDS
Fig. 15. Stress states inferred for models 1428 (above) and 1427 (below). Heavy lines represent cr, (maximum principal stress) trajectories inferred from the observed deformation fabrics (see Fig. 3). Note the presence of multiple beams in model 1428, but the stress state in model 1427 suggests only a single beam, or perhaps two beams in the synclinal area.
model 42 (all lubricated, not shown) is essentially like that observed in the similar model 43. The principal difference between these two models (42 and 43) is that the synform hinge area in model 42 is considerably sheared (as a result of greater structural relief). It is important to note that the multilayer rock models (42, 43 and 44) exhibit alternations in the orientation of a\ through the folded region along transects that are normal to the layering. These reversals of stress state are, equally, inversions of the state of strain, with layer elongations and layer contractions alternating repeatedly. Such inversions are described by Couples et al. (1998) as resulting from the progressive activation of interlayer slip during folding. We will return to this topic in the Discussion.
139
Another comparison also illustrates such strain inversions. Models 1427 (all limestone, no lubrication) and 1428 (all limestone, all lubricated) show the difference between the existence of a potential slip surface and enhancement of a slip surface. Multiple stacked beams are interpreted from strain-pattern- and microfracture-derived stress state reversals in both the antiform and synform portions of model 1428 (Fig. 15a). But in the antiform region of model 1427, there is no evidence for multiple beams (Fig. 15b). This difference suggests that the mere presence of layer interfaces will not automatically lead to strain partitioning during flexural-slip folding, but that the activation of slip (here ensured by lubrication in model 1428) is necessary to make this effect clear. However, the general absence of through-going faults in both of these models (in contrast to the ubiquitous faulting in non-layered models) shows the presence of layering has an effect on the strain distribution even if multiple beams are not obvious. There is, therefore, a compelling basis for arguing that mechanically effective layering, and the related possibility of layer-parallel slip (whether discernible or not), is exerting a major control on fold mechanics. Discussion The multilayered forced-fold models described here provide evidence of a deformation sequence that is distinctly different from that seen in comparable non-layered models (Table 2). There are contrasts in: the shape and size of the folds; the patterns of resulting stress within them; the distribution of their strain, including faulting; and their resistance to bending. Because of the degree of control allowed through the use of an experimental approach, we can be certain that these differences are specifically attributable to
Table 2. Comparison of results
Shape of folds Width of fold pair/interlimb length Curvature of layers Dip of forelimb Throughgoing faults Upthrown Downthrown Bending fabric Interlayer slip
This paper
Previous work*
Narrow/short Low Gentle
Wide/long High Steep
Absent Absent Multiple Measurable if lubricated
Ubiquitous Common Single Not systematically addressed
* Previous work refers directly to Couples et al. (1994), and to Weinberg (1979). In general, the results from Friedman el al. (1976<2, 1980) also meet these descriptors, but the variety of configurations they use preclude making blanket statements.
140
G. D. COUPLES & H. LEWIS
the presence and effectiveness of the layering. In the following sections we focus on the mechanical interpretation of our experimental results, and on how this understanding may be extended to nature.
A flexural-slip paradigm In our experimental models, interlayer slip whether it is measurable or not - exerts a primary control on the shape and state of the resulting folds. When multiple layer interfaces are available for slip, measurable slip is concentrated onto only some of those possible slip surfaces. Because the interface characteristics of the surfaces are identical (within the limits imposed by the fabrication process), the localization of slip must be a product of the folding process, and hence must have a mechanical basis that is associated with the spatially heterogeneous deformation state produced by bending. Segmentation of a layered package into slipbounded mechanical units during the folding must be an action that is energetically more favourable than the faulting response that is observed when layer-parallel slip is not available. Concurrent with the partitioning of the package by layer-parallel slip, heterogeneous states of stress develop with coherent bending patterns observed in each of the slip-bounded mechanical units. In our small experimental models we do not observe significant superposition of these stress states as has been reported for natural folds (Couples et al 1998). We suggest that this situation arises because of the scale. For example, in the limestone layers of our models, any deformation that occurs is expressed at the grain scale, which is of the order of 1 mm, and once such deformation exists within the rock it is likely to preclude any further deformation being expressed from stress domains that are smaller in size. The contractional strains that we observe cannot be attributed to a homogeneous 'compressional' stress state imposed by the loading apparatus because the layered packages of the models are isolated from the apparatus by weak plasticine plugs. Even if we assume that the plasticine is "strong' - strong enough to transmit end loads from the pistons onto the models - this suggestion can be discounted from empirical evidence: the contraction features are not uniformly distributed through the models: they characterize only small domains. The only explanation compatible with the observations is that the contractional deformation is developed during the bending process.
Friedman & Stearns (1971; see also Stearns & Friedman 1972) argued that the deformation state in folds is dominated by the bending process. In the multilayer models reported here, the domains of contraction and elongation are usually bounded by the layer interfaces and, in several examples described in this paper, these boundaries are shown to be active layer-parallel slip surfaces. Ruf el al. (1998) also suggest that the bedding planes bounding active "mechanical units' control the extent of fracturing in nature, and, in their clay-cake models, Withjack el al. (1990) observed fabric changes across layerparallel slip surfaces. The relationship between fabric distribution and slip reinforces the view that the state of strain in many folds is primarily a consequence of the folding process and, particularly, is related to the effects of multiple layers. A similar conclusion was reached by Chester el al. (1988) who illustrated a similar domination of bending deformation in their models of thrust-associated deformation. The evidence suggests that it may be inadvisable to seek to ascertain the tectonic origin of flexuralslip folds based on their contained fabrics. The deformation fabrics that we observe in our models represent both elongations and contractions of the layers. We interpret these strains as being the direct product of the bending stress state that develops in each mechanical unit that is bounded by active slip surfaces. The magnitude of the material damage produced by the folding is everywhere less than the strain magnitude at comparable locations in folds without the layering. Integrated over the region of the folding, the rocks in the multilayer models are substantially less damaged as compared with poorly layered folds. Some of this difference may lie in the different shapes, but most of it is associated with the localized layerparallel slip which is not normally included in strain analyses (such a task is difficult to achieve with any confidence in most cases because of the paucity of displacement data). The deformations produced in our models represent the almost "pure1 bending response of a multilayered sequence undergoing flexure. The major role of layer-parallel slip is made abundantly clear, and the experimental design allows us to be confident that the deformation processes are inherent to the flexure and not related to farfield causes. Because of these 'ideal' circumstances, we believe that the process model reviewed by Couples el al. (1998), and re-iterated above, represents a new flexural-slip paradigm that embodies an understanding of the causal events, and predicts the deformation consequences, of the flexure of layered rock sequences.
INTERLAYER SLIP IN MODEL FORCED FOLDS
Controls on layer-parallel slip In nature, rock successions are almost always layered. The point arising from this paper is that the mechanical effectiveness of that layering is important in terms of controlling the shape of the folding, its spatial extent and the details of the damage imposed onto the rocks. In our experiments, the mechanical effectiveness is controlled by presence/absence of contacts, or by their lubrication/non-lubrication. What controls the mechanical effectiveness of layering in natural rock successions? The real issue here is whether layer-parallel slip is possible, and if it is 'easy'. Our experiments show that mechanical effectiveness is not merely a question of contrasts in strength or ductility. Rather, the primary factor is the resistance (or lack of it) to sliding. Indeed, the contacts between 'stiff and 'weak' layers are not where slip is preferentially localized because, in such a case, the net resistance to motion (expressed as effective friction coefficient) is larger due to the increased 'real' area of contact associated with the ductile behaviour of the weaker material. Such ductile deformation would represent a simple shear of that 'weak' layer and the work performed by that deformation must, apparently, be larger than the work required by frictional sliding (if sliding is possible). In nature, the net frictional resistance is partly related to the true friction and partly to the effective normal stress. If we assume that a given rock succession has some distribution of the coefficients of true layer-interface friction, what other factors can affect the net frictional responses of the layer interfaces? Depth of burial is one possible (and obvious) control on net friction, but elevated fluid pressure can also be important. Less effective layering may be more typical of deeper locations where the normal stress acting on layer interfaces is higher. For that hypothetical rock succession, when it is folded in different locations and at different depths, we predict that there will be variations in fold shape, the extent of folding and the state of the resulting deformation (including the level of damage). Likewise, any folding of that succession that occurs under contrasting conditions of fluid pressure (in or out of an overpressured zone, for example) should also produce these differences. If there are processes that can alter the 'true' friction, such as the introduction of a contrasting and chemically reactive fluid, then similar variations in response might be expected. In regions with laterally consistent stratigraphy, could these differences in fold form be used to 'map' palaeodepths or palaeopressures?
141
If we apply this idea to the Big Horn Basin of northern Wyoming, it might be that narrow forced folds, with steep forelimbs, are more likely in basin-margin settings where erosional unroofing has led to less confinement, and broader, 'more gentle' forced folds may be more typical of the basin centre where syntectonic deposition has led to an increase in confinement. However, if the width of a fold is established early in the deformation, there may be little difference between these sites. Overpressure in the basin centre may complicate this prediction.
Subsidiary folds and their natural counterparts Our models provide insights into other aspects observed in natural forced folds: the presence of subsidiary folds and detachments. These structures have been given a variety of colourful names: back-limb folds, cross-crestal folds and rabbit ear folds (Brown, 1988), and drape subsidiary folds (Weinberg 1978; Weinberg & Stearns 1978). They occur in a variety of locations around large-scale natural folds. They have been attributed to 'room problems', either as the result of the non-concentric shape of the major fold (Weinberg 1978) or as a consequence of excess material being transported 'out of the syncline' (Brown 1988). In our models, similar fold forms occur in the forelimb of the flexed gypsum folds (Fig. 12). The layer-parallel shortening depicted in this photomicrograph - if scaled up to the size of a Wyoming forced fold - would represent a quite spectacular field observation. One could certainly be tempted to derive a causal interpretation of the whole tectonic setting from such a set of features. However, it is worth reemphasizing that the motions occurring in our models do not rely on external processes, but that they are, instead, driven by energies generated within the fold as a consequence of the folding. If the mechanical cause of these subsidiary folds in our experiments has a natural, large-scale counterpart, then the subsidiary folds observed in nature cannot be used by themselves (in isolation) to infer far-field causes. How do the subsidiary folds relate to the flexural-slip paradigm and its emphasis on layer-parallel slip? If we interpret these small folds as representing 'drag' or 'parasitic' folds, then they suggest a sense of motion in which the layers above are moving relatively towards the antiformal crest. In this instance, it seems that the larger-scale layer-parallel slip has been
142
G. D. COUPLES & H. LEWIS
'distributed' across a finite thickness of layers, rather than occurring on a single interface. As a consequence, the zone has been deformed in a simple-shear fashion, and the subsidiary folds have resulted. If this process is extrapolated to natural forced folds, then the impressive structures described by Weinberg & Stearns (1978) and Brown (1988) can be interpreted as proto-disharmonic features related to the movement of large-scale mechanical units. This view has similarities to the kinematic arguments concerning the causes of hinge-zone geometric irregularities proposed by Ramsay (e.g. Ramsay & Huber 1987). It is fully compatible with the interpretation of differential bedding motion made by Weinberg (1978; Weinberg & Stearns 1978).
Competition of deformation processes The principal anomaly of our experimental series is the presence of the throughgoing fault that cuts through model 54 (Fig. 6). This model is quite obviously layered (interleaves of paper and lead), but its response is quite unlike that of all the other multilayered models described here. Does the presence of this fault undermine our view concerning the importance of layer-parallel slip? Quite the contrary. We have shown that where there is potential for layer-parallel slip, the multilayer packages undergo flexure without developing significant faulting or other material damage. In the unlubricated, all-lead case (model 50, Fig. 6), slip is inhibited because the interfaces become welded together. In this case, however, the composite package of lead can flow in response to the imposed loading, and no significant faulting occurs. In the anomalous model 54, neither slip nor flow are possible. Slip is inhibited because of a lack of lubrication between the paper and the lead. Perhaps more importantly, the paper sheets act much like reinforcing bars in concrete, preventing the lateral flow (and lengthening) that would necessarily be associated with flow-caused changes in the thickness of each sheet of lead. With neither of these mechanisms available, this model responds to the uplift with a layerlengthening fault (whose shape becomes distorted as the lead squeezes around the ends of the torn paper sheets, and as the truncated layers' are dragged by the motion of the forcing block). This fault is (if restored to its 'undragged' state) almost exactly as predicted by the analytical models of layer stretching described by Patton & Fletcher (1995), and by the experimental models of Withjack et al. (1990) and Patton
el al. (1998). In the case of model 54, the stretching is related to the diminished flow capacity of the 'basal ductile layer' caused by the insertion of paper sheets very low in the package. We predict that the stretching-fault process would not be needed if we were to create a thicker basal ductile unit in an otherwise-similar model (e.g. not having paper sheets low in the package). If we assume that the most energetically favourable deformation will occur, if possible, and that the circumstances dictate which mechanisms can operate, our results permit us to suggest an ordering of folding mechanisms in terms of energy efficiency. If flexural slip can operate, this mechanism is favoured, and in this case there is the least energy expended in terms of damage imposed onto the rocks. If layer-parallel slip is inhibited (high friction, high effective normal stress), the next most efficient process is flow of the layers (thickness changes, and related length changes), with energy being required to cause the material damage. If neither mechanism can operate, the package becomes faulted. In the background of these statements is an implicit recognition of the importance of the basal ductile unit. Without such a unit, there is no mechanism for lateral transport of the layered package (Couples et al. 1994) and the 'stretching' case dominates. Conclusions Physical models of forced folds demonstrate that the presence and/or enhancement of layering has important mechanical effects. Folds with an increase in the number of layers (as compared to otherwise-similar folds that have single, thick layers) are both weaker and have smoother profiles; they are also more compact. Both elongation and contraction strains occur in the forced folds, as a consequence of the folding. Multilayered folds develop strain distributions that are broader, but less intense, than single-layer folds. Within multilayered folds, strain domains are related to the location of active interlayer slip, and stacked sequences of bending-caused strains are the consequence of the fold-activated flexural slip. These conclusions can be summarized as: • The bending of packages with enhanced layering is easier; • strains are partitioned as a consequence of flexural slip activated by the bending; • both contractional and elongational strains occur throughout the fold pair; • strain intensities are reduced from comparable thick-layer folds;
INTERLAYER SLIP IN MODEL FORCED FOLDS
• layer-layer contacts (between similar materials) that do not slip have a limited effect on the strain distribution, although there is an effect on bending resistance; • enhanced layering makes the resulting deformation more compact. We thank Mel Friedman (deceased), Dave Weinberg, Greg Lynch and an anonymous reviewer for insightful and productive comments. A large number of colleagues have contributed to our thoughts on the folding of multilayers, and we especially acknowledge Dave Stearns and Dave Weinberg in this regard. The experimental work was conducted at the Center for Tectonophysics at Texas A & M University.
References BROWN, W. G. 1988. Deformational style of Laramide uplifts in the Wyoming foreland. In: SCHMIDT, C. J., & PERRY, W. J., Jr. (eds) Interaction of the Rocky Mountains Foreland and the Cordilleran Thrust Belt. Geological Society of America, Memoir, 171, 1-25. CHESTER, J. S., SPANG, J. H., & LOGAN, J. M. 1988. Comparison of thrust fault rock models to basement-cored folds in the Rocky Mountains foreland. In: SCHMIDT, C. J. & PERRY, W. J., JR (eds) Interaction of the Rocky Mountains Foreland and the Cordilleran Thrust Belt. Geological Society of America, Memoir, 171, 65-74. COOKE, M. L. & POLLARD, D. D. 1994. Development of bedding plane faults and fracture localization in a flexed multilayer: a numerical model. In: NELSON, P. P. & LAUBACH, S. E. (eds) Rock Mechanics, Models and Measurements, Challenges from Industry. Proceedings of the First North American Rock Mechanics Symposium. Balkema, Rotterdam, 131-138. & 1997. Bedding plane slip in initial stages of fault-related folding. Journal of Structural Geology, 19, 567-581. , MOLLEMA, P., POLLARD, D. D. & AYDIN, A. 1999. Interlayer slip and joint localization in the East Kaibab monocline, Utah: field evidence and results from numerical modelling. This volume. COUPLES, G. D. 1977. Stress and shear fracture (fault) trajectories resulting from a suite of complicated boundary conditions with applications to the Wind River Mountains. Pure and Applied Geophysics, 115, 113-133. & LEWIS, H. 1998. Lateral variations of strain in experimental forced folds, Tectonophvsics, 295 79-91. , & TANNER, P. W. G. 1996. Hierarchical development of frictional interlayer slip in flexed multilayers. In: BICANIC, N. (ed) Fourth Association for Computational Mechanics in Engineering UK, University of Glasgow, Glasgow, 37-42. , & 1998. Strain partitioning during flexural slip folding. In: COWARD, M. P., DALTABAN, T. S. & JOHNSON, H. (eds) Structural Geology in
143
Reservoir Characterization. Geological Society, London, Special Publications, 127, 149-165. , STEARNS, D. W. & HANDIN, J. W. 1994. Kinematics of experimental forced folds and their relevance to cross-section balancing. Tectonophysics, 233, 193-213. FRIEDMAN, M. & SOWERS, G. M. 1970. Petrofabrics: a critical review. Canadian Journal of Earth Sciences, 7, 477-497. & STEARNS, D. W. 1971. Relations between stresses from calcite twin lamellae and macrofractures, Teton anticline, Montana. Geological Society of America Bulletin, 82, 3151 3162. , HANDIN, J., LOGAN, J. M., MIN, K. D. & STEARNS, D. W. 1976^. Experimental folding of rocks under confining pressure: Part III. Faulted drape folds in multilithologic layered specimens. Geological Society of America Bulletin, 87, 10491066. , HUGMAN, R. H. H. & HANDIN, J. 1980. Experimental folding of rocks under confining pressure, Part VIII - Forced folding of unconsolidated sand and of lubricated layers of limestone and sandstone. Geological Society of America Bulletin, 91,307-312. , TEUFEL, L. W. & MORSE, J. D. \916b. Strain and stress analyses from calcite twin lamellae in experimental buckles and faulted drape-folds. Philosophical Transactions of the Roval Society of London, A283, 87-107. GANGI, A. F., MIN, K. D. & LOGAN, J. M. 1977. Experimental folding of rocks under confining pressure: Part IV. Theoretical analysis of faulted drape folds. Tectonophysics, 42, 227-260. HAFNER, W. 1951, Stress distribution and faulting. Geological Society of America Bulletin, 62, 373-398. HANDIN, J., FRIEDMAN, M., LOGAN, J. M., PATTISON, L. J. & SWOLFS, H. S. 1972. Experimental folding of rocks under confining pressure: Buckling of single-layer rock beams. American Geophysical Union Monograph 16, 1-28. , , MIN, K. D., & PATTISON, L. J. 1976, Experimental folding of rocks under confining pressure: Part II. Buckling of multilayered rock beams. Geological Society of America Bulletin, 87, 1035-1048. HANEBERG, W. C. 1993. Drape folding of compressible elastic layers - II. Matrix solution for two-layer folds. Journal of Structural Geology, 15, 923-932. JAMISON, W. R. 1979. Laramide Deformation of the Wingate Sandstone, Colorado National Monument: A Study of Cataclastic Flow. PhD Thesis, Texas A & M University, College Station. LEWIS, H. & COUPLES, G. D. 1993. Production evidence for geological heterogeneities in the Anschutz Ranch East field, western USA. In: NORTH, C. P. & PROSSER, J.(eds) Characterisation of Fluvial and Aeolian Reservoirs, Geological Society, London, Special Publications, 73, 321-338. PATTON, T. L. & FLETCHER, R. C. 1995. Mathematical block-motion model for deformation of a layer above a buried fault of arbitrary dip and sense of slip. Journal of Structural Geology, 17, 1455-1472.
144
G. D. COUPLES & H. LEWIS
, LOGAN, J. M. & FRIEDMAN, M. 1998. Experimentally generated normal faults in single- and multilayer limestone beams at confining pressure. Tectonophysics* 295, 53-77. RAMSAY, J. G. & HUBER, M. I. 1987. The Techniques of Modern Structural Geology. Volume 2. Folds and Fractures. Academic Press, London. RUE. J. C. RUST. K. A. & ENGELDER, T. 1998. Investigating the effect of mechanical discontinuities on joint spacing. Tectonophysics. 295, 245-257. SANFORD. A. R. 1959. Analytical and experimental study of simple geologic structures. Geological Society of America Bulletin. 70, 19-52. STEARNS. D. W. 1971. Mechanisms of drape folding in the Wyoming province. In: Wyoming Geological Association, 23rd Annual Field Conference Guidebook. Wyoming Geological Association, 125-144. 1978. Faulting and forced folding in the Rocky Mountains foreland, lir. Matthews, V. (ed.) Laramide Folding Associated with Basement Block Faulting in the Western United States. Geological Society of America. Memoir, 151, 1-37. & FRIEDMAN. M. 1972. Reservoirs in fractured rock. Stratigraphic oil and gas fields. American Association of Petroleum Geologists Memoir, 16, 82-106.
& WEINBERG, D. M. 1975, A comparison of experimentally created and naturally deformed drape folds. ///: Wyoming Geological Association. 27th Annual Field Conference Guidebook. Wyoming Geological Association, 159-166. TANNER. P. W^. 1989. The flexural-slip mechanism. Journal of Structural Geology. 11, 635-655. WEINBERG. D. M. 1978. Some two-dimensional kinematic analyses of the drape-fold concept. ///: MATTHEWS. V. (ed.) Laramide Folding Associated with Basement Block Faulting in the Western United States. Geological Society of America. Memoir. 151. 51-78. 1979. Experimental folding of rocks under confining pressure. Part VII - Partially scaled models of drape folds. Tectonophysics. 54. 1-24. & STEARNS. D. W. 1978. Kinematic analyses of drape folds in the Rocky Mountain foreland some geologic implications. ///: Wyoming Geological Association, 30th Annual Field Conference Guidebook. Wyoming Geological Association. 199-212. WITHJACK. M. O.. OLSON. J. & PETERSON. E. 1990. Experimental models of extensional forced folds. AAPG Bulletin. 74. 1038-1054.
Regional tectonics and fracture patterns in the Fall River Formation (Lower Cretaceous) around the Black Hills foreland uplift, western South Dakota and northeastern Wyoming JOHN L. WICKS1, STUART L. DEAN2 & BYRON R. KULANDER 3 l
Dalton and Hanna, 345 N. Market Street, Wooster, OH 44691, USA Department of Geology, University of Toledo, Toledo, OH 43606, USA ^Department of Geological Sciences, Wright State University, Dayton, OH 45435, USA 2
Abstract: The Fall River Formation around the Black Hills uplift is pervasively fractured by layer-perpendicular joints. Systematic joints in the formation maintain consistent orientations over large areas and are commonly abutted by later-formed fractures, resulting in an orthogonal pattern. There are two major systematic sets, trending northeast and northwest, and one minor set trending north-south. The first two sets define two major fracture domains in the study area. The northwest joint set occupies a southern domain where it is the sole systematic fracture set. The northeast joint set is pervasively established throughout the northern domain, where northwest and north south fracture sets are also developed in well-defined sectors. There is no genetic or spatial relationship between joint sets and local Laramide monoclines or folds of the region. Instead, the stratigraphic record indicates that joint development originated early in the lithification history of Fall River sandstones. Jointing occurred in response to local and regional extensional stresses that pervaded the northern and southern domains as a result of recurrent movement on basement faults that parallel the regional lineament system and surface structural zones throughout the region. Major uplift of the Black Hills and local fold development during Laramide time merely resulted in passive rotation of the early formed systematic and non-systematic joints.
Statement of problem Preliminary field work and aerial photograph reconnaissance revealed a pervasive network of joints throughout the Fall River Formation (Lower Cretaceous) outcrop belt around the Black Hills in northeastern Wyoming and western South Dakota. The major objectives of this study were to establish the joint patterns present in sandstones of this formation (Fig. 1) and to determine whether these joints are controlled by this Laramide uplift, its smaller associated folds or deep-seated basement fracture zones that are expressed by a regional lineament system. Systematic jointing in the Fall River Formation of the Black Hills was first documented by N. H. Darton (1904). He described the 'rude columnar appearance' of its massive sandstone cliffs near Newcastle, Wyoming (Fig. 1). Later, Gott et al (1974) attributed changes in horizontal compressive stress directions for different joint trends in the southwestern part of the present investigated area. Other authors, including Bergendahl et al. (1961) and Brobst & Epstein (1963), have commented on the prominent fractures present in the Fall River but have no definite conclusions regarding their overall mode of origin and spatial
relationship to regional or local structure. Overall, the patterns and structural significance of Fall River Formation joints have not been extensively studied by previous workers. The Fall River Formation was chosen for this study because of its areal extent and the comprehensive stratigraphic study by Gott et al. (1974). More recent work by Willis et al. (1995) has further defined the architecture of this formation in the southwestern Black Hills.
Methods of investigation The upper Fall River Formation was chosen for this study because it consists primarily of a 1040-foot (3-12.2-m) thick massive sandstone unit, and caps many hogbacks and divides throughout the Black Hills region. Where this unit was absent or concealed, the middle Fall River Formation was used. One hundred and forty stations were established and 7214 joint attitudes were measured. At each station joint strike was measured, or taken from aerial photographs in inaccessible areas (Fig. 2). All joints examined had an overall perpendicular relationship to bedding. In nearly all circumstances it was possible to distinguish
From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 145-165. 1-86239-060-6/OO/S 15.00 (( The Geological Society of London 2000.
146
J. L. WICKS ET AL.
Fig. 1. Sandstone outcrop 2 miles (3.2km) east of Newcastle. Wyoming on the crest of the Black Hills monocline. Systematic joints are perpendicular to the cliff face. Distance is 825 feet (252m) along the cliff face.
systematic from non-systematic joints based on the terminology of Hodgson (1961). Irregular, curved joints were not measured because no consistent bearing is maintained, and they may represent effects of stress relief rather than tectonic events. Details of other joint features such as abutting relationships, surface morphology (i.e. hackle plumes), mineralization and continuity of the joint faces also were recorded. Inclined bedding was rotated to horizontal on a Schmidt net to place the joints in a vertical position. Mean bearings of systematic and nonsystematic joints are shown as lines in Fig. 3. Where consistent joint trends exist throughout the Fall River outcrop belt, these areas are represented by composite rose diagrams (Fig. 4). All data and diagrams were then analyzed for consistency of regional joint trends, as well as geometric relationships of fractures to folds and surface structural zones. The latter are defined by faults and changes in fold traces and are expressed by major lineaments. Fig. 2. Aerial photograph of systematic (N30 W, i.e. 330 ) and non-systematic joints in Fall River sandstone. Location is 7 miles (11.3km) northeast of Edgemont, South Dakota, between Chilson anticline and Sheep Canyon monocline.
Black Hills; geological setting and structure The Black Hills is a Precambrian-cored anticline that rises out of the Great Plains of western
FRACTURE PATTERNS AROUND FORELAND UPLIFT
147
Fig. 3. Diagram showing systematic (long lines) and non-systematic (short lines) joints in the Fall River Formation around the Black Hills. Major lineaments are shown as heavy dashed lines. The northeastern projection of the crest of the Belle Fourche Arch is shown between the arrows in the northwestern Hills. Geological sources for the map are: Darton (1901, 1902, 1904, 190506), Darton & Smith (1904), Darton & O'Hara (1907, 1909), Darton & Paige (1925), Shapiro (1971), Gott et al. (1974), Wicks (1979), Lisenbee (1985), DeWitt^ al (1989).
148
J. L. WICKS ET AL.
Fig. 4. Diagram of composite joint trends in relation to major structures of the Black Hills. Major lineaments shown as heavy dashed lines. Projected crest of the Belle Fourche Arch is shown between the arrows in the northwestern Hills. (Refer to Fig. 6 for abbreviations of towns; diagram from Lisenbee 1978; courtesy Geological Society of America.)
South Dakota and eastern Wyoming (Fig. 5). It is best described as a doubly plunging fold with a curvilinear axis striking generally north to northwest. The uplift is 180 miles (290km) long, 68 miles (110km) wide and formed during the Laramide orogeny. The Black Hills is the easternmost foreland uplift of the Rocky
Mountain region. It is flanked by the Powder River basin to the west, the Williston basin to the northeast and is an extension of the Chadron Arch, which lies to the southeast. The structure continues to the northwest where it merges with the Miles City Arch in southeastern Montana. To the southwest, the Hartville Uplift connects
FRACTURE PATTERNS AROUND FORELAND UPLIFT
149
Fig. 5. Generalized map of the Black Hills uplift in relation to other regional structural features. The dark pattern shows the area of the Fall River Formation outcrop. The stippled pattern represents the Precambrian core of the foreland uplifts.
with the southwest margin of the uplift across a structural saddle (Lisenbee & DeWitt 1993). Lisenbee (1978, 1985, 1988), DeWitt et al (1989) and Lisenbee & DeWitt (1993) have comprehensively discussed the individual structures associated with the Black Hills. The gross structure of the uplift consists of two large Precambrian basement blocks which are separated by a N-S rending fold and fault zone defined as the Fanny Peak lineament along the South Dakota-Wyoming border (Noble 1952) (Fig. 6).
The eastern (i.e. South Dakota) block is structurally higher and basement rock is exposed. Recent geophysical studies by Robbins (1994) suggest that the western (i.e. Wyoming) block is segmented into a third major block along NESW-trending magnetic and gravity anomalies across northeastern Wyoming and southeastern Montana, presenting an abrupt discontinuity in the northwestern Hills, rather than the usual depiction of northwestward plunge of the Precambrian surface into the Miles City Arch. The
150
J. L. WICKS ET AL.
Fig. 6. Tectonic map of the Black Hills uplift and eastern Powder River basin (revised from Lisenbee 1988. courtesy of the Wyoming Geological Survey and Wyoming Geological Association). Little Missouri (LiM). Dewey (D). Long Mountain (LM) and Edgemont (E) structural zones shown by dotted lines. Projection of the crest of the Belle Fourche Arch shown between the arrows in the northwestern Hills.
FRACTURE PATTERNS AROUND FORELAND UPLIFT
Fig. 7. West-east cross-section between the Powder River basin and the Black Hills uplift across the Fanny Peak monocline just south of Newcastle, Wyoming. (Figure from Lisenbee & DeWitt 1993; courtesy of the Wyoming Geological Survey.) Kp upper part of the Pierre Shale; Kpl - lower part of the Pierre Shale; Klf - Lakota and Fall River Formations; Ps-Jm Spearfish, Sundance and Morrison Formations; Pom - Opeche Shale and Minnekahta Limestone; PPm Minnelusa Formation; Mp - Pahasapa Limestone; PC - Precambrian rocks, undifferentiated.
two Precambrian blocks are bounded by two regionally prominent folds, the Fanny Peak monocline along the western border of the South Dakota block and the Black Hills monocline along the western border of the Wyoming block (Fig. 6). Horizontal strata occupy the tops of the blocks and abruptly dip into monoclines at their margins. Locally, folds, ramps
151
and terraces are developed, with interpreted faults often passing into surface asymmetrical folds. The Fanny Peak monocline forms the southwest boundary between the Black Hills and the Powder River Basin, and extends along the length of the Fanny Peak lineament to the Hartville Uplift. The beds on the monocline are draped over the eastern block creating local monoclines, anticlines and terraces, as shown in Fig. 7. Figure 8 shows part of the lineament east of Newcastle, Wyoming. Figure 9 illustrates the interpreted structure beneath the surface. The Black Hills monocline is the boundary between the Black Hills and the Powder River Basin in the west and northwest sectors of the study area. This fold begins where it intersects with the Fanny Peak monocline at Newcastle and extends from there to the Montana border. Numerous folds and terraces are developed along the extent of this feature. Three structural zones transect the southwestern Black Hills in an east of northeast direction. The Dewey and Long Mountain zones consist of sets of NE-trending normal faults in the region (Figs 3 and 6). The Edgemont zone is defined by the termination of several folds. Collectively, the Dewey-Long Mountain-Edgemont structural zones constitute a major lineament zone across the southern Hills.
Fig. 8. Northeast view of Fanny Peak lineament from Route 16 north of LAK Ranch, 6 miles (9.7 km) eastsoutheast of Newcastle, Wyoming. Fall River Formation is the low hogback in the lower left portion of the photograph. Near-vertical monocline flatirons in the central part of the photograph become horizontal at Fanny Peak in the upper right.
152
J. L. WICKS ET AL.
Fig. 9. Generalized block diagram showing the relationship of surface structures to basement faulting at the junction between the Black Hills monocline and Fanny Peak monocline near Newcastle, Wyoming. (Diagram from Lisenbee & DeWitt 1993; courtesy of the Wyoming Geological Survey.) The west border of the diagram is approximately 9 miles (14.5km) in length.
The eastern flank of the Black Hills uplift has dips ranging from 20° to near vertical. The monocline along the eastern flank is locally interrupted by interpreted faults which die out upward into smaller-scale asymmetrical folds. The northern Black Hills is characterized by sandstone plateaus and several N-plunging folds, such as the Belle Fourche, La Flamme and Albion anticlines. North of Belle Fourche, South Dakota, NW-trending folds cross the La Flamme and Albion anticline trend. The Little Missouri fracture zone also crosses the extreme northwestern Black Hills just north of Devils Tower, Wyoming. This N55E° (055°) trending structural zone is marked by numerous normal faults and folds subparallel to this zone, as well as the terminations of several NW-striking folds. Also of significance in the northern Hills is a group of Eocene plutons which are exposed in a linear trend extending southeastward from Missouri Buttes in eastern Wyoming to the vicinity of Sturgis in western South Dakota. Regional tectonic model A variety of plate tectonics models have been advanced to account for the complex array of uplifts in the Rocky Mountain foreland. Consideration of these is beyond the scope of this paper, although a familiar theme has been the concept of foreland uplift in response to shallow subduction of the east Pacific plate beneath the North American plate (Lowell 1974; Dickenson & Snyder 1978; Jordan & Allmendinger 1986).
The role of basement deformation remains controversial and may involve the effects of basement folding (Snoke 1993) and reactivation of low-dipping, deep-seated Precambrian basement faults (Brown 1988, 1993), which may originate from a detachment at the brittle-ductile transition zone in the middle crust (Schmidt et ai 1985; Kulik & Schmidt 1988). Seismic evidence for thrust faults beneath many Laramide uplifts (Berg 1962; Smithson et al 1978, 1979; Brewer et al. 1982; Allmendinger et al. 1982, 1983; Johnson & Smithson 1985) has weakened the case for the once prevalent view that these structures developed over deep-seated vertical faults and do not require crustal compression and shortening (Prucha et al. 1965; Stearns 1971). Structural evolution of the Black Hills The Black Hills uplift began during Early Palaeocene and movement continued through Late Palaeocene-Early Eocene (Lisenbee & DeWitt 1993; Dutton 1995). It may represent one of the last Laramide uplifts to form, inasmuch as some foreland folds to the west developed earlier than the Black Hills (Brown 1988). We favour the model of Lisenbee & Dewitt (1993), first suggested by DeWitt et al. (1986), which attributes Black Hills uplift to a deep-seated, east-dipping reverse fault which terminates near the basement surface under the western Hills (DeWitt et al. 1989) and separates Precambrian rocks of Proterozoic age on the east from Archean rocks on the west. Several lines of evidence support this interpretation. First is the coincidence of the overall configuration of the Fanny Peak and Black Hills monoclines with the boundary between Precambrian terrains. Secondly, magnetic and gravity anomalies (Kleinkopf & Redden 1975; Robbins 1994) show highs coincident with the postulated eastern upthrown block. Additional evidence is given by the fact that almost all monoclines and anticlines of the western and southern Black Hills have steep limbs facing west and many folds show a paired relationship (DeWitt et al. 1986), with the easternmost fold being the anticline and the western fold the syncline. Large-scale wrench faulting has been postulated for significant control on Laramide structural evolution by various workers, including Sales (1968) and Stone (1969, 1970, 1975). Although extensive surface structural alignments and geophysical lineaments suggest major areal variations in buried Precambrian rock types, limited evidence exists in the Black Hills for significant strike-slip
FRACTURE PATTERNS AROUND FORELAND UPLIFT
faulting during Laramide time. As pointed out by Lisenbee and DeWitt (1993), there is no substantial evidence for strike-slip faulting in the Black Hills. Fall river formation
Stratigraphy The Fall River Formation (Fig. 10) is the uppermost member of the Inyan Kara Group (Lower Cretaceous) and is primarily composed of sandstone with some interbedded siltstone and mudstone. Fall River thickness throughout the Black Hills is highly variable and ranges from 10-200 feet (3-61 m) in thickness. A typical outcrop is characterized by massive fluvial sandstones underlain by interbedded siltstone and mudstone. For purposes of the present study the formation is considered as three units (Waage 1959; Gott, et ai 1974).
153
interbeds are absent and the fluvial sandstone of the upper unit rests directly on the massive sandstone of the middle unit. For consistency, the upper unit was used for joint measurements, except where absent or concealed at a few locations. There, joints from the sandstones of the middle unit were measured.
Joints
All exposures of Fall River Formation sandstones are pervasively fractured (Fig. 11). Joints are perpendicular to bedding, even in steeply inclined strata. Joint spacing ranged from 4 inches to 50 feet (0.1-15.2m), with fractures penetrating strata with bedding partings from 1 inch to 50 feet (0.025-15.2m) in thickness. The fractures in the massive sandstone are large and continuous, had typical spacings of 5 feet (1.5m) or more and penetrated the entire thickness of the bed (Fig. 1). Joint spacing in thinbedded units was typically from 1 inch to 5 feet • Upper unit - massivefluvialsandstone, similar (0.025-1.5m). to that in the middle unit. Underlain by interJoint faces commonly present a morphology beds of sandstone and mudstone, grading that is useful in ascertaining mode of origin and laterally into variegated mudstone. particular aspects of fracture propagation. The • Middle unit - massivefluvialsandstone undermost common feature is the joint plume, which lain by interbedded sandstone and mudstone. is discussed at length by Kulander et al. (1979) • Lower unit - thin interbeds of sandstone, siltand Pollard & Aydin (1988). The present study stone and some mudstone. was mainly concerned simply with the presence This sequence varies locally as well as regionally. or absence of the plume as a means to determine At some localities, the sandstone-mudstone whether or not the fracture formed as an extension joint. Fracture plumes were present on virtually all Fall River Formation sandstone joint faces and verify an extensional origin. The joint surfaces show no evidence of horizontal slickensides. There is no offset on joints. The joints are orthogonal systems and show no evidence for a conjugate shear origin. It is likely that stress relief played a role in fracture development, as it commonly the case with vertical bedrock faces of natural or manmade origin. This possibility exists with regard to the joint exposures on Fall River cliff faces. Commonly, undercutting by erosion causes bedrock to break off toward the unconfined face, with joint initiation at the base of the undercut exposure, resulting in a bottom origin and ensuing upward joint propagation. However, joints originating by this process are more typically curviplanar and show no tendency to fit the local, or regional, fracture pattern. In view of the fact that the joints in the present study have pervasive, planar faces that fit the regional Fig. 10. Generalized stratigraphic section for Lower pattern, it is concluded that they are of tectonic Cretaceous formations (Fig. from Weimer et al. 1982; courtesy of the Colorado School of Mines Press.) origin.
154
J. L. WICKS ET AL.
Fig. 11. Photograph of the west flank of the Cascade anticline 7 miles (11.3 km) south-southwest of Hot Springs. South Dakota. View is perpendicular to strike and the strata dip 60 towards the viewer. Systematic joints trend from upper right to lower left, oblique to strike. Non-systematic joints are orthogonal to systematics. All joints are perpendicular to bedding and were rotated with folding. The distance is 600 feet (183m) along the base of the figure.
Fracture domains
Southern domain
Systematic joint sets and associated nonsystematic sets often occur in distinct domains. The word domain, as used here, is synonymous with the term 'fracture field1 of Hobbs (1911). A fracture domain consists of a regional, uniform, systematic fracture set or sets, with associated non-systematic joints which are pervasive in a rock unit within the boundaries of a particular region. Fracture domain boundaries mark the zone where a regional fracture trend changes or a new set begins. Domain boundaries may be gradational or abrupt and several joint sets can occur within the same domain. In other regions of the Rocky Mountains joint domains are well demonstrated by Kelley & Clinton (1960) and Hodgson (1961, 1965) in the Colorado Plateau. Fractures in Fall River sandstone are not randomly oriented. Systematic and abutting non-systematic joint sets maintain consistent orientations throughout large areas. The fractures in the investigated area can be separated into distinct southern and northern domains (Fig. 12). Each domain has at least one consistent orthogonal system. In the northern domain three systems of systematic and non-systematic joints are present. The domains defined by the fracture sets in the study area are evident in Figs 3, 4 and 12.
The eastern end of the NW-trending Rapid Creek lineament (Pan 1978) serves as an approximate boundary between the northern and southern domains on the eastern flank of the Black Hills (Fig. 13). Here the systematic joint set strikes N75 ; W (285 s ) and parallels the boundary lineament. The strike of the systematic joint set shows a gradual shift to N30"W (330:) as the domain continues southwestward around the Fall River outcrop belt to the Long Mountain structure zone, which is the southern limit of the domain boundary on the southwestern flank of the Black Hills. The joint set in the southern domain is consistent and well defined. These joints were easily identified at every outcrop and showed little variance from field station to field station, or from adjacent aerial photographs. Northern domain The northern domain includes the remainder of the Fall River outcrop belt. This domain contains three distinct systematic fracture sets (Figs 3 and 4). The major set, designated set I, trends N30 D E-N75 3 E (030;-075°) and is
FRACTURE PATTERNS AROUND FORELAND UPLIFT
155
Fig. 12. Diagram showing the major joint domain boundary and systematic fracture trends around the Black Hills, extended where the Fall River Formation is absent. The projected crest of the Belle Fourche Arch is shown between the arrows in the northwestern Hills. (Refer to Fig. 3 for some abbreviations of towns.)
continuous throughout the northern domain. Set II, N50°W-N90°W (310°-270°) coincides with some of the systematic fractures of the southern domain. This set crosses the domain boundary on the eastern flank and is only present northward to the vicinity of Sturgis. On
the northwestern flank, this set appears nearly along the line of extension of the Rapid Creek lineament to the Little Missouri structure zone. Set III trends N10°W-N20°E (350°-020°) and occurs only in the extreme northern Black Hills.
156
J. L. WICKS ET AL.
Fig. 13. Lineament map of the Black Hills area (modified from Pan 1978). The projected crest of the Belle Fourche Arch is shown by arrows in the northwestern Hills.
Domain boundaries The domain boundary is well defined at the southwestern end by the Long Mountain structure zone and at the eastern end by the southeastern termination of the Rapid Creek lineament. Along the Long Mountain structure zone stations on both sides of the boundary
confirm the abrupt change in systematic fracture orientation. There are no stations on either side of the boundary that contain both sets. At the northeastern end of the domain boundary both set I and II joints are present north of the boundary. The NE-trending joints are most pervasive and the northwest joint set is secondary. The domain boundary is expressed further
FRACTURE PATTERNS AROUND FORELAND UPLIFT
into the Black Hills then the Fall River outcrop belt, as shown by fracture-controlled cave passages in Mississippian Pahasapa (i.e. Madison) carbonates (Bakalowicz et al. 1987; Peter et al. 1991). Development of Fall River joints Joint readings were obtained in the Fall River Formation on every major and local flexure around the Black Hills. After bedding was restored to horizontal, comparison of joint trends (Figs 3 and 4) with fold axial traces reveals that the fracture orientations remain constant, although the structural trends of many of these flexures vary considerably (Fig. 3). Clearly the joints formed prior to development of these folds and were merely rotated during later folding. Figures 3 and 4 give the suggestion of radial systematic joint trends around the Black Hills, a configuration characteristic of domal uplifts. However, this interpretation is not supported when the overall joint pattern is considered. In the southern Hills a 90° difference in trends exists between set I and set II, requiring the principal extension direction to change abruptly along the major domain boundary that crosses the southern Black Hills. Additional difficulties arise with this interpretation because of the prevalence and consistency of set I trends around the western and northern parts of the uplift that do not fit a radial pattern. Finally, it has been shown by Lisenbee (1978, 1985), DeWitt et al (1989) and Lisenbee & DeWitt (1993) that the folds around the Black Hills were initiated during the uplift. Fractures pre-date development of these flexures and do not fit an overall radial pattern around the Hills. Lateral compression has been invoked by Gott et al. (1974) to account for jointing in the Inyan Kara Group of the western and southern Black Hills. They proposed that the change in joint trends across the Long Mountain structural zone, which is the major domain boundary of this paper, resulted from a change in an overall northeastern-southwestern compression north of Long Mountain zone, to southeastern-northwestern compression south of this zone. The work of Gries (1983, 1990), Gries et al (1993) and Chapin & Gather (1981, 1983) has also suggested changes in regional compression from a southwesterly to northwesterly direction during the evolution of the Rocky Mountain foreland, which would require NE-trending systematic joints (set I) of the northern domain to form first with cessation of development of this set along the abrupt line of the major domain
157
boundary. Northwest-trending set II systematic joints would then have to form later, principally in the southern domain. Set II joints do continue along the eastern flank of the Black Hills and along the northern hills in a broad band through the northern domain, but with an overall change in strike from approximately N40°W to N60°W (320°-300°) from the southern to northern domains. Several problems exist with this twofold timing of regional lateral compression in systematic joint formation. The change in trend of set II joints from the southern to northern domains would require a major change in regional compression direction and would not account for the complete absence of set II joints in the western hills from Linden, Wyoming, to the Long Mountain structural zone in south Dakota, a distance of some 75 miles (121km). An additional difficulty occurs in the northern hills where another systematic set emerges trending approximately N12°E (012°). Finally, proprietary seismic data reveal a Ntrending, E-dipping Precambrian basement thrust fault that underlies the Old Woman anticline of the same trend well past the major domain boundary in the southern hills. Approximately east-west compression would have to account for the development of the Old Woman anticline, as well as the simultaneous development of sets I and II, even though the systematics trend at 90° different orientations. Regional compression may be a possibility for the Black Hills uplift and some peripheral structures but cannot account for the joint patterns, which have been shown to be of pre-fold origin. The pattern and chronology of Fall River joint development requires an explanation unrelated to domal uplift, formation of local flexures or regional compression. A possible, but speculative, interpretation lies in the major structural zones and lineament systems that cross and border the Black Hills. The lineament systems (Figs 3, 4, 13, 14 and 15) were interpreted primarily from Landsat imagery and isopach maps by several workers. Lineaments of western South Dakota (exclusive of the Black Hills) were interpreted by Shurr (1978). Lineaments across the Black Hills structure were taken from the lineament map by Pan (1978), (Fig. 13). Papers by Thomas (1974) and Brown (1978) provided lineament data for southern Montana and northeastern Wyoming, respectively. The lineaments interpreted by Brown (1978), to some extend overlap those of Pan (1978). Shurr (1978, 1979), Shurr & Rice (1986), and Shurr et al (19890, 1994) have shown from isopach and structure contour maps that this
158
J. L. WICKS ET AL.
Fig. 14. Lineament map of the Black Hills and surrounding area. Data from Thomas (1974), Shurr (1978) and Brown (1978); Black Hills. Pan (1978). The projected crest of the Belle Fourche Arch is shown between the arrows in the northwestern Hills.
lineament system was established well before the Laramide orogeny affected the Black Hills. Shurr (1981), Shurr (1984), Shurr et al. (19890, b, 1994), Shurr & Gay (1992), Shurr & Rice (1986) have documented the inter-relationship of basement block geometry and large-scale lineaments on palaeotectonism, subtle structures and the stratigraphic record in the western and northwestern midcontinent area adjacent to the Black Hills. Slack (1981) (Fig. 15) has shown the dominance of NE-trending lineaments across the Powder River basin and well into the Black Hills uplift. He suggests that zones of subtle movement along basement faults throughout Phanerozoic time have defined the lineament system. Marrs & Raines (1984) (Fig. 15) also show pronounced NE-trending lineaments throughout the Powder River basin, but also a strong northwest trend. North of the Black Hills in the Williston basin, numerous workers, including Thomas (1974) and Gerhard et al.
(1982) have shown NE- and NW-trending regional lineaments and have discussed the role of recurrent movement of basement faults on sedimentation, and development of folds in the sedimentary cover. There are several examples of the correlation between systematic joints, changes in joint trends and structurally defined lineament zones (Figs 3, 6, 13, 14 and 15). The most prominent example is the Dewey-Long MountainEdgemont zone in the southern Hills, which, collectively, suggest a line of deep-seated fundamental weakness in the crust with the same trend as the Colorado lineament (Warner 1978) to the south in Colorado, and the Cheyenne belt in southeastern Wyoming. The Cheyenne belt marks a distinct change between Precambrian rocks of Archean and Proterozoic age, a NE-trending boundary that extends into the southern Black Hills (Lisenbee & DeWitt 1993) (Fig. 16).
FRACTURE PATTERNS AROUND FORELAND UPLIFT
159
Fig. 15. Lineaments across the Powder River basin into the Black Hills. (Diagram modified from Slack (1981) and Marrs & Raines (1984); courtesy of the American Association of Petroleum Geologists.)
Also significant is the northwesterly alignment of Tertiary intrusions, faults in Precambrian rock and the Rapid Creek lineament across the Black Hills. In addition, in the northwestern Hills, the Little Missouri structure zone and the pronounced northeasterly trend of the Belle Fourche River are matched by Landsat lineaments. In this same area the Belle Fourche Arch of Slack (1981) is well defined along its northern border by a major lineament. Figures 13-15 show that the dominant lineament trends across the Black Hills are in a
northeasterly direction. Slack (1981) (Fig. 15) has projected NE-trending lineaments from the Powder River basin into the northwestern Black Hills, and shows the Little Missouri structural zone and the Belle Fourche River zone (i.e. Rozet lineament) to be positioned on the northern limb of the Belle Fourche Arch, a subtle NE-trending basement high during much of Cretaceous time. Dolson et al. (1991) have shown the presence of this arch as a major drainage divide in the northern Black Hills during the time of Muddy deposition (i.e. upper Lower
160
J. L. WICKS ET AL.
Fig. 16. Map showing the generalized boundary between Precambrian terrain in the Black Hills and the surrounding region. Younger Proterozoic rocks to the east and south shown with diagonal patterns. Older Archean rocks of Wyoming Province are shown in white. (Diagram from Lisenbee & DeWitt 1993; courtesy of the Wyoming Geological Survey).
Cretaceous). The lineaments north of the crest of the arch are interpreted to represent the surface expressions of deep-seated faults that were consistently down-thrown to the northwest during Cretaceous time. The most pervasive and consistently formed first-formed joints there are of set I. Southeast of the arch, a series of lineaments of similar trend extend into the Black Hills and have also been interpreted to bound NE-trending basement blocks, that were repeatedly down-faulted to the southeast towards the southern Black Hills. Set I joints are also the most pervasive and consistently first formed there. The eastern projection of the Belle Fourche Arch to a location northwest of Belle Fourche shows a dominance of set III joints, but with some variance about the overall N12°E (012°) mean. Jointing here may have an early association with similarly oriented basement faults that were later reactivated to form the numerous N-S- to NNW-trending anticlines in the extreme northern
hills. The Nesson and Cedar Creek anticlines in the Williston basin of North Dakota and Montana, respectively, have been shown by Gerhard et al. (1982) to be bounded by basement faults that were periodically reactivated during Palaeozoic and Mesozoic time. Geophysical anomalies shown by Robbins (1994) make a convincing argument for a segmented Precambrian basement in the northern and northwestern Black Hills. Magnetic and gravity trends are well developed and have the same directions as all three systematic joint sets in the region. Jointing apparently occurred early in the depositional history of the Fall River Formation, resulting in widespread extension joints. The mechanism shown by Slack (1981) for the evolution of the Belle Fourche Arch across the northern hills and the numerous NE-trending lineaments across the Powder River basin into the uplift provides a plausible explanation for the dominant NE-trending joint set (set I) throughout the Black Hills. Under this scenario, faulting parallel to NE-trending lineaments occurred shortly after Fall River deposition, exerting extensional stresses to the northwest and southeast of the Belle Fourche Arch. Set I fractures formed as extension joints parallel to basement-controlled structural zones, now expressed as lineaments (Fig. 15) Slack (1981) and numerous other authors, including Emme & Weimer (1981), Weimer et al. (1982), Anna (1983), Moore (1983), Tillman & Martinsen (1983), Wheeler & Gustason (1987), Bryan & Petta (1988), Gustason et al. (1988), Weimer et al. (1988), Reinhold & Harwerth (1989). Rice & Keighin (1989), Forster & Home (1994) and Way et al. (1995), have shown the effect of recurrent basement fault movements on depositional patterns in the Powder River basin. Recurrent movement on basement faults, in general, throughout sedimentary depositional history in the Rocky Mountain foreland and mid-continent area is a familiar theme and has been shown by Sonnenberg and Weimer (1982), Anderson et al. (1983), Clement (1983), Weimer (1984, 1992). Lindsay et al. (1988), Weimer & Sonnenberg (1989), Freisatz (1990), Rice et al. (1990), Von Den Bold et al. (1991), May et al. (1992) and other workers to affect and control changes in facies and stratigraphic thickness. The most compelling support for NE-trending extensional jointing in the northern domain is given by Weimer et al. (1982) in a comprehensive treatment of depositional patterns of Lower Cretaceous formations in the southeastern Powder River basin and adjacent to the surface outcrop of the Fall River. They have shown that
FRACTURE PATTERNS AROUND FORELAND UPLIFT
recurrent movement of basement fault blocks during Early Cretaceous time created subtle topographic highs and lows that influenced sedimentation. In the northern joint domain palaeotopographic highs and lows exactly parallel the trends of set I joints to the position of the Long Mountain structure zone and suggest that Fall River jointing occurred before deposition of the Newcastle Formation (Lower Cretaceous). Around the southern Hills Gott et al. (1974) have shown that Early Cretaceous Fall River drainage was dominantly to the northwest and that this palaeodrainage was diverted to the northeast and northwest at the Long Mountain and Dewey structural zones. Variable total thickness of the Lower Cretaceous Inyan Kara Group, and individual fluvial sandstone units (Gott et al. 1974), indicates that sections of present anticlines such as Dudley, Cascade, Chilson and Sheep Canyon probably had slight topographic relief along north to northwest trends during Early Cretaceous time. Major structural relief now present along the Dudley, Cascade, Chilson and Sheep Canyon folds was inherited from later Laramide deformation. The authors feel that set II NWtrending joints also formed before Newcastle deposition and parallel to similarly trending extensional faults that were active during Early Cretaceous time. Although Gott et al. (1974) have indicated that the area south of the Dewey and Long Mountains zones experienced relative down-faulting, it is felt that this faulting occurred in differential fashion along NW-trending blocks, yielding the NW-trending systematic joints of set II. In the northern Black Hills a similar picture may exist for joints of set III (N12°E). Eastwest extension associated with movement on N-S- and NNW-trending basement faults may have initiated joint formation during Early Cretaceous time, with major folds such as the Belle Fourche and Laflamme anticlines forming during Laramide deformation along the same lines of basement weakness. The existence of N-S- and NNW-trending basement faults associated with the Nesson and Cedar Creek anticlines in the Williston Basin to the north (Gerhard et al. 1982) makes basement faulting a plausible explanation for the development of set III joints in the northern Hills. The consistent and pervasive northeast trend of set I joints is interrupted by the broad northwesterly belt of joints across the northern domain parallel to the Rapid Creek lineament. These joints are positioned on the crest of the Belle Fourche Arch on the northwestern edge of the hills and extend to just south of Sturgis on the eastern limb of the uplift. Their timing is
161
secondary on the northwestern and eastern flank of the hills, indicating that they formed after set I joints in both areas. Their overall N75-80°W (285°-280°) trend indicates that they may represent a different joint set than those in the southern domain. They do not appear to have any association with fault-related Early Cretaceous palaeohighs and palaeolows and sedimentation patterns as do sets I and II throughout much of the Black Hills. They may represent minor north-northeast to south-southwest extension that postdated the development of the Belle Fourche Arch and associated NEtrending lineaments and basement faults. Summary Sandstones of the Fall River Formation are pervasively jointed throughout this outcrop belt around the Black Hills uplift, with systematic fractures maintaining consistent orientations over large areas. Two major systematic joint sets (I and II) are present, with a sharp boundary between these sets in the southern hills along the NE-trending Long Mountain structure zone. This feature separates NE-striking systematics (set I) of the northern domain from NW-trending dominant joints (set II) of the southern domain. A third systematic joint (set III) is locally present across the northern hills and trends north-northeast. The stratigraphic record reveals that the Black Hills uplift was initiated late in the Laramide orogeny in Early Palaeocene time, with continued movements until Late Palaeocene-Early Eocene. Uplift is interpreted to have been effected along an east-dipping master thrust fault in the upper crust which may have been localized, in part, by the boundary between Precambrian age Archean rocks to the west and Proteozoic rocks to the east. This fault is interpreted to penetrate into, or near, the Palaeozoic-Precambrian interface at the position of the Fanny Peak and Black Hills monoclines along the western border of the Hills. The numerous monoclines and anticlines around the periphery of the uplift are interpreted to be drape-type folds, formed above steeply dipping Laramide reverse faults, that originated in basement rocks along pre-existing weakness lines inherited from Precambrian time, and were periodically reactivated as extensional faults throughout Palaeozoic and Mesozoic time prior to the Laramide orogeny. Jointing in the Fall River Formation was not associated with the domal uplift of the Black Hills, and was in no way related to Laramide regional compressive stresses or extensional
162
J. L. WICKS ET AL.
effects along the local and regional cover folds in the Fall River around the uplift. Joint development originated early in the lithification history of Fall River sandstones, most probably in late Early Cretaceous time prior to Newcastle deposition. Jointing occurred in response to local and regional extensional stresses that pervaded the northern and southern domains as a result of recurrent movements on basement faults that parallel the regional lineament system and surface structural zones throughout the Black Hills and surrounding region. The authors wish to express their appreciation for the comprehensive reviews of the original manuscript by S. Laubach & D. Pollard.
References ALLMENDINGER. R. W.. BREWER, M. A., BROWN, L. D., KAUFMAN, S.. OLIVER. J. E. & HOUSTON, R. S. 1982. COCORP profiling across the Rocky Mountain front in southern Wyoming, part 2: Precambrian basement structure and its influence on Laramide deformation. Geological Society of America Bulletin. 93, 1253-1263. . BROWN. L. D.. OLIVER, J. E. & KAUFMAN. S. 1983. COCORP deep seismic profiles across the Wind River Mountains. Wyoming. In: BALLV. A. W. (ed.) Seismic expression of structural styles. American Association of Petroleum Geologists, Studies in Geology Series. 15. 3, 3.2. 1.29-1.33. ANDERSON. S. B., GERHARD, L. C. & LEFEVER, J. 1983. Structural and sedimentologic history of Nesson anticline. AAPG Bulletin. 67. 1329. ANNA. L. D. 1983. Structural influence on Lower and Middle Cretaceous sedimentation. Northern Great Plains. AAPG Bulletin. 67, 1329. BAKALOWICZ. M. J., FORD, D. C., MILLER, T. E.. PALMER, A. N. & PALMER. M. V. 1987. Thermal genesis of dissolution caves in the Black Hills, South Dakota. Geological Society of America Bulletin. 99, 729-738. BERG, R. R. 1962. Mountain flank thrusting in Rocky Mountain foreland. Wyoming and Colorado. AAPG Bulletin. 46. 2019-2032. BERGENDAHL, M. H.. DAVIS. R. E. & IZETT, G. A. 1961. Geology and mineral deposits of the Carlile quadrangle. Crook County, Wyoming. United States Geological Survey Bullet in. 1982-J, 613-705. BREWER. J. A.. ALLMENDINGER, R. W., BROWN, L. D., OLIVER. J. E. & KAUFMAN, S. 1982. COCORP profiling across the Rocky Mountain front in southern Wyoming, Part 1. Geological Society of America Bulletin. 93. 1242-1252. BROBST. D. A. & EPSTEIN. J. B. 1963. Geology of the Fanny Peak quadrangle, Wyoming-South Dakota. United States Geological Survey Bulletin. 1063-1. 323 377 BROWN. D. L. 1978. Wrench-style deformational patterns associated with a meridional stress axis recognized in Paleozoic rocks in parts of Montana.
South Dakota, and Wyoming. ///: EconomicGeology of the Williston Basin. 24th Annual Conference of the Montana Geological Society. 17-31. BROWN. W. G. 1988. Deformational style of Laramide uplifts in the Wyoming foreland. In: SCHMIDT. C. J. & PERRY. W. J.. JR (eds) Interaction of the Rocky Mountain Foreland and the Cordilleran Thrust Belt. Geological Society of America. Memoir 171. 1-25. 1993. Structural style of Laramide basementcored uplifts and associated folds. ///: SNOKE. A. W.. STEIDTMANN. J. R. & ROBERTS. S. M. (eds) Geology of Wyoming. Geological Survey of Wyoming. Memoir 5. 374-412. BRYAN. J. G. & PETTA. T. J. 1988. Regional tectonic influence on Early Cretaceous depositional patterns in Powder River Basin. Wyoming and Montana. AAPG Bulletin. 72. 166. CHAPIN. C. E. & CATHER. S. M. 1981. Eocene tectonics and sedimentation in the Colorado Plateau Rocky Mountain area. In: DICKENSO.V W. R. & PAYNE. M. D. (eds) Relations of Tectonics to Ore Deposits in the Southern Cordillera. Arizona Geological Digest. 14. 173-198. 1983. Eocene tectonics and sedimentation in the Colorado Plateau-Rocky Mountain area. ///: LOWELL. J. D. (ed.) Rocky Mountain Foreland Basins and Uplifts. Rocky Mountain Association of Geologists. 9 32. CLEMENT. J. H. 1983. Cedar Creek - a significant paleotectonic feature of the Williston Basin. AAPG Bulletin. 67, 1332. DARTON. N. H. 1901. Preliminary description of the geology and water resources of the southern half of the Black Hills and adjoining regions in South Dakota and Wyoming. ///: United States Geological Survey. Annual Report, part 4. United States Geological Survey. 489-590. 1902. Description of the Oelrichs Quadrangle. South Dakota. United States Geological Survey. Geological Atlas. Folio 85. 1904. Description of the Newcastle Quadrangle. Wyoming-South Dakota. United States Geological Survey. Geological Atlas. Folio 107. 1905a. Description of the Sundance Quadrangle. Wyoming-South Dakota. United States Geological Survey. Geological Atlas. Folio 127. 1905b. Description of the Aladdin Quadrangle. Wvoming-South Dakota. United States Geological Survey. Geological Atlas. Folio 128. (STO'HARA. C. C. 1907. Description of the Devil's To\\~er Quadrangle. Wyoming-South Dakota. United States Geological Survey. Geological Atlas. Folio 150 1909. Description ofthe Belle Fourche Quadrangle. South Dakota. United States Geological Survey. Geological Atlas. Folio 164. & PAIGE, S. 1925. Description of the Central BlackHills. South Dakota. United States Geological Survey. Geological Atlas. Folio 219. & SMITH, W. S. T. 1904. Description of the Edgemont Quadrangle. South Dakota. United States Geological Survey. Geological Atlas. Folio 108.
FRACTURE PATTERNS AROUND FORELAND UPLIFT DEWITT, E., REDDEN, J. A., WILSON, H. B. & BUSCHER, D. 1986. Mineral Resource Potential and Geology of the Black Hills National Forest, South Dakota and Wyoming, With a Section on Salable Commodities by Dersch, J. S. United States Geological Survey, Bulletin 1580. , , BUSCHER, D. & WILSON, A. B. 1989. Geological map of the Black Hills area, South Dakota and Wyoming. United States Geological Survey, Miscellaneous Investigations Series, Map 1-1910. DICKENSON, W. R. & SNYDER, W. S. 1978. Plate tectonics of the Laramide orogeny. In: MATTHEWS, V. Ill (ed.) Laramide Folding Associated with Basement Block Faulting in the Western United States. Geological Society of America, Memoir 151, 355-366. DOLSON, J., MULLER, D., EVETTS, M. J. & STEIN, J. A.
1991. Regional paleotopographic trends and production in Muddy Sandstone (Lower Cretaceous), central and northern Rocky Mountains. AAPG Bulletin, 75, 409-435. DUTTON, S. P. 1995. Interpretation of compaction and quartz cementation processes from integrated petrographic and burial history analyses, Lower Cretaceous Fall River Formation, Wyoming and South Dakota. In: American Association of Petroleum Geologists National Meeting, Abstracts with Program, 24A. EMME, J. J. & WEIMER, R. J. 1981. Tectonic influence on sedimentation of Lower Cretaceous strata, eastcentral Powder River Basin, Wyoming. AAPG Bulletin, 66, 567. FORSTER, J. R. & HORNE, J. C. 1994. The Wright fault zone, Powder River Basin: An example of lineament control on hydrocarbon distribution. In: American Association of Petroleum Geologists National Meeting, Abstracts with Program, 151. FREISATZ, W. B. 1990. Fracture-enhanced porosity and permeability trends in Bakken Formation, Williston Basin, western North Dakota. AAPG Bulletin, 74, 1324. GERHARD, L. C., ANDERSON, S. B., LEFEVER, J. A. & CARLSON, C. G. 1982. Geological development, origin, and energy mineral resources of Williston Basin, North Dakota. AAPG Bulletin, 66, 9891020. GOTT, G. B., WOLCOTT, D. E. & BOWLES, G. C. 1974. Stratigraphy of the Inyan Kara Group and Localization of Uranium Deposits, Southern Black Hills, South Dakota and Wyoming. United States Geological Survey, Professional Paper, 763, 57. GRIES, R. 1983. North-south compression of Rocky Mountain foreland structure. In: LOWELL, J. D. (ed.) Rocky Mountain Foreland Basins and Uplifts. Rocky Mountain Association of Geologists, 9-32. 1990. Rocky Mountain foreland structures: changes in compression through time. In: LETOUZEY, J. (ed.) Petroleum and Tectonics in Mobile Belts. IFP Exploration and Production Research Conferences, Editions Technip, 129-147. , DOLSON, I. C. & RAYNOLDS, R. G. W. 1993. Structural and stratigraphic evolution and hydrocarbon distribution, Rocky Mountain foreland. In: MACQUEEN, R. W. & LECKIE, D. A. (eds)
163
Foreland Basin and Fold Belts. American Association of Petroleum Geologists, Memoir, 55,395-425. GUSTASON, E. R., WHEELER, D.A. & RYER, T. A. 1988. Structural control on paleovalley development, Muddy Sandstone, Powder River Basin, Wyoming. AAPG Bulletin, 72, 871. HOBBS, W. H. 1911. Repeating patterns in the relief and in the structures of the land. Geological Society of America Bulletin, 22, 123-176. HODGSON, R. A. 1961. Regional study of jointing in Comb Ridge Navajo Mountain area, Arizona and Utah. AAPG Bulletin, 45, 1-38. 1965. Genetic and geometric relations between structures in basement and overlying sedimentary rocks, with examples from the Colorado Plateau and Wyoming. AAPG Bulletin, 49, 935-949. JOHNSON, R. A. & SMITHSON, S. B. 1985. Thrust faulting in the Laramie Mountains, Wyoming, from reanalysis of COCORP data. Geology, 13, 534-537. JORDAN, T. E. & ALLMENDINGER, R. W. 1986. The Sierras Pampeanas of Argentina - A modern analog of Rocky Mountain foreland deformation. American Journal of Science, 286, 737-764. KELLEY, V. C. & CLINTON, N. J. 1960. Fracture Systems and Tectonic Elements of the Colorado Plateau. University of New Mexico Publications in Geology, 6. KLEINKOPF, M. D. & REDDEN, J. A. 1975. Bouguer Gravity, Aeromagnetic, and Generalized GeologicMaps of Part of the Black Hills of South Dakota and Wyoming. United States Geological Survey, Geophysical Investigations, Map GP-903. KULANDER, B. R., BARTON, C. C. & DEAN, S. L. 1979. The Application of Fractography to Core and Outcrop Fracture Investigations. United States Department of Energy, Morgantown Energy Tech. Centre, Morgantown, WV. KULIK, D. M., & SCHMIDT, C. J. 1988. Region of overlap and styles of interaction of Cordilleran thrust belt and Rocky Mountain Foreland. In: SCHMIDT, C. J. & PERRY, W. J., J[r. (eds) Interaction of the Rocky Mountain Foreland and the Cordilleran Thrust Belt. Geological Society of America, Memoir, 171, 75-98. LINDSAY, R. F., ANDERSON, S. B., LEFEVER, J. A., GERHARD, L. C. & LEFEVER, R. D. 1988. Structural history and reservoir characteristics (Mississippian) of Nesson anticline, North Dakota. AAPG Bulletin, 72, 875. LISENBEE, A. L. 1978. Laramide structure of the Black Hills uplift, South Dakota—Wyoming-Montana. In: MATTHEWS, V. Ill (ed.) Laramide Folding Associated with Basement Block Faulting In The Western United States. Geological Society of America, Memoir, 151, 165-196. 1985. Tectonic Map of the Black Hills Uplift, Montana, Wyoming, and South Dakota. Geological Survey of Wyoming, Map Series, 13, scale 1:250000. 1988. Tectonic history of the Black Hills uplift. In: DIEDRICH, R P., DYKA, M. A. K. & MILLER, W. R. (eds) Eastern Powder River Basin-Black Hills: Wyoming Geological Association 39th Annual Field Conference Guidebook, 45-52.
164
J. L. WICKS ET AL.
& DEWITT, E. 1993. Laramide evolution of the Black Hills. In: SNOKE, A. W., STEIDTMANN, J. R. & ROBERTS, S. M. (eds) Geology of Wyoming. Geological Survey of Wyoming. Memoir, 5, 374-412. LOWELL. J. D. 1974. Plate tectonics and foreland basement deformation. Geology. 2. 275-278. MARKS. R. W. & RAINES. G. L. 1984. Tectonic framework of Powder River Basin. Wyoming and Montana, interpreted from Landsat Imagery. AAPG Bulletin. 68, 1718-1731. MAY. M. T., PURER. L. C.. MEYERS, J. H.. SUTTNER. L. J.. KVALE. E. P. & DECELLES, P.E. 1992. Complexities in fluvial stratigraphic sequences in medial to distal portions of a developing foreland basin. An example from the Lower Cretaceous of Wyoming and Montana. In: American Association of Petroleum Geologists National Meeting. Abstracts with Program. 84. MOORE. R. W. 1983. Fault control of channel sandstones in the Dakota Formation, southwest Powder River Basin. Wyoming. AAPG Bulletin. 67. 1350. NOBLE. J. A. 1952. Structural features of the Black Hills and adjacent areas developed since Precambrian time. In: Wyoming Geological Association-Billings Geological Society. Third Annual Field Conference Guidebook. 31-37. PAN. K. L. 1978. Regional Geologic Analysis of the Black Hills of South Dakota and Wyoming from Remote Sensing Data. PhD Dissertation. University of Iowa. PETER, k. D.. MILLS, K. R. & LASKET, C. L. 1991. Geologic Structure and Altitude of the Top of the Minnelusa Formation and Orientation of Mapped Cave Passages in the Madison Limestone, Southern Black Hills. South Dakota. United States Geological Survey Water Resources Investigations. Report. 86,4167. POLLARD. D. D. & AYDIN. A. 1988. Progress in understanding jointing over the past century. Geological Society of America Bulletin. 100, 1181-1204. PRUCHA, J. J.. GRAHAM. J. A. & NICKELSEN, R. P. 1965. Basement controlled deformation in Wyoming province of Rocky Mountain foreland. AAPG Bulletin. 49. 966-992. REINHOLD. M. L. & HARWERTH. L. A. 1989. Evidence for local tectonism related to a Mid-Turonian unconformity beneath Upper Cretaceous Turner Sandy Member of Carlisle Formation, northeastern Powder River Basin, Wyoming. AAPG Bullet in. 13. 1171. RICE. D. D. & KEIGHIN. C. W. 1989. Configuration of shelf sandstone oil reservoirs. Upper Cretaceous (Turonian) Turner Sandy Member of Carlisle Shale. Powder River Basin, Wyoming. AAPG Bulletin. 73. 405. . COBBAW. W. A. & SHURR. G. W. 1990. Cretaceous rocks of Cedar Creek anticline area, eastern Montana and southwestern North Dakota. AAPG Bulletin. 74, 1342-1343. ROBBINS, S. L. 1994. Gravity and areomagnetic studies of the Powder River Basin and surrounding areas, southeastern Montana, northeastern Wyoming.
and western South Dakota. United States Geological Survey Bulletin 1917. 17. SALES. J. K. 1968. Cordilleran foreland deformation. AAPG Bulletin, 52. 2000-2015. SCHMIDT. C. J.. EVANS. J. P.. FLETCHER. R. C. & SPANG. J. H. 1985. Spacing of Rocky Mountain foreland arches and Laramide magmatic activity. Geological Society of America. Abstracts with Program. 17. 710. SHAPIRO. L. H. 1971. Structural Geology of the Black Hills Region and Implications for the Origin of the Uplifts in the Middle Rockv Mountain Province. PhD Dissertation. University of Minnesota. SHURR. G. W. 1978. Paleotectonic controls on Cretaceous sedimentation and potential gas occurrences in western South Dakota. In: Economic Geology of the Willistou Basin. 24th Annual Conference of the Montana Geological Society. 283-292. 1979. Lineament control of sedimentary fades in the northern Great Plains. United States. In: PODWYSOCKI. M. H. & EARLE. J. L. (eds) Proceedings of the Second International Conference on Basement Tectonics. Basement Tectonics Committee. 413 422. 1981. Lineaments as basement-block boundaries in western South Dakota. In: Proceedings of Third International Conference on Basement Tectonics. Basement Tectonics Committee. 177 184. 1984. Regional setting of Niobrara Formation in northern Great Plains. AAPG Bulletin. 68. 598 609. . ANNA. L. O. & PETERSON. J. A. 1989a. Zuni Sequence in Williston Basin evidence for Mesozoic Palaeotectonism. AAPG Bulletin. 73. 68-87. . LARSON. B. S. & WATKINS. I.W. 1989b. The basement block mosaic beneath the Montana plains. In: FRENCH. D. E. & GRABB. R. F. (eds) GeologicResources of Montana. Montana Geological Society. Centennial Field Conference Guidebook. 299 309. & GAY. S. P.. JR. 1992. Cretaceous paleofaults on basement block boundaries near Hartzog Draw Field. Wyoming. AAPG Bulletin. 76. 1269. & RICE. D. D. 1986. Paleotectonic controls on depositions of the Niobrana Formation. Eagle sandstone and equivalent rocks (Upper Cretaceous), Montana and South Dakota. ///: PETERSON. J. A. (ed.) Palaeotectonics and sedimentation in the Rocky Mountain region. American Association of Petroleum Geologists Memoir. 41. 193-211. , HAMMOND. R. H. & BRETZ. R. F. 1994. Cretaceous paleotectonism and post depositional tectonism in southcentral South Dakota: an example of epeirogenic tectonism in continental lithosphere. ///: SHI RR. G. W.. LLDVIGSON. G. A. & HAMMOND, R. H. (eds) Perspectives on the Eastern Margin of the Cretaceous Western Interior Basin. Geological Society of America. Special Paper 287. 237-256. SLACK, P. B. 1981. Paleotectonics and hydrocarbon accumulation. Powder River Basin. Wyoming. AAPG Bulletin. 65. 730 743.
FRACTURE PATTERNS AROUND FORELAND UPLIFT SMITHSON, S. B., BREWER, J. A., KAUFMAN, S., OLIVER, J. E. & HURICH, C. H. 1978. Nature of the Wind River thrust, Wyoming, from COCORP deep reflection data and from gravity data. Geology, 6, 648-652. , , , & 1979. Structure of the Laramide Wind River uplift, Wyoming, From COCORP deep reflection data and from gravity data. Journal of Geophysical Research, 84, 59555972. SNOKE, A. W. 1993. Geologic history of Wyoming within the tectonic framework of the North American Cordillera. In: SNOKE, A. W., STEIDTMANN, J. R. & ROBERTS, S. M. (eds) Geology of Wyoming. Geological Survey of Wyoming, Memoir, 5, 2-56. SONNENBERG, S. A. & WEiMER, R. J. 1982. Tectonics, sedimentation and petroleum potential, northern Denver Basin, Colorado, Wyoming and Nebraska. AAPG Bulletin, 66, 634. STEARNS, D. W. 1971. Mechanism of drape folding in the Wyoming province. In: Wyoming Geological Association 23rd Annual Field Conference Guidebook, Wyoming Geological Association, 125-144. STONE, D. S. 1969. Wrench faulting and the Rocky Mountain tectonics. Mountain Geologist, 6, 67-79. 1970. Principal horizontal stress in the central Rocky Mountains versus California. Mountain Geologist, 7, 69-82. 1975. A dynamic analysis of subsurface structure in northwestern Colorado. In: BOLYARD, D. W. (ed.) Symposium on Deep Drilling Frontiers in the Central Rocky Mountains. Rocky Mountain Association of Geologists, 33-40. THOMAS, G. E. 1974. Lineament-block tectonics: Williston-Blood Creek Basin. AAPG Bulletin, 58, 1305-1322. TILLMAN, R. W., & MARTINSEN, R. S. 1983. Possible tectonic influences on and facies distribution of Shannon Ridge sandstones, Wyoming. AAPG Bulletin, 67, 1358. VON DEN BOLD, M. C., JOHNSON, S. R. & GUSTASON, E. R. 1991. Stratigraphy of the Lower Cretaceous Muddy Formation and the drainage pattern of its paleovalleys, Wind River Basin, Wyoming. AAPG Bulletin, 75, 687. WAAGE, K. M. 1959. Stratigraphy of the Inyan Kara Group in the Black Hills. United States Geological Survey Bulletin, 1081-B, 90.
165
WARNER, L. A. 1978. The Colorado lineament: A Middle Precambrian wrench fault system. Geological Society of America Bulletin, 89, 161-171. WAY, N. J., O'MALLEY, P.J. & SUTTNER, L. J. 1995. Tectonic controls on alluvial systems in a distal foreland setting: The Cloverly/Lakota formations (Cretaceous) in Wyoming and South Dakota. In: American Association of Petroleum Geologists National Meeting, Abstracts with Program, 280. WEIMER, R. J. 1984. Relation of unconformities, tectonics, and sea level changes, Cretaceous of western interior, USA. In: SCHLEE, J. S. (ed.) Interregional Unconformities and Hydrocarbon Accumulation. American Association of Petroleum Geologists, Memoir, 36, 7-35. 1992. Developments in sequence stratigraphy: foreland and cratonic basins. AAPG Bulletin, 76, 965-982. & SONNENBERG, S. A. 1989. Sequence stratigraphic analysis, Muddy (J) Sandstone reservoir, Wattenberg Field, Denver Basin, Colorado. In: COALSON, E. (ed.) Sandstone Reservoirs. Rocky Mountain Association of Geologists, 197 220. , EMME, J. J., FARMER, C. L., ANNA, L. O., DAVIS, T. L. & KIDNEY, R. L. 1982. Tectonic influence on sedimentation, Early Cretaceous, east flank Powder River Basin, Wyoming and South Dakota. Colorado School of Mines Quarterly, 77, 4. , REBNE, C. A. & DAVIS, T. L. 1988. Geologic and seismic models, Muddy Sandstone, Lower Cretaceous, Bell Creek-Rocky Point area, Powder River Basin, Montana and Wyoming. AAPG Bulletin, 72, 883-884. WHEELER, D. M. & GUSTASON, E. R. 1987. Sea level and paleotectonic controls on distribution of reservoir sandstone of Lower Cretaceous Muddy Sandstone, Hilight Field, Powder River Basin, Wyoming. AAPG Bulletin, 71, 5, 626. WICKS, J. L. 1979. A regional study of joints in the Fall River Formation (Cretacous) Black Hills of South Dakota and Wyoming. Unpublished MSc thesis. University of Toledo. WILLIS, B., GARDNER, M. H. & DUTTON, S. P. 1995. Architecture of fluvial-dominated valleyfill deposits in the Cretaceous Fall River Formation. In: American Association of Petroleum Geologists National Meeting, Abstracts with Program, A32.
This page intentionally left blank
Geometry of fold arrays in the Silesian-Cracovian region of southern Poland LESLAW TEPER Faculty of Earth Sciences, University of Silesia, ul. Bedzinska 60, PL 41-200 Sosnowiec, Poland, (e-mail:
[email protected]) Abstract: A zone of deformation characterized by en echelon folds extends along the northeast margin of the Upper Silesian Coal basin in southern Poland. The zone is well exposed by mining in numerous collieries where the deformation involving Upper Carboniferous strata can be observed. Folds and associated faults form a structural pattern consistent with an origin in dextral convergent wrenching on both the principal deep-seated fractures and the subsidiary faults affecting crystalline basement beneath this part of the basin. The zone is cross-cut by sublatitudinal fold belts which, according to their geometrical parameters, seem to result from alternate transcurrent shearing along west-east oriented fractures in the basement, sinistral in one case and dextral in the other. Among folds caused by the concentrated operation of a force couple whose vectors lie in the horizontal plane, there are examples of bending folds (forced folds) in each array resulting from vertical components of movement of deep-seated fault blocks. The dimensions of the folds change with depth, and the axial trend and the inclination of axial surfaces differ with distance from the basement fault trace. The length of fold segments strongly depends on axes orientation. Widely distributed evidence of interlayer slip demonstrates a variable spatial distribution of tectonic-transport direction in the sedimentary sequence. The above-mentioned features and other geometrical attributes of the fold arrays enable the dynamics of the forcing structures during Variscan tectogenesis to be determined and illustrate a hierarchical structure of the basement boundary zones. Examples from the Silesian-Cracovian region suggest that bending folds (forced folds) can sometimes originate in conditions provided by a strike-slip regime.
The Upper Silesian Coal basin (the USCB) lies in southern Poland next to the Polish-Czech border (Fig. 1). The basin contains several thousand metres of coal-bearing Carboniferous rocks made up of numerous sequences of clastic (conglomerates, sandstones, siltstones, claystones) and phytogenic (coaly shales and hard coals) rocks. In the area of about 6000km2 there are 65 collieries operating to a depth of 1000m. Some of them have extracted coal since the 18th century. The number of well-documented coal layers is as large as 523, including 264 economic coal seams. The USCB was formed as a molasse-infilled foredeep located in the foreland of the Moravian-Silesian branch of the Variscan fold belt (e.g. Bukowy 1984). The Carboniferous molasse developed on the crystalline block of the Upper Silesian Massif. The block is considered to be a part of Bruno-Silesicum (Kotas 1985). It is framed by first-order crustal boundary zones and subdivided into small segments by deepseated fractures of second order (Fig. 2). The spatial distribution and activity of the boundary zones since the Variscan epoch have determined the structure in these Polish-Czech coal-bearing deposits. The vast number of the main disjunctive structures in the Carboniferous rock-mass
are secondary faults, following older tectonic directions and reflecting movements of basement blocks beneath the USCB (Teper el al 1992; Teper & Idziak 1995; Teper 1998). The USCB structure has been controlled by deep-rooted faults since the early phases of sedimentation (Kotas 1985; Teper el al 1992). Long-lived fundamental fault systems related to the borders of the Upper Silesian Massif have affected the structural pattern of the SilesianCracovian region at least since the Caledonian tectogenic period (Haranczyk 1989). The history of the second-order fractures cutting the Upper Silesian Massif itself probably goes back to the time of consolidation of the block. Major boundary discontinuities acted as zones of right-lateral strike-slip (Bogacz & Krokowski 1981; Kotas 1985) with successive phases of transpression, wrenching and transtension during late Hercynian times (Teper 1989). According to Bogacz & Krokowski (1981) such kinematics resulted from a sinistral rotation of the Upper Silesian Massif. However, Kotas (1985) suggested that the deformation mechanism of primary importance is the interaction between the left-lateral passive drifting of the USCB basement towards the northwest and its gradual subduction under the Variscan accretion wedge.
From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 167-179. 1-86239-060-6/OO/S 15.00 n The Geological Society of London 2000.
Fig. 1. Major structural features of the Upper Silesian Coal basin (USCB) after Jurec/ka & Kotas (1995). 1, Zone of fold tectonics following s t r u c t u r a l style of the Moravian-Silesian branch of the Variscides; 2, /one of fault-block tectonics; 3, contact /one of Upper Silesian and Malopolska basement blocks (Fig. 2, fault /one I I ) ; 4, boundary of the USCB; 5, overthrusts; 6, main faults; 7, synclines; 8, anticlines.
Fig. 2. Structural sketch of the Upper Silesian Massif after Kotas (1985). 1, Upper Silesian segment, Molasicia represents a part having unknown internal structure because of its position under thick metasediments of the Moravian Silesian branch of the Variscan fold belt (Fig. 1); 2, Brunnia segment; 3, first-order fault zones: I—II, Hamburg-Cracow f. z.; Ill, Moravian suture; IV, Elbe f. z.; V, Carpathian suture; 4, second-order fault zones; 5, third-order fault zones; 6, Polish state border.
Fig. 3. Folds in the northern part of the LJSCB. 1. Border of the USCB; 2, axes of fold arrays A, B and C — traces of deep-seated faults; 3, synelines; 4, anticlines; 5, sense of relative movement; 6, locali/ations of interlayer slip phenomena presented in Figs 10 13; 7, cross-sections of examples of forced (drape) folds (Fig. c)); a, h and c, domains of main forced-fold formation.
FAULT ASSOCIATED FOLDING - SOUTHERN POLAND
171
Fold arrays Data source A zone of fold deformation extends along the northeastern margin of the basin (Fig. 1). In the northern part of the USCB the zone is additionally cross-cut by two sublatitudinal fold belts. The three fold sets are well exposed in the extensively worked grounds existing in the area. The fold pattern in the northeast part of the USCB, projected to sea level, is shown in Fig. 3. Archival records and current data concerning folding details were derived from original large-scale plans produced by mining surveyors for coal seam surfaces. Each feature shown on the map was measured directly by driving roadways or face headings. Extensive workings in many seams over a vertical range of up to 1000m yielded information regarding the threedimensional geometry of folds and enabled fold axes to be mapped accurately. The link between the fold arrays and displacement along basement discontinuities has been suggested by Herbich (1980) and Teper (1989) who compared the Silesian folds with the geometry and spatial organization of'en echelon' folds which appear above a deep-seated wrench fault (e.g. Wilcox et al 1973; Odonne & Vialon 1983). The traces of primary faults were estimated using conclusions drawn from laboratory experiments (Wilcox et al. 1973; Odonne & Vialon 1983), 6-lineation sense measurements (Teper 1989) and structural analysis of the fault array (Teper 1989, 1998; Teper & Idziak 1995). Parameters For a given fold straight-line segments of its axis were isolated. Then the angle, 6, between the fold axis and the basement fault trend, as well as the average distance between the fold and the underlying basement fault, j>d, were measured. Graphs of O vs y& (Fig. 4) illustrate changes in axial trend with distance from the fault trace for each array. The results (Fig. 4a) show that the regional trend in the Q/y^ relationship is disturbed by a local anomaly represented by fold array A, i.e. the fold series which follows the northeastern border of the USCB, II (Figs 1 and 2). The average empirical orientation of the segments, O, varies between 21.6° and 26.6°, while 8 value in the models (Odonne & Vialon 1983) is equal to 23.1°. A plot of average angle vs total length of the fold (Fig. 5) shows that the longest folds are oriented along the 6 value.
Fig. 4. Changes in axial trend 6, with distance from the fault trace yd for fold arrays, (a) The regional trend in the tan S/yd relationship in the fold series following the border II (Figs 1 and 2). The result suggests a peripheral position of array A (Fig. 3) in a broad zone of higher order (Fig. 1). (b) Graphs of tan 9 vs yd for fold arrays A, B and C (Fig. 3).
Fold axes mapping and a study of successive cross-sections show that the axial planes of the folds are inclined. The majority of the folds are inclined away from the basement-fault traces, i.e. have outward dipping axial surfaces (Fig. 6). Inward inclinations can be observed in the vicinity of primary fault traces and the upright folds are situated above the basement faults. Both the proportion and the spatial distribution of the types of inclination are close to those obtained in the experiments of Odonne & Vialon (1983). The amplitude/wavelength ratio, H/s, of the entire population was measured in cross-sections perpendicular to the axes. Dimension parameters were compared for two layers, an 'upper' and 'lower', which were selected individually for each case. The choice was controlled by the configuration of the underlying workings. Values of the H/s ratio were plotted on a frequency histogram (Fig. 7), which shows that the mean ratio is higher in the lower part of folds. This points to a systematic increase of deformation with depth.
172
L. TEPER
Fig. 7. Frequency histogram of amplitude wavelength ratio. H .s. for the entire population of folds. Dimension parameters are compared for two layers: 'upper'. U. and "lower". L. The mean ratio is higher in the lower part of folds.
Fig. 5. Relationship between the average orientation 6 and the total length of folds /: (A) for the zone II (Fig. 2); (B) for the array A (Fig. 3).
This tendency is consistent with the coal quality pattern in the USCB: the greatest changes in the rank of coalification occurs along the depth axis and the spatial variability of carbonification grade is found to depend more or less on the structure of the basin (Jureczka & Kotas 1995).
Kinematics related to folding It is clear that the morphology of the folds in the basin has evolved. Evidence can be found in some structures to indicate that after the initial stages, when folds had been produced by flexuring caused by horizontal compression (buckle
Fig. 6. The distribution of fold axial surface inclination for the entire population of folds. 1, Inward inclined; 2, upright; and 3. outward-dipping axial surfaces.
folds), movement takes place along slip surfaces roughly parallel to the decollement level at the basement-cover interface (Teper & Pieczko 1993). However, shear features are observed in vertical as well as horizontal cross-sections. Progressive shearing on the basement fault produced rotational deformation in the cover rocks with extension nearly parallel to the fold axes (see Watkinson 1975). When the shearing was intense enough fold modifications appeared guided by the average trend of already reoriented folds. In other words, the folding was permanently controlled by the orientation of a pre-existing anisotropy (see Cobbold & Watkinson 1981). Such a mechanism of folding may produce geometries typical of polyphase deformation (or progressive deformation within a rotational strain field) and the geometries observed suggest that the studied folds originated as a result of translation (or transpression, or transtension).
Fig. 8. Features of a strike-slip zone that produces domains of subsidence and uplift in a wrenching regime, after Reading (1980). (A) Curvature of a strike-slip fault; (B) braiding of faults.
FAULT ASSOCIATED FOLDING - SOUTHERN POLAND
173
Fig. 9. Examples of forced (drape) folds from the Upper Carboniferous of the Silesian-Cracovian region (see Fig. 3 for location). 1, Roof of the Carboniferous; 2, coal seams.
Along a strike-slip fault system there may appear small-scale alternate zones of sinking and uplift (Reading 1980) which form as a result of curvature of a strike-slip fault or the braiding of faults (Fig. 8). That is probably why the most obvious motion observed in some domains along wrench
zones in the USCB is dip-slip, although the dominant tectonic transport was horizontal. As a result, the orientation and geometry of certain folds (Fig. 9) in such domains (Fig. 3) exhibit forced folding (both extensional and compressional) features (Doktorowicz-Hrebnicki 1935;
174
L. TEPER
Teper & Pieczko 1993). In these domains the fold axes tend to become parallel to the basement fault trace (Fig. 3). The folds themselves have the geometry of forced folds generated by the bending of cover rocks by a component of vertical motion either upwards or downwards on the basement fault. Fault-overlying flexures, monoclinal folds (see e.g. Woodward et al. 1972) and distinctly asymmetric bending anticlines are observed.
Interlayer slip phenomena Analogue experiments of wrench faulting beneath a layered cover (Tchalenko 1968; Odonne & Vialon 1983; Harris & Cobbold 1984) show that the displacement of the basement along a wrench fault generates slip at layer interfaces in the cover rocks. The direction and sense of layer traction is determined by the sense of movement
Fig. 10. Tectonic mesostructures occurring in the horizontal shear zone related to coal seam 816. the Grodziec coalfield (for location see Fig. 3). 1, Average bedding orientation; 2, main fault position; 3. subhorizontal subsidiary shears; 4, steep subsidiary shears; 5, slickenfibre lineation; 6, asymmetric sinistral slickenfibre lineation; 1, asymmetric dextral slickenfibre lineation; 8, en echelons; 9, tectonic-transport sense along the main fault surface. Steep faults (including the main structure) were formed during the late stage of deformation.
FAULT ASSOCIATED FOLDING — SOUTHERN POLAND
on the basement fault. The relative movement of layers produces horizontal shear planes or shear bands (Tchalenko 1968; Odonne & Vialon 1983; Harris & Cobbold 1984). There is considerable evidence of interlayer slip in the coal-bearing formation of the northern part of the USCB. Horizontal shear bands occur frequently in coal seams where shearing produces zones of intense plastic and brittle deformation, as well as a drastic reduction of layer thickness. In some cases, the reduction of a seam occurs along an obvious dislocation plane accompanied by numerous subsidiary structures (Fig. 10). The shearing is documented by the occurrence of subhorizontal P- and 7?-shears, as well as the conjugate Rf set. D-shears parallel to the seam roof are also observed with slickensides developed on their surfaces. The asymmetric slickenfibre lineation indicates the sense of displacement in the zone. The interconnection of R-, P- and Dshears isolates shear lenses (sensu Tchalenko 1968). Microlithons are found (Fig. 11) in which the original fabric is reoriented and forms a compression texture (sensu Tchalenko 1968). According to their morphology and geometry, the microlithons can be compared with horses of small-scale duplexes formed during regional thrusting (Price & Cosgrove 1990), flexural-slip folding (Tanner 1992) and in strike-slip fault systems (Aydin 1988). They are evidence of
Fig. 11. Horizontal shear zone connected with coal seam 816 (for location see Fig. 3). 1, Coal seam; 2, fabric in clays; 3, shears. Arrows represent kinematic indicators related to the Carboniferous period of activity. Shear lenses (horses of the small-scale duplexes) are visible and are the result of the interconnection of /?-, P- and Z)-shears. Note the reoriented fabric forming a compression texture (sensu Tchalenko 1968). (A) Example from the Grodziec coalfield; (B) example from the Paryz coalfield.
175
Fig. 12. Kink-band (KB) examples from the horizontal shear zone in the roof of coal seam 816 in the Paryz coalfield (for location see Fig. 3). Arrows are signs of kinematics of Carboniferous deformation. Note the thin plastic clay intercalation within the seam which facilitated horizontal tectonic transport.
bedding-parallel slip, and indicate the direction and sense of displacement (see Tchalenko 1968; Stock 1992; Tanner 1992). The kink-bands that are recognized in the horizontal shear zones in the USCB (Fig. 12) are very similar to those obtained experimentally (Tchalenko 1968). Fabric within the kink-bands approximately follows the direction of the /^-shears and the kink-bands generally terminate abruptly on Dshears. Laboratory tests and field observations (Tchalenko 1968; Pratt 1992) demonstrate that such kink-bands form when rocks are submitted to compression parallel to initial bedding planes. There are also thrusts and associated drag folds recorded in the horizontal shear bands (Fig. 13). In some cases, thrusting of the roof of the seam is only apparent on a meso- or microscale (Fig. 13b). Some forms, considered previously to be layer splits, resemble sheath folds (i.e. very non-cylindrical folds) reported from natural shear zones (Henderson 1981; Evans & White 1984; Lacassin & Mattauer 1985; Mies 1993) and those generated in experiments (Cobbold & Quinquis 1980). They have horizontal axial planes, as well as curved fold hinges, and are strongly asymmetric, with their longer axes parallel to a local shear direction. Occurrences of the 'sheath-like' forms are commonly associated with intensive folding of plastic rocks which were taken to be 'intercalations'. These structures are the subject of an ongoing study. Within the USCB the tectonic deformation of coal seams is most probably controlled by interlayer slip. In each case the direction and sense of the slip is determined by the dynamics of a particular deep fracture. The beddingparallel displacements recorded on both sides of the fault traces shows a shear sense reversal from wall to wall throughout the area. It is
176
L. TEPER
Fig. 13. Thrusting produced by interlayer slip (for location see Fig. 3). (A) thrust and associated drag folds in the roof part of coal seam 816 in the Paryz coalfield; (B) minor thrusts causing thickness growth of coal seam 816 in the Grodziec coalfield (in both cases arrows indicate sense of movements during the Carboniferous period of activity).
likely that the interlayer slip occurred at the same time as the axial planes of the en echelon folds became inclined. The vergence of the inclination, opposite on either side of a given wrench fault, follows the trend of the shear on the horizontal planes. This is consistent with the model results (Odonne & Vialon 1983). In the sections of the arrays where drape folds appear, the interlayer slip direction changes and tends to become perpendicular to the trace of a basement fault. Features of local and transregional structural patterns
primary strike-slip movement could produce secondary structural features (including forced folds) as well as structures more typical of a strike-slip provenance. The maximum concentration and intensity of the strike-slip movement (dextral during Variscan times) occurred along the northeastern boundary of the USCB (Fig. 1). This has been deduced from geophysical and geological data (Kurbiel \919ab; Bukowy 1984; Bula et al. 1994) because this part of the basin is unworked by mining. A local discontinuity beneath fold array A (Fig 3) forms the southwest edge of the basin. Movement along this discontinuity has produced a broad border zone II (Fig. 2). The sense of movement along the zone during the Upper Carboniferous was dextral. Curvature and possible braiding of the faults in the zone (Fig. 14, see also Fig. 1) enable horizontal movement to be released by vertical, transpression-related displacements on fault surfaces (Teper 1989). The equations proposed by Ramsay & Graham (1970) for calculating the shear strain, 7, and the minimum displacement, 5, were applied yielding a value of 7 of about 2.6 and a value of s of 73 km for the entire border zone, as well as 7 = 1.5 and s = 7.8km for the local discontinuity A (Fig. 3). The subequatorial fracture marked by the linear gravimetric anomaly beneath the fold array B forms a second-order boundary zone separating two segments of the Upper Silesian Massif (Fig. 2). Fold parameters and interlayer slip phenomena document sinistral movement along the zone. A shear strain value, 7 = 2.0, is obtained from the magnitude of the rotation of the fold axes (Ramsay & Graham 1970). It gives a displacement 5 = 10km. Another gravimetric anomaly underlies the second W-E-trending fold array C. Kotas (1985) suggested the existence of a secondorder boundary zone at this locality (Fig. 2). The structural pattern in the cover rocks indicates a dextral sense of the shearing along this discontinuity. According to the algorithm used for the other basement faults, the shear strain value is calculated from the initial and final angles 6 that the fold axes form with the shear direction. A value of 7 = 1.75 is obtained and a minimum displacement s = 5.2 km.
Local deep-seated fracture-controlled pattern
Transregional deep-seated structures
The structure of the basement in the SilesianCracovian region is reflected by geophysical data (Kurbiel \919ab' Goszcz 1986). Gravimetric and magnetic field anomalies reveal linear bodies occurring just beneath the fold arrays. The lineaments are supposedly the dislocations where
Southern Poland is positioned in the TransEuropean Suture Zone (the TESZ), which acted as a Baltica plate boundary during Cadomian and Caledonian times. The Caledonian development within the TESZ significantly differs from place to place. Several geologic units, now
FAULT ASSOCIATED FOLDING - SOUTHERN POLAND
177
Fig. 14. Map of residual gravimetric anomalies (according to Griffin method, r = 2236m) after Kurbiel (\919a) in the first-order boundary zone II (Fig. 2). 1, Supposed trace of principal deep-seated discontinuity; 2, gravimetric anomaly values encircled by gravity field contours. assembled, cannot be compared immediately in regard to their palaeotectonic histories. The Upper Silesian Massif and the Malopolska Massif (Fig. 2) are thought to be two such 'suspect terranes' of uncertain age, probably Cadomian (Dudek 1980; Kotas 1985; Zelazniewicz & Franke 1994), and of unknown provenance (Grygar 1992; Berthelsen 1994). The lithofacial and tectonic development within the TESZ were gradually adjusted during the Silurian. Collisional processes, combined with strike-slip displacements, which caused intense deformation but only weak metamorphism and subordinate magmatism, resulted in the final amalgamation of the terranes and the oblique collision of the Cadomian and Baltican continents (accretion of the East Avalonian terrane collage
against Baltica) in the Late Silurian-Early Devonian (Zelazniewicz & Franke 1994). The collision was accompanied by subequatorial transtension causing oblique rifting of both the subducted Baltica borderland and the Avalonia accretion wedge, which resulted in left-lateral displacement along mobile grabens, fragmentation, drift and, finally, subduction of detached plate segments under the Variscan front (Grygar 1992).
Conclusions The two above-mentioned 'suspect terranes' form the pre-Devonian basement of the SilesianCracovian region itself. Their common boundary zone II (Fig. 2), which is repeated in their
178
L. TEPER
sedimentary cover as fold zone A (Fig. 3), follows the main northwest-southeast direction of the TESZ. The Upper Palaeozoic oblique fracturing reworked discontinuities were sealed by the Cadomian consolidation. Reactivated latitudinal fractures started acting as intra-plate mobile zones (Fig. 2) which were recorded in the cover by two other fold trends, B and C (Fig. 3). The existence of the fold arrays in the Carboniferous complex demonstrates that local boundary zones were still active or became reactivated due to the approaching erogenic front in Variscan times. A regional study of the structural pattern (Teper 1989, 1998; Grygar 1992; Teper et al 1992) shows that dextral strike-slip motion, treated as a part of transpression-transtension continuum, has been an important long-lived factor of the Upper Palaeozoic deformation in the SilesianCracovian region. Folds and faults form a pattern produced by movement along inter- and intraplate boundary zones of the Upper Silesian Massif. The result is consistent with the models of the Hercynian fold belt introduced by Reading (1980) and Krone (1996), and contrasts with the idea that only late Variscan strike-slip movement had importance in Palaeozoic Europe (e.g. Arthaud & Matte 1977; Ziegler 1986). The type of deformation that has been discussed above can, like other tectonic phenomena, represent the combined effect of several factors, and not just the horizontal mobility of the basement alone. The geometry of folds demonstrates that horizontal and vertical movements both played an essential part in the formation of structures in the Silesian-Cracovian region, and indicates that drape (forced) fold formation is possible in regimes which are dominantly strike-slip (i.e. within the transtension-transpression continuum first mentioned by Harland 1971). I would like to acknowledge the critical advice of P. Bankwitz and M. Nemcok. I also wish to thank E. Teper. who drafted the maps and diagrams. References ARTHAUD. F. & MATTE, P. 1977. Late Palaeozoic strike-slip faulting in southern Europe and northern Africa: result of right-lateral shear zone between the Appalachians and the Urals. Bulletin Geological Society of America, 88, 1305-1320. AYDIN, A. 1988. Discontinuities Along Thrust Faults and the Formation of Cleavage Duplexes. Geological Society of America. Special Paper, 222, 223-232. BERTHELSEN. A. 1994. EUROPROBE'S 2nd 'TransEuropean Suture Zone'. Europrobe News 5, 2-3. BOGACZ. W. & KROKOWSKI, J. 1981. Rotation of the Upper Silesian Coal Basin. Annals of the Geological Society of Poland. 51. 361-381.
BUKOWY, S. 1984. Variscan Structures of the SilesianCracovian Region. Silesian University Press. Katowice. BULA. Z.. ZABA. J.. JACHOWICZ. M.. PERSKI. Z. & SIEWMAK-MADEJ. A. 1994. Lithosratigraphic and tectonic problems of the Lower Palaeozoic of northeast margin of the Upper Silesian Coal Basin. In: Palaeozoic of NE Margin of the Upper Silesian Coal Basin. Silesian University Press. Katowice. 134-172. COBBOLD. P. R. & QUINQUIS. H. 1980. Development of sheath folds in shear regimes. Journal of Structural Geology. 1/2. 119-126. & WATKINSON. A. J. 1981. Bending anisotropy: a mechanical constraint on the orientation of fold axes in an anisotropic medium. Tectonophvsics. 72. T1-T10. DOKTOROWICZ-HREBMCKI. S. 1935. Geological Map of the USCB. 1:25000. Explanations to the Grodziec Sheet. Polish Geological Institute. Warsaw. DUDEK. A. 1980. The crystalline basement block of the Outer Carpathians in Moravia: Bruno-Vistulicum. Papers of Czechoslovak Academy of Sciences. 90. 8. EVANS. D. J. & WHITE. S. H. 1984. Microstructural and fabric studies from the rocks of the Moine Nappe. Eriboll. NW Scotland. Journal of Structural Geology. 6. 369-389. Goszcz, A. 1986. Problems of geodynamics of the Upper Silesian Coal Basin in the light of new geophysic and tectonophysic interpretations. Scientific Paper. Silesian Technical University Gliwice. 149. 183-196. GRYGAR. R. 1992. Kinematics of Lugosilesian orocline accretion wedge in relation to the Brunovistulian foreland. Scientific Paper. Mining and Metallurging University Ostrava. 1. 49-72. HARANCZYK. C. 1989. Two solutions - multipulsations model of genesis of the Silesian-Cracovian Zn-Pb ore deposits. Journal of Ore Geology and Petrology. Zagreb. 1. 11-16. HARLAND. W. B. 1971. Tectonic transpression in Caledonian Spitzbergen. Geological Magazine. 108, 27-42. HARRIS. L. B. & COBBOLD. P. R. 1984. Development of conjugate shear bands during bulk simple shearing. Journal of Structural Geology. 1, 37-44. HENDERSON, J. R. 1981. Structural analysis of sheath folds with horizontal X-axes. northeast Canada. Journal of Structural Geology. 3. 203-210. HERBICH. E. 1980. On the Upper Silesian deep fracture. Geological Review. Poland. 3. 156-159. JURECZKA. J. & KOTAS. A. 1995. Upper Silesian Coal Basin. ///: ZDANOWSKI, A. & ZAKOWA. H. (eds) The Carboniferous System in Poland. Papers of Polish Geological Institute. Warsaw. 148. 164-173. KOTAS, A. 1985. Remarks on the structural evolution of the Upper Silesian Coal Basin. ///: Proceedings of the Conference on Tectonics of the USCB. Silesian University Press. Katowice. 17-46. KROHE, A. 1996. Variscan tectonics of central Europe: Postaccretionary intraplate deformation of weak continental lithosphere. Tectonics. 15. 1364-1388. KURBIEL, H. 1979#. Map of Residual Gravimetric Anomalies (According to Griffin Met hod j in
FAULT ASSOCIATED FOLDING - SOUTHERN POLAND Lubliniec-Krakow region 1:100000. Polish Geological Institute, Warsaw. \919b. Map of Magnetic Anomalies (A2 Values) in Lubliniec-Krakow Region 1:100000. Polish Geological Institute, Warsaw. LACASSIN, R. & MATTAUER, M. 1985. Kilometre-scale sheath folds at Mattmark and implications for transport directions in the Alps. Nature, 315, 739-741. MIES, J. W. 1993. Structural analysis of sheath folds in the Sylacauga Marble Group, Talladega slate belt, southern Appalachians. Journal of Structural Geology, 8, 983-993. ODONNE, F. & VIALON, P. 1983. Analogue models of folds above wrench fault. Tectonophysics, 99, 31-46. PRATT, W. T. 1992. The use of kink bands to constrain fault displacements: an example from the Bala Lineament, Wales. Geological Magazine, 129, 625-632. PRICE, N. J. & COSGROVE, J. W. 1990. Analysis of Geological Structures. Cambridge University Press, Cambridge. RAMSAY, J. G. & GRAHAM, R. H. 1970. Strain variations in shear belts. Canadian Journal of Earth Sciences, 7, 786-813. READING, H. G. 1980. Characteristics and Recognition of Strike-Slip Fault Systems. International Association of Sedimentology, Special Publication, 4, 7-26. STOCK, P. 1992. A strain model for antithetic fabric rotation in shear band structures. Journal of Structural Geology, 10, 1267-1275. TANNER, P. W. G. 1992. Morphology and geometry of duplexes formed during flexural-slip folding. Journal of Structural Geology, 10, 1173-1192. TCHALENKO, J. 1968. The evolution of kink-bands and the development of compression textures in sheared clays. Tectonophysics, 6, 159-174. TEPER, L. 1989. Meso- and macrotectonic indicators of strike-slip movements in crystalline basement of
179
northeast part of the USCB. Papers of Polish Academy of Sciences Katowice, 14, 40-41. 1998. Seismotectonics in the Northern Part of the Upper Silesian Coal Basin: Deep-seated Fractures-controlled Pattern. Silesian University Press, Katowice. & IDZIAK, A. 1995. On fractal geometry in fault systems of the Upper Silesian Coal Basin (Poland). In: ROSSMANITH, H. P. (ed.) Mechanics of Jointed and Faulted Rock. Balkema, Rotterdam, 329-333. & PIECZKO, J. 1993. Tectonics of the northern limb of the Bytom syncline within the 'Miechowice' coalfield, the USCB. In: Proceedings of the 16th Symposium on Geology of Coal-bearing Formations of Poland. University of Mining and Metallurgy, Cracow, 109-114. , IDZIAK, A., SAGAN, G. & ZUBEREK, W. M. 1992. New approach to the studies of the relations between tectonics and mining tremors occurrence on example of the Upper Silesian Coal Basin (Poland). Ada Montana, A2 (88), 161-178. WATKINSON, A. J. 1975. Multilayer folds initiated in bulk plane strain, with the axis of no change perpendicular to the layering. Tectonophysics, 28, T7-T11. WILCOX, R. E., HARDING, T. P. & SEELY, D. R. 1973. Basic wrench tectonics. Bulletin American Association of Petroleum Geologists, 57, 74-96. WOODWARD, L. A., KAUFMAN, W. H. & ANDERSON, J. B. 1972. Nacimiento fault and related structures, northern New Mexico. Bulletin Geological Society of America, 83, 2383-2396. ZELAZNIEWICZ, A. & FRANKE, W. 1994. Discussion on U-Pb-ages from SW-Poland: evidence for a Caledonian suture between Baltica and Gondwana. Journal of the Geological Society, London, 151, 1049-1055. ZIEGLER, P. A. 1986. Geodynamic model for the Palaeozoic crustal consolidation of Western and Central Europe. Tectonophysics, 126, 303-328.
This page intentionally left blank
Primary and 'forced folds' of the Columbia River basalt province, eastern Washington, USA A. J. WATKINSON & P. R. HOOPER Geology Department, Washington State University, Pullman, WA 99164, USA Abstract: The Yakima fold belt is a series of asymmetric, east-west anticlines separated by much broader synclines in the thick and otherwise horizontal sequence of Columbia River basalt flows. The folds have grown progressively over the last 17 Ma in response to regional north-south compression and east-west extension. This stress and strain pattern is reflected on a regional scale by the highly oriented north-northwest-south-southeast basalt feeder dykes, by the WNW- (right-lateral) and NNE- (left-lateral) trending conjugate fault system, and by the presence of a similar strain field measurable in the basalt flows today. Peripheral to the Yakima fold belt, pre-basalt structures of different orientations project under the basalt pile. Some of these structures have continued to develop during and after basalt extrusion and have influenced structures in the basalt flows which are, therefore, 'forced1.
The Columbia River basalts in the northwest United States flooded a complex tectonic region over an area of approximately 165000km (Tolan et al 1989) in a relatively short period from 17 to 6 Ma. During and since that time the flows have been progressively deformed in an active regional stress field (Fig. 1). This scenario provides a unique opportunity to observe the effects of reactivation of pre-basalt structures projected beneath the basalts, as well as document the striking phenomenon of brittle basalt layers folding at the Earth's surface. Deformation of the basalt flows is confined to those parts of the basalt province which overlie accreted oceanic crust, and is absent from flows overlying the older, thicker, continental crust of the North American plate and associated Cretaceous batholiths. Detailed field work has quantified the fold style, provided estimates of the intensity of the strain distribution around the folds and illuminates the style of the fracture systems (Price & Watkinson 1989). However, the style of fold deformation is not distinct enough in itself to distinguish between basement-controlled forced and unforced folding in the basalts. First motion studies and seismic distribution give the current stress field and pattern of active deformation, which is compatible with the in situ stress measurements (Kim et 0/.1986) showing a horizontal, north-south direction of maximum compression. The classic problem of extrapolation of folds downwards is well illustrated on the Columbia Plateau. Velocity inversions make seismic reflection/ refraction interpretation difficult, but there are some drill core data through the basalt which suggest at least local detachment. Balanced local sections suggest the possibility of basement control.
This paper provides a general survey of the major tectonic structures surrounding the plateau, discusses the expression of those structures as they project into the basalt plateau, and then contrasts these forced structures with those structures which appear to reflect directly the stress field of the region in effect over the last 17 Ma. Regional structures surrounding the Columbia River province There is a complex architecture of regional-scale structures surrounding and projecting under the basalts. Figure 2 shows the major structures of the Eocene Colville batholith and associated core complexes and grabens (RG) to the north. The Cascade province is to the west, dominated by the Straight Creek strike-slip fault system (SCF). The Olympic-Wallowa lineament (OWL) and associated structures cut across the southwest corner of the basalt plateau from the Olympic peninsula in the west to the Blue Mountains in the southeast. The southeast corner of the Columbia River basalt province is partly bounded by the Cretaceous suture zone (Fleck & Criss 1985), which separates the old cratonic North American plate to the east and north from the accreted oceanic terranes to the west. The suture zone runs from south to north along the western margin of the Idaho batholith, then turns abruptly west beneath the basalts (Fig. 2). The Hite fault zone (HF) appears to displace the suture zone by left-lateral strike-slip motion (Reidel et al. 1994; Sobczyk 1994; Hooper et al. 1995), before the suture zone turns north again to form the eastern margin of the Pasco basin (PB).
From\ COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 181-186. 1-86239-060-6/OO/S 15.00 (£) The Geological Society of London 2000.
182
A. J. WATKINSON & P. R. HOOPER
Fig. 1. Relief map of Washington, Idaho and Oregon with recent seismic epicentres as white dots). The area of flooding by the Columbia River basalts shows as the flatter, more even relief immediately behind the Cascades. While most seismicity occurs in the western corridor, west of the Cascades, the plateau shows significant seismicity. often associated with visible structures such as the Yakima folds and the southeastern end of the OWL lineament.
The Blue Mountains anticline, trending westsouthwest on the south side of the OWL and south of the Pasco basin, is known to be prebasalt (Riddihough et al. 1986) because the ridge formed a divide which separated the Picture Gorge and Grande Ronde basalts during their contemporaneous eruptions at approximately 16 Ma. Evidence of minor postbasalt strain and displacement is apparent on each of these older structures - the OWL, the Hite fault system, the Blue Mountains anticline and the old suture zone. After all these prebasalt structures were covered by the basalts in the Miocene, active deformation continued up to and including the present day. Structures within the basalt Most Columbia River basalt structures, the eastwest folds, the west-northwest-east-southeast right-lateral and north-northeast-south-southwest left-lateral conjugate strike-slip faults, and the consistently NNW-SSE-oriented basalt
feeder dykes, are confined to the accreted terranes west and south of the suture zone, mainly between the east-west suture and the OWL (Fig. 2). North of the suture zone, across the Palouse slope, the basalts on the craton remain essentially undeformed. This presumably reflects a difference in the thickness and strength of the older craton and the younger accreted terranes (Reidel et al 1989; Hooper & Conrey 1989). The most obvious structures in the basalts are the east-west folds of the Yakima fold system (YFB). These reflect the state of stress that is current today and has been active throughout the basalt eruptions (Reidel 1984). A major question concerning these folds is how much if any 'basement' is involved in their formation. Partial restoration across one particularly wellexposed Yakima fold, Umtanum Ridge (Price & Watkinson 1989) suggests the folds could overlie a fault at depth, with the possibility of 'basement' reactivation. The dominant east-west fold trend is seen to be influenced by the underlying structures. Along the western end of the Yakima fold belt
FORCED FOLDS IN THE COLUMBIA RIVER BASALT
183
Fig. 2. Regional map at the same scale as the relief map showing the main geological structures. RG - Republic grabens-Eocene; FF - Frazer fault; SCF - Straight Creek fault in the Cascades; OWL - Olympic-Wallowa lineament; YFB - Yakima fold belt; PB - Pasco basin; KBML - Klamath-Blue Mountains lineament; HF - Hite fault zone; LF - Lewiston fold zone; BFZ - Brothers fault zone; EDZ Eugene-Denino zone; MZ - McLoughlin zone; LCZ - Lewis and Clarke zone. Stars are Cascade volcanoes.
the folds pass into faulted splays, trending northwest-southeast, of the main Cascade Straight Creek strike-slip fault (Tabor et al. 1984). Here, the basalt folds do not simply mimic underlying folds and normal faults in the pre-basalt rocks in a harmonic way, but have a more complex disharmonic geometry (Campbell 1989). Also, the east-west Yakima folds swing
into the west-northwest-east-southeast trend of the OWL where that lineament cuts across the fold belt (Fig. 3). While there is local debate over how much and exactly what form of reactivation occurs along the OWL (see Mann & Meyer 1993), periclinal folds, faults trending in the same trend as the lineament and extensional grabens occurring on the south side of the zone all suggest
184
A. J. WATKINSON & P. R. HOOPER
Fig. 3. Detail of the relief map showing the E-W-trending Yakima fold belt with the west-northwest trending OWL lineament (SE corner). The relief map is the Army Corps of Engineers map series. Series V502. sheet NL10-1.4. 1:250000. It was dusted with Mount St. Helens 1980 volcanic dust! and obliquely illuminated from the northeast (courtesy of Dr Richard Thiessen. Washington State University).
reactivation of the underlying structure during and after the basalt eruptions. Seen in a regional context, it is also notable that the OWL parallels the family of major west-northwest-east-southeast strike-slip zones in the northern Great basin and north through eastern Oregon - the McLoughlin zone (MZ), the Eugene-Denino zone (EDZ) and the Brothers fault zone (BFZ) (Fig. 2) (Lawrence 1976). As similar structural patterns are observed in each of these zones, we interpret the OWL as a northern counterpart of the northern Great basin displacements, partitioned between a general east-westerly extension and the west-northwest dextral strike-slip components (Lawrence 1976; Hooper and Conrey, 1989). The trends of the Hite fault and OWL have a geometric symmetry about the east-northeastwest-southwest trend of the Klamath-Blue Mountain lineament (KBML) (Fig. 2), anticlockwise from the present strain pattern and known to be of pre-basalt age because the structure separates the contemporaneous eruptions of the Picture Gorge and Grande Ronde Basalt Formations of the Columbia River basalts. The observed reactivation of this structure to fold
Columbia River basalt flows is, therefore, a forced direction in the current regional strain field. Using basalt stratigraphy as a horizontal marker, Kuehn et al. (1996) demonstrated significant post-basalt uplift on the east side of the Hite fault, but that displacement was dominantly left-lateral strike-slip because of the almost universally low angle slickensides observed on the fault planes and the en echelon pattern of the faults to the north (Hooper et al. 1995). The important observation by Reidel et al. (1994) that the basement at the bottom of a drill core west of the Hite fault was of continental (cratonic) origin implies that the old lithospheric suture was displaced many kilometres leftlaterally prior to the basalt eruptions. Thus, the small post-basalt displacement on the Hite fault is a reflection of a much larger pre-basalt displacement. On the eastern end of the Hite fault system a NNE-SSW-trending left-lateral strike-slip fault transfers into the E-W-rending reverse fault of the Lewiston anticlinal structure (Hooper et al. 1995). This post-basalt structure parallels and
FORCED FOLDS IN THE COLUMBIA RIVER BASALT
lies immediately south of the old pre-basalt suture zone (Fig. 2) and the associated strong east-west foliation in the older cratonic granites to the north. The east-west trend of the Lewiston structure, like that of the Yakima fold belt, is compatible with the Miocene to present (postbasalt) regional stress field, but its precise location probably reflects the presence of the older suture zone. In summary, the Blue Mountains anticline south of the OWL, the OWL itself, the Hite fault system, and the suture zone between the older cratonic and younger accreted terranes are all pre-Miocene structures that have been reactivated during the Miocene. The Yakima fold belt is not parallel to any known pre-basalt structure, but indirect evidence suggests that at least some of these folds may overlie earlier basement faults. Discussion A brief discussion of the origin of the regional north-south compressional stress field may be appropriate for the benefit of readers not familiar with the tectonics of northwestern USA. It is, after all, a surprising orientation, given a casual glance at the tectonics of subduction along a Ntrending trench of the Juan de Fuca plate under the North American plate. Even the slight obliquity of subduction towards the east-northeast seems insufficient cause, especially as there are no obvious partitioning structures (trench parallel north-south strike-slip and east-west normal compression or extension structures), such as are seen in the hangingwall of other oblique transpressional margins, for example, Hikurangi, New Zealand (Cashman et al. 1992) and Sumatra (Fitch 1972). At the large tectonic scale (e.g. Humphreys & Weldon 1994), the stress field appears to be related to the effects of the northward-directed displacements from the south due to the broad zone of interaction between the Pacific plate and the North American plate (Atwater 1970) plus the north and westward component of extension across the Great basin, and northerly extensions of it, extending away from a stable craton reference frame, so 'escaping' out towards the 'free' boundary of the northwest subduction zone. The east-west structures of Lewiston and the Yakima folds would then be the expression of transferring displacement from inboard out towards the subduction zone. Pre-existing structures, if they lie in the general east-west, northwest- southeast direction, have been reactivated with a variety of styles of deformation expressed in the basalts, including faulting,
185
block uplift and flexure, but only where the basalts overlie accreted terranes. References ATWATER, T. 1970. Implications of plate tectonics for the Cenozoic tectonic evolution of North America. Geological Society of America Bulletin, 81, 35133536. CAMPBELL, N. P. 1989. Structural and stratigraphic interpretation of rocks under the Yakima fold belt, Columbia Basin, based on recent surface mapping and well data. Geological Society of America, Special Paper, 239, 209-222. CASHMAN, S. M., KELSEY, H. M., ERDMAN, C. F., CUTTEN, H. N. C. & BERRYMAN, K. R. 1992. Strain partitioning between structural domains in the forearc of the Hikurangi subduction zone, New Zealand. Tectonics, 11, 2242-2257. FLECK, R. J. & CRISS, R. E. 1985. Strontium and oxygen isotopic variations in Mesozoic and tertiary plutons of central Idaho. Contributions to Mineralogy and Petrology, 90, 241-318. FITCH, T. J. 1972. Plate convergence, transcurrent faults, and internal deformation adjacent to the southeast Asia and the Western Pacific. Journal of Geophysical Research, 77, 4432-4461. HOOPER, P. R. & CONREY, R. M. 1989. A model for the tectonic setting of the Columbia River basalt eruptions. Geological Society of America, Special Paper, 239, 293-306. , GILLESPIE B. A. & Ross M. E. 1995. The Eckler Mountain Basalt and associated flows, Columbia River Basalt Group. Canadian Journal of Earth Science, 32, 410-423. HUMPHREYS, E. D. & WELDON, R. J. 1994. Deformation across the western United States: a local estimate of Pacific North America transform deformation. Journal of Geophysical Research, 99, 19,975-20,010. KIM, K., DISCHLER, S. A., AGGSON, J, R. & HARDY, M. P. 1986. The State of in situ Stresses Determined by Hydraulic Fracturing at the Hanford Site. Rockwell International, RHO-BW-ST-73 P. KUEHN, S. C., HOOPER, P. R., THIESSEN, R. L. & WATKINSON, A. J. 1996. Structures in the Columbia River basalt associated with the Olympic-Wallowa lineament and Hite Fault. Geological Society of America!, Abstracts with Programs, 28, 5, p.83. LAWRENCE, R. D. 1976. Strike-slip faulting terminates the Basin and Range province in Oregon. Geological Society of America Bulletin, 87, 846-850. MANN, G. M^ & MEYER C. E. 1993. Late Cenozoic structure and correlations to seismicity along the Olympic-Wallowa lineament, northwest United States. Geological Society of America Bulletin, 105,853-871. PRICE, E. H. & WATKINSON, A. J. 1989. Structural geometry and strain distribution within the eastern Umtanum fold ridge, south-central Washington. Geological Society of America, Special Paper, 239, 265-282.
186
A. J. WATKINSON & P. R. HOOPER
REIDEL, S. P. 1984. The Saddle Mountains; evolution of an anticline in the Yakima fold belt. American Journal of Science, 284, 942-978. . CAMPBELL. S. P., FECHT. K. R. & LINDSLEY. K. A. 1994. Late Cenozoic structure and stratigraphy of south-central Washington. In: CHENEY. E. & LASMAMIS. R. (eds) The 2nd Symposium on the Geology of Washington. Washington Division of Geology and Earth Resources. Bulletin. 80, 159-180. . FECHT. K. R.. HAGOOD. M. C. & TOLAN. T. L. 1989. The geological evolution of the central Columbia Plateau. Geological Society of America. Special Paper. 239. 247-264. RIDDIHOUGH. R.. FINN, C. & COUCH, R. 1986. KlamathBlue Mountains lineament, Oregon. Geologv. 14, 528-531.
SOBCZYK, S. M. 1994. Crustal Thickness and Structure of the Columbia Plateau Using Geophysical Methods. PhD dissertation, Washington State University. Pullman. TABOR. R. W.. FRIZZELL. V. A.. VANCE. J. A. & NAESER. C. W. 1984. Ages and stratigraphy of lower and middle Tertiary sedimentary and volcanic rocks of the central Cascades. Washington. Geological Society of America Bulletin. 95. 26-44. TOLAN. T. L.. REIDEL. S. P.. BEESON. M. H.. ANDERSON. J. L.. FECHT. K. R. & SWANSON. D. A. 1989. Revisions to the estimates of the areal extent and volume of the Columbia River Basalt Group. Geological Society of America. Special Paper. 239. f-20.
The interplay of faulting and folding during the evolution of the Zagros deformation belt Y. SATTARZADEH1, J. W. COSGROVE2 & C. VITA-FINZI3 1
Department of Geology, University of Tabriz, Tabriz, 51664, Iran Earth Sciences, T. H. Huxley School of the Environment, Imperial College, London SW7 2BP, UK (e-mail:
[email protected]) 3 Department Earth Sciences, University College, Gower Street, London WC1E 6BT, UK
2
Abstract: In this short paper satellite images, aerial photographs and seismic sections are used to show that pure buckle folds, pure forced folds and folds intermediate between the two have all formed, and are still forming, in association with the compression tectonics currently occurring in the Zagros deformation belt which is situated along the northeastern margin of the Arabian plate. The type of folding and its distribution can be linked directly to the distribution of ancient basement faults and to the rheological profile of the cover sequence.
The Zagros fold thrust belt is situated along the northeastern margin of the Saudi Arabian plate (Fig. 1) and represents a currently active compressional plate margin. At present the deformation of the margin is taking place by a complex interaction of faulting, folding and salt tectonics. In this paper we examine the different types of folding which range from pure buckle folds formed in response to a regional compression, through buckle folds formed in cover rocks over active basement strike-slip faults to true forced folds formed in cover rocks as a result of oblique or pure dipslip movement on thrust, normal or wrench faults in the basement. It can be shown that much of the deformation since the onset of collision in the Late Cretaceous relates directly to the reactivation of old basement faults and to the rheological profile of the sediments which overlie the Precambrian basement. Thus, in order to understand the type and distribution of the deformation that has occurred and which is still occurring along the Zagros belt, it is useful to consider briefly the build up of the stratigraphic succession of the Zagros belt and the tectonic history of the region. Tectonic and stratigraphic evolution of the Zagros fold belt According to Stocklin (1968) early orogenic movements in Iran resulted in the consolidation of a Precambrian, shield-like basement and the formation of a vast Iranian platform as an extension of the Arabian Shield and part of Gondwanaland. The 'Assyntic orogeny' occurred
in Late Precambrian time, and resulted in the intensive folding and metamorphism of the Iranian platform (Iranpanah & Esfandiari 1979), including what was to become the basement of the present-day Zagros Mountain belt. This was followed by a long period of remarkable tectonic calm during most of the interval from the InfraCambrian to Middle Triassic times. The rock sequences deposited during this time interval are shallow-water sediments, and include the InfraCambrian Hormuz salt Fig. 3a (part of the Lower Mobile Group, Fig. 2), and have the characteristics of a true platform cover. No orogeny affected this platform during the Palaeozoic era and only epeirogenic movements occurred. Palaeogeographical maps that have been prepared on the basis of palaeomagnetic data indicate that in Early Mesozoic times the continental masses of Africa, Arabia, central Iran, Australia and Antarctica were all part of a great mass of continental crust named Gondwanaland, which was separated from Laurasia by the oceanic crust of the Tethys Sea. This sea was a wide seaway that spread far over the margins of the continental crust so that much of the crust of Arabia and parts of central Iran, for example, were submerged by it (Shearman 1976). During the Mesozoic Gondwanaland began to break up into smaller continents, and during this time central Iran, which forms part of the Iranian platform, separated from Afro-Arabia by seafloor spreading, an event sometimes referred to as the opening of the Neo-Tethys sea (Fig. 3b & c) (Shearman 1976; Alavi 1980). According to Alavi, rupturing occurred in the Late Triassic but may even have been initiated in the Late
From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 187-196. 1-86239-060-6/OO/S 15.00 r The Geological Society of London 2000.
188
Y. SATTARZADEH ET AL.
Fig. 1. Plate tectonic setting of the Zagros fold belt situated along the northeastern margin of the Arabian Plate.
Permian. The resulting normal faults define several horsts and grabens. On the northeastern (Iranian) part of the continent, the environment changed from marine to continental, but on the southwest (Arabian) part shallow-water marine sedimentation continued in the newly generated 'sedimentary basins'. In Early Jurassic times the main graben deepened, while block faulting was still occurring elsewhere, and developed into a spreading centre (the Neo-Tethys Sea) between the two nowseparating continents of Arabia to the southwest and central Iran to the northeast. Volcanic activity along the rift generated oceanic crust and associated Red Sea structures (Fig. 3b & c). Deep-water sedimentation occurred along the rift and continuous shallow-water marine sedimentation occurred along the marginal basins to the southwest. In Late Jurassic times, the central Iranian and Arabian blocks moved away from each other as the ocean developed, and deep-sea sediments (radiolarites and pelagic limestone) and volcanic and volcano-clastic materials were developed on the ocean floor (Fig. 3c). Two shallow basins were formed as the sea transgressed onto the neighbouring continents. The one on the Iranian block was relatively shallow and received clastic sediments. The basin on the Arabian block was somewhat deeper and received continental shelf-type marine carbonates, which form part of the competent group sandwiched between the underlying Hormuz Salt Series (the Lower Mobile group) and the overlying Upper mobile group (Fig. 2). In Early Cretaceous times, as a result of the opening of the Atlantic Ocean, westward motion of the Arabian plate ceased (Iranpanah & Esfandiari 1979). In addition, the eastward
motion of the central Iranian plate also stopped as a result of the closing of the Tethys Sea. Thus, the relative motion of the Arabian and central Iranian plates reversed, and subduction of the oceanic crust beneath the central Iranian plate was initiated (Fig. 3d). As a consequence an oceanic trench formed in front of the Iranian continent (Fig. 3d), and the Arabian plate moved northeastward as the trailing edge continental attached to the subducting oceanic crust. High-pressure, low-temperature metamorphism occurred below the trench, as indicated by glaucophane-jadeite mineral assemblage in the Zagros ophiolites. Uplift of the southwestern edge of the the central Iran continent occurred because of pressure exerted by the down-going oceanic crust, and a foreland basin formed on its former continental shelf (Fig. 3d). Shallow-water sedimentation on the Arabian continental shelf was continuous during this period (Alavi 1980). In Late Cretaceous times the development of the subduction zone, and the associated progressive loss of oceanic crust along the trench, lead to the early phases of collision of the Arabian block and the Iranian continent (Fig. 3e). Several highangle reverse faults formed along the edge of the Arabian continent, and southwest obduction of slices of oceanic crust along these faults caused ophiolitic melange complexes to form. These events were followed by the intense folding, uplift and short-lived erosion of the deep ocean sediments, more plutonic activity on the edge of the Iranian continent, a second phase of metamorphism (high pressure, high temperature) and the migration of the deformation into the undisturbed Zagros sedimentary basin, causing pre-buckle thickening and some scattered salt intrusions further to the southwest. During the Palaeocene further subduction of the oceanic crust resulted in a late-phase and much stronger collisional event and in the reactivation of several normal fault zones as high-angle reverse faults, including the future Zagros thrust line. These affected the continental shelf sequence and the basement complex on the northeastern part of the Arabian continental shelf area. Uplift of the thrust blocks occurred and a flysh basin developed in front of the elevated area on the shelf sediments (Alavi 1980). In Eocene times intense plutonic and volcanic activity along the edge of the Iranian continent, originating from the subducted oceanic crust of the Neo-Tethys Sea, resulted in the development of the Urmieh-Dokhtar magmatic arc (Takin 1972). As can be seen from the above discussion, what is now the northeastern margin of the
FORCE FOLDING IN THE ZAGROS MOUNTAINS
189
Fig. 2. Schematic stratigraphic subdivision of the cover rocks of the Zagros fold belt (after Coleman-Sadd 1978). The rheological profile of this succession is shown on the left-hand side and is made up of a Lower Mobile Group characterized by the evaporites of the Hormuz Salt Series which respond to plate collision in a ductile manner, the Competent Group characterized by strong carbonates which respond in a brittle and ductile manner, and the overlying Upper Mobile Group which is characterized by marls and evaporites which behave in a ductile manner.
190
Y. SATTARZADEH ET AL.
Fig. 3. Plate tectonic evolution of the Zagros Mountain belt (modified from Alavi 1980). Note the continuous build up of shelf deposits on the Arabian plate margin from the Precambrian to the onset of folding in the Tertiary.
FORCE FOLDING IN THE ZAGROS MOUNTAINS
Arabian plate had been a site of sedimentation since at least the Cambrian (Fig. 3). The Precambrian basement is covered with the thick Hormuz salt and this, together with the associated shallow-water sediments, is known as the Lower Mobile Group. It is followed by a thick sequence of shelf carbonates, known collectively as the 'Competent Group' (Fig. 2), which built up on this margin during the Mesozoic and early Tertiary. During the Miocene an Upper Mobile Group was deposited over the Competent Group. It contains numerous evaporite and shale horizons, and allows the decoupling of deformation in the underlying Competent Group from the overlying Incompetent Group (Upper Miocene-Recent) (Fig. 2). In Pliocene times, the elevated SanandajSirjan zone changed from highlands to lowlands, and Alavi (1980) suggested that this was because of the pull exerted by the down-going attached oceanic crust (Fig. 3f). Sedimentary basins developed on both sides of the much expanded magmatic arc. The high-angle reverse faults that had originated in the Late Cretaceous in the Sanandaj-Sirjan zone were then reactivated, especially along the Main Zagros thrust. The accelerated motion of the underlying basement complex relating to the opening of the Red Sea, initiated in Miocene times (Laughton 1966), resulted in intense folding and faulting in the whole Zagros sedimentary sequence, which overlay a rigid metamorphic and igneous basement complex, in the Pliocene (Fig. 30Types of folding in the Zagros
Buckle folds The Zagros fold belt is often cited as one of the best examples of large-scale buckle folding. Satellite images and aerial photographs show numerous examples of anticlines with aspect ratios (half wavelength to axial length ratio) of between 5:1 and 10:1, which are typical of buckle folds on all scales and which display the characteristic en echelon spatial organization in plan view (see for example Price & Cosgrove 1990), in which as one fold dies out it is replaced by another with a parallel axes but offset from the first fold as shown in the idealized fold distribution pattern in Fig. 4a and in the satellite image of the Zagros folds shown in Fig. 5, where this characteristic aspect ratio and spatial organization can be clearly seen. However, closer inspection of the aspect ratio and/or spatial organization of some of the folds of
191
the fold belt (Fig. 6) show features that are uncharacteristic of buckle folds. These features are discussed in the following sections. It should be noted in passing that one method of forming a buckle fold with an uncharacteristically large aspect ratio is for two separate folds to link. As individual folds grow they increase in length along their hinge lines and in so doing may interact with other folds. Experimental work and field observations have shown that there are a variety of ways in which interaction can occur. For example, if the hinges of two interfering folds are offset but the amount of offset is only a small fraction of the wavelength, the folds link or elide to form a longer fold with a deflection in the hinge line (Fig. 4d(i)). Experiments show that if the folds are separated by more than about half their wavelength but are still close enough to interact, they lock up, each preventing further propagation of the other. In this way they become arranged in an en echelon fashion (Fig. 4d(ii)). The linked anticlines (Fig. 4d(i)) will have an aspect ratio much larger (around twice) that of a single buckle fold. However, such anomolies can easily be detected by the deflection of the hinge line. Many examples of these elided anticlines occur in the Zagros fold belt (Fig. 6).
Buckling associated with basement wrench faults There are two convincing lines of evidence that strike-slip faults exist in the basement underlying the Zagros fold belt and that movement on these faults has resulted in folding of the cover rock. The first comes from the recent seismicity of the Zagros belt (see e.g. Jackson & McKenzie 1984, Ni & Barazangi 1986). Although the fault plane solutions discussed by these authors are predominantly associated with thrust faulting, clear lineaments of wrench faulting can be detected which coincide with known basement lineaments such as the Kazarun line. The second line of evidence comes from experimental work on analogue models designed to study the effect of 'basement' strike-slip faults on 'cover' rocks sequences (see e.g. Oliver 1987; Richard 1990, 1991; Richard & Krantz 1991; Richard et al 1991). This work shows that, depending on the mechanical and rheological properties of the cover rock, they may deform in either a brittle or ductile manner and produce arrays of fractures or folds, or combinations of the two. The patterns and distributions of these structures can be used to determine both the existence of an underlying strike-slip fault and its sense of movement.
192
Y. SATTARZADEH ET AL.
Fig. 4. (a) Idealized spatial organization of buckle folds in plan view, (b) Shear along a basement wrench fault induces compression in the cover rocks which may produce en echelon folds as. for example, along the sinistral Dead Sea fault, (c) After Moody & Hill (1956). (d) (i) The amplification and coalescence of two periclines (showrn in plan) separated by a distance X which is somewhat less than half the wavelength of the structure, (ii). two periclines separated by X 1 which is greater than half the wavelength. The structures overlap and lock up, each preventing further propagation of the other. After Price & Cosgrove (1990).
The folds produced when a cover sequence responds in a ductile manner to basement strike-slip faulting are buckle folds, and as such form with an aspect ratio of between 1:5 and 1:10. They can, however, be distinguished from buckle folds associated with a regional shortening by their spatial organization. In plan view they are arranged so that they are consistently offset from each other, either to the right or to the left. The relationship between the underlying basement strike-slip fault, the sense of movement on the fault, the orientation of the resulting folds in the cover rocks and their spatial organization is shown in Fig. 4b. Numerous well-known
examples have been cited from around the world, including fold arrays along the San Andreas fault (Moody & Hill 1956) and the example from the Dead Sea strike-slip zone shown in Fig. 4c. A study of fold distribution and geometry in the Zagros fold belt (Fig. 6) shows several excellent examples of folding above a basement strike-slip fault zone. One of the most convincing is the E-W-trending en echelon array of anticlines above the Bala Rud line which marks the northern boundary of the Dezful embayment and separates the Bakhtiari and Lorestan fold zones (Fig. 6). The fold distribution indicates
FORCE FOLDING IN THE ZAGROS MOUNTAINS
Fig. 5. Satellite image of part of the Zagros fold belt. The characteristic aspect ratio and spatial organization of buckle folds can be clearly seen.
that the strike-slip motion along this line was sinistral. Another example of en echelon folding above a basement strike-slip fault zone occurs along the N-S-trending Kazerun line, which forms the other margin to the Dezful embayment
Fig. 6. Distribution of folds in parts of the 'simply folded belt" of the Zagros Mountains showing the difference in the distribution of the folds within the different fold zones. After Sattarzadeh-Gadim (1997).
193
and separates the Bakhtiari from the Pars fold zone, left hand side of the upper map, Fig. 6. The fold distribution in the cover rocks indicates dextral movement on the basement fault. In addition to their spatial organization, the buckle folds formed in association with basement strike-slip faulting may have second-order features that enable them to be distinguished from buckle folds formed as a result of a regional compression. For example they may become sigmoidal in plan view as a result of rotation of the earlier (central) part of the fold before the growth of the younger (peripheral) portions, in a manner similar to the formation of sigmoidal tension gashes. In addition, they may develop a more complex profile geometry which changes its symmetry when traced from one end of the fold to the other (see Oliver 1987; Fig. 5 in Cosgrove & Ameen this volume). Forced folding Forced folds have been defined by Stearns 1978 as 'folds in which the final overall shape and trend are dominated by the shape of some forcing member below'. Thus, in order for such folds to form when the forcing member is a fault block, it is necessary for there to be an element of dip-slip movement on the underlying fault. Although the folds mentioned in the preceding section are intimately linked to movement on a basement fault, the movement is purely strike-slip and the resulting folds are buckle folds. However, forced folding does occur in the Zagros fold belt and this is often related, either directly or indirectly, to the reactivation of basement normal faults (Fig. 3a-c) as reverse dip-slip faults during and subsequent to the major plate collision in the Miocene (Fig. 3e andf). Many of the Zagros folds are linked to underlying thrusts, and it seems that these in turn are often initiated by the reactivation of basement normal faults. Movement on these faults produces forced folds in the overlying Hormuz salt which behaves as a weak, ductile material. However, the overlying Competent Group (see Fig. 2) responds to the movement in the basement in a brittle manner and the forced folds in the Hormuz salt gives way to thrusts in the more brittle Competent Group. These thrusts climb through the succession from the Lower to the Upper Mobile Group and result in the formation of fault-bend folds in the upper units of this succession, notably the Asmari limestone (Fig. 7) which probably plays an important role in controlling the wavelength of these folds (see Price
194
Y. SATTARZADEH ET AL.
Fig. 7. Part of a northeast-southwest structural cross-section based on a seismic profile across 'Burunjen" (5112=E-32 59 =N). After Payne (1990).
& Cosgrove 1990, p. 282). The thrusts often remain blind beneath the folds but do occasionally break the surface. Thus, as can be seen from Fig. 7, the forced folding in the Asmari limestone is often the result of fault-bend folding associated with the generation of new thrusts in the cover, whereas deeper in the succession, at the basement-cover contact, the forced folding of the Hormuz salt is the result of the reactivation of old basement normal faults. One of the major differences in the geometry of forced folds and buckle folds is their aspect ratio. Because the forcing members that generate forced folds generally form long linear steps in the basement, the resulting folds frequently have long aspect ratios and are frequently, although not always, asymmetric. The detailed study of the Zagros folds by hydrocarbon geologists using field observations, drill data and seismic sections, has indicated clearly that the deformation of the cover rocks is characteristic of a typical fold-thrust belt in which folds and thrusts occur together. Seismic sections have shown that many of the folds now visible at surface are linked to underlying blind thrusts (Fig. 7). The growth of many of these fault-bend folds has been facilitated by the fact that they are near-surface structures with little or no overburden to inhibit their growth. Thus, many have the profile geometry of buckle folds formed near the surface above a decollement horizon (see, for example, the folds in the Jura Mountains that form in the Jurassic limestones that rest on a decollement horizon of Triassic halite). However, these folds can often be easily distinguished from pure buckle folds by their aspect ratio. Their length is controlled by the length of the thrust with which they are associated and some of them extend
for many tens or even hundreds of kilometres. One of the most impressive is the Kabir Kuh anticline which formed in the Lorestan fold zone (Fig. 6) and has a length of 200 km. Thus, although in profile section these folds have a geometry similar to that found in pure buckle folds, their anomolously long lengths reveal the fact that they are forced folds formed over the linear scarp associated, in these examples, with an underlying thrust. Forced folding above oblique-slip faults In the discussion of fault-related folding in the previous sections the faults have been assumed to be either pure strike-slip (where pure buckle folds form in the overlying cover rocks) or pure dip-slip faults (where pure forced folds occur). In the example of pure strike-slip basement faults, the resulting folds are pure buckle folds, and the only difference between these and the buckle folds associated with a regional compression is in their spatial organization. In order for folds to be forced folds it is necessary for the associated fault to involve dip-slip movement. In the examples discussed above, pure dip-slip faults, either inverted normal faults or newly formed thrust faults, were considered. Clearly it is possible that the parent fault may be an oblique-slip fault, having components of both strike-slip and dip-slip motion. In such an example the resulting folds will possess two geometric characteristics which will declare their genetic link to the parent fault. These are the systematic offset of the adjacent folds and their high aspect ratio. An excellent example of such folding can be found in the southern part of the Zagros fold belt above the Minab basement fault. This fault is thought to
FORCE FOLDING IN THE ZAGROS MOUNTAINS
Fig. 8. (a) folds and (b) folds and associated faults linked to the Minab fault zone adjacent to the Strait of Hormuz. The fold geometry and spatial organization reflect the transpressive nature of this zone. The en echelon arrangement of the folds is the result of the component of strike-slip movement on the fault zone and their spatial organization indicate that the sense of movement is dextral. The long aspect ratio of the folds indicates that an important element of dip-slip movement occurred on the faults. The resulting fault scarps control the aspect ratio of the resulting folds which is too large for the folds to be pure buckle folds.
be associated with the current indentation of the Mussandan Peninsular of the Arabian shield into the Iranian plate. The folding of the cover rock above this transpressive fault can be clearly seen in Figs 6 and 8. In addition to an important dextral component of movement, the fault also has a component of reverse dip-slip motion which has caused the eastern block to be thrust over the western block (Figs 6 and 8). The resulting fault scarp at the basement-cover contact has imposed a high aspect ratio on the resulting en echelon folds. In addition, continued movement along the fault has caused the folds to rotate clockwise and has produced a sigmoidal deflection of the fold axes. This can also be seen in the Mand anticline at present growing above the Kazarun line where it hits the Persian Gulf, near the western end of the Fars fold zone (Fig. 6). Conclusion The type of folding in the Zagros Mountains is controlled primarily by the Theological profile of the sedimentary cover, the reactivation of basement faults (wrench faults and the reverse dip-slip reactivation of normal faults) and the generation of new faults (thrusts) in the cover rocks. All these occur in response to the regional
195
compression associated with the collision of the Saudi Arabian and Iranian plates. Buckle folds form over decollement horizons within the Lower and Upper Mobile Group within the cover rocks (Fig. 2). These folds occur on a large-scale above the lower decollement represented by the thick Hormuz salt. Their wavelength is controlled by the thick competent units of the Asmari limestone. They also occur on a smaller scale above the evaporite detachment horizons within the Upper Mobile Group. The overlying beds are thin limestones and marls (of the Gach Saran Formation) and, as a result, the folds have a smaller wavelength. In addition, strike-slip motion on basement faults systems, such as the Kazarun line and the Bala Rud line which define the Dezful embayment, and the Minab fault system associated with the indentation of the Musandan peninsular into the Iranian Plate, have also produced folds in the cover sequence. These range from pure buckle folds arranged in an en echelon manner above the fault zone, and which characterize folding above pure strike-slip basement faults, to folds formed above strike-slip faults involving an important element of dip-slip motion. These oblique-slip faults produce fault scarps (steps) in the basement-cover contact that have an important impact on the aspect ratio of the resulting folds. Such folds as, for example, those formed over the Minab fault zone, show elements of both buckle and forced folds, i.e. they display the characteristic en echelon arrangement but display aspect ratios much greater than pure buckle folds. Pure forced folds are formed by the reactivation of basement normal faults as reverse dip-slip faults. These form forced folds in the overlying Hormuz Salt Series. The resulting displacements of the overlying more competent units result in the initiation of important thrust faults. The growth of these thrusts generates large-scale fault-bend folds whose wavelength seems to be controlled by the thick competent unit of the Asmari limestone. This shortening of the cover also produces thrusts and associated fault-bend folds in the Upper Mobile Group, the Gach Saran Formation. The wavelength of these folds reflects the thin competent beds of this formation and is correspondingly smaller. References ALAVI, M. 1980. Tectonostratigraphical evolution of the Zargrosides of Iran. Geology, 8, 144-149. COLMAN-SADD, S. P. Fold development in Zagros simple fold belt, SW Iran. Bulletin of American Association of Petroleum Geologists, 62, 984-1003.
196
Y. SATTARZADEH ET AL.
COSGROVE, J. W. & AMEEN, M. S. 1999. A comparison of the geometry, spatial organization and fracture patterns associated with forced folds and buckle fields. This volume. IRANPANAH, A. & ESFANDIARI, B. 1979. Structural evolution and correlation of tectonic events in the Alborz Mountains, the Zagros range and central Iran. Bulletin de la Societe Belgique de Geologie. 88, 285-295. JACKSON. J. A. & MCKENZIE, D. P. 1984. Active tectonics of the Alpine Himalayan belt between western Turkey and Pakistan. Geophysics Journal. Royal Astrological Society, 77, 185-264. LAUGHTON. A. S. 1966. The Gulf of Aden. Philosophical Transactions of the Roval Society, 259, 150-171. MOODY. J. D. & HILL. M. J. 1956. Wrench fault tectonics. Geological Society of America. Bulletin. 67,1207-1246. Ni, J. & BARAZANGI, M. 1986. Seismotectonics of the Zagros continental collision zone and a comparison with the Himalayas. Journal of Geophysical Research. 98, 8205-8218. OLIVER, D. 1987. The Development of Structural Patterns Above Reactivated Basement Faults. PhD Thesis. University of London. PAYNE, A. 1990. A Structural Interpretation of the Zagros Fold Belt, SW Iran-NE Iraq. British Petroleum Company Ltd. internal report.
PRICE. N. J. & COSGROVE. J. W. 1990. Analysis of Geological Structures. Cambridge University Press. Cambridge. RICHARD. P. 1990. Champs de Failles Audessus d'un Decrochement de Socle: Moderation Experimentale. PhD Thesis. University of Rennes. 1991. Experiments on faulting in a two layer cover sequence overlying a reactive basement fault with oblique slip. Journal of Structural Geologv. 13. 459-470. & KRANTZ. R.W. 1991. Experiments on fault reactivation in strike-slip mode. Tectonophvsics. 188.117-131. . MOCQUET. B. & COBBOLD. P. R. 1991. Experiments on simultaneous faulting and folding above a basement wrench fault. Tectonophvsics. 188. 133 141. SATTAZADEH-GADIM. Y. 1997. Active tectonics in the Zagros Mountains. Iran. PhD Thesis. University of London. SHEARMAN, D. J. 1976. The geological evolution of southern Iran: the report of the Iranian Makran expedition. Geographical Journal. 142. 393-413. STEARNS. D. W. 1978. Faulting and forced folding in the Rocky Mountains foreland. Bulletin of the Geological Society oj America. 151. 1-37. STOCKLIN. J. 1968. Structural history and tectonics of Iran. AAPG Bulletin. 52. 1229-1258. TAKIN. M. 1972. Iranian geology and continental drift in the Middle East. Nature. 235. 147-150.
Complex metamorphic zonation of the Thaya dome: result of buckling and gravitational collapse of an imbricated nappe sequence P. STIPSKA12, K. SCHULMANN11 & V. HOCK3 1
Institute of Petrology and Structural Geology, Charles University, Albertov 6, 12843 Prague, Czech Republic 2 Geophysical Institute, Czech Academy of Science, Bocni II11401, 14131, Prague, Czech Republic 3 Institute of Geology and Palaeontology, University of Salzburg Abstract: The metamorphic isograd geometry and dome structure of the Thaya tectonic window, which emerges through the Moldanubian nappe pile at the eastern margin of the Bohemian Massif, is interpreted as a result of large-scale buckling of an imbricated nappe sequence. This large-scale mechanical instability was initiated by the blocking of passively transported hot viscous rocks to shallow crustal levels. The wavelength and shape of the buckle fold is controlled by the greatest thickness of the uppermost nappe and by the low ratio of strong orthogneiss to weak micaschists in individual nappes. The steep inclinations of the thrust surfaces and the oblique movement of the crustal multilayer minimizes the role of gravity on fold generation. Medium-scale gravitational folds originated at the end of the buckling of the nappes and are associated with the late sliding of the nappes away from the core of the anticline during late exhumation. Thermal, rheological and fold calculations are presented which document and explain the mechanism of folding of large-scale crystalline nappes.
Over the past two decades an increasing number of examples of thrust-related inverted metamorphic zonations have been reported, for example, from the Caledonides (Mason 1984), from the Variscides (Burg et al. 1989) and from the Tertiary Himalayan chain (Mohan et al. 1989; Treloar era/. 1989). Several models have been proposed to explain the inversion of metamorphic zones. These are: (1) large-scale syn-metamorphic folding of isotherms; (2) conductive heating from the hot upper plate to a colder lower plate - the 'hot iron effect' (Hubbard 1989); (3) shear heating; (4) imbrication and tectonic stacking of metamorphic rocks with higher-grade rocks in the hangingwall and lower-grade rocks in the footwall (Treloar et al. 1989; Brunei & Kienast 1986); and (5) inversion of metamorphic isograds by a superimposed zone of ductile shear (Mason 1984). Whatever the mechanism of inversion, the metamorphic isograds should be subparallel to lithotectonic boundaries in trend and often also in dip. The close geometric relationship of isograds with tectonic boundaries is reported from many case studies (Mohan et al. 1989; Treloar et al. 1989). The main feature of the eastern margin of the Bohemian Massif is the eastward thrusting of a hot Moldanubian unit over the easterly Brunovistulian continent. The deformation and
metamorphism of the lower Brunovistulian plate led to the formation of a shear zone called the Moravian zone. This zone consists of a complicated nappe system derived from the Brunovistulian plate which emerges through the Moldanubian nappes in the form of northnortheast elongated tectonic windows (Fig. 1). The Brunovistulian nappe pile in the southern part of the Thaya window shows an inverted metamorphic zonation ranging from biotite to sillimanite zones (Hock 1975, 1995; Stipska & Schulmann 1995). An explanation of the inverted metamorphic zonation in this region, which involves the imbrication of metamorphic zones and their passive deformation (i.e. a combination of models 4 and 5) has been suggested by Stipska & Schulmann (1995). However, the NE-trending mineral zones are only parallel to the boundary of the window in its central part. In the south and in the north they cut the regional structure (Hock 1975, 1995). None of the abovementioned models of inversion of metamorphic zones is able to explain this geometry. This paper is not intended to explain the mechanism of inversion of metamorphic zones, but focuses on the elucidation of the oblique geometric relationship between the lithotectonic boundaries and metamorphic zonation in the dome-like structure of the Thaya tectonic window. This geometry is explained in terms of 'buckling' of a multilayer nappe pile during the
From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 197-211. 1-86239-060-6/OO/S 15.00 © The Geological Society of London 2000.
198
P. STIPSKA ET AL.
Fig. 1. Tectonic sketch map of the eastern margin of the Bohemian Massif. The Thaya window is situated to the south and the Svratka window to the north.
late stages of oblique convergence. The proposed mechanism of crustal-scale nappe buckling is supported by simple one-dimensional thermal and rheological modelling.
Geological setting The lithology, structure and metamorphism of the Thaya window has been studied in detail since the end of the last century and the nappe structure, as well as the inversion of metamorphic zonation, had already been recognized by Suess (1912, 1926). Detailed descriptions of the geology of the Thaya window from the Austrian and Bohemian parts were presented by a number of workers (Preclik 1927; Waldmann 1930; Frasl 1970). In addition the description of the structural evolution of the Thaya window (the polyphase structural history, the orientation of structures, the finite strain and the kinematics), has been the objective of detailed studies carried out over the past 10 years (Schulmann 1990; Fritz 1991; Fritz & Neubauer 1993; Schulmann et al 1994; Kolafikova et al 1997; Lobkowicz et al 1998).
The metamorphic evolution of the Thaya dome (the mapping of the metamorphic isograds, the polyphase metamorphic evolution and the P-T estimates) has also been described in numerous papers (Frasl 1970; Hock 1975, 1995; Stipska & Schulmann 1995). The interested reader is referred to these works and only a short summary of the metamorphic and structural evolution is presented here.
Lithology The Thaya tectonic window (Fig. 2), which is situated at the southeastern margin of the Bohemian Massif, consists of a para-autochthonous Brunovistulian basement (Dudek 1980) in the core rimmed by two basement-derived nappes. They show the following lithology and structure. The structurally deepest Thaya granite of Cadomian age (550 Ma, Rb-Sr, Scharbert & Batik 1980) is overlain by a paraautochthonous Upper Proterozoic metapelitic to metapsammitic sequence forming the original roof of the intrusion. The next, higher unit is a basement-derived Lower Moravian nappe (LMN) which has
LARGE-SCALE FOLDING DURING PLATE COLLISION
199
Fig. 2. Structural map and a northeast-southwest cross-section of the Thaya window, (a) Tectonic and structural map showing the major nappe structure of the window. The orientation diagrams of lineations and foliations show the structural pattern in the central strike-slip and northern and southern dip-slip domains (data from Fritz 1991; Schulmann 1990; Schulmann et al. 1994). (b) Schematic northeast-southwest cross-section of the Thaya window showing syn-metamorphic kinematic indicators after Fritz (1991), Fritz & Neubauer (1993) and Schulmann et al. (1994).
strongly sheared orthogneiss (the Weitersfeld orthogneiss) at the base overlain by metasedimentary cover sequences composed of metapelites, marbles and calcsilicate rocks. The Weitersfeld orthogneiss is restricted to the central part of the Thaya dome (Fig. 2). The uppermost basement-derived Upper Moravian nappe (UMN) also has an orthogneiss body at its base (the Bites orthogneiss) and has a meta-volcano-sedimentary cover. The boundary between the Moldanubian nappes and the Brunovistulian basement-derived nappes is
formed by the so-called Micaschist zone (MZ), a monotonous sequence of micaschists and subordinate amphibolites. Structures The main structural feature of the Thaya dome is the syn-metamorphic foliation S1 which parallels the lithotectonic boundaries and forms an Sshaped structure (Fig. 2). The SI foliation bears a stretching and mineral lineation LI, and is
Fig. 3. Photographs of laic gravitational structures, (a) West-facing late folds indicate the westward sliding of nappes in the upper part of the Bites orthogneiss (central domain. Vranov nad Dyji). (b) Greenschist facies normal shear zone in the Bites orthogneiss indicating westward sliding (central domain, Vranov nad Dyji). (c) Normal shear /one in the upper part of the Bites orthogneiss (southern dip-slip domain close to Horn), (d) Normal greenschist facies shear /one in M o l d a n u h i a n inigmalitcs associated with large-scale normal fault separating the Thuya window from the M o l d a n u h i a n rocks (Bitov, north of Drosendorf).
LARGE-SCALE FOLDING DURING PLATE COLLISION
associated with syn-schistose isoclinal sheath folds Fl. The syn-metamorphic stretching lineation LI trends north-northeast-south-southwest (Fig. 2) cutting the lithotectonic boundaries at various angles (Schulmann 1990; Fritz 1991). Recent structural investigations (Fritz 1991; Fritz & Neubauer 1993; Schulmann et al 1994) showed that the main deformation phase involved a top-to-the-north (north-northeast) inclined transpressional movement. This is indicated by the mineral stretching lineations and numerous shear-sense indicators which can be traced over large parts of the Moravian zone. The deformations have taken place in the central section in amphibolite facies conditions, which is in accordance with the mineral assemblages in the metapelites and metacarbonate rocks. Fritz (1991) has shown that in the southern dip-slip domain the kinematics exhibit the northeast thrust movements, whereas in the western part of the Thaya window dextral strike-slip movements occur (Schulmann 1990). Schulmann et al. (1994) reported that in the northern termination of the Thaya window the kinematic indicators reveal top-to-the-northeast dip-slip movements (Fig. 2). Detailed finite strain studies (Schulmann 1990; Fritz 1991) have shown oblate fabrics in both the southern and northern dip-slip domains, and plane strain fabric in the western strike-slip domain. The syn-metamorphic structures are refolded by late-metamorphic W-facing recumbent F2 folds (Fig. 3a) developed as a result of the later exhumation of the nappe sequence (Schulmann et al. 1994). These folds vary in size from several metres to 100m, and their hinges change orientation when traced around the dome
201
Fig. 5. Tectonic sketch map of the Thaya window and orientation diagrams showing the orientation of normal greenschist facies shear zones. Lowerhemisphere, equal area projection.
(Fig. 4) from northeast-southwest to almost east-west at the northern termination of the Thaya window (Schulmann 1990). Gravitational folding is associated with the development of medium-scale normal shear zones at the central W-dipping part of both the Upper and Lower Moravian nappes (Figs 3b and 5) which also indicate late westward gravitational sliding (Schulmann et al. 1994; Kolafikova et al. 1997). Numerous normal greenschist facies shear zones are symmetrically developed at the northern and southern terminations of the dome and in the overlying Moldanubian migmatites (Figs 3 and 5). Normal greenschist facies shear zones developed in all levels of the nappe pile as well as the F2 folds, indicate gravitational sliding of a thick Moldanubian migmatitic sequence and Moravian nappes away from the core of the Thaya window (Lobkowicz et al. 1998). This gravitational sliding occurred at a relatively shallow crustal level as indicated by the associated retrogression which occurred under greenschist facies conditions (Kolafikova et al. 1997). Metamorphic evolution Metamorphic zonation
Fig. 4. Block diagram of the northern termination of the Thaya dome showing hinge orientations of late gravitational folds swinging around the Thaya dome.
Two distinct metamorphic events have been recognized (Frasl 1970; Hock 1995). The first is related to the intrusion of the Thaya batholith and is of Upper Proterozoic age. The main regional metamorphism is of Variscan age. It is of
202
P. STIPSKA ET AL.
Fig. 6. (a) Metamorphic zonation map of the Thaya window after Hock (1975. 1995). The P- T estimates after Bernroider (1989), Hock (1995) and Stipska & Schulmann (1995). (b) Schematic northwest-southeast crosssection of the Thaya window (location in Fig. 2) showing the main mineral assemblages using AFM topologies in individual nappes and metamorphic zones. The stars in the AFM diagrams indicate whole rock compositions. Mineral compositions after Stipska & Schulmann (1995). Minerals: bt biotite; chl chlorite; grt garnet; ky kyanite; sil silmanite; st staurolite.
Barrovian type and was followed by greenschist fades retrogression. 40 Ar/ 39 Ar cooling ages of amphiboles from amphibolites in the Bites orthogneiss give 328.7 ±3.3 Ma, muscovites from the Bites orthogneiss 328.7 ± 0.8 Ma and the muscovite of Weitersfeld orthogneiss yields 325.5 ±0.7 Ma (Dallmeyer el al 1992). These data reflect the intersection of approximately 500°C (amphibole) and 350°C (muscovite) isotherms, respectively during a rapid uplift and exhumation. As Suess (1912), Preclik (1927) and Frasl (1970) pointed out, the metamorphic grade is highest in the western part and decreases towards the south, east and north. Therefore, the metamorphism is inverse, i.e. the structurally highest units exhibit the highest grade of metamorphism. In 1975, Hock identified four mineral zones with different assemblages in the metapelitic rocks striking parallel to the mineral stretching lineation (southwest-northeast), and cutting the regional structure (south-north to southwest-northeast), especially in the southern part of the window (Fig. 6a). Recent investigations by the present authors in the northern part of the Thaya window show that the mineral zonation also cuts the regional structure in this area.
The metamorphic zonation is defined by: (1) the albite-oligoclase boundary lying mainly in the autochthonous Thaya granite; (2) a garnetbiotite zone; and (3) a garnet-biotite-staurolite zone (Hock 1995). Kyanite and sillimanite zones (4) occur in the upper part of metapelites of the Upper Moravian nappe (Fig. 6). The peak metamorphic conditions attained in each unit are recorded in the central sector of the Thaya window where the sequence of mineral assemblages can be defined as follows: (1) grt + bt or grt + chl + bt in the lower part of the para-autochthonous unit and st + grt + bt in its upper part; (2) grt + st + bt in the Lower Moravian nappe; (3) grt -h ky -h st -h bt in the Upper Moravian nappe and sil + bt -h grt in its uppermost part (see Fig. 6) (Stipska & Schulmann 1995). These mineral assemblages are syn-tectonic with mainly top-to-the-northeast shearing (Schulmann el al. 1994). The garnet zoning in all tectonic units reflects the prograde temperature evolution (Hock 1995; Stipska & Schulmann 1995). As noted above, the mineral zones strike north-northeast, obliquely cutting the lithotectonic boundaries in the south, running parallel in the central sector and cutting these boundaries obliquely in the north (Fig. 6).
LARGE-SCALE FOLDING DURING PLATE COLLISION
Pressure-temperature (P—T) estimates We have used the published pressure and^temperature estimates of Bernroider (1989) and Stipska & Schulmann (1995) for the northern part of the Thaya window and those of Hock (1995) and Hock et al (1990) for its southern part. Detailed thermobarometry was performed in all tectonic units and includes estimates of maximum P-T conditions and, in some cases, also conditions of retrogression. Temperature estimates in the metapelites were obtained using both garnet-biotite and garnet-staurolite thermometers. Pressures were estimated using garnet-plagioclase geobarometers in the metapelites and phengite barometry in the orthogneisses. Pressure-temperature estimates of retrogression were obtained using late Mn-rich garnet rims compositions (Hock 1995) and fluid inclusions in late quartz veins (Fritz & Loitzenbauer 1994).
Peak metamorphic conditions The peak metamorphic conditions are summarized in Figs 6 and 7. These conditions in the kyanite zone of the Upper Moravian nappe and the Micaschist zone were estimated at 630650°C and 8-10kbar in the northern section and 600-630°C and 9-10kbar at the south. In the southernmost part of the Upper Moravian nappe Hock (1995) estimated pressures of 4.55 kbar at temperatures of 450-500°C. Calculated peak metamorphic conditions of the staurolite zone of the Lower Moravian nappe and the para-autochthonous unit are approximately 600°C and 7-8kbars. Pressure-temperature estimates for the garnet zone of the paraautochthonous unit correspond to approximately 580°C and 7-8kbars in the south, and about 600°C and 7-9kbars in the northern
Fig. 7. The P- T estimates from metapelites of the Thaya window with indicated P-T paths. Calculations taken from Bernroider (1989), Hock (1995) and Stipska & Schulmann (1995).
203
part. Peak pressure conditions in the autochthonous deformed Thaya granite were estimated at approximately 6 kbar ( Kolafikova et al. 1997).
Retrograde part of the pressure-temperature evolution The retrograde stage in the Micaschist zone and the Upper Moravian nappe is marked by the growth of sillimanite at the expense of garnet under conditions of 600-630°C at 5.5-7 kbars and subsequent greenschist facies retrogression (Kolafikova et al. 1997). In the Lower Moravian nappe, small Mn-rich rims around the garnets reflect a stage in which 400-450°C was attained in both the southern and northern parts of the Thaya window (Hock 1995; Stipska & Schulmann 1995). In this unit, Fritz & Loitzenbauer (1994) calculated a pressure of 3.3 kbars and a temperature of 300°C from fluid inclusions in late quartz veins. In the para-autochthonous unit the minimum pressure and temperature of 4 kbar and 450°C were obtained for the retrograde part of the P-Tpath.
Inversion of metamorphic isograds Extensive field observations and P-T calculations on the pelitic assemblages suggest a threestage model for the Variscan evolution of the complex metamorphic zonation pattern. The prograde part of the P-T evolution is documented by the prograde zonation of garnets in all metamorphic zones (Stipska & Schulmann 1995). Based on garnet core compositions for both the Upper and Lower Moravian nappes, Hock (1995) suggests conditions of 400-500°C and 4-5.5 kbar for this stage. This stage is connected with continuous underthrusting of the Brunovistuljan block beneath the hot Moldanubian block (Stipska & Schulmann 1995). Continental underthrusting leads to complete equilibration of the continental geotherm after underthrusting, and to the restoration of the subhorizontal and parallel position of the isotherms (isograds) after a certain period of thermal relaxation. We suggest that this stage is connected with the maximum attained P-T conditions in the deepest part of the underthrust plate. At this time, the isograds formed an acute angle with the major thrust boundary (Fig. 8a). The second stage is associated with the blocking of underthrusting as a result of buoyancy forces and shear stress at the thrust surface leading to a break up of the deepest part of the Brunovistulian plate. In this way, the first
204
P. STIPSKA ET AL.
foliation planes follow the shape of the western part of the Thaya window (Fig. 2a). The synmetamorphic kinematic indicators show top-tothe-northeast thrust movements in the S-dipping dip-slip domain (near Horn), dextral movements in the western strike-slip domain (Fritz 1991) and down-slip movements in the northern termination of the Thaya window (Schulmann et al. 1994) (Fig. 2b). This suggests that the early kinematics have been refolded by a large-scale antiformal fold with an axis gently plunging to the west. The model of progressive folding of the Variscan nappes and associated kinematics is shown in Fig. 9. Earlier in the paper we described how inverted metamorphic zonation occurs within individual tectonic units, i.e. the Upper Moravian and Lower Moravian nappes. The kinematic analysis
Fig. 8. Schematic diagram illustrating two stages of the mechanical inversion of metamorphic isograds by transportation of metamorphic zones within nappes and their heterogeneous simple shear deformation (modified from Stipska & Schulmann 1995).
high-grade nappes (the Micaschist zone and the Upper Moravian nappe) formed and were transported upwards together with the metamorphic zones. The rest of the Brunovistulian basement was being continuously underthrust. The Lower Moravian nappe originated in an identical manner and was thrust together with attached Upper Moravian nappe over the para-autochthon and autochthonous Thaya granite. This process caused an apparent parallelism of the metamorphic zonation with the thrust boundaries. This parallelism was further enhanced by the heterogeneous simple shear deformation of fossil isograds (Fig. 8b) transported within the nappe. Refolding of nappes and isograds The evolution of inverted metamorphic zonation discussed above implies that the oblique angular relationship between metamorphic zones and lithotectonic boundaries that appears in the present erosional section must have been produced by a later post-peak metamorphic effect. We suggest a model involving the folding of the Moravian nappes that is supported by the following structural and metamorphic evidence. The stretching lineations exhibit a constant southwest-northeast direction, whereas the
Fig. 9. Diagrammatic schematic evolution of the refolding of the Variscan nappes and their kinematic indicators around a gentle, W-plunging axis. Arrows mark syn-metamorphic kinematic indicators.
LARGE-SCALE FOLDING DURING PLATE COLLISION
205
shows that the nappes were subsequently folded into a large-scale antiform with an axis plunging gently to the west. Figure lOa shows individual nappes with their inverted metamorphic zonation. If this nappe sequence is folded and subsequently cut by a horizontal section, it can be seen that the high-grade rocks would occur in the hinge zone and lower-grade rocks in the limb regions (Fig. lOb and c). This is because the metamorphic zones cut the individual lithotectonic boundaries obliquely. The same geometry is developed in both the Upper Moravian and the Lower Moravian nappes. The more or less continuous trends of the isograds crossing the entire window (Fig. 6a) can be explained by small differences in metamorphic grade between individual nappes and by the small thickness of the nappes. When both folded nappes are placed together, then the resulting trend of metamorphic isograds is achieved (Figs 6a and lOc). From the observations presented above, we can conclude that the structure of the Thaya window can be viewed as a large-scale antiform plunging gently to the west with a wavelength of 40-45 km.
Thermal and rheological considerations
Fig. 10. Schematic diagram showing the refolding of the metamorphic isograds. (a) Inverted metamorphism in the individual nappes before folding.( b) Folding of the individual nappes with previously developed isograds. (c) The effect of isograds 'cross-cutting' tectonic boundaries in a horizontal section through this fold. Note lower metamorphism of the Lower Moravian nappe in the erosional section.
The basic assumption for the model presented above is that the high-grade nappes were transported upwards at a velocity sufficient to retain the temperature necessary for their viscous behaviour. England & Thompson (1984) have shown that the thermal budget of exhumed rocks depends on the rate of their vertical elevation. To estimate the rate of elevation of rocks within individual nappes, a numerical onedimensional thermal model was used. This enabled the simulation of the P—T evolution of the Moravian nappes during their exhumation. The main parameters of the thermal numerical model are given in Table 1. The P-T conditions likely to occur in the subducted crust, after the underthrusting of the Brunovistulian slab below the Moldanubian domain, can be approximated using a model of thickening of the crust by thrusting proposed by England & Thompson (1984). The thermal evolution in the Upper and Lower Moravian nappes were modelled separately and their vertical elevation corresponds to an erosion rate of 1.5 mm year"1. The model representing the thermal evolution of the Upper Moravian nappe was assumed to be instantaneously buried from a depth of 1km to a depth of 36km, and the model representing the metamorphic evolution
206
P. STIPSKA ET AL.
Table 1. Parameters for the thermal numerical model presented in Fig. 1 la Thermal diffusivity Thermal conductivity Heat flow at MOHO Radioactive heat Depth of heat production Time lag
0.9 x K^mV 1 2.25WmK~ 1 40mWrrr : 2UW1TT-1 15km 25 Ma
of the Lower Moravian nappe was assumed to be instantaneously buried to a depth of 29km. The effect of continuous underthrusting was simulated by assuming a time lag of 25 Ma for thermal relaxation. According to this simple thermal model a temperature of 630°C is achieved at a depth of 36 km and the temperature of 580CC at a depth of 29 km, which corresponds approximately to the peak metamorphic conditions determined using geothermobarometry in the Upper and Lower Moravian nappes, respectively (Fig. 11 a). The model predicts that after 6-7 Ma, the Upper Moravian nappe would have been exhumed to a depth of 28-25 km and that the temperature would have increased to 650°C (conditions of re-equilibration of the Upper Moravian nappe). At this depth (29 km, 580°C) the Lower Moravian nappe becomes detached and starts to be exhumed with the Upper Moravian nappe. After the next 8 Ma of joint exhumation the nappes rise to a depth of 13-15 km and the temperature decreases in both nappes to approximately the same temperature of 530-550°C (Fig. lla). In order to assess the mechanical behaviour of crustal nappes during the exhumation, it is
important to know their rheological properties. Taking into account the lithological composition of Moravian nappes, the rheological properties of the Bites and Weitersfeld orthogneisses have been assumed to be similar to those of albitic rocks, and those of the surrounding metapelites assumed to be similar to the rheology of wet quartzitic rocks. The quartzite rheology was selected for the metapelites because the rheological data for the rheology for micaceous rocks are lacking and because quartz is significantly weaker than albite at all temperatures. The variations in creep strength, a, with temperature and depth we have calculated for both albite and quartz (Ranalli & Murphy 1987) using equation (1) for a strain rate of ^HT'V 1 : where A is the pre-exponential function, n the creep exponent, Q the activation enthalpy of creep and R the universal gas constant. The calculated shear stress values and viscosities for albite and quartzite for temperature range from 500 to 650CC are presented in Table 2. Calculated viscosity contrasts between albite and quartzite in individual nappes for temperatures taken from a thermal model (Fig. lla) are shown in Table 3. An important feature is the decrease in viscosity contrast with decreasing temperature and increasing absolute values of shear stress for both the albitic rock and quartzite (Fig. lib). The calculated shear stress values for albitic rock and quartzite from Table 2 were used to constrain the rheological evolution of the Upper and Lower Moravian nappes
Fig. 11. Results of thermal modelling using a one-dimensional computer program showing P-Tand rheological evolution of the Upper and Lower Moravian nappes during their exhumation after 25 Ma of thermal relaxation, (a) Model of P-T-t paths of the Lower Moravian nappe which starts after 7 Ma of exhumation of the Upper Moravian nappe. Both nappes reach the same thermal conditions after 8 Ma of joint exhumation, (b) Shear stress calculated for albite-rich rock (orthogneiss) and for quartzite (metapelites) using T values from the thermal model.
LARGE-SCALE FOLDING DURING PLATE COLLISION
207
Table 2. Calculated shear stress values and viscosities for albitic (orthogneiss) and quartzitic (metapelite) rock over the temperature range of 500-650°C
T (C)
^ab
^qtz
A^ab
12777 9630 7337 5648 4390 3443 2725 2174 1748 1416 1155
1316
6.39 4.81 3.67 2.82 2.19 1.72 1.36 1.09 8.74 7.08 5.77 4.74 3.92 3.26 2.72 2.29
(MPa)
(MPa)
500 510 520 530 540 550 560 570 580 590 600 610 620 630 640 650
969 721 543 414 318 247 193 152 121 97 79 64 52 43 36
948 784 651 545 458
/^qtz
(Pa s)
(Pas) x x x x x x x x x x x x x x x x
1023 1023 1023 1023 1023 1023 1023 1023 1022 1022 1022 1022 1022 1022 1022 10"
6.58 4.84 3.61 2.72 2.07 1.59 1.23 9.66 7.62 6.07 4.87 3.93 3.20 2.62 2.16 1.79
x x x x x x x x x x x x x x x x
1022 1022 1022 1022 1022 10" 1022 1021 1021 1021 1021 1021 1021 1021 1021 1021
Ab, albite; qtz, quartzite. Rheological parameters are taken from the experiments of Jaoul et al. (1984) for hydrous quartzite: n = 2.4, 2 = 1 6 3 k J m ~ 1 , A = \Q~5 MPa~" s~l and of Shelton & Tullis (1981) for albitic plagioclase: n = 3.9, e = 234kJm~ ! ; A = 2.51 x 10~6 MPa"" s'1.
during their exhumation (Fig. lib). The ternperature values used for calculation of shear stress at different depths for each nappe were taken from the P-T model (Fig. 1 la). Our thermal modelling shows that the nappes during vertical elevation are transporting heat, and that the temperature during the early stages of exhumation increases. The temperature evolution is coupled with rheological behaviour of the crustal nappes during their exhumation. Consequently, the shear stress values decrease with increasing temperature in both nappes (Fig. 1 Ib),
which is followed by hardening with decreasing temperature at supracrustal levels. The Upper Moravian nappe, which was hotter than the Lower Moravian nappe, was significantly weaker for both albite and quartzitic lithologies. When the temperature converges (at 520-540°C) in both nappes at depths approximately corresponding to 15-16.5 km, the shear stress attains the same values for quartzite (414 MPa) and albitic rock (4390MPa) in both units (Table 2, Fig. lib), Our thermal models show that rocks in the individual nappes were transported upwards to
Table 3. Calculated viscosity ratios for albitic (orthogneiss) and quartzitic (metapelite) rock in the Upper and Lower Moravian nappes (UMU, LMU) during their exhumation Depth (km)
T CQ UMN
T (°C) LMN
AW^qtz
29.5
655 660 660 655 650 630 620 600 570 540
580 585 590 630 590 590 585 580 560 540
12.9 13.0 13.0 12.9 12.8 12.4 12.3 11.9 11.3 10.6
27
25.5 24.0 22.5
21
19.5
18
16.5
15
UMN
/W^qtz
LMN
2.8 2.8 3.2 3.5 3.8 5.4 5.8 7.6 8.8
10.6
W-, (km) for // a b (UMN)/// q t z UMN
/WUMN)//VLMN
W{ (km) for
40.5 40.6 40.6 40.5 40.4 40.0 39.8 39.4 38.7 38.0
24.2 24.5 25.4 26.2 26.9 29.0 31.0 33.9 35.7 38.0
Ab, albite; qtz, quartzite. Temperatures and depths are taken from the numerical model presented in Fig. 1 la. The theoretical wavelength for the Bites orthogneiss embedded in UMN metapelites (/i a b UMN/// q t z UMN) and LMN metapelites (/^ a b UMN/// q t z LMN) has also been calculated (see Fig. 12b).
208
P. STIPSKA ET AL.
Fig. 12. Results of calculations of fold wavelength (using the Biot formula) for viscosities of albitic rock and quartzite and for different thickness of competent layers, (a) The wavelength is calculated for the Bites and Weitersfeld orthogneiss at temperatures ranging from 500 to 750 C. (b) Wavelength of the Bites orthogneiss layer calculated for viscosity contrast between the Bites orthogneiss and the overlying and underlying pelites. The viscosities of albite rock and quartzite were calculated using T and stress values from Fig. 11.
supracrustal levels rapidly enough to avoid significant temperature loss (Fig. 11 a). Therefore, they could deform or fold in a ductile manner when they reached a relatively rigid and cold obstacle such as the autochthonous Thaya granite. It is important to point out that at a certain depth, which is controlled by the rate of exhumation, the rocks making up the nappes attain the same rheological properties (Fig. lib). Crustal multilayer The development of folds in a multilayer made up of alternating competent and incompetent layers is controlled by the ratio of the thickness of the incompetent and competent layers (n = d^/d\) and their viscosity contrast //!///•> (Ramberg 1962, 1964). The Upper Moravian nappe is marked by the presence of an orthogneiss body 4-5 km thick and a 2-3 km thick micaschist sequence in the hangingwall. The Lower Moravian nappe contains significantly less orthogneiss than micaschist, the former is about 2 km thick and the latter 3 km thick. The metasediments of the para-autochthonous unit have a thickness of 2 km and contain relatively small amounts of orthogneiss. Thus, the proportion of incompetent to competent layer thickness for the Moravian nappes approaches a value slightly higher than n = 1. This means that the multilayer system will fold harmonically, i.e. will behave mechanically as a multilayer where all layers develop the same wavelength rather than a series of single layers in which each competent layer develops its own wavelength
according to the single-layer buckling equation (see equation 2) (see Price & Cosgrove 1990). The folding of simplest multilayer system requires similar rheological properties of competent layers embedded in soft matrix with unique viscosity. The wavelength W\ of this multilayer system is controlled by the thickest competent layer (Ramberg 1970), and is calculated using the Biot (1961) and Ramberg (1964) formula: where Wx is the initial fold wavelength, d is the thickness of competent layer and \JL\J\JL^ is the viscosity contrast. However, the requirement of similar viscosity of competent layers is not fulfilled for the large part of exhumation history of the Moravian nappes. Figure l i b shows that the thicker Bites orthogneiss is Theologically weaker than the underlying Weitersfeld orthogneiss for depths deeper than 16.5km. If the folding is driven by stronger Weitersfeld orthogneiss, then the wavelength of the fold would reach 18-20 km only (Fig. 12a) which is not consistent with the observed wavelength of 40-45 km of the Thaya antiform (Fig. 2). However, when the fold wavelength is calculated for the Bites orthogneiss for the temperature range 500-650°C and viscosity contrast // a b/M q tz m the range 10-13 (Table 3), then the W, values vary between 37 and 40km (Fig. 12a, Table 3), i.e. the approximate wavelength exhibited by the Thaya antiform. We therefore suggest that the folding was driven by this thick competent layer. This is possible only when the Bites orthogneiss is of equal viscosity or is rheologically stronger than the underlying Weitersfeld orthogneiss, i.e. at depths of less than 16.5km.
LARGE-SCALE FOLDING DURING PLATE COLLISION
209
thrusting and enhanced the buckling of the multilayer system. The style of folding of the Moravian nappes multilayer sequence is controlled by relatively low ductility contrast (^ a ibite//Vartz) anc* moderate ratio of incompetent to competent layer thickness. This relatively close spacing of the Bites and Weitersfeld orthogneiss ensures that a harmonic fold assemblage develops and that the zone of contact strain (Ramberg 1962) of the Bites orthogneiss is sufficiently large to control the folding of underlying sequences. The relatively low viscosity contrast may have been responsible for important modification of fold geometry by homogeneous flattening which would lead to a thickening of both the competent and incompetent layers in the hinge zone and thinning in the fold limbs. The flattening of this megafold is well documented by its similar fold geometry and by the finite strain studies of Schulmann (1990) and Fritz (1991), which show oblate strain ellipsoids in the limb regions and plane strain fabrics in the fold hinge. The folding of crustal nappes along an inclined surface is responsible for the rapid elevation of fold hinge into supracrustal levels in response Discussion to horizontal shortening (Fig. 2b). The overlying The complex geometry of the inverted meta- thick nappe of Moldanubian migmatites cannot morphic isograds and their oblique relationship be folded, and is therefore affected by extensional with respect to the main lithotectonic boundaries faulting where it lies above the fold limbs (Figs 2b have been discussed, and detailed petrological and 3d). These huge extensional to transtensional data and structural observations used to argue faults place high-grade Moldanubian rocks in that the complex pattern of metamorphic iso- direct contact with low-grade Moravian rocks grads is in part the result of large-scale buckling (Schulmann et al. 1994). Extensional faulting of the metamorphic nappe sequence. Pressure- also affects the Moravian units and results in a temperature data in combination with thermal reduction of their thickness in the northern and modelling have been used to estimate the rheolo- southern parts of the window (Figs 2b, 3b and gical evolution of the nappe sequence during c and 5). The extensional faulting is accompanied exhumation. by the development of medium-scale gravitaIt is suggest that the buckling of the Moravian tional folds in the highly anisotropic parts of nappe pile started at a depth of 15km during Moravian nappes (Schulmann et al. 1994) exhumation and that its wavelength was which do not affect the whole crustal multilayer controlled by the Bites orthogneiss the thickest (Figs 3a and 4). competent layer. At this depth the viscosities of Ramberg (1968, 1970) demonstrated that the Bites and the Weitersfeld orthogneiss gravity folding results in significantly greater become equal and the viscosity of medium fold wavelengths than buckling for horizontal embedding the Bites orthogneiss (the underlying layers thicker than 1 km. However, in the case and overlying metasediments) becomes homo- of the Moravian nappes, the folding starts geneous. The calculated fold wavelength of along a relatively steeply inclined surface 38km is in good agreement with the observed (Figs 6b and 9). Therefore, it is argued that wavelength of the Thaya antiform, 40-45 km. gravity did not play a major role in the formation The estimated depth of 15km, at which folding of large-scale buckle folds. Instead, we suggest was initiated, roughly coincides with the depth that in this example gravity influenced the sliding estimate of 18km (5-6kbar, Kolafikova et al of nappes along extensional faults which devel1997) for the burial of the autochthonous oped on the limbs of the Thaya antiform. We Thaya granite. It is argued that, at this depth, interpret the extensional structures as secondary the autochthonous Thaya granite acted as a products of large-scale buckling of a nappe rigid buttress which inhibited further nappe sequence. Another requirement of this buckling model is the unique viscosity of the weak medium embedding orthogneiss bodies. This requirement is not fulfilled at depths exceeding 16km where the temperatures of both nappes become equilibrated. During joint exhumation of the Upper and Lower Moravian nappes, the viscosity ratio between the Bites orthogneiss and the overlying metasediments (/^ ab UMN//^ qtz UMN) is different from the viscosity ratio between the Bites orthogneiss and the underlying metasediments of the Lower Moravian nappe (/^ ab UMN//u qtz LMN) (Table 3). We have calculated the fold wavelength for the Bites orthogneiss with respect to overlying (UMN) and underlying (LMN) metasediment viscosities, respectively (Table 3, Fig. 12b). The fold wavelength of the Bites orthogneiss if embedded in a medium of the underlying metasediments is 25km at a depth of 29.5km and steadily increases to W{ = 38km at a depth of 15km. At this depth, the temperature, as well as the viscosity of both the underlying and overlying metapelites, converge.
210
P. STIPSKA ET AL.
Financial support of Czech National Grant Agency (No. 205 96 0277) and Ministry of Education (No. 2431 3005) are gratefully acknowledged.
References BERNROIDER. M. 1989. Zur Petrogenese prekambrischer Metasedimente und cadomischer Magmatite im Moravikum. Jahrbuch des Geologischen Bundesanstalt. 132, 349-373. BIOT. M. A. 1961. Theory of folding of stratified, viscoelastic media and its applications in tectonics and orogenesis. Geological Society of America Bulletin. 12, 1595-1632. BRUNEL. M. & KIENAST, J. R. 1986. Etude petro-structurale des chevauchements ductiles himalayens sur la transversale de TEverest-Makalu (Nepal oriental). Canadian Journal of Earth Sciences, 23, 1117-1137. BURG. J. P.. DELOR. C. P.. LEYRELOUP. A. F. & ROMNEY. F. 1989. Inverted metamorphic zonation and Variscan thrust tectonics in the Rouergue area (Massif Central, France): P-T-t record from mineral to regional scale. In: CLIFF, R. A., DALY. J. S. & YARDLEY. B. W. D. (eds) Evolution of Metamorphic Belts. Geological Society. London. Special Publications, 43. 423-439. DALLMEYER. R. D.. NEUBAUER. F. & HOCK, V. 1992. Chronology of late Paleozoic tectonothermal activity in the southeastern Bohemian massif, Austria (Moldanubian and Moravo-Silesian zones) - 40Ar 39Ar mineral age controls. Tectonophysics. 210. 135-153. DUDEK. A. 1980. The crystalline basement block of the Outer Carpathians in Moravia. Rozpravy Ceskoslovenske akademie red, 90. 1-85. ENGLAND. P. C. & THOMPSON. A. B. 1984. Pressuretemperature-time paths of regional metamorphism. part I: heat transfer during the evolution of regions of thickened continental crust. Journal of Petrology, 25. 894-928. FRASL. G. 1970. Zur Metamorphose und Abgrenzung der Moravischen Zone im niederoesterreichischen Waldviertel. Nachrichtungen Deutche Geologische Gesselschaft, 2, 55-60 FRITZ. H. 1991. Structures and kinematics along the Moravian-Moldanubian boundary: preliminary results. Osterreichische Beitrage ~u Meteorologie und Geophysik, 3. 77-96. & LOITZENBAUER, J. 1994. Fluid activity during late stage of Variscan deformation in the Moravian nappe complex. Mitteilungen der Osterreichischen Mineralogischen Gesselschaft, 139, 47-49. & NEUBAUER, F. 1993. Kinematics of crustal stacking and dispersion in the south-eastern Bohemian Massif. Geologische Rundschau, 82, 556-565. HUBBARD. M.S. 1989. Thermobarometric constraints on thermal history of the Main Central Thrust Zone and Tibetan Slab, eastern Nepal Himalaya. Journal of Metamorphic Geology, 7, 19-30. HOCK. V. 1975. Mineralzonen in Metapeliten und Metapsammiten der Moravischen Zone in
Niederoesterreich. Mitteilungen der Geologischen Gesselschaft, 66-67. 49-60. 1995. Moravian zone: Metamorphic evolution. In: DALLMEYER, R. D., FRANKE. W. & WEBER. K. (eds) Pre-Permian Geology of Central and Eastern Europe, Springer-Verlag, Berlin. 541-553. . MARSCHALLINGER, R. & TOPA. D. 1990. GranatBiotit-Geothermometrie in Metapeliten der Moravischen Zone in Osterreich. Osterreichische Beitrage -u Meteorologie und Geophvsik, 3. 149167. JAOUL. O.. TULLIS. J. & KRONENBERG A. 1984. The effect of varying water content on the behaviour of Heavitree quartzite. Journal of Geophysical Research, 89, 4298-4312. KOLARIKOVA. A.. MARQUER. D. & SCHULMANN, K. 1997. Evolution of mass-transfer during progressive oblique under-thrusting of the Variscan foreland: eastern Bohemian Massif. Geodinamica Acta, 10.81-93. LOBKOWICZ. M., SCHULMANN. K. & MELKA. R. 1998.
Variscan deformation, microstructural zonation and extensional exhumation of the Moravian Cadomian basement. Geodinamica Ada, 11(2 3). 119-137. MASON. R. 1984. Inverted isogrades at Sulitljema. Norway: the result of shear zone deformation. Journal of Metamorphic Geology. 2, 77-82. MOHAN. A.. WINDLEY. B. F. & SEARLE. M. P. 1989. Geothermobarometry and development of inverted metamorphism in the Darjeeling-Sikkim region of the eastern Himalaya. Journal of Metamorphic Geology, 7, 95-110. PRECLIK. K. 1927. Zur Tektonik und Metamorphose der moravischen Aufwolbung am Ostrand der Bohmischen Masse. Geologische Rundschau. 18. 81-103. PRICE. N. J. & COSGROVE. J. W. 1990. Analysis of Geological Structures. Cambridge University Press. Cambridge. RAMBERG, H. 1962. Contact strain and folding instability of a multilayered body under compression. Geologische Rundschau, 51. 405-439. 1964. Selective buckling of composite layers with contrasted rheological properties, a theory for simultaneous formation of several orders of folds. Tectonophysics, 4, 307-341. 1968. Instability of layered systems in the field of gravity. I, II. Physics of the Earth and Planetary Interiors, 1. 427-447. 1970. Folding of compressed multilayers in the field of gravity. I. II. Physics of the Earth and Planetary Interiors. 2. 203-232. RANALLI. G. & MURPHY. D. C. 1987. Rheological stratification of the lithosphere. Tectonophysics. 132,281-295. SCHARBERT, S. & BATIK. P. 1980. The Age of the Thaya (Dyje) Pluton. I'erhandlungen des Geologischen Bundesanstalt. 3. 325 331. SCHULMANN, K. 1990. Fabric and kinematic study of the Bites orthogneiss (southwestern Moravia): Result of large-scale northeastward shearing parallel to the Moldanubian Moravian boundary. Tectonophysics. 111. 229-244.
LARGE-SCALE FOLDING DURING PLATE COLLISION , MELKA, R., LOBKOWICZ, M., LEDRU, P., LARDEAUX, J. M. & AUTRAN, A. 1994. Contrasting styles of deformation during progressive nappe stacking at the southeastern margin of the Bohemian Massif (Thaya Dome). Journal of Structural Geology, 16, 355-370. SHELTON, G. & TULLIS, J. 1981. Experimental flow laws for crustal rocks. Eos, Transactions of America! Geophysical Union, 62, 396. SxipSKA, P. & SCHULMANN, K. 1995. Inverted metamorphic zonation in a basement-derived nappe sequence, eastern margin of the Bohemian Massif. Geological Journal, 30, 385-413. SUESS, F. E. 1912. Die moravischen Fenster und ihre Beziehung zum Grundgebirge des Hohen
211
Gesenke. Akademie der Wissenschaften, Denkschrift Matematisch-Natunvissenschaftliche Klasse, 83,541-631. 1926. Intrusionstektonik und Wandertektonik im variszichen Grundgebirge. Bornttrager, Berlin. TRELOAR, P. J., BROUGHTON, R. D., WILLIAMS, M. P., COWARD, M. P. & WINDLEY, B. F. 1989. Deformation, metamorphism and imbrication of the Indian plate, south of the Main Mantle Thrust, north Pakistan. Journal of Metamorphic Geolog\\ 7, 111-125. WALDMANN, L. 1930. Zum geologischen Bau der Thayakuppel und ihre Metamorphose. Mitteilungen der Geologischen Gesselschaft, 21, 133-152.
This page intentionally left blank
Predicting patterns of strain from three-dimensional fold geometries: neutral surface folds and forced folds RICHARD J. LISLE Laboratory for Strain Analysis, Department of Earth Sciences, University of Wales, Cardiff CF1 3 YE, UK Abstract: The geometries and densities of fractures associated with fold structures can be predicted by assuming that the strains accommodated by fractures mimic the bulk strains induced in the strata during folding. This paper examines, from a theoretical standpoint, the distributions of bedding-plane strains expected in folds formed by various folding mechanisms. The relationship between the state of bedding-plane strain and fold-surface geometry is found to vary according to different fold types, distinguished on the basis of their curvature properties. The first type are developable fold surfaces, which have Gaussian curvature equal to zero. Folding mechanisms which are dominated by the mechanical strength of the layering, such as buckling, produce surfaces of this type. Folds of this type allow the possibility of estimating the bedding-plane strains from the geometrical features of the folded layer. Neutral surface folds and flexural-slip folds are discussed as examples. The other main class of folds have non-developable surfaces, which have nonzero Gaussian curvature. Folded surfaces with this form arise predominantly from mechanisms that involve the passive deflection of the layering in response to displacement gradients originating outside of the layer, e.g. drape folding. Although the geometry of these surfaces implies the presence of bedding plane strains, the quantification of these strains cannot be made from the fold geometry but requires additional information on these displacement patterns.
Rock strains accommodated by brittle fracturing are known to affect profoundly the bulk physical properties of the rock volume. In the case of some hydrocarbon reservoirs, extensional strain expressed by sets of open fractures can have a dominant influence on the potential for the storage and transport of fluids. In consequence, the identification of zones of increased fracturing is of great value to the modelling of reservoir properties. As such strains can be recorded directly only in a small proportion of a reservoir's volume (i.e. in boreholes), it becomes necessary to devise criteria for the prediction of the distribution of strains in the reservoir based on indirect evidence. An example of this is to attempt to establish a relationship between the frequency of occurrence of fractures of different sizes, in order to allow the prediction of the number of fractures from the number of fractures large enough to be mapped seismically (Heffer & Bevan 1990; Yielding et ai 1992). Another approach, adopted in this paper, is to explore relationships that exist between strain and the geometrical attributes of folds for the purpose of allowing the prediction of fracture densities and orientations from the geometry of folded horizons within the reservoir. A number of field studies have demonstrated a link between fold geometry and fracturing, in spite of the observation that joints also occur in
unfolded strata. In many cases the orientation of joints have been related to the directions of geometrical features of folds in the same areas (e.g. Harrison & Moench 1961; Badgley 1965, pp. 104-105; Norris 1967; Stearns 1968; Engelder & Oertel 1985). Cases have also been documented where the density of fractures correlates with the curvature of the folded beds (e.g. Harris et al. 1960; Murray 1968; Gorham et al 1979; Woodward 1984; Padgett & Nesler 1991). This paper examines theoretically the nature of the two-dimensional strains induced on the bedding planes during folding. Because no unique relationship exists between fold shape and the state of strains within folded beds (see Ramsay 1967, p. 344; Hobbs 1971), relationships have to be established separately for a number of specific models of folding. Previous analyses of the strain distributions which characterize different models (Ramsay 1967 pp. 391-447; Hobbs 1971) are mostly two-dimensional and are, therefore, of restricted relevance to the general, non-cylindrical structures typically found in hydrocarbon reservoirs. For this reason, emphasis is placed on three-dimensional models of folding in this paper. The consideration of strains associated with various folding models is restricted to the two-dimensional strains in the surface of the bedding. The present or absence of such strains will be of crucial
From: COSGROVE, J. W. & AMEEN, M. S. (eds). Forced Folds and Fractures. Geological Society, London, Special Publications, 169, 213-221. 1-86239-060-6/OO/S 15.00 © The Geological Society of London 2000.
214
R. J. LISLE
importance for the development of fractures which enhance permeabilities in the direction normal to the stratification. Three-dimensional flexural-slip folding model From the point of view of bedding-plane strain, this can be considered the simplest model, as a fundamental feature of this folding mechanism is that there is no distortion in the surface being folded. Fold growth requires interbed slip and, therefore, flexural slip is an important mechanism in systems consisting of mechanically strong layers separated by beds or interfaces of easy slip, such as thinly bedded sandstoneshale sequences. The systems prone to deformation in this mode are anisotropic, which exhibit a greater resistance to layer-parallel stretching than to layer-parallel shear. The model can also only operate if the slip surfaces show a spacing which is small compared to the radius of curvature of the beds. If the mechanically active bedding planes are widely spaced, the mode of deformation will deviate from flexural slip because buckling effects will produce beddingplane distortion. From a geometrical point of view, the folded surfaces resulting from flexural slip are developable surfaces (Lisle 1992). These are defined as ruled surfaces in which adjacent straight rulings intersect each other (Fig. 1A), if extended far enough. As the surfaces are folded isometrically (without the development of bedding-plane strains) the final geometry of these surfaces is constrained by Gauss's theorem which states: Tsometric deformation does not change the Gaussian curvature at points on the surface'. The Gaussian curvature, K, is a measure of the degree of double curvature present at a given point on a surface (Fig. 1) and is calculated by the product of the principal curvatures, k}, k< A flat bedding surface (with K = 0 at all points) will, therefore, deform by flexural slip to give a folded surface consisting of points with K — 0. It has been suggested that analysis of Gaussian curvature in real folds provides a way of assessing the validity of the isometric folding mechanism and of highlighting those parts of a horizon which are likely to have undergone bed stretching or contraction (Lisle 1992, 1994; Stewart & Podolski 1998). On a Gaussian curvature map, such as the example in Fig. 2a, portions of the structure where K is non-zero are interpreted as regions where the folded
Fig. 1. Gaussian curvature K at a given point on a surface is the product of the principal curvatures k\ and k2. (a) Developable surface consisting of points with K = 0. (b) Non-developable surface, with nonzero Gaussian curvatures. Such a surface cannot be developed from a flat sheet without straining of the surface.
horizon has suffered bedding-plane strains. Support for this interpretation comes from the observation that fracture densities increase where there is a change in the direction of fold hinges (Muehlberger 1961; Narr 1991). Figure 3 illustrates this with a field example and shows a concentration of fractures on a folded surface at points of high Gaussian curvature. An analysis of Gaussian curvature of natural folds allows the testing of the viability of the flexural-slip mechanism for each patch of the folded surface. Under the structural conditions mentioned above, where flexural slip is a likely folding mechanism, this method can serve to identify anomalous zones within a structure where the mechanism has not operated, i.e. where folding is non-isometric. The method is useful in the overall context of the flexural-slip mechanism but not as a general method of estimating the magnitudes of the strains. This is because there exists no general relationship between the Gaussian curvature and strain. For
STRAIN PATTERNS FROM 3D FOLD GEOMETRY
215
Fig. 2. (a) Example of a Gaussian curvature map. A folded surface, represented by contours, has been analysed using the method of Lisle & Robinson (1995). Dark shading indicates parts of the folded structure with non-zero Gaussian curvature corresponding to regions of implied stretching of the folded surface. Gaussian curvature analysis (Lisle 1992, 1994) uses this as a basis of predicting regions of greater fracture densities, (b) Curvature trajectory map. Dashes show one of the principal curvature directions, k\ or k2, depending on which has the least absolute magnitude; the equivalents of the fold axis for a cylindrical fold. If neutral surface folding is assumed, these directions could be used for predicting fracture directions.
example, all cylindrical folds consist of points where one of the principal curvature values is always zero. The zero K values associated with such a structure only imply that such a fold shape is consistent with the flexural-slip mechanism. It should not be taken to mean that the bed strains are necessarily zero.
In spite of the difficulties posed by the lack of a universal link between curvature of a surface and its distortion, it may still be possible to establish the relationship between curvature and strain for specific folding mechanisms. This is explored in the sections below. Interfering kink folds
Fig. 3. Curvature-related fracturing in a small fold (Forth Dafarch, Holy Island, Anglesey, Wales SH233800). The fracturing is most intense in regions where the fold hinge line is itself curved; the regions where the Gaussian curvature is significantly nonzero.
This is a modification of the flexural-slip model to permit the formation of non-developable fold shapes. This simple model considers the geometry of the surface produced by the superimposition of two kink folds. Individually the kink folds are cylindrical, conforming to the flexural-slip (no distortion in the bedding plane) mechanism. Whether the two folds develop synchronously or sequentially, a non-developable geometry is produced by their superimposition (Fig. 4). According to Gauss's theorem the folded surface with its dome-and-basin form must have involved some distortion of the surface during folding. In the present model this distortion is restricted to a lozenge-shaped patch at the intersection of the two kinks (Fig. 4). The stain history of this patch can be calculated from the fact that edges PQ and PR are
216
R. J. LISLE
Fig. 5. Relationship between Gaussian curvature and strain in the intersecting kink-fold model.
(1986) shows that this discrepancy from 360 , the so-called defect angle, —, is a direct measure of the Gaussian curvature, K, at a given point like P, namely
Fig. 4. Simple kink-fold model used to explore the relationship between Gaussian curvature and straining of the surface.
material lines which remain lines of no finite elongation at every stage of deformation. The principal axes of the strain ellipse, therefore, always coincide with material lines that bisect the angle between the lines PQ and PR. The strain history is, therefore, coaxial. During folding the original angle 0 between edges PQ and PR is reduced by an amount 6. It can readily be shown that for orthogonal kinks (0 = 90°) the angle 6 is given by: where a and (3 are the angles of deflection of the two kink folds (Fig. 4). The amount of rotation (6/2) of the edges PQ and PR towards the maximum elongation direction allows the strain ratio (Rs) to be calculated, using Marker's equation (Ramsay 1967, equation 3-34), as
where 0 is the angle expressing the oblique superimposition of the two kinks (Fig. 4). The distortion of the lozenge-shaped patch means that the surface shape in the vicinity of point P is not developable (K is not zero), i.e. it cannot be simply unrolled to a flat surface. This follows from the fact that the angles of the faces meeting at P no longer add up to 360°. In fact, Calladine
where 6 is in radians and A equals the area of the surface associated with the point X (Lisle 1994). The relationship between strain and Gaussian curvature is obtained by combining equations (l)and(2).
This is displayed graphically in Fig. 5. As pointed out by Stewart (1993) there is also area change associated with the straining of the patch. This is an area decrease given by
where A, A1 are the original and new areas, respectively. Ramsay (1974) has suggested mechanical reasons in which cylindrical kink folds cease to grow once a certain degree of tightness is reached. For interfering kink folds the problem of accommodating this area change is likely to cause 'locking up' at an earlier stage of development than for non-interfering kink folds. The relationship is specific to this model but this simple model serves as an illustration of the fact that, within a structure of general developable structures, local non-zero Gaussian curvature will be indicative of local bed straining. Three-dimensional tangential longitudinal strain model The flexural-slip fold mechanism characterized by zero bedding-plane strain will cease to operate where potential slip horizons are widely spaced
STRAIN PATTERNS FROM 3D FOLD GEOMETRY
compared to the radius of curvature of the fold (Donath & Parker 1964). Thick beds will be subject to stretching of bedding planes on the fold outer arcs and shortening of bedding surfaces on the inner arc according to the tangential longitudinal strain (TLS) folding model (Ramsay 1967, pp. 397-403). If the shear strength of the slip surfaces is very low compared to that of the beds, the bedding surfaces will approximate to principal surfaces of stress and strain throughout the folding history. As a result, each of the three families of principal strain trajectories (X, Y, Z) will show an arrangement which is either close to parallel or approximately perpendicular to the bedding surfaces. The relationship between strain and curvature has been derived for a two-dimensional TLS model (Ramsay 1967) which can be applied to cylindrical folds. Tangential longitudinal strain folding is sometimes referred to as neutral surface folding. The first term refers to the orientation pattern of the strains, and the second to the assumed presence of a zero-strain horizon within the buckled layer, close to its middle plane. In the threedimensional TLS model these two attributes may not go together. According to Gauss's theorem, the presence of a neutral surface in a TLS fold is compatible only with developable fold geometry. Zhao et al. (1997) have applied elastic thin plate theory to calculate strains and stresses to double curvature folds. The predictions of the theory are that the principal strains within the folded layer are governed by the principal curvatures of the layer and by the distance from the middle plane of the layer. The theory they apply assumes, however, the presence of a neutral surface, a surface within the layer of zero strain. The theory is, therefore, inappropriate for folds which have large amplitudes (relative to layer thickness), unless the middle plane is folded into a developable surface, i.e. a surface with one principal curvature equal to zero (Timoshenko & Woinowsky-Krieger 1959, pp. 1-3). A step towards understanding the strains associated with non-cylindrical TLS folds is made possible by use of a geometrical property of orthogonal families of curved surfaces, encapsulated in Dupin's theorem (Hilbert & CohnVossen 1932), which states: 'Orthogonal families of surfaces intersect each other along principal curvature directions on those surfaces'. Therefore in situations where continuous XY, YZ and XZ principal surfaces of strain exist and one of these follows the bedding surface, the principal curvature directions on the latter must coincide with the intersection with the two families of surfaces which are perpendicular to bedding. The principal curvature directions at each point
217
Fig. 6. Dupin's theorem can be used to predict strain from curvature directions of folded surfaces (see text for details). Assuming a tangential longitudinal strain model, i.e. the principal surfaces of strain are arranged parallel and perpendicular to the bedding surfaces, the principal curvature directions in the bedding are principal strain directions.
on the folded bedding are therefore parallel to a pair of principal strain directions (Fig. 6). By using this device the approximate strain orientations for non-cylindrical folds with special geometries can be predicted, e.g. conical folds (Fig. 7a) and ellipsoidal domes (Fig. 7b).
Fig. 7. Tangential longitudinal strain model. Examples of non-cylindrical surfaces and principal curvature trajectories. If strains distribution within the layer accords with the tangential longitudinal strain model, the principal curvature directions coincide with the principal curvature directions in the bedding surfaces, (a) Conical surface; (b) ellipsoidal surface with umbilical point, points where both principal survatures are equal and where curvature trajectories are locally undefined.
218
R. J. LISLE
With some fold geometries this technique for determining the strain trajectories fails because the principal planes of strain do not necessarily form continuous surfaces (Mandl 1987; Treagus & Lisle 1997). For such folds it has not yet been proved that strain trajectories track the lines of curvature on the folded surface. Pencil box shear folding The simple-shear folding mechanism (Ramsay 1967, p. 425) has been invoked to explain geometrical features of folds formed in the hangingwalls of normal faults (Verrall 1981; Gibbs 1983). Of the various mechanisms proposed to explain such structures, the simple shear mechanism appears to predict most successfully the form of the fault surface from the observed geometry of the hangingwall anticline in natural examples (White & Yielding 1991). This folding mechanism is considered because it yields forced folds; folds in which the layering is distorted in a passive manner in response to externally imposed displacement patterns. The model considered above is used to explain features in cross-sections and is, therefore, twodimensional. Roll-over folds frequently depart from cylindrical geometry. For analysis of these a three-dimensional equivalent of the simpleshear mechanism would be more appropriate, similar to that considered by Kerr et al (1993). In three dimensions the deformation accords with 'pencil box shear' (Fig. 8), where differential displacement of the pencils parallel to the r
direction transforms points according to the coordinate transformation equations:
In the case where the shear direction (r) is vertical and orthogonal to the pre-deformation attitude of bedding, the heterogeneous deformation pattern can be analysed in terms of slender prismatic domains parallel to r within which homogeneous simple shear occurs. The plane of shear within each prism is defined by the local strike of the folded surface (Fig. 8b) and the common shear direction, r. The amount of shear in the prism and the stretching of the bedding both relate directly to the deformed angle of dip of bedding. The strain ratio /?s in the plane of the bedding is given by: In the case of inclined shear, where r is not perpendicular to bedding in the undeformed state, the relationship between strain state and geometry of the produced fold is less direct. Nevertheless, the state of strain in the bedding plane here again depends on the attitude of the bedding rather than on its curvature. In summary, the simple-shear model allows the estimation of bedding-plane strain in different parts of a non-cylindrical fold provided the shear direction is known. However, as proponents of this folding mechanism point out, it should be realized that simple-shear assumption is a simplification which provides a convenient description of the bulk deformation pattern associated with collapse of the hangingwall of faults. In this regard it would be unfair to test the model by comparing the predictions with actual fracture densities on individual beds. Discussion
Fig. 8. The pencil box shear model, (a) The displacement pattern, for the case where the shear direction r is vertical and orthogonal to the undeformed bedding, (b) For slender prism parallel to the shear durection. the deformation is simple shear with a shear plane parallel to i and the bedding strike.
The discussion above highlights the difficulties associated with the deduction of strains from fold geometry. The strain patterns expected differ according to the folding mechanism involved. This is consistent with the observation that strain-related structures (e.g. fracture distributions) associated with folds from different settings correlate with different geometrical attributes of those folds. For a given fold structure, the task of predicting the distribution of bedding-plane strain is made easier if the folding mechanism can be established. In general, the
STRAIN PATTERNS FROM 3D FOLD GEOMETRY
219
Table 1. Classification of three-dimensional fold geometries and the folding mechanisms which produce them Developable folds (K = 0)
Non-developable folds (0 < K > 0)
Neutral surface folds
Flattened neutral surface folds
Forced or passive folds
Mechanical influence of strong layering:
Neutral surface folds modified by homogeneous strains, e.g. flattened buckles No fixed relationship between curvature and strains
Geometry of fold governed largely by imposed displacement gradients, e.g. folds in incompetent layers, folds above diapirs, thin competent beds adjacent to thick competent beds Non-zero K implies lack of neutral surface but strains not closely related to curvature
(a) Thin beds in relation to radius of curvature, e.g. flexural-slip folding No strain on bedding planes Gaussian curvature analysis useful for detecting local strains (b) Thick beds in relation to radius of curvature Strains relate to principal curvature values k^, k2 and vary across the bed
identification of folding mechanism is based on sets of geometrical attributes (e.g. Ramsay 1967, pp. 372-456). It is useful in the present discussion to classify folding mechanisms according to the scheme which has neutral surface folding and forced folding as end-members (Table 1). This classification is based on similar criteria to that of Donath & Parker (1964) who distinguished flexural and passive folding, respectively. In neutral surface folding it is the mechanical strength of the layering which exerts controls on the geometry of the folds. These competent layers, which may equate with individual beds or packets of beds which function mechanically as a single unit, are folded into parallel-style folds. The resistance to stretching of this layer means that a neutral surface exists within the competent layers. The range of geometry of these folds is restricted by the fact that the neutral surface must be developable, i.e. the folded surfaces possess cylindrical, conical or other shapes characterized by zero Gaussian curvature. For these folds the detailed pattern of straining of the bedding planes is controlled by the ratio of layer thickness relative to the radius of curvature. If this ratio is sufficiently small, the bedding-plane strain is close to zero throughout the sequence and mechanism approaches that of flexural slip. For larger ratios, the strain on the bedding surface is not negligible. The strain is proportional to the greatest absolute principal curvature and distance from the neutral surface. An important aspect of the strain distribution is that its configuration changes across the neutral surface. Strain-induced fractures will, therefore, not extend across the mechanical layer (see Price & Cosgrove 1990) to the detriment of permeability pathways. Local
deviations from developable surface geometry, brought out by analysis of Gaussian curvature, indicate regions of departure from this folding mechanism and straining of the middle surface. The other end-member mechanism is forced folding (Table 1). Folds develop from deflections of the layer brought about by external displacement gradients. The layer responds in a passive manner to these displacements and the folded surfaces adopt geometries which are controlled by the pattern of imposed displacements. Examples are folds developed in incompetent layers adjacent to buckled competent layers, the folds induced by the displacement of the hangingwalls of faults and the forced folds developed above daipiric intrusions. The middle surface, the surface defined by points which lay in the centre of the layer before folding, is likely to have undergone stretching. For this reason this surface is not constrained to be developable. The folded surfaces will, therefore, have a greater range of geometries than for neutral surface folds. The strain magnitudes in the bedding surfaces will not generally correlate with the principal curvatures of those surfaces. The pattern of distribution of strain will be largely independent of the layering. Fracturing, and other structures expressing the rock strains, will not be strata-bound, but will transect the layering system. Natural folding mechanisms inevitably will fall somewhere in between these two end-members. For example, we may consider an elastic layer bulged by pressure exerted from one side (Withjack & Scheiner 1982). For small defections, the strain distribution is close to that expected for a neutral surface fold. For large deflections there will be a ballooning contribution which will
220
R. J. LISLE
eventually dominate the strain pattern relating to the bending of the sheet. The author is grateful to Shell Research, Rijswik. The Netherlands, for support of work on curvature properties of geological structures.
References BADGLEY, P. C. 1965. Structural and Tectonic Principles. Harper & Row, New York. CALLADINE, C. R. 1986. Gaussian curvature and shell structures. ///: GREGORY J. A. (ed.) The Mathematics of Surfaces. Clarendon Press, Oxford. DONATH, F. A. & PARKER. R. B. 1964. Folds and folding. Geological Societv of America Bulletin, 75. 45-62 ENGELDER, T. & OERTEL, G. 1985. The correlation between undercompaction and tectonic jointing within the Devonian Catskill Delta. Geologv, 13, 863-866. GIBBS, A. D. 1983. Balanced cross-section construction from seismic sections in areas of extensional tectonics. Journal of Structural Geology, 5, 153-160. GORHAM. F. D., WOODLAND, L. A., CALLENDER, J. F. & GREER, A. R. 1979. Fractures in Cretaceous rocks from selected areas of the San Juan Basin, New Mexico - Exploration implications. AAPG Bulletin, 63, 598-607. HARRIS. J. F.. TAYLOR, G. L. & WALPER. J. L. 1960. Relation of deformational fractures in sedimentary rocks to regional and local structure. AAPG Bulletin, 44. 1853-1873. HARRISON, J. E. & MOENCH, R. H. 1961. Joints in Precambrian Rocks, Central City-Idaho Springs Area, Colorado. United States Geological Survey, Professional Paper, 374B HEFFER. K. J. & BEVAN, T. G. 1990. Scaling relationships in natural fractures - data, theory and applications. Proceedings of the European Petroleum Conference, 2, 367-376. HILBERT. D & COHN-VOSSEN, S. 1932. Geometry and the Imagination. Chelsea Publishing, New York (reprinted in 1983). HOBBS, B. E. 1971. The analysis of strain in folded layers. Tectonophysics, 11, 329-375. KERR. H. G., WHITE, N. & BRUN, J.-P. 1993. An automatic method for determining three-dimensional normal fault geometries. Journal of Geophysical Research, 98. 17,837-17,857. LISLE. R. J. 1992. Constant bed-length folding: threedimensional geometrical implications. Journal of Structural Geology, 14. 245-252. 1994. Detection of zones of abnormal strains in structures using Gaussian curvature analysis. AAPG Bulletin, 78, 1811-1819. & ROBINSON, J. M. 1995. The Mohr circle for curvature and its application to fold description. Journal of Structural Geology. 17, 739-750. LOVE. A. E. H. 1934. A Treatise on the Mathematical Theory of Elasticity. Cambridge University Press, Cambridge.
MANDL, G. 1987. Discontinuous fault zones. Journal of Structural Geology, 9. 105-110. MUEHLBERGER, W. R. 1961. Conjugate joint sets of small dihedral angle. Journal of Geologv. 69. 211-219. MURRAY. G. H. 1968. Quantitative fracture study. Sanish Pool. McKenzie County. North Dakota. AAPG Bulletin. 52. 57-65. Narr, W. 1991. Fracture density in the deep subsurface: Techniques with application to the Point Arguello Oil Field. AAPG Bulletin. 75. 1300-1323. NORRIS, D. K. 1967. Structural analysis of the Queensway folds. Ottawa, Canada. Canadian Journal of Earth Sciences. 4, 299-321. PADGETT. M. J. & NESTER. D. C. 1991. Fracture evaluation of block P-0315, Point Arguello field, offshore California, using core, outcrop, seismic data and curved space analysis. In, The integration of geology, geophysics, petrophysics and petroleum engineering in reservoir delineation, description and management. /;;: Proceedings of the First Archie Conference. 22-25 October 1990. Houston. Texas. American Association of Petroleum Geologists. PRICE. N. J. & COSGROVE. J.W. 1990. Analysis of Geological Structures. Cambridge University Press. Cambridge. RAMSAY. J. G. 1967. Folding and Fracturing of Rocks. McGraw-Hill. New York 1974. Development of chevron folds. AAPG Bulletin. 85. 1741-1754. STEARNS. D. W. 1968. Certain aspects of fracture in naturally deformed rocks. In RIECKER. R. E. (ed.) NSF Advanced Science Seminar in Rock Mechanics. Air Force Cambridge Research Laboratories. STEWART. S. A. 1993. Fold interference structures in thrust systems. Tectonophysics. 225. 444-456. STEWART. S.A. & PODOLSKI. R. 1998. Curvature analysis of gridded surfaces. In: COWARD. M. P.. JOHNSON. H. & DALTABAN. T. S. (eds) Structural Geologv in Reservoir Characterization. Geological Society. London. Special Publication. 127. 133147 TIMOSHENKO. S. & WOINOWSKY-KRIEGER. S. 1959. Theorv of plates and shells. McGraw-Hill. New York." TREAGUS. S. H. & LISLE. R. J. 1997. Do principal surfaces of stress and strain always exist? Journal of Structural Geology. 19. 997-1 f 10. VERRALL, P. 1981. Structural interpretation \\~ith applications to North Sea problems. Course Notes No. 3, Joint Association for Petroleum Exploration Courses (UK). WHITE, N. & YIELDING. G. 1991. Calculating normal fault geometries at depth: theory and examples. In: ROBERTS, A. M., YIELDING. G. & FREEMAN. B. (eds) The Geometry of Normal Faults. Geological Society. London. Special Publications. 56. 251-260 WITHJACK. M. O. & SCHEINER. C. 1982. Fault patterns associated with domes - an experimental and analytical study. AAPG Bulletin. 66. 302316.
STRAIN PATTERNS FROM 3D FOLD GEOMETRY WOODWARD, L. A. 1984. Potential of significant oil and gas fracture reservoirs in Cretaceous rocks of Raton Basin, New Mexico. AAPG Bulletin, 68, 628-636. YIELDING, G., WALSH, J. & WATTERSON, J. 1992. The prediction of small-scale faulting in reservoirs. First Break, 10, 449-460.
221
ZHAO, P., Liu, J. & POLLARD, D. D. 1977. Prediction of subsurface fracture density using curvature and composite plate method. Stanford Rock Fracture Project (Stanford University, CA), 8, H1-H17.
This page intentionally left blank
Index Ainslie detachment, Cape Breton Island, Hollow Fault 88-96 Alba Field, UKCS, differential compaction and post-depositional sandstone remobilization 61-71 Anglesey, Wales, curvature-related fracturing 215 Arabian Plate tectonic and stratigraphic evolution 187-96 Zagros deformation belt 187-96 basement faulting, Silesian Cracovian Region, geometry of fold arrays 167-79 bedding-plane faulting curvature-related joints 23-4 fault-related joints 24 Black Hills foreland uplift structural evolution 152-3 W South Dakota and NE Wyoming 145-66 geological setting 146-52 Bohemian Massif, Thaya dome buckling and gravitational collapse of imbricated nappe sequence 197-212 metamorphic evolution 201-4 refolding of nappes and isograds 204-9 boundary element method (BEM) experiments 33-44 Brunovistulian continent, Moldanubian nappe sequence 197-212 buckle folds 3-D geometry and spatial organization 7-8 associated fractures 11-17 defined 1 and forced folds, temporal/spatial relationship 187-96 Zagros Mountains, Iran 7 8 , 191-3 caldera resurgence, Ischia drape folds 112-13 monoclinal folding 103-14 Canada, Maritimes basin, Nova Scotia, forced folding, and displacement transfer 87-101 CitaraTuff, Ischia 105-9 coal beds Poland, Upper Silesian Coal basin (USCB), geometry of fold arrays 167-79 Wyoming, US, interbedded with sandstone, differential compaction and fractures 51-60 compactional forced folds Alba Field, UKCS 61-71 Kaibab Monocline, Utah 23-49 Wyoming, coal interbedded with sandstone, differential compaction and fractures 51 -60 compressional and strike-slip environments Columbia River Basalt Province, Washington, USA, primary and forced folds 181-6
Fall River Formation, Black Hills foreland uplift, W South Dakota and NE Wyoming 145-66 inter-layer slip, model of forced folds 129-44 Silesian Cracovian Region, Poland, geometry of fold arrays 167-79 curvature-related joints, bedding-plane faulting 23-4 deformation processes competition 142 en echelon folds 167-79 developable fold surfaces, neutral surface folds and forced folds 213-21 displacement transfer, and forced folding, Maritimes basin, Nova Scotia, eastern Canada 87-101 Dogger strata, Rhine graben, extensional forced folding and decollement 73-86 drag folds (stratal folding), Mexico 115-28 comparison with throw patterns 124 distribution 120-4 drape folds and caldera resurgence 112-13 coal, USCB 173 dune boundaries, Navajo Formation boundary element method (BEM) experiments 33-44 defined 28 field data 32-3 joints associated 30 slip, location and extent 30-3 slip and fold curvature, East Kaibab Monocline, Utah 23-49 en echelon folds 167-79 extensional forced folding caldera resurgence at Ischia 103-14 and decollement, Rhine graben 73-86 Maritimes basin, Nova Scotia 87-101 Mississippi Delta tectonics, Gulf of Mexico 115-28 Fall River Formation Black Hills foreland uplift, W South Dakota and NE Wyoming 145-66 fracture domains 153-7 joint readings 157-61 stratigraphy 153 fault-bend folds 11 flexural flow folds, inter-layer slip, model 129-44 flexural slip folding 140 model 214-15 footwall and hangingwall 121 Rhine graben 73-86
224
INDEX
forced folds 3-D geometry and spatial organization 8-11 adjacent to normal faults, Gulf of Mexico 115-28 Columbia River Basalt Province, Washington 181-6 defined 1, 193 extensional and decollement, pre-rift series Rhine graben 73-86 Maritimes basin, Nova Scotia 87-101 fold arrays in Silesian Cracovian Region, southern Poland 167-80 fold curvature, and fracture density 17-19 Maritimes basin, Nova Scotia 87-101 model, inter-layer slip 129-44 retrodeformable kinematic model, sequential development 97 Rhine graben 73-86 Tertiary sediments of Alba Field, UKCS 61-71 Zagros Mountains, Iran 7-8, 193-5 fracture density, and fold curvature 17-19 fracture formation, timing 19-20 fracture patterns, W South Dakota and NE Wyoming 145-66 Gaussian curvature, 3-D flexural slip folding model 214-15 Green River Basin, Mesaverde Group, sandstone and coal 51-60 growth faults, Gulf of Mexico 115-28 Gulf of Mexico, fold patterns adjacent to normal faults 115-28 Gulf of St Lawrence, Maritimes basin, Nova Scotia 87-101 hangingwall and footwall, drag folds 121 Rhine graben 73-86 inter-layer slip model of forced folds 129-44 wrench faulting 174-6 Iran, Zagros Mountains 7-8, 187-96 Iranian Block, tectonic and stratigraphic evolution 187-96 Ischia, Italy caldera resurgence 103-14 drape folds 110-11 Kaibab Monocline, Utah curvature-related joints 23-4 field evidence 27 interlayer slip 27 numerical modelling 23-49
kink folds 220-1 layer-parallel slip 141 Louisiana, Gulf of Mexico, fold patterns adjacent to normal faults 115-28 Maritimes basin, Nova Scotia 87-101 Mesaverde Group, differential compaction of interbedded sandstone and coal 51-60 metamorphic zonation, Thaya dome, imbricated nappe sequence 197-212 Mississippi Delta tectonics, Gulf of Mexico 115-28 models of forced folds 129-44 composition 130-3 Gaussian curvature 214-11 interpretation of fracture fabrics 137-9 results 133-7 Moldanubian nappe sequence, Bohemian Massif, Thaya dome 197-212 monoclinal folding, induced by caldera resurgence at Ischia 103-14 Moray Firth, Outer, Alba Field, differential compaction and post-depositional sandstone remobilization 61-71 Mount Epomeo Green Tuff (MEGT), Ischia 103-13 multilayer flexures, interlayer slip and joint localization 23-49 Navajo Formation, Utah 28-49 neutral surface folds, and forced folds, threedimensional fold geometries, and strain patterns 213-21 normal faults, vs reverse faults 115-28 North Sea, Alba Field, UKCS, differential compaction and post-depositional sandstone remobilization 61-71 Olympic-Wallowa lineament (OWL), Washington 181-6 pencil box shear folding 218 Poland, Silesian Cracovian Region, geometry of fold arrays 167-79 Powder River basin, and Black Hills foreland uplift, Wyoming 151 Priabonian, Rhine graben 74-7 profile geometry of forced folds and buckle folds 7-21 Quaternary, Late, monoclinal folding, caldera resurgence at Ischia 103-14
INDEX Rattlesnake Mountain anticline, Rocky Mountains, monoclinal structure 96 regional tectonics, and fracture patterns, Fall River Formation, W South Dakota and NE Wyoming 145-66 reverse folding, adjacent to normal faults 119 Rhine graben extensional forced folds and decollement, pre-rift series, influence on geometry of syn-rift sequences 73-86 geology and tectonics 73-4 Rocky Mountains, Rattlesnake Mountain anticline 96 Rupelian, Rhine graben 74-7 sandstone Bedded and Injected Sandstone Facies, Alba Field 63-4 interbedded with coal 51-60 remobilization, post-depositional 61-71 Saudi Arabian Plate, Zagros deformation belt 187-96 shear folding, pencil box 218 Silesian Cracovian Region, Poland, geometry of fold arrays 167-79 South Dakota, and NE Wyoming, Fall River Formation, regional tectonics 145-66 strain patterns, prediction, 3-D fold geometries, neutral surface folds and forced folds 2 1 3 2 1 stratal folding (drag folds) 115-28 strike-slip faulting, Silesian Cracovian Region, geometry of fold arrays 167-79 strike-slip zone, wrenching regime, kinematics 172 syn-rift sequences, Rhine graben 73-86 temporal/spatial relationship forced folds, and buckle folds 187-96 prediction of 3-D fold geometries, neutral surface folds and forced folds 213-21 Thaya dome, Bohemian Massif, buckling of nappe sequence 197-212
225
Tethys and Neo-Tethys Sea 187-8 Thaya dome, Bohemian Massif, buckling and gravitational collapse of imbricated nappe sequence 197-212 three-dimensional fold geometries classification and folding mechanism 219 flexural slip folding model, Gaussian curvature 214-15 and spatial organization, forced folds and buckle folds 7-8 underthrusting, Brunovistulian continent, Moldanubian nappe sequence 197-212 Urmieh-Dokhtar magmatic arc 188 Utah, East Kaibab Monocline, Hackberry Canyon 23-49 Vosges Massif, Rhine graben 73-4 Washington, Columbia River Basalt Province, primary and forced folds 181-6 wrench faulting associated buckling 191-3 inter-layer slip 174-6 strike-slip zone 172 Wyoming, Rock Springs Uplift, interbedded coal with sandstone 51-60 Wyoming, NE, and South Dakota, Fall River Formation 145-66 Yakima fold belt, Columbia River Basalt Province, Washington 181-6 Zagros deformation belt, Saudi Arabian Plate 187-96 Zagros Mountains, Iran, buckle folds and forced folds 7-8
