Continental transpressional and transtensional tectonics

Continental Transpressional and TranstensionaI Tectonics

Geological Society Special Publications Series Editors A. J. FLEET A. C. MORTON A. M. ROBERTS

It is recommended that reference to all or part of this book should be made in one of the following ways.

HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. E (eds) 1998. Continental Transpressional and Transtensional Tectonics. Geological Society, London, Special Publications, 135.

LIN, S., JIANG,D. & WILLIAMS,P. E 1998. Transpression (-transtension) zones of triclinic symmetry: natural example and theoretical modelling. In: HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. E (eds) Continental Transpressional and Transtensional Tectonics. Geological Society, London, Special Publications, 135, 41-58.

G E O L O G I C A L S O C I E T Y S P E C I A L P U B L I C A T I O N NO. 135

Continental Transpressional and Transtensional Tectonics EDITED

BY

R. E. H O L D S W O R T H University of Durham, UK R. A. S T R A C H A N Oxford Brookes University, UK AND J. F. D E W E Y University of Oxford, UK

1998 Published by The Geological Society London

THE GEOLOGICAL SOCIETY

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

Distributors

USA AAPG Bookstore PO Box 979 Tulsa OK 74101-0979 USA (Orders: Tel. (918) 584-2555 Fax (918) 560-2652)

First published 1998

Australia The publishers make no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility for any errors or omissions that may be made. 9 The Geological Society 1998. 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 0305-8719/98/$10.00.

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

India Affiliated East-West Press PVT Ltd G-1/16 Ansari Road New Delhi 110 002 India (Orders: Tel. (11) 327-9113 Fax (11) 326--0538)

Japan British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. ISBN1-86239-007-X Typeset by Type Study, Scarborough, UK Printed in Great Britain by The Alden Press, Osney Mead, Oxford, UK.

Kanda Book Trading Co. Tanikawa Building 3-2 Kanda Surugadai Chiyoda-Ku Tokyo 101 Japan (Orders: Tel. (03) 3255-3497 Fax (03) 3255-3495)

Contents

Preface Acknowledgements

DEWEY, J. E, HOLDSWORTH,R. E. & STRACHAN,R. A. Transpression and transtension

xii xiii 1

zones

Modelling Transpression and Transtension FOSSEN, H. & TIKOFF, B. Extended models of transpression and transtension, and application to tectonic settings

15

JONES, R. R. & HOLDSWORTH,R. E. Oblique simple shear in transpression zones

35

LIN, S., JIANG, D. & WILLIAMS,P. E Transpression (or transtension) zones of triclinic symmetry: natural example and theoretical modelling

41

SCHREURS, G. & COLLETTA,g. Analogue modelling of faulting in zones of continental transpression and transtension

59

Continental Transform Zones BUTLER, R. W. H., SPENCER,S. & GRIFFITHS,H. M. The structural response to evolving

81

plate kinematics during transpression: evolution of the Lebanese restraining bend of the Dead Sea Transform TAVARNELLI,E. Tectonic evolution of the Northern Salinian Block, California, USA: Paleogene to Recent shortening in a transform fault-bounded continental fragment

107

RUST, D. Contractional and extensional structures in the transpressive 'Big Bend' of the San Andreas fault, southern California

119

REIJS, J. & MCCLAY, K. Salar Grande pull-apart basin, Atacama Fault System, northern Chile

127

TEYSSIER, C. & TIKOFF,B. Strike-slip partitioned transpression of the San Andreas fault system: a lithospheric-scale approach

143

Oblique Divergence Zones KRABBENDAM,M. & DEWEY,J. F. Exhumation of U H P rocks by transtension in the Western Gneiss Region, Scandanavian Caledonides

159

DOKKA, R. K., ROSS, T. M. & Lu, G. The Trans Mojave-Sierran shear zone and its role in Early Miocene collapse of southwestern North America

183

WATKEYS,M. K. & SOKOUTIS,D. Transtension in southeastern Africa associated with

203

Gondwana break-up ALLEN, M. B., MACDONALD,D. I. M., ZHAO NUN.,VINCENT,S. J. & BROUET-MENZIES,C. Transtensional deformation in the evolution of the Bohai Basin, northern China

215

Oblique Convergence Zones EBERT, H. D. & HASUI,Y. Transpressional tectonics and strain partitioning during oblique collision between three plates in the Precambrian of southeast Brazil

231

GAYER,R., HATHAWAY,Z. & NEMCOK,M. Transpressionally driven rotation in the external orogenic zones of the Western Carpathians and the SW British Variscides

253

GLEIZES, G., LEBLANC,D. & BOUCHEZ,J. L. The main phase of the Hercynian Orogeny in the Pyrenees is a dextral transpression

267

TANNER, D. C., BEHRMANN, J. H., ONCKEN, O. & WEBER, K. Three-dimensional retro-modelling of transpression on a linked fault system: the Upper Cretaceous deformation on the western border of the Bohemian Massif, Germany

275

CURTIS,M. L. Development of kinematic partitioning within a pure-shear dominated

289

dextral transpression zone: the southern Ellsworth Mountains, Antarctica

SEARLE,M. R, WEINBERG,R. E & DUNLAP,W. J. Transpressional tectonics along the Karakoram Fault Zone, northern Ladakh: constraints on Tibetan extrusion

307

SAINT BLANQUAT,M., TIKOFF, B., TEYSSIER, C. • VIGNERESSE, J. L. Transpressional kinematics and magmatic arcs

327

SCHIATTARELLA,M. Quaternary tectonics of the Pollino Ridge, Calabria-Lucania boundary, southern Italy

341

Index

355

Preface

This volume contains a broad spectrum of papers that summarize recent advances in the understanding of continental transpressional and transtensional tectonics. The papers include theoretical and case studies from a global set of contributors. The volume contains 22 papers. The opening contribution by Dewey et al. is an overview of the basic features of transpressional and transtensional deformation zones aimed at setting the scene for the more detailed papers to follow. These are grouped into four sections. The first, Modelling Transpression and Transtension, includes a series of papers which discuss theoretical strain models in the context of field examples and analogue experiments (Fossen & Tikoff, Jones & Holdsworth, Lin et al. and Schreus & Colletta). The second section details the tectonic evolution of Continental Transform Zones and includes papers on the Dead Sea Transform (Buffer et al.), the San Francisco Bay area (Tavarnelli), the Transverse Ranges of southern California (Rust), the Atacama Fault System of Chile (Rjeiis & McClay) and a lithosphere-scale view of the San Andreas fault system (Teyssier & Tikoff). The third section is entitled Oblique Divergence Zones. The first two papers are concerned with transtensional structures developed during gravitational collapse in the Caledonides of western Norway (Krabbendam & Dewey) and in southwestern North America (Dokka et al.). The two following papers describe the transtensional structures developed in South Africa during break-up of Gondwana (Watkeys & Soukoutis) and the evolution of the combined pull-apart/transtensional Bonai Basin, northern China (AHen et al.). The fourth section is concerned with Oblique Convergence Zones and includes case studies from the Precambrian basement of Brazil (Ebert & Hasui), the European Variscides (Gayer et al., Gleizes el al. and Tanner et al.), the Permo-Triassic Gondwanian orogen of west Antarctica (Curtis), the Himalayas (Searle et al.) and the Sierra Nevada batholith (St Blanquat et al.), and concludes with a discussion of Quaternary tectonics in southern Italy (Shiatterella). The impetus for this volume was a conference held in March 1997 at Burlington House, London, under the auspices of the Tectonic Studies Group of the Geological Society of London. The editors would like to thank all the staff at Burlington House who helped ensure the smooth running of the conference, including projection facilities and refreshments. Bob Holdsworth, Durham, UK Rob Strachan, Oxford, UK John Dewey, Oxford, UK

Acknowledgements The editors thank the following colleagues and friends who kindly helped with the reviewing of the papers submitted for this volume: Ian Alsop, St Andrews, UK Arild Andreasson, Oslo, Norway Jim Andrews, Southampton, UK Rob Butler, Leeds, UK Peter Cobbold, Rennes, France Sandy Cruden, Toronto, Canada Dickson Cunningham, Leicester, UK Mike Curtis, BAS, UK Richard D'Lemos, Oxford Brookes, UK Roy Dokka, Louisiana, USA Tim Dooley, London, UK Bill Fitches, Aberystwyth, UK Mary Ford, ETH-Zurich, Switzerland Haakon Fossen, Bergen, Norway Rod Gayer, Cardiff, UK Laurel Goodwin, New Mexico, USA Reinhard Greiling, Heidelberg, Germany John Grocott, Kingston-on-Thames, UK Bob Hatcher, Tennessee, USA Becky Jamieson, Dalhousie, Canada Richard Jones, Halden Norway Steve Knott, Aberdeen, UK Peter Koons, Otago, New Zealand Shoufa Lin, GSC Ottawa, Canada

Geoff Manby, Greenwich, UK Ken McCaffrey, Kingston-on-Thames, UK Andy McCaig, Leeds, UK Tim Needham, Spilsby, UK Franz Neubauer, Salzburg, Austria Clyde Northrup, Boston, USA Terry Pavlis, New Orleans, USA John Ridley, ETH-Zurich, Switzerland Jurriaan Rjeiis, London, UK Gerald Roberts, London, UK Paul Ryan, Galway, Eire Bryan Storey, BAS, UK Brian Sturt, Trondheim, Norway Art Sylvester, California, USA Enrico Tavarnelli, Potenza, Italy Christian Teyssier, Minnesota, USA Basil Tikoff, Minnesota, USA Andrew Tomlinson, Santiago, Chile Pete Treloar, Kingston-on-Thames, UK Mike Watkeys, Durban, South Africa Alastair Welbon, Statoil, Norway John Whalley, Portsmouth, UK Nigel Woodcock, Cambridge, UK

Transpression and transtension zones J. F. D E W E Y 1, R. E. H O L D S W O R T H

2 & R. A . S T R A C H A N 3

1Department of Earth Sciences, University of Oxford, Parks Road, Oxford OX1 3PR, UK 2Department of Geological Sciences, University of Durham, Durham DH1 3LE, UK 3Geology and Cartography Division, School of Construction and Earth Sciences, Oxford Brookes University, Gypsy Lane, Headington, Oxford OX3 0BP, UK Abstract: Transpression and transtension are strike-slip deformations that deviate from simple shear because of a component of, respectively, shortening or extension orthogonal to the deformation zone. These three-dimensional non-coaxial strains develop principally in response to obliquely convergent or divergent relative motions across plate boundary and other crustal deformation zones at various scales. The basic constant-volume strain model with a vertical stretch can be modified to allow for volume change, lateral stretch, an oblique simple shear component, heterogeneous strain and steady-state transpression and transtension. The more sophisticated triclinic models may be more realistic but their mathematical complexity may limit their general application when interpreting geological examples. Most transpression zones generate flattening (k < 1) and transtension zones constrictional (k > 1) finite strains, although exceptions can occur in certain situations. Relative plate motion vectors, instantaneous strain (or stress) axes and finite strain axes are all oblique to one another in transpression and transtension zones. Kinematic partitioning of non-coaxial strike-slip and coaxial strains appears to be a characteristic feature of many such zones, especially where the far-field (plate) displacement direction is markedly oblique (<20 ~ to the plate or deformation zone boundary. Complex foliation, lineation and other structural patterns are also expected in such settings, resulting from switching or progressive rotation of finite strain axes. The variation in style and kinematic linkage of transpressional and transtensional structures at different crustal depths is poorly understood at present but may be of central importance to understanding the relationship between deformation in the lithospheric mantle and crust. Existing analyses of obliquely convergent and divergent zones highlight the importance of kinematic boundary conditions and imply that stress may be of secondary importance in controlling the dynamics of deformation in the crust and lithosphere.

Transpression (TP) and transtension (TT) (Harland 1971) occur on a wide variety of scales during deformation of the Earth's lithosphere. On the largest scale, this is an inevitable consequence of relative plate motion on a spherical surface: plate convergence and divergence slip vectors are not commonly precisely orthogonal to plate boundaries and other deformation zones. Plate b o u n d a r y zones will, therefore, experience oblique relative motions at some time during their history along some part of their length (Dewey 1975). Within a plate boundary zone, strain is focused generally into displacement zones that bound units of less-deformed material on several scales. This is particularly evident in continental orogens, where broad deformation belts develop in which fault- and shear-zone-bounded blocks partition strains into a series of complex displacements, internal strains and rotations in response to far-field plate tectonic stresses and large-scale body forces (Dewey et al. 1986). Here again, block convergence and divergence slip vectors are not commonly precisely orthogonal to or parallel to plate

margins or to smaller-scale deformation zone boundaries. In many cases, this arises because block margins are inherited features that act as zones of weakness, repeatedly reactivated during successive crustal strains, often in preference to the formation of new zones of displacement (Holdsworth et al. 1997). Any displacement zone margin that is significantly curvilinear or irregular is bound to exhibit oblique convergence and/or divergence unless it follows exactly a small circle of rotation. In addition to collisional orogenic belts, TP and TT occur widely in a large range of other tectonic settings: oblique subduction margins in the forearc (TP), arc (TP and TT) and back-arc (TT) regions; 'restraining' (TP) and 'releasing' (TT) bends of transform and other strike-slip displacement zones; continental rift zones (TT), especially during the early stages of continental break-up and formation of new oceanic lithosphere; during late orogenic extension (TT) and in slate belts (TP), where deformation may be accompanied by large-scale volume loss. This paper sets out to introduce some of the basic features of transpressional and

J. E DEWEY,R. E. HOLDSWORTH& R. A. STRACHAN.1998. Transpression and transtension zones. In: HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. E (eds) 1998. ContinentalTranspressional and TranstensionalTectonics.Geological Society, London, Special Publications, 135, 1-14.

2

J.F. DEWEY ETAL.

transtensional deformation zones and to highlight briefly some important problems that may be encountered in such regions and their implications for geologists. Our review is not exhaustive and is intended to set the scene for the papers in this volume.

Properties of transpression and transtension zones Basic definitions Relative plate, or block, motion vectors and the orientation of the plate or block margins are the principal boundary conditions during lithospheric deformation. Therefore, in contrast to the original definitions discussed by Harland (1971), we suggest that the terms oblique convergence and oblique divergence should be used to indicate the relative motion of boundary plates or blocks. These are measured using the angle between the horizontal far-field (plate) displacement vector and the boundary of the deformation zone (Fig. 1; Tikoff & Teyssier 1994). We suggest that the terms transpression and transtension be restricted to the resulting combinations of non-coaxial and coaxial strains. Thus transpression and transtension can be defined as strike-slip deformations that deviate from simple shear because of a component of, respectively, shortening or extension orthogonal to the deformation zone (Fig. 1). The deformation zone is commonly steeply dipping or sub-vertical.

Strain patterns Existing studies have modelled transpression and transtension zones using either finite and incremental strain (e.g. Sanderson & Marchini 1984; Fossen & Tikoff 1993; Tikoff & Teyssier 1994) or strain rate (e.g. Ramberg 1975). Models based on strain have been shown to be effective starting points in the analysis of three-dimensional deformation zones and examples are shown in Fig. l a - g (for transpression zones), ordered approximately in terms of increasing realism and complexity. The basic model of Sanderson & Marchini (1984) (e.g. Fig. la) involves a constantvolume, homogeneous strain in a vertical zone in which horizontal shortening (or extension) across the zone was accommodated by vertical extension (or shortening). It is possible to modify this basic model by changing each of the boundary conditions to allow for volume change (e.g. Fig. lb; Fossen & Tikoff 1993), lateral stretch (Fig. lc; Dias & Ribeiro 1994; Jones etal. 1997) or an oblique simple shear component in which the bounding blocks are displaced vertically and laterally (Fig. ld; Robin & Cruden 1994; Jones &

Fig. 1. Some examples of transpressional strain models. (Note that the arrows and angle ~ are omitted in (f) and (g) for simplicity.) Holdsworth this volume). All these homogeneous strain models are idealized compared with the strain patterns in naturally occurring shear zones. In particular, the deformation zone boundaries are unlikely to be unconstrained, as they cannot simultaneously allow free slip in all directions while transmitting the shear stress imparted by the component of simple shear (Schwerdtner 1989; Robin & Cruden 1994). A more realistic model would display a strain gradient developed from zero slip at the boundary walls to maximum vertical extension in the centre of the zone, yielding a heterogeneous strain (Fig. le; Robin & Cruden 1994). Unfortunately, such models are extremely complex and do not easily allow generalizations to be made concerning finite strains; therefore, they may be of limited use in the analysis of most naturally deformed rocks. In all these models, the deformation zone has a fixed width, so the strain rates will exponentially increase (transpression) or decrease (transtension). Further variants are steady-state transpressional or transtensional models, either symmetrical (Fig. lf) or asymmetrical (Fig. lg), that yield a constant strain rate (Dutton 1997).

TRANSPRESSION AND TRANSTENSION ZONES

3

Transpression/Transtension with vertical stretch a) J Volume [ constant I ' ,

~--~,~-j~ k=co I ~ ' ~ ~ ' ~ k = 8

~~~STRICTIkON/~transpressi~ k=1.35

z

.k=n97 = ~--~

J~, I

k=O

b)

,o3

prolate ~ \/10~ sJrains~L..../'~

(zvalues~ ~

-I Dlanestrains J

;0~)cJ~trSt~i~Sia~~j~r

/~ 1

o, [// ~ ~.,~v50~tra~ssi~onal ~

~.~15

10~ lO

5

I0 ~

101

LogY/Z

10 2

oblatestrains k=0

10 3

Fig. 2. (a) Flinn plot to illustrate examples of transpressional flattening and transtensional constrictional deformations. The double-headed arrows are incremental stretching directions. (b) Logarithmic Flinn plot to illustrate some transpressional (flattening) and transtensional (constrictional) strain paths. This is important because the correlation of flow parameters, such as vorticity, with finite strains will be valid only if the deformation is steady state (Jiang & White 1995). In the more straightforward strain models for transpression and transtension (e.g. Fig. 1a-c), one of the principal axes of finite strain remains fixed and vertical as deformation progresses, and the other two axes rotate in the horizontal

plane because of the non-coaxial wrench simple shear component. This produces strains that, in common with simple shear, have a monoclinic symmetry. In contrast, more complex models, requiring an oblique simple shear component (e.g. Fig. ld-g) generally have a triclinic symmetry in which all three axes of finite strain rotate relative to a fixed external reference frame. Figure 2a and b shows examples of finite

4

J.F. DEWEY E T A L .

SV

AA

AA

~: ,.:.7

Insubric

i r

Line

Lepontine(Penninic)core 3implon Line

N

A I

100km

t

Fig. 3. Simplified tectonic map of the Central Alps. Lines with open arrowheads are Eocene-Oligocene lineations; lines with filled arrowheads are Miocene lineations; AA, Austro-Alpine zones. Inset map shows a schematic illustration of the Miocene trans-Alps slip vector partitioning (H, Helvetics; L, Lepontine; IL, Insubric Line; RL, Rh6ne Line; UZ, Urserenzone; SV, slip vector). strains and strain paths using the basic constantvolume, vertical stretch model of Sanderson & Marchini (1984) (e.g. Fig. la). These show that transpression generates flattening (k < 1) and transtension constrictional (k > 1) bulk strains. This is also generally the case in constantvolume deformation zones where lateral extrusion (Fig. lc) or oblique simple shear (Fig. ld) or heterogeneous transpression (Fig. le) occur, although the range of k values is reduced, clustering increasingly towards plane strains as the lateral stretch or vertical simple shear components of finite strain are increased, respectively (Robin & Cruden 1994; Jones et al. 1997; Jones & Holdsworth this volume). However, it is possible to develop constrictional or prolate strains in transpression zones and flattening or oblate strains in transtension zones in special circumstances. For example, constrictional strains will develop in transpression zones where there is a component of lateral stretch and vertical shortening (or volume loss) (e.g. Dias & Ribeiro 1994; Fossen & Tikoff this volume). A good example of this developed during Miocene oblique-dextral convergence in the Swiss Alps (Fig. 3), where partitioning developed between roughly zone-orthogonal Helvetic thrusting and

zone-parallel dextral slip on the Insubric Line, which passed northwestwards into the extensional Simplon Line and then westward into the dextral Rh6ne Line. During dextral slip, the Lepontine Alps were extensionally unroofed and vertically shortened simultaneously with north-south horizontal shortening and east-west extension, thus leading to a bulk constrictional deformation in a plate-boundary zone of oblique convergence. In constant-volume transpression or transtension zones with a vertical stretch, finite strain paths at low angles of convergence (c~ < 20~ 'wrench-dominated' models of Tikoff & Teyssier 1994) are strongly non-linear and complex with 'bouncing' off k = 0 and k = oo axes (Fig. 2b). This coincides with a switching (or 'swapping') in the orientation of the finite strain axes (e.g. Fig. 4; Sanderson & Marchini 1984; Tikoff & Teyssier 1994; Tikoff & Greene 1997). In transpression zones, an initially horizontal xaxis swaps orientation with the vertical y-axis with increasing finite strain, whereas in transtension zones, the y-axis swaps with an initially vertical z-axis. Far-field plate slip vectors will be parallel to the axes of instantaneous or finite strain in a

TRANSPRESSION AND TRANSTENSION ZONES

5

T r a n s p r e s s i o n with vertical stretch 10000 -='G" ,'-E ~ 1000 -I-,

.m

.___~

~

~

100

.c: "~ 0

C 0

o N

~o

ri onta,

10

x vertical,Y & Z horizontal: retching lineation

,

,

10

15

rr

1

0

5

20

I

25

Angle of convergence (o0 Increasing coaxial component Fig. 4. Plot of angle of convergence (~) v. horizontal finite strain ellipse ratio for homogeneous transpression (after Tikoff & Greene 1997, fig. 5). A line of oblate strain separates the field in which the long axis of finite strain is horizontal from the field in which it is vertical.

deformation zone only when they are oriented orthogonally to the zone boundaries (~ = 90~ Thus, in all zones of transpression and transtension, plate motions, instantaneous strain (or stress) axes and finite strain axes are oblique to one another (e.g. see top of blocks in Fig. 2a). However, the horizontal plate m o t i o n will always correspond to one of three flow apophyses that define the maximum, intermediate and minimum rates of particle movement in the zone (Fossen et al. 1994). Such flow apophyses critically control the passive rotation of marker structures in the deformation zone even though these structures may have formed initially in response to instantaneous or finite strains. This emphasizes that the relationships between plate motions and resulting strain patterns are complex. Strain p a r t i t i o n i n g Strains may be kinematically non-partitioned or partitioned in transpression and transtension zones (Fig. 5a and b). There are many ways in which partitioning may occur (e.g. Fig. 3 and 6a-f; see also Oldow et al. 1990). Fitch (1972) showed how oblique plate convergence in the Indonesian arc from Java to the A n d a m a n

Islands is partitioned between orthogonal subduction and intra-arc (or volcanic axis) strikeslip, with the strike-slip component becoming more important northwestwards. In the Andes, a fairly constant 100 mm/a E N E slip vector between the Nazca and South American Plates is accommodated and partitioned in various ways along the mountain belt (Dewey & Lamb 1992). Depending upon the sense (sinistral or dextral) of obliquity, left-lateral or right-lateral intra-arc strike-slip faulting, mainly along the weaker arc volcanic zone (where present), is combined with orthogonal Benioff Zone slip and/or back-arc thrusting. Molnar (1992) has provided the best rationale yet for plate boundary-scale partitioning. Molnar's argument is that in a strong, yet ductile, continental upper mantle, a viscous continuum should generate principal stresses and strain rates parallel to or perpendicular to the Earth's surface. Because upper-crustal block rotation about vertical axes is coupled, via a viscous lower crust, to the vorticity field of the upper mantle, oblique-slip faults should not be stable during bulk finite strain. Strain partitioning may also be facilitated by the reactivation of preexisting structural weaknesses that are in suitable orientations to minimize work done (e.g.

6

J.F. DEWEY ETAL. angles, there is strong partitioning into thrust and strike-slip components. This is well illustrated by the low degree of partitioning in the South Island of New Zealand and the strong partitioning along the San Andreas zone, where obliquity is, respectively, high and low (Teyssier et al. 1995). Tikoff & Teyssier (1994) further suggested that partitioning is especially likely where the angles of convergence or divergence are low (<20 ~ because of non-parallelism of incremental and finite strain axes. Fabric patterns a n d structural style

Fig. 5. (a) Non-partitioned transpression. (b) Partitioned transpression in which a significant component of the wrench component is accommodated by a discrete strike-slip fault. Jones & Tanner 1995) or at angles at which the coefficient of sliding friction is less than the coefficient of internal friction for new faults. Michael (1990) demonstrated that the rate of energy release on oblique-slip faults is greater than that on pairs of strike-parallel dip-slip and strike-slip faults, by invoking Hamilton's Principle that systems choose configurations that minimize work done. This could apply, particularly, to simple bimodal partitioning at island arcs where a very weak zone along the magmatic axis of the arc represents a suitable surface to accommodate the strike-slip component. Platt (1993) showed, by modelling a variety of accretionary-prism rheologies, that a linear viscous model generates a continuously partitioned strain from orthogonal near the Benioff Zone with a rapidly increasing strike-slip component near the back-stop. However, a plastic rheological model demonstrated strong dependence upon the angle of convergence; at high angles, there is little or no partitioning, whereas at low

The angle of obliquity (a), intensity of finite strain and degree of kinematic partitioning principally control fabric orientation in both transpression and transtension zones (e.g. Fig. 6a-f; McCoss 1986; Fossen & Tikoff 1993; Fossen et al. 1994; Tikoff & Teyssier 1994; Tikoff & Greene 1997). In monoclinic transpression zones where the bulk deformation follows the vertical stretch model of Sanderson & Marchini (1984), the strikes of the principal flattening surface (cleavage, schistosity, gneissosity) will vary with the non-coaxial component of the strain (Fig. 6a-f) but the dip is always vertical. Characteristically, associated stretching lineations can switch from vertical to horizontal (Fig. 4) where a <20 ~ ('wrench-dominated transpression'). If no kinematic partitioning occurs, stretching lineations will be initially horizontal but, as the finite strain increases, they will switch to vertical; the exact strain threshold required depends on the value of t~ (Fig. 4). If, however, the wrench non-coaxial component is kinematically partitioned into boundary-parallel highstrain zones, vertical lineations may switch back into sub-horizontal orientations. Possible field examples of such switching behaviour have been described by Hudleston et al. (1988), Holdsworth (1989) and Tikoff & Greene (1997). Conversely, in transtension, the stretching direction is always horizontal, but the shortening direction switches from vertical to horizontal where ct < 20 ~ with increasing finite strain (McCoss 1986). Thus, if no kinematic partitioning occurs, any planar fabric switches from vertical to horizontal with increasing strain or vice versa if boundary-parallel wrench simple sheardominated high-strain zones form as a result of partitioning. Switching behaviour is suppressed in transpression and transtension zones where there is a significant component of lateral stretch and lineations are mainly horizontal (e.g. Jones et al. 1997). The relationship between planar and linear fabrics in triclinic transpression and transtension

TRANSPRESSION AND TRANSTENSION ZONES

I Cross section

7

Plan view I 9

9t

9

l

simple partitioning

\,,.

9

9

9

A,,

'"

v

9

D ~ e

\

9

-,

\ transport direction

9

no partitioning

.

D

.

.

.

.

.

.

.

continuous partitioning

r

o .qo

F discontinuous partitioning

I Zero Reference Datum] Fig. 6. Plan view of six transpressional zones showing various degrees and styles of partitioning, particle paths in the transpression zone (arrows on sides) and cleavage traces (fine lines). D, tectonic transport direction; a, angle between D and zone margin; e, zone-normal direction. Topographic profiles and possible gravity-driven thrust movements (half arrows) are indicated in the cross-sections. (a) Simple partitioning into marginal thrust belts and a central strike-slip zone. (b) A non-partitioned zone of oblique cleavage (particle paths parallel to D). (c) Continuous symmetrical partitioning of coaxial and non-coaxial components towards the edges and centre of the zone, respectively. (d) Continuous symmetrical partitioning of the coaxial and non-coaxial components towards the centre and edges of the zone, respectively. (e) Continuous asymmetrical partitioning of the coaxial and non-coaxial components to either side of the zone. (f) Discontinuous partitioning of the coaxial and non-coaxial components.

8

J.F. DEWEY E T A L .

/•

LATE DUCTILE (LD) FOLDS LD folds + fabrics

AND FABRICS .

~ ,,,-"

LD folds

+fabricsrare~

~"

F *OVERPRINTED

i

~ -SIDE UP

. "~

L

UP,,,,,,"

_~,N

LDSbr:Ctntres

[ NW--SIDE UP

EARLY FOLDS AND FABRICS Fig. 7. Three-dimensional diagram showing foliation, lineation, shear sense and associated folding patterns from a transpressional terrane boundary shear zone developed in the Avalon Zone, NE Newfoundland Appalachians (redrawn after Holdsworth 1994). Two phases of ductile transpressional structures, 'early' and 'late ductile' (LD), are recognized, both of which appear to have an overall triclinic symmetry. (Note that in both cases, the wrench component is concentrated adjacent to the present-day terrane boundary, the Dover Fault.) zones (e.g. Fig. 1d-g) is poorly understood. The preliminary transpressional models of Robin & Cruden (1994) suggest that complex and systematic variations in the orientation of both foliations and lineations will occur depending upon the intensity of finite strain, the obliquity of the simple shear component and the nature of any kinematic partitioning within the deformation zone. This seems to be consistent with the field observations made in steeply dipping transpression zones where there is evidence for a vertical simple shear component (e.g. Fig. 7 based on Holdsworth (1994); other examples have been given by Robin & Cruden (1994) and Goodwin & Williams (1996)). Figure 8a-e illustrates schematically in cross-section what may happen to foliation trajectories in vertically extruding, heterogeneous, triclinic transpression zones with a range of boundary wall morphologies. It should be noted that the broadening or narrowing of the zones will lead to significant changes

in both the intensity of finite strain and the strain rate. The depth variation in the style of transpressional and, especially, transtensional deformation structures and how such systems may be kinematically linked is very poorly understood. The shallower crustal levels are generally characterized by brittle deformation among complex domains of rotating crustal blocks. Exhumed examples of middle-crustal transpression zones display similar configurations except that the blocks may be internally deformed and the block-bounding structures are now ductile shear zones (e.g. Hudleston et al. 1988; D'Lemos et al. 1992). The nature of transpressional and transtensional structures in the weak ductile lower crust and much stronger ductile upper mantle is uncertain but is of some importance if the coupled model proposed by Molnar (1992) for zones of oblique convergence and divergence is correct. Transpressional slate belts also present

TRANSPRESSION AND TRANSTENSION ZONES

t Cross Sections t

z.--

//J 1111

1"///,/,I

y

I,"

.~////li',,

/z////i /// \

/

9

L Fig. 8. Transpression with various boundary wall orientations showing possible cleavage trajectories in section. (a) Vertical walls with vertical escape. (b) Zone widening upwards between inward-dipping walls. (c) Zone narrowing upwards between outward-dipping walls. (d) Inclined walls dipping in same direction. (e) Inclined walls, more steeply dipping extensional detachment above more gently dipping basal thrust. Possible gravitydriven thrust movements are also indicated by the shaded zones and arrows. a special problem. These are, structurally, highlevel anchimetamorphic zones in which transecting cleavages and cleavage sequences are developed often with very large volume losses (up to 55%) because of fluid-assisted diffusive

mass transfer mechanisms (e.g. Wood 1973; Cox & Etheridge 1989). It is not yet clear where modern slate belts develop and how they maintain compatibility with subjacent high-grade crustal levels.

10

J.F. DEWEY ETAL.

Lateral stretch, extrusion and escape There is some semantic and kinematic confusion in the literature arising from failure to distinguish between lateral extrusion (or lateral escape) and lengthening of a transpression or transtension zone developed in other ways. There are three ways in which lateral stretches may develop in plate boundary or other deformation zones: lateral extrusion, non-coaxial stretching, and radial spreading.

Lateral extrusion. Lateral extrusion (or escape) sensu stricto involves a stretch in the horizontal direction that causes the deformation zone to lengthen relative to the undeformed or lessdeformed rocks that form the zone margins (e.g. Fig. lc; Jones et al. 1997). Homogeneous lateral extrusion necessarily involves slippage along zone walls with changing slip along strike, but heterogeneous extrusion may be boundary wall compatible where a zone-orthogonal strain gradient exists, although the amount of extrusion will be severely limited. Lateral extrusion, therefore, involves horizontal, along-strike mass movement of material towards the end (or ends) of the zone where compatibility problems of material extruding from the zone are solved by creating space at the end of the zone (Harland 1971; Ramsay & Huber 1987). In theory, this should be kinematically possible at certain types of triple junction or at a transform-bounded subduction zone. These are rare situations and escape is, therefore, extremely unlikely at a large plate-boundary zone scale. However, exceptions are known to exist in some transpressional and transtensional settings at smaller scales, where specific geometric and mechanical boundary conditions and internal theologies in the deformation zone may favour lateral extrusion. It should be noted that, to apply such models, it is necessary to account for the lateral space problem at the terminations of the deformation zone. (1) Transpression involving lateral extrusion affects deformed serpentinites intruding a complex suture zone in SW Cyprus (Bailey 1997; Jones et al. 1997). In this case, the serpentinites were intruded as irregular and possibly isolated bodies into a pre-existing fault system. Subsequent transpressional deformation then redistibuted the rheologically weak material laterally and, to a lesser extent, vertically within the fault zones. Similar deformation patterns may be common in plutonic bodies emplaced syn-tectonically along pre-existing faults and shear zones in active transpressional arcs. (2) Small but significant components of

deformation that may be attributed to a component of lateral extrusion have been documented in continental strike-slip duplex systems (Laney & Gates 1996) and broader regions of transpressional deformation, e.g. mid-Devonian, central Scotland (Jones et al. 1997). The complex and irregular nature of plate and 'block' boundaries and their mutual interaction means that small components of lateral extrusion may be possible in continental deformation zones.

Non-coaxial components of stretching. These can lead to horizontal zone elongation at low angles to the plate or deformation zone boundary in transpression and transtension zones. This stretch is boundary wall compatible and does not lead to extrusion. On a plate boundary scale, the arc-parallel along-strike extension in Sumatra (McCaffrey 1991) and the Aleutians (Ekstr6m & Engdahl 1989), and the transpressional zone-parallel extension of northeast Venezuela that may be an important mechanism for the exhumation of high-pressure-low-temperature metamorphic rocks in accretionary wedges (Av6 Lallemant & Guth 1990), are probable examples of this non-coaxial stretching. Radial spreading. Radial spreading driven by body forces in curved transpressional arcs, accretionary prisms and thrust wedges may lead to zone-parallel plate-boundary scale extension. Good examples occur by radial thrusting in the Himalayas, which leads to orogen-parallel stretching, and the radial back-arc spreading of the Aegean, which is causing arc-parallel extension in the Cretan fore-arc. In such radial systems, body-force driven coaxial strains are combined with the coaxial and non-coaxial strains of transpression caused by relative plate motion across oblique portions of the plate boundary zone. Compatibility is not required along the free-slipping basal thrust or Benioff Zone but is maintained with the zone-orthogonal extending thrust sheet or upper plate. Transpression, topography and exhumation Large-scale steeply dipping or vertical transpression zones are likely to develop significant surface topographies both above and at the lateral terminations of the deformation zones. These topographies may be sufficient to generate regionally important episodes of gravitydriven deformation that may aid the exhumation of deep crustal rocks. Variations in strain intensity, kinematic partitioning and orientation of the deformation zone boundaries will all affect

TRANSPRESSION AND TRANSTENSION ZONES

11

Fig. 9. Three possible modes of termination of transpressional zones. (a) Termination in coaxial zones where the zone termination boundaries do not rotate and where there is no compatibility problem between the surface topography of the coaxial and transpressional zones. The only compatibility problem exists where zone orthogonal, gravity-driven thrusting is in different directions at the coaxial-transpressional boundary. (b) Termination in non-coaxial zones where the termination boundaries rotate with non-coaxial zone vorticity and where a compatibility problem exists between the growing topography of the transpressional zone and the 'non-topography' of the non-coaxial zone. (c) One end terminated by a non-coaxial zone, the other by a coaxial zone. Gravity-driven lateral flow will be towards the non-coaxial zone. the surface topography of transpression zones and resulting geometries and distributions of gravity-driven deformation (Fig. 6a-f and 8a-e). The terminations of transpression (and transtension) zones are rarely considered. They may be terminated by coaxial or non-coaxial zones in one of three ways (Fig. 9a-c). Where coaxial zones terminate the transpressional zone (Fig. 9a), the transpressional boundary does not rotate and there are no compatibility problems b e t w e e n transpressional and coaxial zones. Where the transpressional zone is terminated by non-coaxial simple shear zones (Fig. 9b), not only do the n o n - c o a x i a l - t r a n s p r e s s i o n zone boundaries rotate with the non-coaxial rotational sense but there is also a topographic

mismatch b e t w e e n the vertically t h i c k e n e d transpression zone and the non-thickened noncoaxial zone that is likely to engender high-level gravity-driven flow. Third, the transpressional zone may be terminated by both non-coaxial and coaxial strain zones, allowing unidirectional gravity-driven flow towards the non-coaxial zone (Fig. 9c).

Implications All transpression and transtension zones display three-dimensional, non-coaxial strains; existing and future attempts to model their development represent part of a broader and long overdue attempt to deal with geological deformation

12

J.F. DEWEY E T A L .

zones in a realistic manner. There are important practical difficulties. In particular, it is not clear whether strain models need to strive for realism, making them complex and unwieldy, or simplicity, meaning that they may only be applied qualitatively to finite bulk strain patterns. On a b r o a d e r note, the strain patterns deduced in transpression and transtension zones have several i m p o r t a n t implications for the broader geological community: (1) Balanced cross-section techniques. Many quantitative studies of crustal deformation use cross-sections and explicitly or implicitly assume two-dimensional (i.e. k = 1 plane) strains. Examples include most palinspastic or balanced section methods in convergent and divergent settings (e.g. Buchanan & Neiuwland (1996) and references therein) and some forward-modelling techniques employed in the analysis of sedimentary basins (e.g. Kusznir et al. 1991). If the region being considered has suffered any component of wrenching, extension (transpression) or contraction (transtension) occurs normal to the line of section (Jamison 1991). Used in isolation, such two-dimensional cross-sectionbased methods may yield misleading results, and the development and use of three-dimensional techniques is required (e.g. Ma & Kuzsnir 1992, 1993). (2) Strain ellipses and faulting. Wilcox et al. (1973) and Harding (1974) used a two-dimensional finite strain ellipse to account for the wide range of second-order contractional, extensional and strike-slip structures that formed in analogue models of strike-slip d e f o r m a t i o n zones viewed in plan. These workers demonstrated that such models may be used qualitatively to explain deformation patterns within strike-slip zones where the shear plane is well defined and the deformation approximates to a steeply dipping wrench simple shear. However, such ellipse models have also been very widely used in the interpretation of faulting patterns in offshore sedimentary basins and, in some cases, they are used to support strike-slip or oblique opening hypotheses (e.g. Gibbs 1986; Fossen 1989; Dor6 & Lundin 1996). The use of such two-dimensional, simple shear models in regions of three-dimensional transtensional strain is inappropriate and may lead to serious errors in the interpretation of crustal deformation patterns. (3) Structural complexity. Complex and, as yet, poorly understood deformation patterns are likely within many transpression and transtension zones. Existing field studies (e.g. Fig. 7) show that structures that differ significantly in orientation may form simultaneously. Many of

these features would previously have been and, in some areas, still are interpreted as the products of polyphase deformation events. A more fundamental difficulty exists in all transpression zones because there is not, in general, a simple relationship between stretching lineations and the direction of tectonic transport. Marked changes in lineation pattern may occur because of spatial variations in finite strain and/or kinematic partitioning (Tikoff & Greene 1997). This problem is particularly acute in triclinic transpression zones, where the precise relationship between the development of geological structures and deformation is very poorly understood. (4) Deduction o f plate motions. Zones of transpression and transtension present particular problems when trying to relate crustal deformation patterns to relative plate motions. If kinematic partitioning occurs, as is fairly common, the deformation seen in one region may not be representative of the system as a whole. This is particularly i m p o r t a n t in the recognition of major strike-parallel motions in crustal deformation zones because structures recording such movements very commonly occur in narrow zones of high strain of limited areal extent that may be poorly exposed (Goodwin & Williams 1996). In addition, attempts to relate regional lineation patterns to plate motions (e.g. Shackleton & Ries 1984; Ellis & Watkinson 1987) may also be unwise in transpression zones. (5) Boundary conditions versus stress. The analysis of transpression and transtension zones has highlighted the importance of kinematic boundary conditions during deformation of the crust and lithosphere (e.g. Molnar 1992; Tikoff & Teyssier 1994). Many traditional approaches in deformation zones follow Anderson (1951) in considering stress to be the main deformation control, particularly in the brittle crust. If, however, the development of most geological structures is controlled by the strain imposed by the boundary conditions, as appears to be the case in transpression zones, then there may not be a simple or significant relationship between large-scale crustal deformation structures and stress. The authors would like to sincerely thank all the participants for helping to make the meeting in London such an enormous success. We would also like to thank R. Jones and P. Ryan for their constructive reviews of this paper, and A. Roberts, who allowed us to see his unpublished manuscript on the abuses of strain ellipse models in the hydrocarbon industry. K. Atkinson at Durham drafted the diagrams and showed great patience when the authors dithered over their ~xangles and other matters.

TRANSPRESSION AND TRANSTENSION ZONES --

References ANDERSON, E. M. 1951. The Dynamics of Faulting. Oliver & Boyd, Edinburgh. Avt~ LALLEMANT,H. G. & GUTH, L. R. 1990. Role of extensional tectonics in exhumation of eclogites and blueschists in an oblique subduction setting, northwest Venezuela. Geology, 18, 950-953. BAILEY, W. R. 1997. The structural evolution of a microplate suture zone, SW Cyprus. PhD thesis, University of Durham. BUCHANAN, R. G. & NIEUWLAND, D. A. (eds) 1996.

Modern Developments in Structural Interpretation, Validation and Modelling. Geological Society, London, Special Publications, Cox, S. A. & ETHERIDGE, M. A. 1989. Coupled grainscale dilatancy and mass transfer during deformation at high fluid pressures: examples from Mount Lyell, Tasmania. Journal of Structural Geology, 11, 147-162. DEWEY, J. E 1975. Finite plate evolution: some implications for the evolution of rock masses at plate margins. American Journal of Science, 275-A, 260-284. & LAMB, S. H. 1992. Active tectonics of the Andes. Tectonophysics, 205, 79-95. - - , HEMPTON, M. R., KIDD, W. S. E, SAROGLU,E & SENGOR,A. M. C. 1986. Shortening of continental lithosphere: the neotectonics of Eastern Anatolia a young collision zone. In: COWARD, M. P. & RIES, A. C. (eds) Collision Tectonics. Geological Society, London, Special Publications, 19, 3-36. DIAS, R. & RIBEIRO, A. 1994. Constriction in a transpressive regime: an example in the Iberian branch of the Ibero-Armorican arc. Journal of Structural Geology, 16, 1543-1554. D'LEMOS, R. S., BROWN, M. & STRACHAN,R. A. 1992. Granite magma generation, ascent and emplacement within a transpressional orogen. Journal of the Geological Society, London, 149, 487-490. DORE,A. G. & LUNDIN,E. R. 1996. Cenozoic compressional structures on the NE Atlantic continental margin: nature, origin and potential significance for hydrocarbon exploration. Petroleum Geosciences, 2, 299-312. DuTroN, B. J. 1997. Finite strains in transpression zones with no boundary slip. Journal of Structural Geology, 19, 1189-1200. EKSTROM, G. & ENGDAHL, E. R. 1989. Earthquake source parameters and stress distribution in the Adak Island region of the central Aleutian Islands, Alaska. Journal of Geophysical Research, 94, 15499-15519. ELLIS, M. & WATKINSON,A. J. 1987. Orogen-parallel extension and oblique tectonics: the relation between stretching lineations and relative plate motions. Geology, 15, 1022-1026. FITCH, T. J. 1972. Plate convergence, transcurrent faults, and internal deformation adjacent to southeast Asia and the western Pacific. Journal of Geophysical Research, 77, 4432-4460. FOSSEN, H. 1989. Indication of transpressional tectonics in the Gullfaks oil-field, northern North Sea. Marine and Petroleum Geology, 6, 22-30. 9 9 .

-

13

& TIKOW, B. 1993. The deformation matrix for simultaneous simple shearing, pure shearing and volume change, and its application to transpression-transtension tectonics. Journal of Structural Geology, 15, 413-422. & -1998. Extended models of transpression and transtension, and application to tectonic settings. This volume. & TEYSSIER, C. 1994. Strain modeling of transpressional and transtensional deformation. Norsk Geologisk Tidsskrift, 74, 134-145. GIBBS, A.D. 1986. Strike-slip basins and inversion: a possible model for the Southern North Sea areas. In: BROOKS,J., GOFF, J. C. & VAN HOORN, B. (eds) Habitat of Palaeozoic Gas in NW Europe. Geological Society, London, Special Publications, 23, 23-35. GOODWIN, L. B. & WILLIAMS, P. E 1996. Deformation path partitioning within a transpressive shear zone, Marble Cove, Newfoundland. Journal of Structural Geology, 18, 975-990. HARDING, T. P. 1974. Petroleum traps associated with wrench faults. Bulletin, American Association of Petroleum Geologists, 58, 1290-1304. HARLAND, W. B. 1971. Tectonic transpression in Caledonian Spitzbergen. Geological Magazine, 108, 27-42. HOLDSWORTH, R. E. 1989. The Start-Perranporth Line: a Devonian terrane boundary in the Variscan orogen of SW England? Journal of the Geological Society, London, 146, 419-421. 1994. The structural evolution of the Gander-Avalon terrane boundary: a reactivated transpression zone in the NE Newfoundland Appalachians. Journal of the Geological Society, London, 151, 629-646. - - - , BUTLER, C. A. & ROBERTS, A. M. 1997. The recognition of reactivation during continental deformation. Journal of the Geological Society, London, 154, 73-78. HUDLESTON, P. J., SCHULTZ-ELA,D. & SOUTHWICK,D. L. 1988. Transpression in an Archaean greenstone belt, northern Minnesota. Canadian Journal of Earth Sciences, 25,1060-1068. JAMISON,W. R. 1991. Kinematics of compressional fold development in convergent wrench terrains. Tectonophysics, 190, 209-232. JIANG, D. & WHITE, J. C. 1995. Kinematics of rock flow and the interpretation of geological structures with particular reference to shear zones. Journal of Structural Geology, 17, 1249-1265. JONES, R. R. & HOLDSWORTH, R. E. 1998. Oblique simple shear in transpression zones. This volume. & TANNER, P. W. G. 1995. Strain partitioning in transpression zones. Journal of Structural Geology, 17, 793-802. , HOLDSWORTH,R. E. & BAILEY,W. 1997. Lateral extrusion in transpression zones: the importance of boundary conditions. Journal of Structural Geology, 19, 1201-1218. KUSZNIR,N. J., MARSDEN,G. & EGAN, S. S. 1991. A flexural cantilever simple-shear/pure-shear model of continental lithosphere extension: application to the Jeanne d'Arc Basin and Viking Graben. In: -

-

-

-

-

-

14

-

J. E D E W E Y E T A L .

ROBERTS, A.M., YIELDING, G. & FREEMAN, B. (eds) The Geometry of Normal Faults. Geological Society, London, Special Publications, 56, 41-60. LANEY, S. E. & GATES,A. E. 1996. Three-dimensional shuffling of horses in a strike-slip duplex: an example from the Lambertville sill, New Jersey. Tectonophysics, 258, 53-70. MA, X. Q. & KUSZNIR,N. J. 1992.3-D subsurface displacement and strain fields for faults and fault arrays in a layered elastic half-space. Geophysical Journal International, 111, 542-558. & -1993. Modelling of near-field subsurface displacements for generalized faults and fault arrays. Journal of Structural Geology, 15, 1471-1484. MCCAFFREY,R. 1991. Slip vectors and stretching of the Sumatran forearc. Geology, 19, 881-884. 1992. Oblique plate convergence, slip vectors, and forearc deformation. Journal of Geophysical Research, 97, 8905-8915. McCoss, A. M. 1986. Simple constructions for deformation in transpression/transtension zones. Journal of Structural Geology, 8, 715-718. MICHAEL,A. J. 1990. Energy constraints on kinematic models of oblique faulting: Loma Prieta versus Parkfield-Coalinga: Geophysical Research Letters, 17, 1453-1456. MOLNAR, E 1992. Brace-Goetze strength profiles, the partitioning of strike-slip and thrust faulting at zones of oblique convergence, and the stress-heat flow paradox of the San Andreas Fault. In: EVANS, B. & WONG,T.-E (eds) Fault Mechanics and Transport Properties of Rocks. Academic Press, London, 435-459. OLDOW, J. S., BALLY,A. W. & AVE LALLEMENT,H. G. 1990. Transpression, orogenic float, and lithospheric balance. Geology, 18, 991-994. PLATr, J. P. 1993. Mechanics of oblique convergence. Journal of Geophysical Research, 98, 1623916256. RAMBERG, H. 1975. Particle paths, displacement and -

-

-

progressive strain applicable to rocks. Tectono-

physics, 28, 1-37. RAMSAY,J. G. & HUBER, M. I. 1987. The Techniques of

Modern Structural Geology. Volume 2: Folds and Fractures. Academic Press, London. ROBIN, P.-Y. E & CRUDEN,A. R. 1994. Strain and vorticity patterns in ideally ductile transpression zones. Journal of Structural Geology, 16, 447-466. SANDERSON,D. J. & MARCHINI,W. R. D. 1984. Transpression. Journal of Structural Geology, 6, 449-458. SCHWERDTNER, W.M. 1989. The solid-body tilt of deformed palaeohorizontal planes: application to an Archean transpression zone, southern Canadian Shield. Journal of Structural Geology, 11, 1021-1027. SHACKLETON,R. M. & RIES, A. C. 1984. The relation between regionally consistent stretching lineations and plate motions. Journal of Structural Geology, 6, 111-117. TEYSSIER,C., TIKOFF,B. & MARKLEY,M. 1995. Oblique plate motion and continental tectonics. Geology, 23, 447-450. TIKOFF, B. & FOSSEN,H. 1993. Simultaneous pure and simple shear: the unifying deformation matrix. Tectonophysics, 217, 267-283. & GREENE, D. 1997. Stretching lineations in transpressional shear zones: an example from the Sierra Nevada Batholith, California. Journal of Structural Geology, 19, 29-39. & TEYSSlER,C. 1994. Strain modeling of displacement-field partitioning in transpressional orogens. Journal of Structural Geology, 16, 1575-1588. WILCOX, R. E., HARDING, T. E & SEELY,D. R. 1973. Basic wrench tectonics. Bulletin, American Association of Petroleum Geologists, 57, 74-96. WOOD, D. S. 1973. Patterns and magnitudes of natural strain in rocks. Philosophical Transactions of the Royal Society of London, Series A, A283, 373-382.

Extended models of transpression and transtension, and application to tectonic settings HAAKON

FOSSEN 1 & BASIL TIKOFF 2

1Department of Geology, University of Bergen, AllOgaten 41, N-5007 Bergen, Norway (e-maik haakon.fossen@geol, uib.no) 2Department of Geology and Geophysics, Rice University, Houston, TX 77005, USA (e-maik btikoff@geophysics, rice.edu) Abstract: We introduce a spectrum of transpressional and transtensional deformations that potentially result from oblique plate interaction. Five separate types of deformation are designated, in which a simple shear deformation is combined with an orthogonal coaxial deformation. The types vary in the amount of extension v. contraction, both parallel to the margin and vertically. The interaction between the angle of convergence, kinematic vorticity, infinitesimal strain axes, finite strain, and rotation of material lines and planes is investigated. Quantification of the finite strain indicates that the orientation, magnitude, and geometry (flattening, constriction, etc.) change continually during steady-state transpression. These results are then applied to the cases of transpression, particularly resulting from oblique plate convergence of terranes. The obliquity of plate motion and the geometry of the plate margin determine which of the types of transpression or transtension is favoured. A component of margin-parallel stretching also potentially causes terrane motion to locally exceed oblique plate motion, or move opposite to the general direction of movement between the converging plate boundaries. The kinematic models also suggest that the boundaries between converging terranes are likely to exhibit vertical foliation, but either vertical or horizontal lineation. Finally, narrow transpressional zones between colliding blocks may have very high uplift rates, resulting in exhumation of high-grade metamorphic fabrics. Transpression and transtension are broadly defined as steep strike-slip influenced deformation zones that deviate from simple shear by a c o m p o n e n t of shortening (transpression) or extension (transtension) across the zone (Fig. 1). The result is a spectrum of three-dimensional

deformation which involves complex histories of fabrics, strain, and rotation. Even for homogeneous models for transpression, such as a perpendicular combination of a pure shear and a simple shear (Sanderson & Marchini 1984), the resulting d e f o r m a t i o n path is m u c h m o r e complex than for a simple shear zone. In this paper we discuss different models for transpression and transtension, and explore some of the deformation patterns and strains that can emerge. We then apply these models to simple tectonic analyses involving oblique convergence and divergence of terranes.

Approach

Fig. 1. Definition of transpression as a combination of a simple shear (wrench) component and a simultaneous coaxial shortening component perpendicular to the vertical shear plane. For volume-conservingdeformations, the latter component must be accommodated by vertical and/or horizontal extension or contraction, and variations of these variables give rise to the spectrum of transpressional deformations illustrated in Fig. 2.

The deformation within a steep deformation zone in the crust is primarily g o v e r n e d by external or boundary conditions. On the scale of plate tectonics, the boundary conditions include the relative movements of the plates involved and w h e t h e r the d e f o r m a t i o n zone extends (lengthens) vertically and/or horizontally. Vertical extension is, for example, possible if the zone reaches the surface of the crust, in which case tectonic and erosional processes allow the zone to grow upwards. Vertical contraction is feasible

15 H. FOSSEN& B. TIKOFF. 1998. Extended models of transpression and transtension, and application to tectonic settings. In: HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. E (eds) 1998. Continental Transpressionaland TranstensionalTectonics.Geological Society, London, Special Publications, 135, 15-33.

16

H. FOSSEN & B. TIKOFF

Fig. 2. The spectrum of transpressional and transtensional deformations discussed in this paper. Five reference deformations are here named types A-E, and show decreasing vertical extension (for transpression) or shortening (for transtension) from left to right. The five reference deformations differ in terms of coaxial deformation components only. Transpression or transtension as defined by Sanderson & Marchini (1984) is represented by type B. Particle flow patterns and the oblique, horizontal flow apophysis (o) are also shown.

if the zone extends to depths governed by ductile flow. Horizontal contraction or extension is controlled by the margin geometry and the divergence of plate motion (e.g. McCaffrey 1992). In this work we will assume that the zone is able to grow or shrink vertically, and locally also horizontally along strike. For simplicity, we will consider two rigid blocks with different relative velocities, and an intermediate zone of deformation. The deformation is homogeneous and determined by the relative motions of the rigid blocks.

Modelling The definition of a transpressional or transtensional zone outlined above is similar to, but more general than that of Sanderson & Marchini (1984). Sanderson & Marchini discussed the case where a vertical shear zone with horizontal displacement was modified by an additional

pure shear in a plane perpendicular to the shear plane. It is useful to consider such simple models of transpression and transtension for general reference, and then modify these if more complex models are needed. However, Sanderson & Marchini's model is only one of several related models. In this work we define five such reference deformations for transpression, and five corresponding ones for transtension. They all contain a shear system and a shortening (extension) perpendicular to the shear plane (Fig. 1), and thus fulfil the above definition of transpression (transtension). It is emphasized that slip is allowed along the margins of the transpression zone, as in the model by Sanderson & Marchini (1984). This condition is a simplification of actual b o u n d a r y conditions in many transpression zones, although less so if one considers basal boundary conditions (Teyssier & Tikoff, this volume). The results of the simple models of transpression and transtension are

TRANSPRESSION AND TRANSTENSION MODELS easily applicable in a qualitative, if not quantitative, way to geological examples.

The five reference models of oblique con vergence We consider constant-volume transpression (transtension) where the deformation is separated into a single simple shear component with vertical shear plane and a coaxial component with axes oriented perpendicular and parallel to that shear plane (Fig. 1). We keep the nature of the shear component the same for all deformations and allow for variations in the relative magnitudes of the infinitesimal stretching axes of the coaxial components. In this scheme, a spectrum of volume-conserving deformations emerges, in which we sample five transpressional and transtensional reference deformations, labelled from A to E (Fig. 2). Each of these classes of transpression implies shortening perpendicular to the shear plane (along the yaxis), and they differ only in the coaxial components in the other two directions (x- and z-axes). Type A transpression involves an equal amount of shortening along the x- and y-axes and stretching along the z-axis, type B also has vertical extension, but no stretching along the shear direction (x-axis), type C involves flattening of the shear plane (equal stretching in x and z directions), type D has constant vertical height, and in type E, the coaxial component causes a vertical shortening of equal magnitude to the shear-zone normal shortening, compensated by stretching in the x direction. Transtensional deformations involve a component of extension perpendicular to the shear plane (in the y direction). The transtensional reference deformations correspond directly to the transpressional deformations, except that the directions of the coaxial components of deformations are simply reversed. This simple relationship allows us to use the same labels (A-E) for the transtensional reference deformations. Two of the deformations (B and D) are special in that they are combinations of two plane strain deformations (a simple shear and a pure shear), and D is the only plane strain deformation in this spectrum of transpressional and transtensional deformations. Type B is identical to the model studied by Sanderson & Marchini (1984), and type D has been referred to as sub-simple shear by Simpson & De Paor (1993) (also Weijermars 1991). Furthermore, type E transpression was discussed in terms of finite strain by Dias & Ribeiro (1994).

17

The difference between the different types of transpression (transtension) lies in how the shear zone perpendicular shortening is accommodated along the shear plane. For example, in the Sanderson & Marchini model (B), the shear zone perpendicular shortening is accommodated by vertical extension (shortening). The more general cases include additional components of horizontal shortening or extension. However, all of the classes of transpression or transtension (Fig. 2) have a vertical simple shear plane and component of horizontal shortening perpendicular to the simple shear component of deformation. Having defined the deformations (Fig. 2), we investigate the related movement, kinematic vorticity, finite strain, and material line and plane rotation patterns. The mathematical relations have been outlined by Tikoff & Fossen (1993, 1995) and may be visualized for twodimensional sections using available computer programs (Tikoff & Fossen 1996).

Flow apophyses and angle of convergence If we fix a coordinate system to one of the two margins of a transpressional zone, then the observed movement direction of the opposite margin will define the direction of convergence. The angle between this direction and the strike of the transpression zone is defined as the angle of convergence, and governs the degree of noncoaxiality in the horizontal plane. An important observation is the parallelism between the convergence vector and one of the three axes called flow apophyses (or flow asymptotes). The flow apophyses are directions of maximum, intermediate and minimum rate of particle movement of the flow. In transpression and transtension, the one flow apophysis which is horizontal and (in general) oblique to the x-axis marks the convergence direction. For transpressional and transtensional deformations, the flow apophyses are described by the vectors (1, 0, 0), {~/[ln(kx/ky)], 1, 0}, (0, 0, 1)

(1)

where -/is the simple shear component (shear strain), and kx and ky are the pure shear components in the x and y directions, respectively (for derivation, see Tikoff & Fossen (1993)). The expressions show that one of the flow apophyses is always vertical (z-axis), one is always horizontal and parallel to the shear direction (x-axis), and the third is also horizontal but generally oblique to the shear direction. Particle paths are always straight along the

18

H. FOSSEN & B. TIKOFF

oblique apophysis for all types of transpression and transtension (Fig. 2). The convergence angle, o~, is the angle between the oblique flow apophysis and the x-axis, and is described by the formula o~= tan -1 [ln(kxlky)lT]

(2)

Thus, the relative components of pure and simple shear deformation are described by the convergence angle (o0 of the system. If the coaxial component is missing, the deformation zone is a simple shear wrench zone with oL= 0 ~ If the simple shear component is zero, we have a pure contractional zone (oL = 90~ As discussed by Tikoff & Fossen (1993), equation (2) is not applicable if kx = ky. This case, which is found for perfect type A transpression or transtension, is mathematically degenerate as two flow apopohyses are parallel to the x direction (e.g. Simpson & DePaor 1993). In this case, the coaxial deformation in the xyplane leads to area loss, compensated by flow in the z direction. Any change in relative magnitude of the finite strain axes in the xy-plane is caused by the simple shear component. Thus, we need a more general way to describe the ratio between the coaxial and simple shear components of any type of transpression or transtension.

Kinematic vorticity The kinematic vorticity number (Wk; Truesdell 1953) quantifies the relative rate of stretching to rotation for any three-dimensional flow. Thus, Wk can be considered a non-linear ratio between simple shear and coaxial components of deformation (Tikoff & Fossen 1995). Wk can vary between zero and one for any of the deformation types in Fig. 2. Consequently, we choose to investigate the differences of the types of transpression and transtension with respect to Wk, rather than orientation of the flow apophyses. However, for types B-E, there is a direct correlation between these two quantities (Fig. 3): oL= 90 ~ corresponds to Wk = 0 and oL= 0 ~ corresponds to Wk = 1.

Infinitesimal stretching axes During deformation, there are three directions called infinitesimal stretching (or strain) axes (ISAs). These orientations represent the directions of maximum, intermediate, and minimum elongation of material lines for an infinitesimal deformation, and are mutually perpendicular. Both the magnitude and orientation of these

90

9

70 60

"~\~?

50 a40 ~

B&E~ C,XD;,~~

30-

20 100 0.0

i

1

0.2

i

i

0.4

0.6

0.8

1.0

WR

Fig. 3. Relation between angle of convergence and the kinematic vorticity number (Wk). It should be noted that equation (2), from which the curves are calculated, is not applicable if kx = Icy (type A transpression or transtension). This case is mathematically degenerate, as both horizontal flow apophyses are parallel to the x direction (no convergence direction defined).

axes depend on the type of flow during deformation. They vary within the spectrum of deformations discussed in this paper (also within each of types A - E in Fig. 2), but by definition do not change during steady-state deformation. Both boundary conditions (Fig. 2) and kinematic vorticity number (Wk) control the orientations of the ISAs. Because Wk can vary within any of the deformation types in Fig. 2, so can the orientations of the ISAs. One of the ISAs is always vertical (parallel to x-axis), implying that the other two always lie in the horizontal (xy) plane. The two horizontal ISAs are oblique to the coordinate axes (x and z), as a result of the simple shear component. They are parallel to the coordinate axes only if deformation is perfectly coaxial. If the d e f o r m a t i o n is perfect simple shear in the horizontal plane, the largest horizontal ISA is inclined at a 45 ~ angle to the shear direction (x-axis). For intermediate cases (0<Wk 0.87

TRANSPRESSION AND TRANSTENSION MODELS

19

Fig. 4. Orientation of infinitesimal strain axes (ISAs) and finite strain axes (X) for the spectrum of transpressional or transtensional deformations discussed in the text. ISA> largest infinitesimal stretching direction; kl, maximum finite stretch direction. It should be noted that the ISAs have a constant orientation during steady-state deformations, whereas the finite strain axes rotate (for Wk>O) and may exchange positions. Wk = 1: simple shear (top). Wk = O: coaxial deformation (bottom). ISA2 becomes vertical (e.g. simple shear dominated). Similarly, for Sanderson & Marachini's model of transpression (type B), the orientation of ISA1 is in the horizontal plane only for 1>Wk>0.81 (simple shear dominated), and vertical for Wk<0.81 (pure shear dominated). If a stretch is added to the horizontal shear direction (x), the shift between ISA1 and ISA2 occurs at lower Wk values, and when the stretching along the x-axis matches or is greater than that along the z-axis (type C transpression), ISA2 is always vertical for any Wk. ISA2 is also vertical for type D to E transpression. The orientations of the ISAs depend on W k and the boundary conditions for transtensional deformations in a similar way, as mapped in Fig. 4 (right-hand half). The distinction is that for type A and B transtension, ISA 3 (rather than ISA1) is oriented parallel to the z-axis for coaxially dominated deformation. Change in the orientation of ISAs during a single deformation history implies that deformation is non-steady state. Hence, Wk, the angle of convergence or divergence, and the orientation and magnitude of ISAs remain constant throughout steady-state deformations. Finite strain

As opposed to the ISAs, the orientation, magnitude and shape (flattening, constriction, etc.) of the finite strain ellipsoid change continually during steady-state transpression. The change in orientation is a result of the simple shear component of deformation, and the change in

geometry reflects the interaction of the simple shear and coaxial components of deformation. The initial orientations of the finite strain axes are identical to those of the ISAs (Fig. 4). Hence, for all of the transpression and transtension classes in Fig. 2, one of the finite strain axes will always be vertical, parallel to one of the flow apophyses. The other two finite strain axes lie in the horizontal xy-plane, with the larger horizontal strain axis (hi or he) rotating towards the shear (x) direction (also a flow apophysis) during progressive deformation. The long and intermediate finite strain axes rotate into parallelism with the flow apophyses, which act as attractors of the finite strain axes (Passchier 1997). Thus, for transpression, the horizontal finite strain axes rotate towards the coordinate axes. For transtension, the horizontal finite strain axes may be oblique to coordinate axes with one axis parallel to the oblique flow apophysis. The orientation of the strain ellipsoid at any point during deformation depends on deformation type (A-E), angle of convergence, and the non-coaxiality of deformation (Wk). In terms of the pure shear components in the horizontal plane and the simple shear component, the angle dObetween the larger horizontal strain axis (Xl or Xe) and the x-axis is given by the formula qb = arccos {[-kyF/(C 2 + kx 2 - X)]}

(3)

where F = y(kx-ky)/[ln(kx/ky)] (Tikoff & Fossen 1993). Equation (3) holds for the entire spectrum of deformations discussed in this work.

20

H. FOSSEN & B. TIKOFF

For the subspectrum of deformations between types C, D, and E, the intermediate principal strain axis (X2) is vertical for any W k (this applies to both transpression and transtension). For the other deformations, the axes of the finite strain ellipsoid may switch positions for high values of Wk. The occurrence of such switches depends on W k and the interplay between the simple shear and the coaxial deformation components. The simple shearing component produces maximum stretching in the horizontal plane (initially at 45 ~ to the shear direction). If the vertical extension imposed by the coaxial deformation is strong (left of type C in Fig. 4, including types A and B), two possibilities exist. The first is that the finite strain axes are subparallel to the coaxial components throughout deformation. Alternatively, for high-Wk (simple shear-dominated) transpression, kl still may initiate with a horizontal orientation, but does not grow as quickly as the vertical principal strain axis (X2). Thus, if deformation proceeds long enough, the result is that kl and k 2 switch positions. In steady-state type A and B transpressions, the switch may only occur if W k > 0.87 (type A) or 0.81 (type B; Fig. 4). Similarly, k2 changes to k3 for the corresponding transtensional reference deformations (types A and B). The reason for such switches is that coaxial deformations accumulate strain faster than simple shear (Tikoff & Fossen 1995). Consequently, stretching caused by the simple shear component may dominate the strain in the early history, but eventually is overcome by the increasingly effective coaxial deformation at a later stage. Switches in strain axes were first described for type B transpression by Sanderson & Marchini (1984), and occur as the strain exhibits a perfect flattening (transpression) or constriction (transtension) geometry. This coincides with the strain path touching the horizontal axis of the Flinn diagram (see A and B transpression in Fig. 5). Furthermore, the boundaries between deformations where finite strain axes do and do not switch positions during progressive steady-state deformation trace deformations where the first strain increment produces a perfect flattening geometry (constrictional geometry for transtension). In general, transpressional (transtensional) deformations in the vicinity of these lines develop strains very close to pure flattening (constriction). Fig. 5. Steady-state (constant Wk and e0 strain path in Flinn space for the five reference transpressional deformations A-E. (Note the large variations in paths and strain geometry. See text for discussion.)

TRANSPRESSION AND TRANSTENSION MODELS

21

Fig. 6. Variation in shape of strain ellipsoid for transpression or transtension. Perfect coaxial strains along bottom (Wk = 0) line. Simple shear along upper, horizontal boundary. The strain ellipsoids at the bottom apply for Wk = 0 and illustrate the shape and orientation of the coaxial components of three of the reference deformations (A, C, and E) of transpression and transtension.

More important than switching strain axes is perhaps the geometry of finite strain and the strain path in general for the different types of transpression. Figure 5 shows a selection of steady-state (constant Wk) transpression strain paths (Wk = 0.98, 0.95, 0.85, 0.5, and 0.2) (the corresponding paths for transtension are found by reflecting the transpression paths about the Wk = 1 line). The simplest pattern is developed during type D deformations, which is a plane strain sub-simple shearing. It is well known that plane strain deformations develop along the line with slope of one (Flinn k value of one), regardless of the degree of non-coaxiality of the deformation. Type C deformations develop a spectrum of strain paths between simple shear (plane strain; Wk = 1) and perfect flattening (Wk = 0). However, strain becomes increasingly oblate during deformation, and strong flattening fabrics are characteristic even for relatively low strains (unless deformation is very close to simple shear). Type E deformations behave identically to type C, except that strains are constrictional instead of flattening. Type B transpression develops flattening strains, particularly for Wk values between c. 0.6 and 0.95. Type A transpression is different because (1) any strain geometry (oblate or prolate) can be achieved (depends only on Wk), and (2) steady-state paths with Wk>0.5 may

cross the diagonal Wk = 1 line. If so, strain develops from a flattening fabric to a constrictional one for transpression (but never from constriction to flattening) and vice versa for transtension. The geometry of the strain ellipsoid is qualitatively shown in Fig. 6 for the spectrum of deformations discussed in this work. It should be noted how type D transpression and transtension separate deformations that produce oblate and prolate strain shapes, and how perfect constriction and flattening throughout deformation only occur along the base of the diagram (Wk = 0). Finite strain has implications for fabric development in deformed rocks. It is generally true that constrictional strain leads to strongly lineated rocks (L-fabrics) whereas flattening strain results in strongly foliated rocks without a lineation (S-fabrics). Intermediate cases give rise to LS- and SL-fabrics. Although it is shown above that the geometry of finite strain changes during deformation in most cases, it is possible to predict the finite fabric in rocks deformed by the various types of transpression and transtension discussed in this work. Such fabric variations are shown in Fig. 7 for reasonable amounts of finite strain. In this figure, which is based on Figs 4-6, type D transpression and transtension separate fields of S- and L-dominated fabrics, and the upper Wk=l (simple shear) boundary

22

H. FOSSEN & B. TIKOFF

Fig. 7. Variation in type of fabric for the different types of transpression and transtension discussed in the text. marks fabrics where planar and linear fabrics are equally well represented in the ideal case.

Rotation of lines Analytical work has revealed that lines and planes rotate away from a certain direction (usually a line orientation, less commonly a plane), commonly referred to as the source, fabric repellor or repulsor. Lines and planes rotate towards a different direction (or plane) known as the sink or fabric attractor (Passchier 1997). These directions are governed by the orientations of flow apophyses (see above), which are directions of maximum, intermediate and minimum rate of particle movement of the flow. Passive rotation of linear and planar markers can be modelled mathematically by use of the deformation matrix, as described by Flinn (1979) and explored for type B transpression or transtension by Fossen et al. (1994). The result of such modelling (Fig. 8) shows a range of patterns that characterize the different classes of transpression. For type A transpression, a rotational movement of line markers towards vertical is characteristic. This 'down-the-drain' pattern is only perfectly developed if the two coaxial components along the shear plane are of equal magnitude. If not, an oblique flow apophysis emerges and modifies this pattern somewhat. However, for all transpressional deformations between types A and C (Fig. 8), lines will always rotate towards the vertical direction. This pattern changes across type C transpression, where the two coaxial components along the shear plane are of equal magnitude. Here material lines rotate from an oblique, horizontal orientation along great circles until they meet the vertical shear plane and essentially stop

rotating (fly-paper effect). Hence, line orientations are expected to be distributed along the shear plane for high strains. However, if the vertical stretching is reduced or reversed in magnitude (including types D and E transpression), lines are finally parallel to the horizontal shear direction. Type E transpression is peculiar insofar as lines rotate away from an oblique plane defined by two of the flow apophyses (planar source). In general, lines are finally vertical in transpression if the vertical stretch is significant, but rotate towards the horizontal shear direction if the vertical stretch is small or negative. For transtension, on the other hand, lines tend to rotate towards a horizontal direction which is governed by the oblique flow apophysis. The exception to this general rule occurs in the case when an additional vertical component of stretching occurs (type E in Fig. 8). In this case lines are attracted to, and eventually distributed along, a plane containing the oblique flow apophysis and the z-axis (fly-paper effect). An important difference between transpression and transtension is that whereas lines rotate towards parallelism with the x- or z-axis in transpression, lines generally rotate towards a horizontal direction oblique to the shear direction (x) in transtension (compare type B, C and D transpression with transtension in Fig. 8). In both cases a large angle is expected between rotated linear structures and the shear direction.

Rotation of planes Rotation patterns for passive planar markers (poles to planes) in transpression are also shown in Fig. 8. The shortening perpendicular to the shear zone generally causes planes to rotate towards a vertical position, eventually parallel to

TRANSPRESSION AND TRANSTENSION MODELS LINES

x

PLANES

x

fU

X

x

x

@

I LINES

P

23

x

x

x

x

x

Y

Fig. 8. Rotation patterns of lines and poles to planes for Wk = 0.75. Squares are sources; triangles are sinks or attractors. Planar sources and sinks are shown by stippled lines. (See text for discussion.) the shear zone. A n exception to this is (rare) type E transpressional zones, which shorten vertically and extend in the shear direction. In this case planes end up with a common strike direction (parallel to the shear direction) but with any amount of dip. In transtension (Fig. 8, lower part), planes rotate towards a horizontal position if the shear zone perpendicular extension is mainly compensated by vertical shortening (left of type C). If shortening in the shear direction is significant (right of type C), planes rather rotate into a vertical orientation which is oblique to the shear plane.

Complicating factors The characteristics explored above for various types of transpression (and transtension), such

as finite strain, rotation patterns and fabrics, are not unique to a certain type of deformation. However, a combination of several characteristics may help to constrain the boundary conditions and kinematic vorticity. Table 1 shows the characteristic features for type A - E transpression. Although natural deformation zones may have additional complications that may restrict the usage of this table, it is useful for a general classification of the deformation in question. C o m m o n complications include heterogeneous deformation, non-steady-state deformation, and strain partitioning (see below). Strain is not likely to be h o m o g e n e o u s throughout the zone, but rather: (1) increases in intensity towards the central part of the zone, and/or (2) preferentially partitions a component of the simple shear deformation into a high

24

H. FOSSEN & B. TIKOFF

Fig. 9. Heterogeneous strain across a transpression zone can be modelled by use of several deformation boxes with different finite strain. The example shown models heterogeneous, steady-state type B transpression with maximum strain in the central part of the zone. strain zone. In the first case, the kinematic vorticity number and boundary conditions may be constant. The deformation zone can be modelled by using multiple deforming elements instead of just one (Fig. 9), and Table 1 is still applicable. However, the shape and the orientation of the finite strain ellipsoid generally change during deformation, even if W k and the angle of convergence do not. Of particular interest are the cases where k2 and ~k1 switch positions during progressive deformation (e.g. type A and B transpression; see Fig. 4). In the extreme case (type A transpression with W k around 0.6), oblate marginal strain ellipsoids may change to prolate ellipsoids towards the centre. Similarly, low-strain parts of the deformation zone may show horizontal lineations whereas more intensely deformed parts exhibit vertical lineations (switch of k2 and kl) (Fig. 4). For transtension, low-strain zones may have vertical foliation and high-strain zones have horizontal foliation (switch of )t2 and k3). Any partitioning

of the simple shear component will tend to accentuate such effects (Tikoff & Teyssier 1994; Tikoff & Greene 1997). Non-steady-state transpressional deformations are easily modelled as a number of incremental deformations with different kinematic vorticity number and/or different boundary conditions (e.g. a change from type B to type C transpression). However, the problem requires knowledge of the change in external boundary conditions, which is difficult or impossible in most cases (see, however, Fossen & Tikoff 1997). A change in convergence angle along an obliquely convergent plate margin is the simplest case that could be modelled very easily (as sequential deformations). Evidence of strain partitioning is easily observed in the field as domains of low kinematic vorticity numbers (strong component of shortening across the shear zone) between strike-slip dominated faults or narrow shear zones of high shear strain (e.g. Tikoff & Teyssier 1994; Jones & Tanner 1995; Kirkwood 1995; Goodwin & Williams 1996; Northrup & Burchfie11996).

Tectonic application of transpression and transtension Obliquely convergent plate boundaries are generally characterized by a contractional orogenic wedge tens to hundreds of kilometres wide (e.g. Vauchez & Nicolas 1991). Relative plate motion and the plate boundary are proposed to provide the primarily boundary conditions for continental deformation (e.g. McKenzie & Jackson 1983; Molnar 1992; Teyssier et al. 1995). Proof of this relation is the generally good correlation between neotectonic strain rates from geodetic surveys and relative plate motions. This relation is relatively straightforward in continental interiors and orthogonal collisional belts,

Table 1. Characteristics for the five transpressional reference deformations in Fig. 2; a similar scheme can be made for transtension, using the results" presented in the other figures and in the text Type

Strain

A transpr,

any

B transpr,

oblate

C transpr, D transpr, E transpr,

oblate plane prolate

kl (lineation) vertical (horizontal) vertical (horizontal) horizontal horizontal horizontal

Fabric

ISA 1

L, S

vertical, horizontal vertical, horizontal horizontal horizontal horizontal

LS, S S, SL S=L SL, L

Linear markers

Planar markers

--)vertical

--~vertical (any strike)

~vertical

~vertical (Hshear plane)

---)yz-plane --)horizontal ~horizontal

---~vertical(][shear plane) -~vertical 0[shear plane) any dip, strike normal to oblique flow apophysis

TRANSPRESSION AND TRANSTENSION MODELS

25

but also applies in more complex areas of oblique convergence (e.g. the San Andreas fault system; see Teyssier & Tikoff, this volume). Thus, we investigate the effect of the boundaries of obliquely convergent systems on the style of transpression. Movement

a n d slip rates

D e f o r m a t i o n within an obliquely collisional zone depends primarily on the direction of the plate motion relative to the plate margin orientation. Plate motion is commonly decomposed into a normal and tangential component (e.g. Engebretson et al. 1985). However, the relation between the plate motion and the deformation is complex, except for the case where the movement is perfectly orthogonal to the plate boundary. In regions of oblique convergence or divergence, the motion, the infinitesimal strain axes (ISAs), and the finite strain axes are not parallel (Fig. 10b). In all models of transpression (types A - E ) with any wrench component, there is an angular difference between the plate motion (oblique flow apophysis) and the fastest horizontal shortening direction (ISAHmin, equal to ISA 3 if ISA 3 is horizontal or ISA2 if ISA 3 is vertical). The finite strain long axis is originally parallel to the fastest horizontal stretching direction (ISAHmax, equal to ISA1 or ISA2), but rotates toward a margin-parallel orientation. Progressive simple shearing (wrenching) is a good end-member example of this behaviour. Experimental deformation, even applied to fracturing, for example, by Withjack & Jamison (1986), corroborates the numerical modelling. In their transtensional model (type B), normal faults form slip parallel to the ISAHmax, not the plate motion. It is a relatively common misconception in the geological and geophysical literature that the slip direction on faults (representing the infinitesimal contraction direction or ISA) is parallel to the plate motion vector (e.g. Mount & Suppe 1992; Yu et al. 1993) (Fig. 10a). This idea of 'uniaxial' deformation is not applicable to obliquely convergent boundaries, because it fails to account for the wrench component of deformation. Parallelism of the plate motion and the slip vectors only applies if (1) the deforming zone is orthogonal to the motion (oL -- 90 ~ or (2) the deforming zone is offset along a strike-slip fault so that the margin locally is orthogonal to the motion vector (Fig. 10c). Typically, the orientation of the slip vectors is at a higher angle to the plate boundary than the plate motion vector (e.g. Yu et al. 1993) (Fig. 10b), similar to

v

v

v

v

9

L . 4 ~ ~ Y Y ~ S,/~HSi AHmax b.

v',~--vPlatemotion

c

d.

\

%

x

Fig. 10. The relationship between plate motion (convergence vector), plate boundary orientation, infinitesimal strain axes (ISAHmaxand ISAHrnin), and finite strain axes in the horizontal section for a transpressional system. (a) The common misconception is that the plate motion vector is parallel to the infinitesimal contraction direction (ISAHmin) or the finite contraction direction, a geometry that cannot accommodate the wrench component. In general, the plate motion, ISAs, and finite strain axes are all differently oriented (b) except for special cases where the margin is locally orthogonal to the motion vector (c). (d) Strike-slip partitioning increases the misfit between the plate motion and the ISAHmin.

26

H. FOSSEN & B. TIKOFF

numerical predictions for transpression. For example, on the South Island of New Zealand, an obliquely convergent setting which does not exhibit strike-slip partitioning, the slip on increasingly recent faults is oriented at a systematically higher angle than that predicted from the plate motion direction (Norris et al. 1990; Teyssier et al. 1995). If strike-slip partitioning occurs, the angle of contractional slip vectors should be oriented at an even higher angle to the plate margin (Fig. 10d). Conditions leading to type B transpression or transtension The orientation of the relative plate motion, its along-strike variation, and the geometry of the plate margin will have major effects on the resultant deformation (e.g. Beck 1991; McCaffrey 1992). In a straight plate margin with a constant angle of plate convergence, type B transpression is the most likely deformation. The reasoning is based on the nature of the flow apophyses and particle paths discussed above. The map-view flow lines created in such a tectonic setting must move directly toward each other, as no gradient of deformation will occur along the collisional boundary. Rather, deformation will accumulate by vertical movement of material. In type B transpression, flow lines are straight when viewed in the horizontal plane (Fig. 2). However, vertical movement is accommodated by a pure-shear component of deformation. It should be noted that, at high crustal levels, this 'bulk' pure shear component of deformation is attained along opposite dipping thrust faults and related folds, such as in a flower structure. Conditions leading to type A , (7, D a n d E transpression or transtension A component of arc parallel stretching (Fig. 11) occurs if: (1) the plate margin has a convex orientation or (2) the angle of relative plate motion changes along strike, such that the tangential component of the relative plate motion increases along the plate margin. This situation results if the oceanic plate has a nearby pole of rotation, located within the oceanic plate (Av6 Lallemant & Guth 1990; Beck 1991; McCaffrey 1992,1996). In these cases, the material is forced to stretch parallel to the margin, to maintain compatibility. The resulting types of deformation are, with an increasing component of arc-parallel stretching, types C, D, and E. The amount of divergence and curvature will ultimately determine which deformation occurs. In

Fig. 11. The effect of divergent or convergent displacement fields and non-planar boundaries (margins) of transpression zones on the resulting type of transpression. Type B transpression would be expected from a setting with straight margins and a homogeneous displacement field (not shown). Based on McCaffrey (1992). terms of movement in the horizontal plane, curved particle paths are necessary to move material away from the colliding unit. An extreme case of margin-parallel extension occurs as a collision, either of a crustal salient or a triple-junction. The effects of such a plate margin will probably be most extreme in the fore-arc region, in which the stretching will be maximum (Av6 Lallemant & Guth 1990). Very prominent, margin-perpendicular normal faults are expected and may involve structures similar to core-complexes. Rapid uplift of deep-seated material, such as blueschists, has been proposed, as a result of such kinematics (Av6 Lallement & Guth 1990). If type D transpression is suggested, the plane-strain aspect of this deformation is easily accommodated by conjugate strike-slip faults (Weijermars 1993). The reverse case, involving a component of arc-parallel stretching, occurs if (1) the plate margin has a concave orientation, such as a

TRANSPRESSION AND TRANSTENSION MODELS large-scale salient or (2) the tangential component of the relative plate motion decreases along the plate margin, caused by a nearby pole of rotation within the adjacent continental plate (Fig. 11; Beck 1991; McCaffrey 1992, 1996). In these cases, a transition from type B to type A transpression is expected, and all of the normal component of oblique plate motion and a percentage of the transcurrent motion (resulting in orogen-parallel shortening) is accommodated by vertical uplift. Thus, this type of plate interaction is extremely efficient at uplifting rocks quickly. In the brittle regime, the margin-parallel and margin-normal shortening will both result in contractional structures. Thus, dome and basin structures may result from folding, or two separate, sub-perpendicular orientations of thrust faulting may occur. Strike-slip p a r t i t i o n i n g transpression

Fitch (1972) originally observed that the oblique plate motion is commonly accommodated by a large strike-slip fault (e.g. Great Sumatra fault) and regions undergoing primarily contractional deformation. However, as was routinely noted (e.g. Oldow et al. 1989), regions on both sides of these major structures commonly contain a large component of wrench deformation. In particular, Jamison (1991) demonstrated that a significant percentage of wrench motion is accommodated within the structures adjacent to major strike-slip faults. This process of strikeslip partitioning has been kinematically modelled for type B transpression (Fossen et al. 1994; Tikoff & Teyssier 1994; Jones & Tanner 1995; Krantz 1995) and type B transtension (Fossen et al. 1994; Teyssier et al. 1995). Any of the models of transpression or transtension (A-E) could exhibit strike-slip partitioning behaviours if the partitioning is caused by a tendency to localize deformation. However, in a dynamic explanation of the process, Tikoff & Teyssier (1994) suggested that strike-slip partitioning is caused by a major misorientation of infinitesimal strain and finite strain axes. If this reasoning is correct, strike-slip partitioning is limited to types A and B transpression (e.g. Tikoff & Teyssier 1994).

Transpressional terrane tectonics A common observation in orogenic belts is the accumulation of small fragments of oceanic or continental material, or tectonic terranes, within areas of oblique convergence (e.g. Coney et al. 1980). The western edge of North America, for instance, is a well-known example

27

of terrane accretion. Because these terranes collide obliquely with the irregular margin of North America, the kinematic models of transpression are relevant in describing their deformation. This analysis follows the concept of transpressional terranes proposed by Oldow et al. (1989). Terrane b o u n d i n g faults a n d shear z o n e s

It is commonly considered that major transcurrent motion only occurs on vertical faults or shear zones with sub-horizontal lineations. It has been repeatedly pointed out (Fossen & Tikoff 1993; Fossen et al. 1994; Robin & Cruden 1994; Tikoff & Teyssier 1994) and documented in geological field examples (Hudleston et al. 1988; Tikoff & Greene 1997) that the lineation direction in transpressional shear zones does not reflect the tectonic movement direction of the simple shear component of deformation. Strain modelling indicates that it takes a relatively low angle of convergence (c~<20~ to cause a vertical lineation within a type B transpressional shear zone. If the transpressional deformation is intermediate between type B and type A transpression, the angle of convergence is even lower for vertical lineations to form. These observations are particularly appropriate for terrane boundaries, which typically involve components of both contraction and strike-slip motion. Alternatively stated, terrane boundaries are generally transpressional. Finite strain patterns for the types of transpressional deformations are (1) vertical foliation with horizontal lineation or (2) vertical foliation with vertical lineation. Hence, the orientation of the foliation is a more appropriate guide for recognizing transpressional shear zones than is the lineation. In fact, models of transpressional deformation which include compatibility restrictions (Robin & Cruden 1994) suggest an even more complicated pattern of lineation development but a comparatively straightforward pattern of foliation development. Therefore, the rotation of the foliation associated with a strain gradient may be the most useful guide to the vorticity. A problem is that vertical lineations may also result from highangle reverse faulting. However, the existence of vertical to steeply oriented shear zones commonly found near terrane collisions is not consistent with thrust tectonics. Relative movement on a vertical dip-slip fault does not achieve crustal contraction, it merely juxtaposes different crustal levels. In short, no tectonic effect is accomplished on vertical dip-slip faults except local accommodation.

28

H. FOSSEN & B. TIKOFF

Case example: coast shear zone and Baja-BC The terrane movement of the Baja-British Columbia (Baja-BC) terrane against North America serves as a useful example for the significance of lineations in tectonic interpretation (Fig. 12). A major debate in the evolution of the North American Cordillera involves differences in geological-based small-scale versus palaeomagnetic-based large-scale estimates of marginparallel terrane transport. Palaeomagnetic constraints on the Baja-BC terrane require 3000 km of dextral movement of the Baja-BC terrane relative to North America, and 2000 km with respect to the intermediate Intermontane terrane, in the 40 Ma interval between 90 and 50 Ma (e.g. Irving et al. 1995; Wynne et al. 1995). Geologically derived estimates on observed strike-slip faults are of the order of 500-1100 km (e.g. Gabrielse 1985; Struik 1993). The Late Cretaceous-early Tertiary Coast shear zone is the best candidate for accommodation of large-scale motion of the Baja-BC terrane. Extending over 1000 km along the western flank of the Coast Mountains in southeastern Alaska and coastal British Columbia, the Coast shear zone must record 2000 km of strike-slip motion if the palaeomagnetic data are correct (Cowan 1994). This interpretation is supported by plate motion reconstructions that suggest a strong oblique component to subduction along the northeast Pacific margin from c. 85 to 50 Ma (Engebretson et al. 1985). The Coast shear zone is a 5-15 km wide, steeply oriented shear zone with sub-vertical foliation and steep, down-dip lineations (Crawford & Crawford 1991; McClelland et al. 1992). The Coast shear zone displays ambiguous senseof-shear indicators parallel to the lineation direction, although the movement is interpreted as an east-side-up thrust fault derived from the sense-of-shear indicators parallel to the lineation (Ingram & Hutton 1994). This interpretation is at odds locally with the metamorphic geology, which exposes deeper rocks on the west side of the fault compared with the east side (e.g. McClelland et al. 1992). In contrast, sense-ofshear indicators perpendicular to the lineation are dextral, although domainal (e.g. Stowell & Hooper 1990). The existence of only steeply oriented lineations has been used to rule out the possibility of large-scale horizontal movement (e.g. Ingram & Hutton 1994). This evaluation is in direct contrast to strain modelling of transpressional shear zones, which suggests vertical lineations particularly in high-strain zones. Although the Coast shear zone records some

Fig. 12. Geological map of the Canadian Cordillera, indicating the location of the Coast shear zone and the Baja-BC terrane discussed in the text. component of east-side-up movement (Ingram & Hutton 1994), it may have accommodated much larger amounts of strike-slip motion. The existence of steeply oriented, down-dip lineations does not rule out the possibility of major (hundreds to thousands of kilometres) strikeslip motion on the Coast shear zone.

Faster than plate motion? Opposite to plate motion? The nature of margin-parallel movement of terranes depends on (1) the kinematic model of oblique convergence and (2) the amount of strike-slip partitioning between the terrane and the adjacent continent. For transpressional deformations that involve a component of margin-parallel extension, an interesting effect

TRANSPRESSION AND TRANSTENSION MODELS occurs (Fig. 13). In this case, the motion of the terrane results in a combination of the transcurrent motion and the extrusion caused by the coaxial component of deformation. Parts of the terrane actually move faster than the relative plate motion, where these two effects are combined. Other sections of the terrane move much slower than, or even in an antithetical direction to (opposite to the tangential component of) plate motion. The latter case occurs if the coaxial deformation is larger than the transcurrent component. This effect could occur for either homogeneous or partitioned transpression (Fig. 13). Strike-slip partitioning enhances the effect of the coaxial component of deformation because the transcurrent component is localized. Thus, a larger percentage of the terrane moves faster or opposite to plate motion. The upper-crustal structural manifestation of margin-parallel extension is a conjugate set of strike-slip faults or margin-perpendicular normal faults (e.g. McCaffrey 1992). However, these same structures may also result from type B transpression or transtension (Withjack & Jamison 1986; Jamison 1991), with no margin-parallel extension, as a result of horizontal elongation caused by wrenching. An example of this behaviour is recorded by palaeomagnetic data from the Baja-BC terrane. Southern parts of the terrane indicate dextral transcurrent offsets of c. 2900 km (e.g. Mt Stuart; Irving et al. 1995; Wynne et al. 1995). Recent analysis of Cretaceous-Paleogene plate motions adjacent to North America (e.g. Kelley 1993; Kelley & Engebretson 1994) suggested less offset than previous reconstructions (Engebretson et al. 1985). Assuming that the Kula plate was adjacent to North America, 2900 km of offset is consistent with plate motion. However, palaeomagnetic data from the northern part of the terrane suggest c. 4000 km of dextral translation, which is approximately equal to or slightly higher than plate motion (Irving et al. 1996). The difference in offset between the southern and northern part of the terrane suggests margin-parallel stretching. Thus, we suggest that the observation of the northern part of the terrane moving faster than plate motion is geologically permissible, particularly given the evidence for strike-slip partitioning during northward movement of the Baja-BC terrane.

E x h u m a t i o n o f deep crustal rocks Strike-slip tectonism is commonly discounted as an effective mechanism for uplifting deep crustal rocks (e.g. Platt 1993). This idea overlooks the transpressional or three-dimensional nature of

29

Fig. 13. The effect of margin-parallel extension on terrane motion. Modelling suggests that terrane motion faster than, or opposite to, the tangential component of plate motion is geologically acceptable. strike-slip zones, i.e. a component of wrench and normal convergence. The three-dimensional nature of transpression provides the two components needed to uplift deep crustal rocks: (1) a vertical conduit supplied by the wrench component; and (2) vertical flow from the contractional component. Transpressional systems are inferred to extend to deeper crustal levels as nearly vertical structures, a geometry that allows the wrench component of movement. Using this conduit, upward movement of lower-crustal rocks from deep to shallow settings is possible during a single deformation. If the lower-crustal rocks are able to flow, an analogy is possible to magma emplacement in transpressional or strike-slip settings. Similar to magmas, lowercrustal rocks may be 'emplaced' in structural voids in an overall transpressional environment, such as between en echelon P-shears (Tikoff & Teyssier 1992). If there are rheological contrasts and the flow field departs from homogeneity, pods of higher-pressure (and higher-density) rocks may be brought up by the flow in the transpressional zone, analogous to entrained xenoliths. The contractional aspect is responsible for the rapid upward movement of lower-crustal rocks. In all of the transpressional models (types A - E ) , this upward movement is accommodated by the coaxial component which is more efficient than

30

H. FOSSEN & B. TIKOFF

the wrench component (Tikoff & Teyssier 1994). Type C, D, and E transpression causes marginparallel extension, which is capable of exhuming crustal rocks by causing extension perpendicular to the margin trend (Av6 Lallement & Guth 1990). An equally effective mechanism of exhumation is type A and B transpression combined with erosion. In these cases, the normal component of plate movement is accommodated by vertical uplift of rocks. This vertical movement is visible in particle path movement and material line rotation (Fig. 8), and also indicated by a vertical axis of maximum finite strain. If the deformation zone has a relatively narrow width, erosion rates can keep pace with tectonic rates, leading to effective exhumation. As discussed above, type A and B deformation is favoured by straight to concave margins. Modern analogies of this setting are the 'bend' in the San Andreas system and in the Bolivian segment of the Andes, with attendant uplift of the Tranverse Ranges and Altiplano. E x h u m a t i o n o f blueschist belts by terrane collision A good example of the role of transpressional tectonics in the exhumation of deep crustal rocks involves the presence of high-pressure rocks (e.g. blueschists). Because of the density of these rocks, it is commonly recognized that buoyancy did not result in their upward movement (Hs0 1991). Many models have been developed for the exhumation of these assemblages to the surface of the Earth, most involving flow in subduction zones (e.g. Cloos 1982; Platt 1986). Whereas the blueschist mineralization clearly formed in a subduction-zone setting, this is not necessarily the cause of its exhumation. The outboard existence of colliding terranes potentially explains the uplift of long, linear trends of blueschist commonly found in orogenic zones. Although linear, margin-parallel zones of high-pressure mineral assemblages presumably formed in subduction zone settings, they are commonly found in areas of collisional belts. Thus, the uplift of high-pressure rocks, and blueschists in particular, is broadly related to collision dynamics (Hsa 1991). Terrane collision would effectively stop subduction and, if oblique, result in transpressional kinematics. As discussed above, transpression zones are effective at uplifting material because of the wrench-induced vertical conduit and the contraction-induced uplift. Further, in a terrane collision setting, the tectonic thickening is potentially accommodated in a narrow zone. A

good example of this behaviour is thickening and extrusion in the form of flower structures, particularly along the San Andreas fault (e.g. Harding 1985). Because the zone is relatively narrow, erosion may be very efficient, leading to increased exhumation (e.g. Beaumont et al. 1992). In this way, metamorphic reactions from the subduction setting do not equilibrate, and high-grade mineral assemblages are preserved by rapid uplift (Thompson et al. 1997). For example, a transpressional terrane collision may apply to the classic Franciscan complex of California, which is at present bounded on its west side by the Salinian block. 4~ and K/Ar ages indicate the last phase of blueshist metamorphism at c. 92 Ma (Cloos 1985), which corresponds generally to cessation of the Sierra Nevada magmatic arc (83 Ma; Stern et al. 1981). The uplift of the Franciscan complex is Late Cretaceous to early Paleogene, as indicated by pebbles of the Franciscan blueschists found in the Paleocene Great Valley sequence (Berkland 1973; Harms et al. 1992). This time also corresponds to proposed dextral strike-slip faulting of the Salinian block (proto San Andreas fault; Nilsen & Clarke 1975) and the inferred northward movement of Baja-British Columbia (Beck 1986). Thus, uplift of the Franciscan blueschists is hypothesized to be a result of terrane collision, rather than 'business-as-usual' subduction.

Conclusions We have defined a spectrum of transpressional (and transtensional) deformations which have in common a simple shear (wrench) component and a perpendicular shortening component. Five reference deformations in this spectrum are named A - E , in which type B involves no extension or shortening along strike of the zone (transpression or transtension of Sanderson & Marchini (1984)). Type B is probably the most common type of transpression or transtension, but irregularities such as curved plate boundaries may give rise to other types. Simple mathematical modelling shows that the orientation and geometry of the finite strain ellipsoid, fabric type (L, S, SL, etc.), fabric orientation, maximum infinitesimal stretching direction, kinematic vorticity, and/or rotation pattern of linear and planar features can together be used to distinguish between the different types of transpression (transtension) (Table 1), and to determine the angle of convergence. During heterogeneous deformation, several of the parameters listed in Table 1 will show large variations as W k and finite strain vary across the zone

TRANSPRESSION AND TRANSTENSION MODELS (e.g. strain and fabrics for type A transpression). Also, partitioning of strain in transpression or t r a n s t e n s i o n z o n e s implies t h a t o b s e r v a t i o n s within one part of the zone m a y not be representative for the entire system. T h e s e m o d e l s have implications for tectonic analysis in areas of oblique c o n v e r g e n c e and divergence. In areas of oblique c o n v e r g e n c e , the plate m o t i o n vector is not parallel to the infinitesimal c o n t r a c t i o n (stress) direction, as repres e n t e d by the seismic slip vector. F u r t h e r m o r e , the finite strain features are not parallel to either the plate m o t i o n or the slip vectors. E v i d e n c e of e x h u m a t i o n or tectonic uplift of rocks within transpression zones, resulting f r o m a r c - n o r m a l extension, results from type C - E transpression. If erosion is active, type A and B transpression can lead to very rapid e x h u m a t i o n . C o r r e s p o n d ingly, a n o m a l o u s crustal t h i n n i n g w i t h i n the zone suggests type A - C transtension. T h e s e results are also applicable to transpressional settings resulting f r o m oblique plate conv e r g e n c e of t e r r a n e s . T h e exact t y p e of transpression (types A - E ) is a result of the plate m o t i o n vectors and g e o m e t r y of plate margins. The boundaries between converging terranes are likely to exhibit vertical foliation, but either vertical or h o r i z o n t a l lineation. Margin-parallel stretching m a y result in t e r r a n e m o t i o n w h i c h locally exceeds, or m o v e s in an opposite direction to, t h e t a n g e n t i a l c o m p o n e n t of p l a t e motion. Finally, kinematics of transpression can cause rapid uplift in n a r r o w , m a r g i n - p a r a l l e l zones of d e f o r m a t i o n , which m a y result in the e x h u m a t i o n of high-pressure rocks. C. Teyssier is acknowledged for both conversations and insights. B. T. wishes to thank B. McClelland for helpful discussions about the problems of Baja-BC, and K. Basset for providing the basis for the Cordillera diagram. We are grateful for thorough and helpful reviews by R. Jones and an anonymous referee. B. T. was supported by NSF Grant E A R 9628381. References

AVE LALLEMANT,H. G. & GUTH, L. R. 1990. Role of extensional tectonics in exhumation of eclogites and blueshists in an oblique subduction setting: Northern Venezuela. Geology, 18, 950-953. BEAUMONT, C., FULLSACK, P. & HAMILTON, J. 1992. Erosional control of active compressional orogens. In: MCCLAY, K. R. (ed.) Thrust Tectonics. Chapman & Hall, New York. BECK, M. E. 1986. Model for Late Mesozoic-Early Tertiary tectonics of coastal California and western Mexico and speculations on the origin of the San Andreas fault. Tectonics, 5, 49-64. - 1991. Coastwise transport reconsidered: lateral displacements in oblique subduction zones, and

31

tectonic consequences. Physics of the Earth Planetary Interiors, 68, 1-8. BERKLAND, J. O. 1973. Rice Valley outlier - new sequence of Cretaceous-Paleocene strata in northern Coast Ranges, California. Geological Society of America Bulletin, 84, 2389-2406. CLOOS, M. 1982. Flow melanges: numerical modelling and geological constraints on their origin in the Franciscan subduction complex, California. Geological Society of America Bulletin, 93, 330-345. 1985. Thermal evolution of convergent plate margins: thermal modeling and reevaluation of isotopic Ar-ages for blueschists in the Franciscan complex of California. Tectonics, 4, 421-433. CONEY,P. L., JONES,D. L. & MONGER,J. W. H. 1980. Cordilleran suspect terranes. Nature, 299, 329-333. COWAN, D. S. 1994. Alternative hypotheses for the mid-Cretaceous paleogeography of the western Cordillera. Geological Society of America Today, 4, 181-186. CRAWFORD,M. L. & CRAWFORD,W. A. 1991. Magma emplacement in a convergent tectonic orogen, southern Revillagigedo Island, southeastern Alaska. Canadian Journal of Earth Sciences, 28, 929-938. DIAS, R. & RIBEIRO, A. 1994. Constriction in a transpressive regime: an example in the Iberian branch of the Ibero-Armorican arc. Journal of Structural Geology, 16, 1543-1554. ENGEBRETSON,D. C., Cox, A. & GORDON, R. G. 1985. Relative Relations between Oceanic and Continental Plates in the Pacific Basin. Geological Society of America, Special Paper, 206. FITCH, J. T. 1972. Plate convergence, transcurrent faults, and internal deformation adjacent to Southeast Asia and the Western Pacific. Journal of Geophysical Research 77(23), 4432-4460. FLINN,D. 1979. The deformation matrix and the deformation ellipsoid. Journal of Structural Geology, 1, 299-307. FOSSEN,H. & TIKOFF,B. 1993. The deformation matrix for simultaneous simple shearing, pure shearing, and volume change, and its application to transpression/transtension tectonics. Journal of Structural Geology, lg, 413-422. -& -1997. Forward modeling of non steadystate deformations and the 'minimum strain path'. Journal of Structural Geology, 19, 987-996. & TEYSSIER,C. T. 1994. Strain modeling of transpressional and transtensional deformation. Norsk Geologisk Tidsskrifl, 74, 134-145. GABRIELSE, H. 1985. Major dextral transcurrent displacements along the Northern Rocky Mountain Trench and related lineaments in north-central British Columbia. Geological Society of America Bulletin, 96, 1-14. GOODWIN,L. B. & WILLIAMS,P. F. 1996. Deformation path partitioning within a transpressive shear zone, Marble Cove, Newfoundland. Journal of Structural Geology, 18, 975-990. HARDING, T. P. 1985. Seismic characteristics and identification of negative flower structures, positive flower structures, and positive structural

32

H. FOSSEN & B. TIKOFF inversion. Bulletin, American Association of Petroleum Geologists, 69, 582--600.

HARMS,T. m., JAYKO,A. S. & BLAKE, M. C. 1992. Kinematic evidence for extensional unroofing of the Franciscan complex along the Coast Range fault, northern Diablo Range, California. Tectonics, 11, 228-241. Hsr3, K. J. 1991. Exhumation of high-pressure rocks. Geology, 19, 107-110. HUDLESTON, P. J., SCHULTZ-ELA,D. & SOUTHWICK,D. L. 1988. Transpression in an Archean greenstone belt, northern Minnesota. Canadian Journal of Earth Sciences, 25, 1060-1068. INGRAM, G. M. & HUTrON, D. H. W. 1994. The Great Tonalite Sill: emplacement into a contractional shear zone and implications for Late Cretaceous to early Eocene tectonics in southeastern Alaska and British Columbia. Geological Society of America Bulletin, 106, 715-728. IRVING, E., THORKELSON, D. J., WHEADON, P. M. & ENKIN, R. J. 1995. Paleomagnetism of the Spences Bridge Group and northward displacement of the Intermontane Belt, British Columbia: a second look. Journal of Geophysical Research, 100, 6057-6071. ,9WYNNE, P. J., THORKELSON,D. J. & SCHIARIZZA, P. 1996. Large (1000 to 4000 km) northward movements of tectonic domains in the northern Cordillera, 83 to 45 Ma. Journal of Geophysical Research, 101, 17901-17916. JAMISON, W. R. 1991. Kinematics of compressional fold development in convergent wrench terranes. Tectonophysics, 190, 209-232. JONES, R. R. & Tanner, P. W. G. 1995. Strain partitioning in transpression zones. Journal of Structural Geology, 17, 793-802. KELLEY, K. 1993. Relative motions between North

America and oceanic plates of the Pacific basin during the last 130 million years. MSc thesis, Western Washington University, Bellingham. KELLEY, K. P. & ENGEBRETSON, D. C. 1994. Updated relative motions and terrane trajectories for North America and oceanic plates: Cretaceous to present. Geological Society of America, Abstracts with Programs, 26, A459. KIRKWOOD, D. 1995. Strain partitioning and progressive deformation history in a transpressive belt, northern Appalachians. Tectonophysics, 241, 15-34. KRANTZ, R. W. 1995. The transpressional strain model applied to strike-slip, oblique-convergent and oblique-divergent deformation. Journal of Structural Geology, 17, 1125-1137. MCCAFFREY, R. 1992. Oblique plate convergence, slip vectors, and forearc deformation. Journal of Geophysical Research, 97, 8905-8915. 1996. Estimates of modern arc-parallel strain rates in fore arcs. Geology, 24, 27-30. MCCLELLAND,W. C., GEHRELS, G. E., SAMSON,S. D. & PATCHETT, P. J. 1992. Structural and geochronologic relations along the western flank of the Coast Mountains batholith: Stikine River to Cape Fanshaw, central southeastern Alaska. Journal of Structural Geology, 14, 475-489.

-

-

MCKENZIE, D. & JACKSON, J. 1983. The relationship between strain rates, crustal thickening, paleomagnetism, finite strain and fault movements within a deforming zone. Earth and Planetary Science Letters, 65,182-202. MOLNAR, P. 1992. Brace-Goetze strength-profiles, the partitioning of strike-slip and thrust faulting at zones of oblique convergence, and the stress-heat flow paradox of the San Andreas fault. In: EVANS, B. & WONG, T.-F. (eds) Fault Mechanics and Transport Properties of Rocks. Academic Press, New York, 435--459. MOUNT,V. S. & SUPPE,J. 1992. Present-day stress orientations adjacent to active strike-slip faults: California and Sumatra. Journal of Geophysical Research, 97, 11995-12013. NILSEN, T. S. & CLARKE, S. H., Jr 1975. Sedimentation

and Tectonics in the Early Tertiary Continental Borderland of Central California. US Geological Survey Professional Paper, 925. NORRIS, R. J., KOONS, R O. & COOPER,A. F. 1990. The obliquely-convergent plate boundary in the South Island of New Zealand. Journal of Structural Geology, 12, 715-725. NORTHRUP,C. J. & BURCHFIEL,B. C. 1996. Orogen-parallel transport and vertical partitioning of strain during oblique collision, Efjorden, north Norway. Journal of Structural Geology, 18, 1231-1244. OLDOW, J. S., BALLY,A. W., AvE LALLEMANT,H. G. & LEEMAN,W. P. 1989. Phanerozoic evolution of the North American Cordillera: United States and Canada. In: BALLY,A. W. & PALMER,A. R. (eds)

The Geology of North America: an Overview. Geological Society of America, Boulder, CO, 139-232. PASSCHIER, C. W. 1988. The use of Mohr circles to describe non-coaxial progressive deformation. Tectonophysics, 149, 323-338. PASSCHIER,C. W. 1997. The fabric attractor. Journal of Structural Geology 19, 113-127. PLArr, J. P. 1986. Dynamics of orogenic wedges and uplift of high-pressure metamorphic rocks. Geological Society of America Bulletin, 97,1037-1053. 1993. Exhumation of high-pressure rocks: a review of concepts and processes. Terra Nova, 5, 119-133. ROBIN, P.-Y. E & CRUDEN, A. R. 1994. Strain and vorticity patterns in ideally ductile transpression zones. Journal of Structural Geology, 16, 447-466. SANDERSON, D. & MARCHINI, R. O. 1984. Transpression. Journal of Structural Geology, 6, 449-458. SIMPSON, C. & DE PAOR, D. G. 1993. Strain and kinematic analysis in general shear zones. Journal of Structural Geology, 15, 1-20. STERN,Y. W., BATEMAN,P. C., MORGAN, B. A., NEWELL, M. E & PECK, D. L. 1981. Isotopic U-Pb Ages of

Zircon from Granitoids of the Central Sierra Nevada, California. US Geological Survey Professional Paper, 1185. STOWELL, H. H. & HOOFER, R. J. 1990. Structural development of the western metamorphic belt adjacent to the Coast Plutonic Complex, southeastern Alaska: evidence from Holkham Bay. Tectonics, 9, 391-407.

TRANSPRESSION AND TRANSTENSION MODELS STRUIK,L.C. 1993. Intersecting intracontinental transform fault systems in the North American Cordillera. Canadian Journal of Earth Sciences, 30, 1267-1274. TEVSSIER,C. & TIKOFF,B. 1998. Strike-slip partitioned transpression of the San Andreas fault system: a lithospheric-scale approach. This volume. --,-& MARKLEY,M. 1995. Oblique plate motion and continental tectonics. Geology, 23, 447-450. THOMPSON, A. B., SCHULMANN,K. & JEZEK, J. 1997. Extrusion tectonics and elevation of lower crustal metamorphic rocks in convergent orogens. Geology, 25, 491-494. TIKOFF,B. & FOSSEN, H. 1993. Simultaneous pure and simple shear: the unified deformation matrix. Tectonophysics, 217, 267-283. & ~ 1995. Limitations of three-dimensional kinematic vorticity analyses. Journal of Structural Geology, 17, 1771-1784. & -1996. Visualization and calculation of deformation: computer applications for teaching. In: DEPAOR, D. (ed.) Microcomputers and Structural Geology. Elsevier, Amsterdam. & GREENE, D. 1997. Stretching lineations in transpressional shear zones. Journal of Structural Geology, 19, 29-40. - - & TEYSSmR,C. 1992. Crustal-scale, en echelon 'Pshear' tensional bridges: a possible solution to the batholithic room problem. Geology, 20, 927-930. -

-

-

-

33

& -1994. Strain modeling of displacementfield partitioning in transpressional orogens. Journal of Structural Geology, 16, 1575-1588. TRUESDELL, C. 1953. Two measures of vorticity.

-

-

Journal of Rational Mechanical Analysis, 2, 173-217. VAUCHZZ,A. & NICOLAS,A. 1991. Mountain building: strike-parallel displacements and mantle anisotropy. Tectonophysics, 185, 183-201. WEIJERMARS, R. 1991. The role of stress in ductile deformation. Journal of Structural Geology, 13, 1061-1078. 1993. Estimation of paleostress orientation within deformation zones between two mobile plates. Geological Society of America Bulletin, 105, 1491-1510. WITHJACK,M. O. & JAMISON,W. R. 1986. Deformation produced by oblique rifting. Tectonophysics, 126, 99-124. WYNNE, P. J., IRVING,E., MAXSON,J. A. & KLEINSPEHN, K.L. 1995. Paleomagnetism of the mid-Cretaceous Silverquick and Powell Creek formations: evidence for 3000 km of northward displacement of the eastern Coast Belt, British Columbia. Journal of Geophysical Research, 100, 6073~5091. Yu, G.,WESNOUSKY,S.G. & EKSTROM,G. 1993. Slip partitioning along major convergent plate boundaries. Pageoph, 140, 183-210.

Oblique simple shear in transpression zones RICHARD

R. J O N E S 1 & R O B E R T E. H O L D S W O R T H

2

lOstfold Research Foundation, Postboks 573, 1754 Halden, Norway Present address: CognlT a.s., Postboks 610 Busterud, 1754 Halden, Norway 2Department o f Geological Sciences, University o f Durham, Durham DH1 3LE, UK Abstract: A mathematical and conceptual framework is presented for extending existing models of transpression and transtension to consider vertical and oblique displacements of the deformation zone bounding blocks. This allows transpression and transtension to be considered in terms of the interplay between four biaxial 'end-member' strain components, which are shown graphically as the four vertices of a 'strain tetrahedron'. Transpressional and transtensional strains represented by the strain tetrahedron generally display triclinic symmetry, although the end-member biaxial strains are orthorhombic or monoclinic. Deformation in which vertical simple shear dominates over horizontal simple shear gives rise to finite strains that tend towards plane strain. Finite strains caused by deformation involving oblique simple shear are generally orientated non-parallel to all three Cartesian reference axes. Geologists attempting to analyse complex zones of three-dimensional deformation in the Earth's crust are often faced with a level of structural complexity that makes coherent interpretation extremely challenging. To improve our understanding of such zones, mathematical and analogue models have been utilized in which individual parameters that affect deformation can be isolated and varied, and their effect on resultant strain measured. Although all such models are greatly idealized and simplified in comparison with most naturally occurring shear zones, the models can nevertheless be extremely useful in allowing the interplay of variables in complicated 3D strain zones to be visualized, and thereby more easily understood. The mathematical framework presented by Sanderson & Marchini (1984) is the simplest, most idealized description for the analysis of transpression and transtension zones. Their model describes constant-volume, homogeneous finite strain in a basally and laterally confined deformation zone, bounded by parallel, vertical zone walls and flanked by undeformed blocks beyond the zone margins. Deformation is fully defined in terms of just two strain factors; a horizontal stretch orthogonal to the zone, and a simple shear parallel to the zone. Subsequent workers have adopted Sanderson & Marchini's model, and by changing one or more of the assumptions and boundary conditions of the original model, have sought to increase its applicability to documented field examples. In this way, the effects of volume change (Fossen & Tikoff 1993), heterogeneity (Robin & Cruden 1994), strain partitioning (Fossen et al. 1994; Tikoff & Teyssier 1994; Jones & Tanner 1995; Teyssier et al. 1995), simultaneous pure and

simple shearing and progressive deformation (Fossen & Tikoff 1993, Tikoff & Fossen 1993), and basal and lateral stretch (Jones et al. 1997) have all been discussed. This paper addresses a further boundary condition inherent in the Sanderson & Marchini model; that the far-field displacement vector acting across the transpression or transtension zone must lie in the horizontal plane (the xyplane of Fig. la), thereby precluding the possibility of relative vertical m o v e m e n t of the bounding blocks that flank the deformation zone. By relaxing this boundary condition, we are able to analyse deformation in which the farfield displacement vector is oblique (non-parallel and non-orthogonal) to all three Cartesian axes, causing one zone bounding block to experience a component of vertical displacement with respect to the other bounding block (Fig. l b and c). Structural, metamorphic and thermochronological studies of naturally occurring transpression and transtension zones have shown that zone bounding blocks are sometimes displaced vertically with respect to one another (e.g. Newfoundland, Holdsworth 1994; Southern Alps, Koons 1987). The following mathematical analysis shows how a component of vertical simple shear can give rise to the dipping foliations and plunging lineations commonly observed in such zones.

Strain analysis Strain matrix The mathematical description of the transpressional and transtensional deformation shown in

JONES,R. R. & HOLDSWORTH,R. E. 1998. Oblique simple shear in transpression zones. In: HOLDSWORTH, 35 R. E., STRACHAN,R. A. & DEWEY,J. E (eds) 1998. Continental Transpressionaland TranstensionalTectonics. Geological Society, London, Special Publications, 135, 35-40.

36

R.R. JONES & R. E. HOLDSWORTH

Fig. 1. End-member models of transpression showing the homogeneous transformation of a unit cube,

assuming constant volume. Xo yc and Zc are the reference Cartesian axes, ~ is angular shear strain, ~ is shear strain, and az is the vertical stretch. (a) Transpression zone with one component of simple shear in the xyplane parallel to the x-axis (based on fig. 1 of Sanderson & Marchini (1984)). (b) Transpression and (c) transtension zones displaying oblique simple shear, which can be factorized into two separate simple shear components, one acting in the xy-plane parallel to the x-axis (%y), the other acting in the yz-plane parallel to the z-axis (~lzy).

Fig. lb and c requires the addition of a second component of simple shear to the transpression strain matrix of Sanderson & Marchini (1984), so that in addition to simple shear acting in the horizontal (xy)-plane parallel to the x-axis, we must also include a simple shear component in the vertical (yz)-plane parallel to the z-axis: D

( \ 0 simple shear in xy parallel to x

1

~~ ~lzy

1

0 a z

pure shear in yz (vertical stretch)

simple shear in yz parallel to z

-~ 0

(1)

Otz "Yzy ~ z -1 Otz

transpression with oblique simple shear

As with previous studies, such a factorization is conceptually convenient, because it allows us to visualize complex three-dimensional finite strains in terms of the interaction between less complicated biaxial plane strains. In this way we can extend the graphical representation we have used previously (Jones et al. 1997), in which three e n d - m e m b e r biaxial strains form the apices of a strain triangle, the inside of which represents unconfined transpressional deformation. Addition of a fourth end-member strain (the component of simple shear acting in the yzplane, as introduced in equation (1)), can be visualized in a strain tetrahedron, with four vertices corresponding to four biaxial strain components (Fig. 2). The resultant strain matrix, D, in equation (1) represents just one face of the tetrahedron, whereas the inside of the tetrahedron describes a vast range of complex triaxial

strains produced by transpressional transtensional deformation.

and

Simultaneous pure and simple shearing Fossen & Tikoff (1993) and Tikoff & Fossen (1993) have discussed the limitations of the factorization scheme presented by Sanderson & Marchini, and have developed a more rigorous approach involving simultaneous (rather than sequential) pure and simple shearing. We can rewrite equation (1) to adopt this approach as follows:

O = ( l'lxy(1-az-1)/ln(~ -lzl 0 ) 0 "yzy(1 otz- )/ln(az) a z

(2)

Because the analysis of deformation described by equation (1) is qualitatively similar to that in equation (2), for simplicity in this conceptual study we use the Sanderson & Marchini factorization. Nevertheless, the quantitative differences between the two methods are significant, and the deformation matrix for simultaneous shearing should be used when modelling real transpression and transtension zones.

Shape of the finite strain ellipsoid Following the technique of Sanderson & Marchini (1984) it is possible to analyse the shape and orientation of the strain ellipsoid by deriving the eigenvalues and eigenvectors of the Finger tensor (DDT). In practice, algebraic derivation of eigenvalues from equation (1) is

OBLIQUE SIMPLE SHEAR IN TRANSPRESSION ZONES

37

Fig. 2. Schematic diagram showing how homogeneous, constant-volume transpression can be expressed in terms of four separate plane strain components, represented by the four vertices of the strain tetrahedron. Each edge of the tetrahedron represents a limited range of deformations between two end-member strain components, and each face represents deformation involving three end-member components. The interior of the tetrahedron represents a full spectrum of transpressional and transtensional strains displaying triaxial symmetries. Deformation involving oblique simple shear (described by equation (1) in the text) is represented by the right-hand visible face of the tetrahedron; the rear-most (hidden) face represents transpression in which material is extruded laterally (Jones et al. 1997). The deformed cubes depicted here show deformation in which CX-1 z _<1 (i.e. transpression not transtension) and simple shears are positive (~/xyis sinistral, ~zy is 'front block up'). However, mathematically, the tetrahedron describes a much wider spectrum of deformation (e.g. transtension when CX-1 z > 1; dextral simple shear when ~/xyis negative, 'front block down' when ~zy is negative, etc.), including all five reference types shown in fig. 2 of Fossen & Tikoff (this volume).

extremely complex, involving very large, highorder polynomial expressions, so an iterative computer-based numerical approach has been used. Finite transpressional strain can be shown graphically in a Flinn diagram, in which the shape of the strain ellipsoid varies according to the amounts of simple shear along the zone and shortening or extension across it (see fig. 2 of Sanderson & Marchini (1984)). Because we have added an additional simple shear component to our analysis, a whole spectrum of Flinn diagrams is n e e d e d to show the interrelationship of all three strain factors. By way of example, in Fig. 3 two such diagrams (Fig. 3b and Fig. 3c) are compared with normal Sanderson & Marchini deformation (Fig. 3a). The following observations regarding finite strain can be made from Fig. 3 and other similar plots: (1) The shape of the finite strain ellipsoid is heavily dependent not only upon the amount of pure shear and horizontal simple shear, but also upon the amount of vertical simple shear.

(2) Deformations in which the component of vertical simple shear ('Yzy) is greater than horizontal simple shear (~xy) increasingly tend towards plane strain. When ~/xy= 0 and ~zy ~ O, deformation is plane strain, and all finite strains plot on the 'k = 1' diagonal of the Flinn plot. (3) When vertical simple shear is not zero, deformation cannot give rise to purely prolate and oblate finite strains. (4) In each graph the range of data points depicts a surface that is folded upon itself. In Fig. 3a this occurs at the abscissa and ordinate of the Flinn plot, and is associated with the 'switching' of orientation of the principal strain axes, as described by Sanderson & Marchini (1984, p. 450). When a component of vertical simple shear is present no such 'switching' of finite strain axes occurs, although in other respects the graphs in Fig. 3b and Fig. 3c are conceptually similar to that in Fig. 3a: when the surface of the graph is 'folded', a given shape of finite strain ellipsoid can be the result of one of two entirely different combinations of pure and simple finite shears (though whereas the shape of the ellipsoid is the

38

R.R. JONES & R. E. HOLDSWORTH

-40

same for the two combinations, the orientation is different; see below).

a. ~zY = 0 (no vertical simple shear) "fxY = 10 Otz-1 =2.0 "[xy'x= U --9

-30

~

',

= .8

,• -"_ ~~,,~z_1=1.6

5

-20

YxY= 4

", ",

"\ \

",

Strain s y m m e t r y and the orientation o f the finite strain ellipsoid

'\, "., "\

',

\

,,

",

\ (Zz-1 = 1.2

9

",

" O(z'l = 1.0

\ ,, , "\ ~ " O~z-1= 0.9

,"\< ',

'

; b.

,~z =0.6

",

\

' \ ",0.4

'

\0

] t z y = ~txy ( v e r t i c a l s i m p l e s h e a r = h o r i z o n t a l s i m p l e s h e a r )

-40 -30

Y = "/xu = 10. O~z-1 = 2.0 . "/xv,=UO~z-1 = 1.8 o

++

,+ . ,+

,; +

/ / / / , ' ~ i, ~,,=o+

• _1o

ii

Y'~=~

]7 -6

Yxv=, ,,

-5

,,'

,'

:'

i

i i

t

'

;

i:

)~,-~oo8

", i ", ", " ~,, ~

" ',,

',,

, ',

~ ~ ",

",

= 0.6 ,.Otz-1 = 0.5 " O~z-1 = 0.4

C(z-1 = 0.2 "/'~', ",", '\" '<'." O~ -1 = 0.

/

i

c.

i

i

i

i

i

i

i

I

i

yzv = 2 7xy

.40

,+,:,o ,+ =y///~z~ = o ~ ~,, ~ ~ ] - ; :-o7(

_2o .+

,

_8-9

~~!!!!" ,",;

-10

)'xy=2

-7 -6

,,

. ' ,'

-5

i

~,+: 2o, y~+=~ - /;~.e Yx+=u/~ .6 Y• = 8~'/':1.4

-30

(• = "tzv=2,~xy=l

,,'

"

,;

,' i

"",

=0.5

'; " ~' "'

.3

+

J,r

,2

,3

4, ,s 678,,

910I

A d d i t i o n a l strain c o m p o n e n t s Although, mathematically, a n u m b e r of additional strain components can be added to increase the complexity of the transpressional strain matrix, it is difficult to visualize the interplay of more than four separate strain factors using a three-dimensional object such as a tetrahedron (higher-order phase-space is needed to analyse more than four independent factors). In any case, the four end-member strains that mark the vertices of the tetrahedron shown in Fig. 2

," " ,' :O(z-1 = 0.6

,' ," ," ,'! : ,,'. !%-1

; : ! ;; ', Otzq = 0.4 i

In the type of transpression and transtension considered by Sanderson & Marchini, one principal axis of the finite strain ellipsoid always remains vertical during progressive deformation, whereas the non-coaxiality of the simple shear component acting in the xy plane causes the other two principal strain axes to rotate in the horizontal plane. Such strains display monoclinic symmetry (Fig. la). In contrast, deformation involving two components of simple shear equation (1) causes rotation of all three principal axes of the strain ellipsoid, so that none lies parallel to any of the three Cartesian reference axes (unless shear strain approaches infinity). Figure 4 is a lowerhemisphere stereonet showing the three-dimensional orientation of the finite strain ellipsoid for different combinations of pure and simple shear, in this case when "Yxy = 'Yzy" The symmetry of this type of strain is triclinic. All points inside the tetrahedron of Fig. 2 describe strains that have triclinic symmetry, as do points lying within the two visible faces of the tetrahedron (all four vertices, the two hidden faces, and all edges have orthorhombic or monoclinic symmetries).

YdZs

20,

30

Fig. 3. Series of deformation plots (Flinn diagrams) showing the shape of the finite strain ellipsoid produced by constant-volume transpression and transtension, for a range of values of orthogonal stretch (oLz-~)and simple shears ('Yxyand ~/zv).In each plot the ratio of horizontal simple shear ("Yxy)to vertical simple shear ('Yzy) is contant. (a) ~/zy= 0 (see fig. 2 of Sanderson & Marchini (1984)). (b) "Yzy= "Yxy. (c) ~zy -- 2"yxy.It should be noted that the lines on the Flinn plots are not strain paths: they represent lines joining points of equal pure shear or simple shear strain.

OBLIQUE SIMPLE SHEAR IN TRANSPRESSION ZONES "'~' = '~ZY = yXY (vertical simple shear = horizontal simple shear)

39

Xc

i intermediate / '. axis y=~/

=0.8 \

y=5)

\O~z-1= 1.0 (all values of 7)

\ -

y=1oY lc

Zc

/ ~ ......................... ~.-l=ZO / J"~ C(z-1 = 0.2

..

".... ".,... '...., '.... IZz-1 = 2.0 '"'....

major axis

..-,=o.

/

.... ~/~y=o

minor

' 'I I ".-'=~

I'-"

~;;~y=,Ot]"lO

..Yl/I/~

axis -"////r Y.--1"'/ / ////(Zz-1 = 0.6 I

( 1~/=5

.......Y~~ ~='~ ! ............"~;2f z z z ~ - = ~ - ' . ~

0~z-I=1~0"" ~,=I~//~-'I-/ 0

7= ""

--'"

(Zz-'=0"8

y=5

Fig. 4. Example lower-hemisphere equal-angle stereonet showing the orientation of the principal axes of the finite strain ellipsoid for a range of values of orthogonal stretch (~z-1) and simple shears (~xy and ~zy). For this stereonet, ~zy -- "Yxy-Traces for transpression are shown as continuous lines, oblique simple shear (i.e. o~z-1 = 1) as dashed lines, and transtension as dotted lines. Xc,Ycand zc are the reference Cartesian axes (see Fig. 1). It should be noted that the traces are not rotation paths: they represent lines joining points of equal pure shear. are the most obvious strain c o m p o n e n t s to include in models in which deformation occurs between two rigid, undeformed bounding blocks (i.e. models developed from Sanderson & Marchini (1984)), because these strain components act in the xy- and yz-planes parallel to the zone boundaries. In contrast, the four additional offdiagonal terms of the strain matrix that are not considered in Fig. 2 are inconsistent with the boundary conditions that we have specified, as they affect the geometry of the xy Cartesian plane (either implying deformation of the zone bounding blocks, or requiring additional farfield shear displacements related to lateral bounding blocks; e.g. Dias & Ribeiro 1994).

Conclusions The Sanderson & Marchini (1984) model of transpression and transtension, upon which many subsequent mathematical models have

I

........................................................................... ~7~ [

been based, assumes that the non-coaxial component of deformation is described by a simple shear acting in the horizontal xy Cartesian plane, parallel to the x-axis. We can allow for vertical and oblique zone boundary displacements by supplementing the transpressional strain matrix with a second component of simple shear. The two simple shear components, together with components of vertical stretch (Sanderson & Marchini 1984) and lateral stretch (Jones et al. 1997), fully describe finite, homogeneous, constant-volume transpression and transtension in which the boundaries of the deformation zone remain undeformed. The interplay of the four end-member strain components can be visualized in a strain tetrahedron. Generally, transpressional and transtensional strains of this type display triaxial symmetry, and the orientation of the strain ellipsoid during progressive deformation changes about a non-fixed, non-vertical pole of rotation.

I

40

R.R. JONES & R. E. HOLDSWORTH

Our description of transpressional

Thanks are due to W. Bailey for previous discussions regarding this work, and to C. Teyssier and H. Fossen for helpful reviews.

References DIAS, R. & RIBEIRO,A. 1994. Constriction in a transpressive regime: an example in the Iberian branch of the Ibero-Armorican arc. Journal of Structural Geology, 16, 1543-1554. FOSSEN,H. & TIKOFF,B. 1993. The deformation matrix for simultaneous simple shearing, pure shearing and volume change, and its application to transpression-transtension tectonics. Journal of Structural Geology, 15, 413-422. & -1998. Extended models of transpression and transtension, and application to tectonic settings. This volume. --, - - & TEYSSIER, C. 1994. Strain modelling of -

-

transpressional and transtensional deformation.

and

t r a n s t e n s i o n a l strains f o r m s p a r t of a m o r e g e n e r a l way of analysing t h r e e - d i m e n s i o n a l noncoaxial finite d e f o r m a t i o n . P l a n e strains and coaxial strains are n o t h i n g m o r e than individual strain states in a m u c h b r o a d e r s p e c t r u m of 3D d e f o r m a t i o n . T h e i n t e r a c t i o n of c u r v i - p l a n a r tectonic plates m o v i n g over a sphere m u s t give rise to strains that are i n h e r e n t l y 3D in character, and t h e r e f o r e this a p p r o a c h is m o r e r e p r e sentative of d e f o r m e d rocks f o u n d in the E a r t h ' s crust.

Norsk Geologisk Tidsskrifi, 74, 134-145. HOLDSWORTH,R. E. 1994. Structural evolution of the Gander-Avalon terrane boundary: a reactivated transpression zone in the NE Newfoundland Appalachians. Journal of the Geological Society, London, 151, 629-646. JONES, R. R. & TANNER,P. W. G. 1995. Strain partitioning in transpression zones. Journal of Structural Geology, 17, 793-802. , HOLDSWORTH,R. E. & BAILEY,W. 1997. Lateral extrusion in transpression zones: the importance of boundary conditions. Journal of Structural Geology, 19, 1201-1217. KOONS, P. O. 1987. Some thermal and mechanical consequences of rapid uplift: an example from the Southern Alps, New Zealand. Earth and Planetary Science Letters, 86, 307-319. ROBIN, P.-Y. E & CRUDEN,A. R. 1994. Strain and vorticity patterns in ideally ductile transpression zones. Journal of Structural Geology, 16, 447-466. SANDERSON, D. J. & MARCHINI, W. R. D. 1984. Transpression. Journal of Structural Geology, 6, 449-458. TEYSSIER,C.,TIKOFF,B. & MARKLEY,M. 1995. Oblique plate motion and continental tectonics. Geology, 23, 447-450. TIKOFF, B. & FOSSEN,H. 1993. Simultaneous pure and simple shear: the unifying deformation matrix. Tectonophysics, 217, 267-283. & TEYSSIER, C. 1994. Strain modelling of displacement-field partitioning in transpressional orogens. Journal of Structural Geology, 16, 1575-1588. -

-

Transpression (or transtension) zones of triclinic symmetry: natural example and theoretical modelling S H O U F A L I N 1, D A Z H I J I A N G 2 & P A U L F. W I L L I A M S 2

1Geological Survey of Canada, 601 Booth Street, Ottawa, ON, Canada, K1A OE8 Present address: Manitoba Energy & Mines, 25-59 Elisabeth Drive, Thompson, MB, Canada R8N 1X4. (e-maik slin@norcom, mb. ca) 2Department of Geology, University of New Brunswick, Fredericton, NB, Canada, E3B 5A3 Abstract: We describe a natural shear zone with triclinic symmetry, present a general model for triclinic shear zones based on natural examples, and investigate the kinematics and strain geometry within such zones. In the Roper Lake shear zone in the Canadian Appalachians, the orientation of a stretching lineation is oriented approximately down-dip near the shear zone boundary and becomes gradually shallower towards the centre. The structures in the central portion of the shear zone exhibit approximately monoclinic symmetry where the poles to both the S- and C-surfaces, the stretching lineation on the S-surfaces and the striations on the C-surfaces all plot in a great circle girdle. However, the lineations from the marginal portion do not plot in the same girdle, and the bulk symmetry of the shear zone is triclinic. Theoretical modellingshows that the observed strain geometry can be interpreted by an oblique transpression with a larger ratio of simple shear to pure shear in the centre of the shear zone than in the margin. The latter suggests a higher degree of localization of the zone boundary-parallel movement component relative to the boundary-normal compression component. We emphasize that, as the imposed boundary displacements for most natural shear zones lie between dip-slip and strike-slip, their movement pictures are generally triclinic; monoclinic shear zones are special end members. Structural data that exhibit monoclinic symmetry do not necessarily mean that they resulted from a monoclinic movement picture; the present modelling demonstrates that a triclinic movement picture with a high ratio of boundary-parallel movement to boundary-normal movement can result in apparent monoclinic structural geometry. The results of the modelling also show that the simple statement made for simple shear zones that stretching lineations will align with, and therefore indicate, the shear direction cannot be extrapolated to three-dimensional transpressional (or transtensional) shear zones.

Zones of large strain (shear zones) separating relatively weakly d e f o r m e d or u n d e f o r m e d domains are a c o m m o n feature of the Earth's lithosphere at scaling dimensions varying from h a n d s p e c i m e n to the lithospheric plate. Because relative motions of adjacent domains (at the largest scale, lithospheric plates) generally contain b o t h a zone boundary-parallel ('trans') and a b o u n d a r y - n o r m a l ('press' or 'tens') component, the resulting m o v e m e n t on approximately vertical shear zones is generally transpressional or transtensional. Current theoretical models of shear zones are formulated either in terms of finite strain or in terms of instantaneous flow. Models based on finite strain include the plane-strain models of Ramsay & Graham (1970) and Ramsay (1980), and the transpression models of Sanderson & Marchini (1984) and its revised forms (Jones & Tanner 1995; Krantz 1995). Models based on

instantaneous flow include the plane-strain models of R a m b e r g (1975), the transpression-transtension models of Fossen & Tikoff (1993; see also Tikoff & Fossen 1993), Robin & Cruden (1994) and Tikoff & Teyssier (1994). The finite strain approach is purely geometrical. The approach of instantaneous flow has the potential to deal with t i m e - d e p e n d e n t deformation by considering the flow as a function of time. The commonly practised correlation of flow parameters (such as vorticity) with finite strain, however, is valid only within the assumption of a homogeneous and steady-state deformation (Jiang & White 1995). Most classical shear zone models (e.g. Ramsay & G r a h a m 1970; R a m b e r g 1975) and some transpression models (e.g. Sanderson & Marchini 1984) have monoclinic symmetry where the rotation axis of the finite deformation or the vorticity vector of the instantaneous flow is set to

LIN, S., JIANG,D. & WILLIAMS,P. E 1998. Transpression (or transtension) zones of triclinic symmetry: natural example and theoretical modelling. In: HOLDSWORTH,R. E., STRACHAN,R. A. 8r DEWEY,J. E (eds) 1998. ContinentalTranspressionaland TranstensionalTectonics.Geological Society, London, Special Publications, 135, 41-57.

41

42

S. LIN ETAL.

be parallel to one of the principal strain or principal strain rate axes. These monoclinic models have been very popular and have been applied to many natural shear zones with varying degrees of success. However, they cannot explain stretching lineations (sensu stricto) in a shear zone that vary between strike-parallel and dip-parallel (see discussion by Lin & Williams (1992a)), a p h e n o m e n o n observed in recent years in many natural shear zones, e.g. in the Palaeozoic Canadian Appalachians (e.g. Caron & Williams 1988; Lin 1992; Lin & Williams 1993; Goodwin & Williams 1996) and in the Proterozoic Sveconorwegian orogenic province in southwest Sweden (Robin & Cruden 1994). Understanding the kinematic and geometric relationships of the structures and fabrics developed in these shear zones and their evolution with time requires a general triclinic model. The transpression models of Fossen & Tikoff (1993) and Robin & Cruden (1994) include deformation with triclinic symmetries. Fossen & Tikoff (1993) assumed a steady-state deformation and described a general zonal deformation by a few finite strain parameters. How geological structures such as lineations and foliations would vary across such a deformation zone and evolve with time was not investigated. The model of Robin & Cruden (1994) only investigates the instantaneous flow; it cannot be readily applied to interpret structures recording a finite amount of strain as do most structures in shear zones, as acknowledged by those workers. The model also assumes that the zone boundary-parallel motion, resulting in the simple shear component of the deformation, is accommodated within the same volume of rock as the boundarynormal motion, leading to the thickening (or thinning) of the crust. This assumption is not supported by studies of ancient shear zones, observations of current plate-boundary deformation or theoretical models (see later discussion). In this paper, an example of a triclinic transpressional shear zone, the Roper Lake shear zone, is described, and a general model, which successfully reproduces the observed geometry, is presented. The natural example caused us to look at transpression, but both transpression and transtension are presented in the modelling because they are part of a continuum.

The Roper Lake shear zone

Geological setting The Roper Lake shear zone is in northeastern Cape Breton Island, in the Canadian Appalachians (Fig. 1). The geology of the area has been

Fig. 1. (a) Location map showing northeastern Cape Breton Island in the Canadian Appalachians. (b) Simplified geological map of northeastern Cape Breton Island (simplified after Lin 1993, 1995), showing the geological setting of the Roper Lake shear zone, indicated by a black arrow. The Eastern Highlands shear zone (D1 shear zone) is shaded. Cs, Carboniferous sedimentary rocks; Dg, Devonian granitic plutons; OSs, Ordovician-Silurian metasedimentary-volcanic rocks; Pd, Late Precambrian dioritic-tonalitic plutons; Ps, Late Precambrian metasedimentary rocks. described by Raeside & Barr (1992) and Lin (1993, 1995). Detailed structural analysis of the area (Lin 1992,1995) reveals two major episodes of ductile shearing, D1 and D2. D1 deformation is Late Silurian to Early Devonian in age. It is concentrated in, but not confined to, the Eastern Highlands shear zone (D1 shear zone in Fig. lb). The shear zone is characterized by an amphibolitefacies deformation zone over 5 km wide (Dla) with late-stage greenschist-facies deformation

TRANSPRESSION ZONES OF TRICLINIC SYMMETRY "

-r +

Granite

~-" [

,,n,4,~,-,rm,~, .............

] '

9

'~

t

,

+

+

,

+

+

+

S-lobation

+

,

+

+

+

,

+

+

+

+ +

Stretchinglineation

4-

++

++

.

+++ C-surface

+

+

+

+++

§

+ + +

,

{

+

+

+

+

+ {

+

{

+

+

~

+ +

+ +

~ +

[{

+

+

+

+ +

{

+ +

+

+

+

+ + +

+

4-

+

+

+

+

/

2

+ . ....... +

/ -[

1/'67 3_.

, +

/

.

. ~

! .

.

+-' +

+ +.'" ~ ~ ~ ~ , +," + + 4+ .: + ,'/ i ! ~. . . .

+,'+

......

+++("4-'+' +../ , , ..... ".'.-. § ,4" r r

~ z

,

.

4

/

. t

+

/

+ !

Roper

4:': / ,'

Lake

( ...... ~'" ?

+ z

+

....

t

, ........... ." +

t

-' ~

.: /

; ..... ,

+

+

l

/

+

+

4-

.

.

+

+

+

+

+

.

.

.

.

+

+

, *

4

/

!

.

+

/

•T

.

+

+

+

" ~ ;" t." ! +'" + "' ! ~ . . . . "+ 186+

"!

~ " .... ,at.

" ~1

:

771~

, +

!

/

+

,

~

~

+

/

+

+.

!

+

/ ," +,'+ ,.',

+ "+

+

+

! ,

+

+

!

+

!

~ !

+

,

+

+

+

+

+

!

+

,

+

+

+I

+

+ I

~

+[I

..'/ / / t / l /+ + + 4+ + +

55 ~ ,( % 43

t72 ~1 ~58

.

~

! ~

4 ~ 4t ~ + +

,

t1421+ / ~

l~

,

* Z /

.

.

+

.

, +

4-

, +

+

/

, +

+

, +

t

! +

/

|

4

.

+ I

,

'

+

/

/

I

+

,

+

/

/

. . . . . + + / ~ ' ! ! ! ! ! / + >F + + + + / ~ 70 / ! J ! / .~/ . . . . / 59

+

~ z95

,

!

+

l~ §

Z63 {

/

.

+

/ .

!

~52 /~

.

+

+

i

p,4

I ......

~

a_ / +

I

+ ',"'

*

,

/

+ ! + ! l/ + +

,7

+

/

.

/

+/ !/

7 8 3 ~ /

.

--

L/ ! , ~ 3 ~, ! # .~,~(" KOper Brook _ / ~ > .......... -.-" , - - ' "................. ..". . . . § . +. . . ~ . . + " .... ~ 1--. ~,._.' , "F t 4 ~ ................. ~ 4zs"/] + ~, ~ ./ ..... ~ -/ ~ 71 ./r ..... / ! ~r ' '/ + + 5b 48 ~/" ~ " ! ~ ! ! + 4+ + +

!

'

,', 4-

,

,

I I

+

Centraldomain

!

2. ! {

,.'

4,

!

/,~4 Lu-. b, ~37 + +

.

: , , / \ ' , ~

+

'

+

.

+/ + + ! " ! / "" ~ 6 4

,,!

,"

"

:/ ~

46~

+

, ";

,',

•T

.

4-

/-i-,

t

..'+ !

+

/ -T

.

,': ,.r

."

! {

! +

;

t

! +

/

~

Z ~

, +

Z60 L\ ~'

!

~

//65

!

+

, 4-

57 . . . . , , I ! r + + + + / ! J ! / . . . . . .

4~

~

, § .... _r

+

!

~-

-

,

+

,

+

z .

.

764 ~

+

!

,

+

/

+ !

+

,

I

+

,

)

+." -,

~

+

.

.z60 L~.

/

+ +

+

t .

.

+

',

.

-

+

+

. . §. .+ . ~.

+

+

,

+

. ___ ~u3 ~

+ !

!

4-

!

{

!

r e,. ::.T

+

.

4-

,

+ !

+

// / t i , , , t , %," ~ + + + + + + + +

!

/ {

/

,

+ /

+

+'-/+

t

~

~

t

,

+ /

,'+'+'+'+Marginaldomain

+

+

,% ..........

+

!

+

+

.

+

~

+

+

+

+,"

,,~'~'+,4-,+'+,4-,

? '

+' +, /

r

1

!

+

+

+

+/+

.

+

!

+ /

+

+ .'! / / I / / ! / ! ! / + .'+ + + + + + + + + + + +

+

+

52 +

/ +

~

+ /

+

~;"

.

.

+

/

+ /

+

+

+

+

,

+"! z ~ / :" ' ~=7 . . I + + + + + " I ,Ju.J ~ /-64 /.~' + ~ + + + + ,/+ + '' , ',+, ,,~t- L \ ..... ~ + + + .t t f / - " 9 k~ ! / / .... + + + 4 4~ 4; .... : '=8"" + " + + + + + ,r" ~ ~ ! ! ' ! : ' + + 4- + + + + + .~.,'.:; + + + + + . . .... . :. " . . . . +

I

+

+ +

+

"

~

+

. . . . .

+ 4-

+

+

.. //../+ .I "

~"

+.,

]"

+ !

+

+

+

/

/ .

+ t

§

! ~

!

+ + + .,~-

+

+.,;

§ ,."

+" "+

+

+

+

+

+

.."

+.

.

,

t

§

t/.+

.# =~ + ,~.oo +..; ~ "~

+

.,

. . . . . bmauononu

+

++ . . . . +

1-

+

~+

+

+

+

+

+

-t+

+

+

1-

+

+

1+

+

+

1-

-f

+

-t+

+

+

rock

t+

+

deformed

Country

+ +

4

43

/

7.82'

~'

r

<. - z ~ o ~ ~ ~[ "~tt~% o~ + ~, ! z , 6,~ + + +"

;":+ +~'4+7 / l @ . . ~ t ' : ' + ! / ". . . . . . . . 4 ~ , /

+

~......../ §

- - - -

9

500 ~

m

~' ~

+," ..... + + + + +1 ,...,,,,.a~ ~- / " z ," r " i t t t t t t

Fig. 2. Geological map of the Roper Lake area, showing internal structures of the Roper Lake shear zone. The location of the map area is indicated in Fig. lb. Sample locations are indicated by the three-digit sample numbers. localized in a zone about I km wide in the centre ( D l b ) (Lin 1995). D2 deformation took place in Late Devonian to Early Carboniferous time (Lin 1992). It occurred under greenschist-facies conditions and was concentrated in several northeast-striking shear zones along lithological contacts (Fig. lb). Mylonitic foliation in the shear zones strikes northeasterly and dips steeply to the southeast. Stretching lineations pitch moderately to the southwest on the foliation. Movement on the D2 shear zones is oblique with a dextral horizontal c o m p o n e n t and west-side-up dip-slip component (Lin 1992). The Roper Lake shear zone of this study is one of the D2 shear zones.

Structural geometry The Roper Lake shear zone is developed along the southeastern contact of the Middle Devonian Black Brook granite (U-Pb monazite age 375+~ Ma, Dunning et al. 1990) (Figs lb and 2). Structures related to the shear zone occur both in the granite and in the country rocks (metadiorite and metasedimentary rocks) near the contact. The deformation is strongest at the contact (the centre of the shear zone) and becomes gradually weaker away from it. The Black Brook granite is medium grained

and largely composed of quartz, K-feldspar, plagioclase and small amounts of muscovite and biotite (partially altered to chlorite in the shear zone). Outside the Roper Lake shear zone, the granite is undeformed and isotropic, suggesting that all anisotropic fabrics observed in the granite in the shear zone are related to the (D2) shear zone deformation. In contrast, the country rocks to the southeast were intensely deformed in the Eastern Highlands shear zone during D1 and had strong D1 fabrics before being deformed in the Roper Lake shear zone during D2. Here, the effects of D2 cannot always be differentiated from those of D1, especially in areas where D2 deformation is relatively weak. Therefore, the country rocks are too complex to study the strain geometry of the Roper Lake shear zone. For this reason, the following description and interpretation of structures in the shear zone is based on observations in the granite from the centre of the shear zone (i.e. the granite-country rock contact) to its northwestern margin. Similar structures are observed in the other half of the shear zone in the country rocks, but detailed analysis is impossible for the reason stated above. Deformation in the granite is heterogeneous and two transitional domains with different structural features can be recognized (Fig. 2). In

44

S. LIN E T A L .

Fig. 3. (a) and (b) Photograph and photomicrograph, respectively, of sample 142 from the central domain with well-developed S-C structure. (e) Photograph of sample 583 from the marginal domain. Only S-surfaces are generally developed in the latter domain. Sections cut perpendicular to the foliation and parallel to the lineation. (See Fig. 2 for sample locations.) the domain near the centre of the shear zone (the central domain), the d e f o r m a t i o n is intense. Quartz is generally dynamically recrys-

tallized and both S- and C-surfaces are developed (Fig. 3a and b). C-Surfaces are approximately planar. They are narrow zones of high

TRANSPRESSION ZONES OF TRICLINIC SYMMETRY shear strain. S-Surfaces curve asymptotically into the C-surfaces. Both the S- and C-surfaces strike northeasterly and dip steeply to the southeast. The S-surface generally has a more northerly strike and a shallower dip than the Csurface (Fig. 4a). On the S-surface, a stretching lineation (Ls) defined by elongate quartz crystals and quartz aggregates is well developed. It pitches moderately to the southwest (Fig. 4a and b). On the C-surface, a slickenside striation (Lc) of the 'ridge-in-groove' type, as described by Lin & Williams (1992b), is well developed. The S-C structures consistently indicate an oblique movement with dextral strike-slip and west-side-up dip-slip components (Figs 3a and b and 4). In the domain near the margin of the shear zone (the marginal domain), deformation is less intense. Quartz is only partially recrystallized and only S-surfaces are generally developed (Fig. 3c). The S-surfaces strike northeasterly and dip steeply to the southeast (Fig. 4a). Incipient C-surfaces are observed locally in thin sections. A stretching lineation (Ls) similar to that in the central domain is developed. It pitches steeply to the southwest (Fig. 4a and b). Quartz c-axis measurements were made on representative samples from both domains. Samples from the central domain show c-axis fabrics similar to those reported by Krohe (1990) for S-C mylonites (Fig. 5a-c). Where grains inside and outside the C-zones are not differentiated, the overall c-axis fabric is an asymmetrical type I crossed girdle (Lister & Williams 1979) (Fig. 5a); where differentiated, the c-axis fabric from outside the C-zones tends to be a type I crossed girdle still (Fig. 5b), but that from inside the C-zones shows a well-defined single girdle (Fig. 5c), indicating significant concentration of simple shear deformation in the Czones (Schmid & Casey 1986). The sample from the marginal domain also shows a type I crossed girdle (Fig. 5d). Several lines of evidence indicate that structures in the two domains were coeval. As described above, regional structural mapping demonstrates that only one episode of deformation is responsible for the ductile deformation of the granite. The two domains are transitional (Figs 2 and 4). No overprinting relationship between structures of the two domains is observed and no structural break exists between them. D e f o r m a t i o n in both domains occurred in the solid state and under greenschist-facies conditions, after the cooling of the Middle Devonian granite but before the deposition of a Carboniferous conglomerate which unconformably overlies the deformed

45

jN

// \ \

-.

/

(a)

SE

NW ~"

80

~"

60

80

9

ii

.a i

60

A

A A

8 40

AA~,A A

9

40

A

'8

~

20 0

(b)

Marginal domain

20

Central domain

I

I

I

I

I

I

I

I

1.6

1.4

1.2

1.0

.8

.6

.4

.2

0 0

Distance from shear zone centre (km)

Fig. 4. (a) Equal-area lower-hemisphere projection of foliations and lineations in the deformed granite in the Roper Lake shear zone. Data from the central domain are indicated by open symbols (including stars) and data from the marginal domain by solid symbols. It should be noted that structures in the central domain approximate a monoclinic symmetry, and poles to S-surfaces (a-S, circles), poles to Csurfaces (-=C, stars), the stretching lineation on S (L~, open triangles) and the striation on C (Lc, squares) all plot in a great circle girdle (G); also, the stretching lineation (L~) from the marginal domain (solid triangles) has a much steeper pitch than in the central domain and does not plot in the girdle. The bulk symmetry of the shear zone is therefore triclinic. (b) Plot of pitch of stretching lineations on S-surfaces v. distance from the centre of the Roper Lake shear zone. The transition from a steep lineation near the shear zone boundary to a much shallower lineation towards the centre of the shear zone is clearly shown.

granite and contains clasts of granitic mylonite derived from the shear zone (Lin 1992). The metamorphic conditions of the deformation are indicated by the widespread evidence of intracrystalline deformation and dynamic

S. LIN ETAL.

46

Ls

s

6421 / . / - ~

~

Fig. 5. Quartz c-axis fabric diagrams oriented perpendicular to the foliations and parallel to the lineations. Orientations of the C-surface, the S-surface, the striation on C (Lc) and the stretching lineation on S (Ls) are indicated where applicable. (a) Sample 142 from the central domain. Grains inside and outside the C-zones are not differentiated. (b) and (c) sample 584 from the central domain. Grains measured from outside and inside the C-zones, respectively. (d) Sample 583 from the marginal domain. Number of grains measured is indicated on each diagram. Contoured with program FABRIC (Starkey 1977). Contours in multiples of uniform distribution with contour intervals of one. (See Fig. 2 for sample locations.)

recrystallization of quartz and mica, in contrast to the lack of intracrystalline deformation in feldspar, and by the stable mineral assemblage quartz + feldspar + white mica + chlorite. Structures in the central domain approximate monoclinic symmetry, as the poles to both S- and C-surfaces, the stretching lineation on S-surfaces and the striations on C-surfaces all plot in a great circle girdle (Fig. 4a). However, although S-surfaces in both domains have a similar attitude, the stretching lineation in the marginal domain has a greater pitch and does not plot in the girdle (Fig. 4a). Consequently, the bulk symmetry of the shear zone is triclinic.

Theoretical modelling The velocity gradient tensor We describe the flow within a shear zone in terms of the instantaneous velocity field. A coordinate frame is established such that xl is parallel to the strike of the zone boundary, x2 normal to the boundary and x3 along the dip of the boundary (Fig. 6). The boundary velocity as a vector, v, can be resolved into a boundarynormal component, v~ and a boundary-parallel component, vii (Fig. 6). The boundary-normal component will thicken or thin the zone. Because such thickening or thinning is widely

TRANSPRESSION ZONES OF TRICLINIC SYMMETRY

Lp: (Oe8)

47

andLT=(~

~ -6 8 )(1)

where 4 = v~_/h,v~_and h being the magnitude of v~_ and the instantaneous width of the zone, respectively. The strike-slip component of the velocity is VllCOS4) and the dip-slip component is Vllsin4) (Fig. 6). Thus, the velocity gradient tensor related to simple shear component Ls is Ls :

~gs4) 8 ~in4) 0

(2)

where -) = vll/h. The total velocity gradient tensors, being the sum of the pure shear and the simple shear components, are as follows, Lps for transpression and LTS for transtension:

Lps:(a~c~

\ 0 ~in4)

Fig. 6. A homogeneous domain in an oblique transpressional zone. The imposed boundary displacement vector V is resolved into a boundarynormal component V• leading to pure shear (~), and a boundary-parallel component V,, leading to simple shear (~). The obliquity of the transpression is measured by the angle 4) between Vii and the strike of the zone. Across a transpressional zone, both 4//4 and qbcan vary. Coordinate frame xl x2 x3. spread, as will be discussed later, the flow component caused by the boundary-normal motion component can be considered as pure shear. The boundary-parallel component results in a simple shear flow component. Thus, flow within the shear zone is the sum of the pure shear and simple shear components. In the present analysis, it is assumed that length parallel to the strike of the zone is constant during deformation. The vertical stretching (or more precisely, stretching parallel to the dip of the zone, for an inclined zone) or pure shear is denoted by ~ (with a positive sign for transpression and a negative sign for transtension). The simple shear component resulting from the boundary-parallel motion is denoted by ~. The imposed boundary velocity for a zonal deformation is generally oblique (see discussion below), and 4), the transpression obliquity, is defined as the angle between vlt and the strike of the zone (xl-axis) (Fig. 6); 4) = 0~ for 'transcurrent' transpression (transtension) and 0 ~ < 4) < 90 ~ for oblique transpression (transtension). The velocity gradient tensors related to the pure shear component of transpression Lp and transtension LT are

andLTs:(aqlc~

8 ) (3)

0 4/sin4) -4

It is readily seen that the flow described by (3) is triclinic except when 4) = 0 ~or 90 ~ When 4) = 0 ~ the model is reduced to the monoclinic endm e m b e r transpression-transtension model of Fossen & Tikoff (1993) and Tikoff & Teyssier (1994). When 4) = 90 ~ the model becomes the two-dimensional monoclinic shear zone of Ramberg (1975).

Kinematic vorticity The velocity gradient tensors (3) can be partitioned into a stretching tensor and a vorticity tensor (Truesdell & Toupin 1960; Means et al. 1980). A straightforward but somewhat tedious derivation demonstrates that Truesdell's kinematic vorticity number (Truesdell 1953) Wk is a function of the ratio ~/4., but is independent of 4), the obliquity of the shear (Fig. 6): 1

Wk = ")/414+ (41/4)2]-~

(4)

Being independent of +, Wk is less useful for measuring the kinematics in oblique transpressional or transtensional zones, because the same Wk can describe an infinite number of transpressions or transtensions with different qb values. A more useful parameter may be the two- dimensional sectional kinematic vorticity defined by Robin & Cruden (1994). The vorticity-normal section (VNS) is the one with the maximum noncoaxiality. The sectional kinematic vorticity on this section, W~, can be obtained by a Mohr circle construction (Fig. 7). Unlike Wk, W~ is a

S. LIN ETAL.

48

Fig. 7. Mohr circle construction for the flow described in equation (3). (a) Geometrical relationship between the vorticity vector, W, the section normal to W (shaded), transpression obliquity (+) and the reference frame. (b) and (e) Mohr circles for transpression and transtension zones. Although Truesdell's kinematic vorticity is independent of +, the sectional kinematic vorticity on the vorticity-normal section (VNS), W], is dependent on +. The two-dimensional flow on the xlx3-plane is represented by the circle M N O centred on the horizontal axis, and O M = 4. The line L1, the intersection of the VNS and the zone boundary, making an angle + with respect to xl on the Xax3-plane, is plotted at N on circle MNO. Therefore stretching along L1 is represented by OP on the horizontal axis, where NP is perpendicular to the horizontal axis. The stretching of L2 for transpression and transtension is respectively - 4 and 4, where I-,z is a line perpendicular to L1 on the VNS. This, together with the shearing q/, gives a point Q. PQ is a diameter of the Mohr circle for the two-dimensional flow on the VNS, from which the sectional kinematic vorticity can be obtained (W~ = cos o0. It is readily seen that as qb varies from 0 ~ to 90 ~ the two-dimensional flow on the VNS varies within the shaded area. Co, C+ and C90 are centres of the Mohr circles for qb = 0~ 0 ~ < + < 90 ~ and qb = 90 ~ respectively. W] reaches a m a x i m u m when qb = 0 ~ and a m i n i m u m when qb = 90 ~

f u n c t i o n o f qb. It is m a x i m u m w h e n ~b = 0 ~ a n d m i n i m u m w h e n + = 90~ W~ (90 ~ < W~ (~b) < W~ (0 ~

(5)

w h e r e W~ (90~ T r u e s d e U ' s k i n e m a t i c vorticity, a n d W~ (0 ~ = [1+(4_/q/)2]-1/2 (Fig. 8).

Finite strain geometry T h e e v o l u t i o n o f t h e finite strain g e o m e t r y will d e p e n d o n t h e a c t u a l d e f o r m a t i o n p a t h , i.e. h o w qb, ~ a n d ~/vary w i t h t i m e . B e c a u s e o f difficulties in c o n s t r a i n i n g t h e s e p a r a m e t e r s , w e h a v e assumed a steady-state deformation path.

TRANSPRESSION ZONES OF TRICLINIC SYMMETRY

49

f

1,0

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

.4 c"

~2

.2

~f

[

I

2

4

I

I

i

6

8

10

Fig. 8. Sectional kinematic vorticity W~(+) as a function of ~/~.

Comparison of the theoretical results with the observed data (see later) suggest that the Roper Lake shear zone can be interpreted as resulting from a steady period of deformation. Within a homogeneous domain and for a steady period (Jiang & White 1995), the differential equations in (3) can be solved to give the following deformation gradient tensors, Fes and FTs for transpression and transtension, respectively:

Fps =

FTs =

(i

exp(--~) ~in+sinh(~/) ~'-c~

exp(~t) ~sin+sinh(~t)

0 exp(~t)

00 exp(-~t)

)

(6a)

(6b)

The finite strain can be calculated by taking the eigenvalues and eigenvectors of the 'Cauchy-Green tensor', F F r (Truesdell & Toupin 1960). The eigenvalues are the magnitudes of the squares of the three principal stretches (ha, h2 and h3) whereas the corresponding eigenvectors are their orientations in the deformed state. The pattern and time evolution of hi is used to represent the variation across the shear zone of finite-strain-related lincations, such as most stretching lineations. Similarly, ~kl~k2 defines the finite-strain-related foliations and X3 the poles to the foliations. Figure 9 presents lower-hemisphere projections showing the evolution with time of the orientation of hi and k3 for various ~//~ and + values, both of which may vary across a shear

zone. It is clear that the strain geometry of transpressional zones (Fig. 9a) is significantly different from that of transtensional zones (Fig. 9b). The main features of the finite strain geometry of general transpression and transtension zones with 0~ < + < 90~are summarized below 9The end -member situations of + = 0~ and 90~ have been discussed by previous workers (e.g. Ramberg 1975; Sanderson & Marchini 1984). A main feature of transpression zones with + = 0~ is that the lineations are either horizontal or down-dip; when ~//~ < 2.76 (or Wk < 0.81), a lineation is always down-dip; when ~/~ > 2.76 (or Wk > 0.81), it is horizontal when the strain is low and may 'switch' to down-dip when the finite strain increases to a critical value (Fig. 9a; see Sanderson & Marchini (1984) and Fossen & Tikoff (1993) for discussion). In transpressional zones with 0~ < + < 90~ the lineations plot away from the vorticity-normal section (VNS) and close to the dip line of the zone, where the values of ~//~are low (this, for a general transcurrent zone, means steeper lineations, because the dip line of a transcurrent zone is nearly vertical). The lineations approach the VNS as "9/~increases (which means shallower lineations for a transcurrent zone). When the value of -~/~is high (>10) and for low to intermediate finite strain magnitudes (hal/z/h31/2 less than c.100), the locus of the lineations is virtually on the VNS (Fig. 9a), a feature to be expected of a monoclinic shear zone. Therefore a monoclinic structural geometry can result from a triclinic movement picture with a large boundary-parallel movement component. As finite strain accumulates, the lineations on all "9/~ paths and for various + values converge towards the dip line of

50

S. LIN E T A L .

k3

~6

k.-..~-;-----~...'il .....

77

133

,=40 ~

X2

~

1

X3

~N~~.~,,qt4-S

J,~2

~ . ,~,,~

-~

rdire

x~

(a) Transpression 133

nn

I x~ ~--- V N S ------r~\.,,--

(~ ....

I

~, I

)

:_~

>~

v . ~ .... :,_~ >1 Xs

TRANSPRESSION ZONES OF TRICLINIC SYMMETRY

51

(b) Transtension

.

oO:,,,~ ~,~__/n/ ,"Z'" ~~"~-~% % I~~"se~dNct,o

.

shear direct/ion

Fig. 9. Equal-area lower-hemisphere projection of variation and evolution with time of finite-strain-related lineations (ha) and poles to finite-strain-related foliations ()~3) in a transpressional (a) or transtensional (b) zone with varying ~/~ and +. Numerals 1, 2, 4, 6 and 20 are values of ~/~. White arrows indicate the evolution with time (increasing finite strain) for various ~/~ values. The dashed white arrow on the diagram of qb = 0~ indicates 'switch' of orientation of stretching lineations from horizontal to vertical. VNS, Vorticity-normal section; SZB, shear zone boundary, xl, x2, x3 correspond to those in Fig. 6. Let us take a transpressional zone with + = 20 ~ as an example. When ~/~ is low (e.g. < 3), the lineation is nearly down-dip. As "~/~increases, the lineation is closer to lying on the VNS. On an average foliation, the lineation across a transpressional zone swings from nearly down-dip where ~/~ is low to nearly strike-parallel where 4//~is high. As finite strain increases the lineations converge to being steeper though with significantly different rates of convergence. When ~/~ is high (>10) and for low to intermediate finite strain magnitudes (~k11/2/~.31/2less than c.100), the lineations essentially lie on the VNS, and the symmetry of the structures becomes monoclinic within the resolution of observation, although the movement picture is trielinic. The shear zone boundary is shown as vertical here. Geometry for non-vertical shear zones can be obtained by rotating these diagrams (e.g. Fig. 11). (See text for further discussion.) the shear zone. Thus, the e n d orientation of the lineation is not the shear direction of the simple shear c o m p o n e n t but the dip line of the zone (close to a vertical line for a general transcurrent

zone). For a general transcurrent zone (+ is low but not 0 ~ and the zone b o u n d a r y is n e a r vertical), the lineations swing from nearly down-dip w h e r e the ~/~ ratio is low to nearly strike-parallel w h e r e

52

S. LIN E T A L .

Fig. 10. Plots of the shapes of the finite strain ellipsoids for transpressional or transtensional zones of constant strike length for various qbvalues. Transpression produces oblate strains (K <1) whereas transtension produces prolate strains (K >1). Simple shear produces plane strain (K = 1). Solid square, solid circle, cross and open square represent, respectively, "~/4= 2, 4, 6 and 10.

the "~/4ratio is high on an average foliation (see the diagram of + = 20~ in Fig. 9a). The poles to foliations in the transpressional situation, however, always plot near the VNS (Fig. 9a). In contrast, in transtensional zones, the lineations plot close to the VNS, very close to it when the finite strain is low and slightly away from it as the finite strain increases. Poles to foliations in transtension are variable and similar to the lineations in transpressional situations (Fig. 9b). Figure 9a shows that stretching lineations in transpressional zones can (1) lie on or very close to the VNS and, as strain increases, approach the

shear direction for monoclinic or apparent monoclinic (~?/~> 10) situations; (2) be perpendicular to the shear direction when + = 0~ and hi is parallel to x3; (3) generally be inclined to both the VNS and the shear direction in triclinic situations. The simple statement made for simple shear zones that stretching lineations will align with, and therefore indicate, the shear direction cannot be extrapolated to three-dimensional transpressional (or transtensional, Fig. 9b) shear zones. Figure 10 presents plots on Flinn diagrams of the shapes of the finite strain ellipsoids in transpressional and transtensional zones of constant

TRANSPRESSION ZONES OF TRICLINIC SYMMETRY

53

strike length. They show the evolution with time of the shape of the finite strain ellipsoids for various values of ~/~ and dO.It is readily seen that transpression produces oblate strains (K <1) and transtension produces prolate strains (K >1).

Interpretation and discussion Application of theoretical modelling results The results of the theoretical modelling should be applied to natural shear zones with care and the following points should be kept in mind. (1) The present model assumes both isochoric deformation and constant strike length of the shear zone. A more general model that includes triclinic flow, volume change and strikelength change is given in Jiang & Williams (in press). (2) The theoretical model is for a homogeneous domain and a steady period, and the more closely these conditions are approximated by a natural shear zone the better the comparison will be. Heterogeneously deformed shear zones are best treated by dividing them into domains that better approximate the homogeneous condition, as is generally done in structural analysis, and the results of the present modelling can be readily applied to such shear zones. The heterogeneity arises in a number of ways. Not only strain magnitude, but also do and -)/~, may vary from point to point. Thus different parts of a shear zone may have different do value paths, different -9/~paths, and/or have been at different positions along any one of these paths. Structural data from natural shear zones do not necessarily, and generally do not, define complete paths as shown in Fig. 9. (3) We have assumed here that lineation and foliation are approximately parallel to the finite strain axes (kl and )tl)t2, respectively). This is not always true and the theoretical models should only be applied where there is good reason to believe that the assumption is valid. Taking the Roper Lake shear zone as an example (Fig. 4a), finite strain-related stretching lineations (Ls) were carefully differentiated from the 'ridge-ingroove'-type striations (Lc). Only Ls and poles to S-foliations are compared with the theoretical results. However, C-surfaces give an approximation of the shear zone boundary and plots of Lc indicate the shear direction (Figs 4a and 11). The scatter of Ls plots reflects the heterogeneity of the shear zone. The most significant point of these data is that the statistically defined mean for the stretching lineation of the marginal domain deviates from the great circle girdle defined by the data of the central domain.

Fig. 11. Comparison of the strain geometry of the Roper Lake shear zone (as shown in Fig. 4a) with the strain geometry predicted for a dextral oblique transpression zone with + = 50~ that strikes 45~ and dips 75~ The Roper Lake shear zone is interpreted as such a transpression zone with a higher ~/~ ratio in the central domain than in the marginal domain. Comparison of the Ls data with the evolution of ~k1 with time indicates that the central domain is more evolved (more deformed) than the marginal domain, consistent with the observations described above. The shaded areas are plots of mS and Ls from Fig. 4a. G, C, and Lc also correspond to those in Fig. 4a. Numerals 2, 4, 6 and 10 are values of "~/~..White arrows indicate the evolution with time (increasing finite strain) for the various "~/~values. VNS, Vorticity-normal section; SZB, shear zone boundary. Because lineations in a transpression zone may rotate as passive lines towards being steeper after their formation (Fossen et al. 1994), a + value obtained by comparing the observed with the predicted geometry may only be a maximum estimate.

Interpretation of the Roper Lake shear zone The structural geometry of the R o p e r Lake shear zone is characterized by variable orientations of the lineations and relatively constant orientations of foliations. Comparison with the results of the theoretical modelling indicates that the shear zone can be interpreted as a transpression zone with oblique boundary-parallel motion (i.e. oblique transpression) and a higher ~/~ ratio in the central domain than in the marginal domain (Figs 11 and 12). The latter explains why the stretching lineation is shallower in the former domain and steeper in the

54

S. LIN E T A L . quartz c-axis fabrics are stronger, in the central domain. These all indicate localization of the simple shear component (4/) there. Thus, we suggest that variation in the 4//~ratio is due more to the localization of the simple shear (4/) than to variation in the pure shear (~) across the shear zone (Fig. 12). As discussed below, oblique transpression (0 ~ < + < 90~ (and thus a triclinic movement picture) and variation in the 4//~ratio as a result of localization of 4/ are probably common features. O b l i q u e transpression a n d triclinic m o v e m e n t picture

Fig. 12. Schematic diagrams showing a general model of transpression zones. (a) Pre- deformational geometry. (e) Geometry after transpressional deformation. In the model, boundary-normal compression (~) and boundary-parallel shearing (4/) take place simultaneously. To assist visualization, the deformation component related to ~ is shown as (b). The narrow zone in (b) indicated by the dashed lines corresponds to that in (c) in which 4/is concentrated. The main features of the model are that (1) boundaryparallel shearing (4/) is oblique and contains both strike-slip (ss) and dip-slip (ds) components, and (2) boundary-parallel shearing (4/) is much more localized than boundary-normal compression (~). The movement picture is of triclinic symmetry. It should be noted that the narrow zone in which boundaryparallel shearing (4/) is localized does not have to be in the centre of the wider zone under boundary-normal compression (~) as shown here. latter domain. As described earlier, C-surfaces, asymmetrical mica fish and an oblique shape fabric in recrystallized quartz grains are much more intensely developed or only observed, and

Observations of present plate motions suggest that most convergent plate boundaries have an oblique displacement vector (e.g. Liu et al. 1995). Such an oblique convergence is often heterogeneously accommodated as a result of slip partitioning; i.e. the net oblique displacement is partitioned into a more dip-slip c o m p o n e n t accommodated at the subduction zone and a component accommodated in the overriding and underriding plates via intraplate deformation (e.g. Fitch 1972; DeMets 1992; McCaffrey 1992; Shen-Tu et al. 1995). The intraplate d e f o r m a t i o n shows evidence of further slip partitioning (e.g. Gao & Wallace 1995). However, a complete partitioning where the net oblique slip is partitioned into two end members: pure dip-slip and pure strike-slip components, as argued by, for example, Tikoff & Teyssier (1994, and references therein), may be rare. Liu et al. (1995) defined a parameter K to measure the degree of slip partitioning for a subduction zone. K = 0 (e.g. in northeastern Japan) and K = 1 (e.g. in New Hebrides) represent zero and complete partitioning, respectively (Liu et aL 1995, fig. 6). It is probable that K generally lies between zero and one, as for example in the Aleutians where K --~ 0.34 (Liu et al. 1995, fig. 6). The general incompleteness of slip partitioning implies that the imposed boundary displacements for many deformation zones are oblique; the boundary displacement vector may lie anywhere in the spectrum from dip-slip (+ = 90~ to strike-slip (+ = 0~ This gives rise to triclinic movement pictures for the internal deformation. Shear zones with triclinic movement pictures are probably far more common than reported in the literature. The observed monoclinic symmetry in natural deformation zones could well be reflecting a high 4//~ratio rather than a true monoclinic m o v e m e n t picture. As d e m o n s t r a t e d above, when the ratio of 4//~ is high, the structures and fabrics in an oblique transpressional zone (having a triclinic movement picture) will

TRANSPRESSION ZONES OF TRICLINIC SYMMETRY

55

Fig. 13. Schematic diagram showing the distribution, localization and partitioning of deformation across a convergent plate boundary. The boundary-normal motion results in crustal thickening (~) which is distributed over a wide region. The boundary-parallel motion resulting in shearing (9) is often localized into narrow deformation zones or faults. These zones can be transpressional if there is an appreciable ~ component, bulk simple shear if the ~ component is negligible relative to 9, or discrete faults where the deformation is extremely localized. It should be noted that these zones or faults may have the same or different degrees of movement obliquity (+), depending on the plate motion obliquity, slip partitioning and distribution. Integration of the motions accommodated by all these zones gives the relative motion of plates. exhibit monoclinic symmetry within the resolution of observation. It is likely that many natural shear zones described in the literature are only the equivalents of the central domain of the Roper Lake shear zone described above, in the sense that they only represent more intensely deformed portions (with higher ratios of "9/~) of much wider shear zones. W i t h o u t the less intensely deformed portions (or the equivalents of the marginal domain of the Roper Lake shear zone) being considered together with the more deformed portions, the potential triclinicity of these shear zones cannot be easily recognized, as exemplified by the Roper Lake shear zone.

Localization o f simple shear Studies of ancient shear zones (including the Roper Lake shear zone described above) and observations of current plate-boundary deformation indicate that in shear zones and orogens alike, the c o m p o n e n t of boundary-parallel motion (simple shear component -~) tends to be localized whereas the component of boundarynormal motion (pure shear component ~) tends to be widely distributed (see Gordon (1995) and references therein). Shear zones are localized features of much wider orogenic belts (Fig. 13). Theoretical models lead to the same conclusion. For example, for power law rheology with stress

exponents n = 3 and n = 10, the length/width ratio for a pure compression or extension zone is one and two, respectively, whereas it is five and ten, respectively, for a simple shear zone (England et al. 1985; England & Jackson 1989; Sonder & England 1986). This means that boundary-parallel motions are approximately five times more localized than boundary-normal motions. There are other factors that may contribute to this phenomenon. (1) Boundary-normal motion results in thickening or thinning of the crust leading to buoyancy forces that make further thickening or thinning more difficult (see also England et al. 1985; England & Jackson 1989). (2) Simple shear is believed to be a fabric weakening process whereas pure shear is believed to be a fabric hardening process (Williams & Price 1990; P. F. Williams unpublished). Much of deformation at the granular scale tends to take place along weak grain boundaries. Simple shear aligns elongate minerals and therefore makes grain boundaries ever closer to the shear plane orientation, thus making grain boundary sliding (diffusion dependent or frictional) an increasingly effective deformation mechanism. Pure shear, on the other hand, aligns grain boundaries progressively more perpendicular to the shortening direction. Thus grain boundary sliding becomes increasingly

S. LIN E T A L .

56

difficult a n d f u r t h e r d e f o r m a t i o n r e q u i r e s s t r o n g e r i n t r a g r a n u l a r m e c h a n i s m s to o p e r a t e . We t h e r e f o r e c o n c l u d e t h a t l o c a l i z a t i o n of b o u n d a r y - p a r a l l e l m o t i o n w i t h i n a w i d e r z o n e of c o m p r e s s i o n or e x t e n s i o n p e r p e n d i c u l a r to t h e z o n e b o u n d a r y is a g e n e r a l p h e n o m e n o n . L o c a l i z a t i o n d o e s n o t r e q u i r e p r e - e x i s t i n g surfaces of w e a k n e s s (e.g. faults, s h e a r zones, lithological b o u n d a r i e s , weak layers and/or r h e o l o g i c a l a n i s o t r o p y ) (cf. J o n e s & T a n n e r 1995), a l t h o u g h t h e p r e s e n c e of s u c h surfaces or zones will facilitate t h e process. The work is supported by the Geological Survey of Canada through funding to S. Lin and a contract to D. Jiang from the Canada-Nova Scotia Co-operation Agreement on Mineral Development (1992-1995), and by the Natural Sciences and Engineering Research Council of Canada through a research grant to P. F. Williams and a Post Doctoral Fellowship to D. Jiang. The manuscript was improved by reviews from H. Fossen, L. B. Goodwin, S. Hanmer and S. B. Lucas. This paper is Geological Survey of Canada Contribution 1996467.

R

e

f

e

r

e

n

c

e

s

CARON,A. & WILLIAMS,P. E 1988. The kinematic indicators of the Love Cove Group in the northeastern Newfoundland. Geological Association of Canada Program with Abstracts, 13, A17. DEMETS, C. 1992. Oblique convergence and deformation along the Kuril and Japan trenches. Journal of Geophysical Research, 97, 17615-17625. DUNNING, G. R., BARR, S. M., RAESIDE, R. P. & JAMIESON, R. A. 1990. U - P b zircon, titanite, and monazite ages in the Bras d'Or and Aspy terranes of Cape Breton Island, Nova Scotia: implications for igneous and metamorphic history. Geological Society of America Bulletin, 102, 322-330. ENGLAND,P. C. & JACKSON,J. 1989. Active deformation of the continents. Annual Review of Earth and Planetary Sciences, 17, 197-226. - - , HOUSEMAN, G. A. & SONDER, L. J. 1985. Length scales for continental deformation in convergent, divergent, and strike-slip environments: analytical and approximate solutions for a thin viscous sheet model. Journal of Geophysical Research, 90, 3551-3557. FITCH, T. J. 1972. Plate convergence, transcurrent faults, and internal deformation adjacent to Southeast Asia and the western Pacific. Journal of Geophysical Research, 77, 4432--4460. FOSSEN, H. & TIKOFF,B. 1993. The deformation matrix for simultaneous simple shearing, pure shearing and volume change, and its application to transpression-transtension tectonics. Journal of Structural Geology, 15, 413-422. , & TEYSSIER, C. 1994. Strain modeling of transpressional and transtensional deformation. Norsk Geologisk Tidsskrift, 74, 134-145.

GAO, L. & WALLACE, Y. C. 1995. The 1990 Rudbar-Tarom Iranian earthquake sequence: evidence for slip partitioning. Journal of Geophysical Research, 100, 15317-15332. GOODWIN, L. B. & WILLIAMS,P. E 1996. Deformation path partitioning within a transpressive shear zone, Marble Cove, Newfoundland. Journal of Structural Geology, 18, 975-990. GORDON, R. G. 1995. Plate motions, crustal and lithospheric mobility, and paleomagnetism: prospective viewpoint. Journal of Geophysical Research, 100, 24367-24392. JIANG, D. & WHITE, J. C. 1995. Kinematics of rock flow and the interpretation of geological structures, with particular reference to shear zones. Journal of Structural Geology, 17, 1249-1265. JONES, R. R. & TANNER, P. W. G. 1995. Strain partitioning in transpression zones. Journal of Structural Geology, 17, 793-802. • WILLIAMS, P. F. High-strain zones: a unified model. Journal of Structural Geology, in press. KRANTZ, R. W. 1995. The transpressional strain model applied to strike-slip, oblique-convergent and oblique-divergent deformation. Journal of Structural Geology, 17, 1125-1137. KROHE, A. 1990. Local variations in quartz C-axis orientations in non-coaxial regimes and their significance for the mechanics of S-C fabrics. Journal of Structural Geology, 12, 995-1004. LIN, S. 1992. The stratigraphy and structural geology of -

-

the southeastern Cape Breton Highlands National Park and its implications for the tectonic evolution of Cape Breton Island, Nova Scotia, with emphasis on lineations in shear zones. PhD thesis, University of New Brunswick, Fredericton. 1993. Relationship between the Aspy and Bras d'Or 'terranes' in the northeastern Cape Breton Highlands, Nova Scotia. Canadian Journal of Earth Sciences, 30, 1773-1781. 1995. Structural evolution and tectonic significance of the Eastern Highlands shear zone in Cape Breton Island, the Canadian Appalachians. Canadian Journal of Earth Sciences, 32, 545-554. & WILLIAMS, R E 1992a. The geometrical relationship between the stretching lineation and the movement direction of shear zones. Journal of Structural Geology, 14, 491-497. --& -1992b. The origin of ridge-in-groove slickenside striae and associated steps in an S-C mylonite. Journal of Structural Geology, 14, 315-321. & -1993. A transpressional model for interpreting the variation in the pitch of the stretching lineation across a shear zone. Geological Associ-

-

-

-

-

-

-

ation of Canada - Mineral Association of Canada Program with Abstracts, 18, A60. LISTER, G. S. & WILLIAMS, P. E 1979. Fabric development in shear zones: theoretical controls and observed phenomena. Journal of Structural Geology, 1, 283-297. LIU, X., MCNALLY,K. C. & SHEN, Z.-K. 1995. Evidence for a role of the downgoing slab in earthquake slip partitioning at oblique subduction zones.

TRANSPRESSION ZONES OF TRICLINIC SYMMETRY

Journal of Geophysical Research, 100, 1535115372. MCCAFFREY, R. 1992. Oblique plate convergence, slip vectors, and forearc deformation. Journal of Geophysical Research, 97, 8905-8915. MEANS,W. D., HOBBS, B. E., LISTER, G. S. & WILLIAMS, E E 1980. Vorticity and non-coaxiatity in progressive deformations. Journal of Structural Geology, 2, 371-378. RAMBERG, H. 1975. Particle paths, displacement and progressive strain applicable to rocks. Tectonophysics, 28, 1-37. RAMSAY, J. G. 1980. Shear zone geometry: a review. Journal of Structural Geology, 2, 83-89. & GRAHAM,R. H. 1970. Strain variation in shear belts. Canadian Journal of Earth Sciences, 7, 786-813. RAESIDE, R. P. & BARR, S. M. 1992. Geology of the northern and eastern Cape Breton Highlands, Nova Scotia. Geological Survey of Canada Paper 89-14.

ROBIN, P.-Y. E & CRUDEN,A. R. 1994. Strain and vorticity patterns in ideally ductile transpressional zones. Journal of Structural Geology, 16, 447-466. SANDERSON, D. J. & MARCHINI,W. R. D. 1984. Transpression. Journal of Structural Geology, 6, 449-458. SCHMID, S. M. & CASEY, M. 1986. Complete fabric analysis of some commonly observed quartz Caxis patterns. In: HOBBS, B. E. & HEARD, H. C. (eds) Mineral and Rock Deformation: Laboratory Studies. Geophysical Monograph, American

57

Geophysical Union, 36, 263-286. SHEN-TU, B., HOLT, W. E. & HAINES, A. J. 1995. Intraplate deformation in the Japanese Islands: a kinematic study of intraplate deformation at a convergent plate margin. Journal of Geophysical Research, 100, 24275-24293. SONDER,L. J. & ENGLAND,P. C. 1986. Vertical averages of rheology of the continental lithosphere: relation to thin sheet parameters. Earth and Planetary Science Letters, 77, 81-90. STARKEr, J. 1977. The contouring of orientation data represented in spherical projection. Canadian Journal of Earth Sciences, 14, 268-277. TIKOVF,B. & FOSSEN,H. 1993. Simultaneous pure and simple shear: the unifying deformation matrix. Tectonophysics, 217, 267-283. - & TEVSSIER, C. 1994. Strain modeling of displacement-field partitioning in transpressional orogens. Journal of Structural Geology, 16, 1575-1588. TRUESDELL, C. A. 1953. Two measures of vorticity. Journal of Rational Mechanical Analysis, 2, 173-217. & TOUPIN,R.A. 1960. The classic field theory. In: FL~3GGE,S. (ed.) Encyclopedia of Physics, Volume III: Principles of Classical Mechanics and Field Theory. Springer-Verlag, Berlin, 226-793. WILLIAMS, P. E & PRICE, G. P. 1990. Origin of kinkbands and shear-band cleavage in shear zones: an experimental study. Journal of Structural Geology, 12, 145-164.

Analogue modelling of faulting in zones of continental transpression and transtension GUIDO

SCHREURS t & BERNARD

COLLETTA 2

1Geological Institute, University o f Bern, Baltzerstrasse 1, CH-3012 Bern, Switzerland (e-mail: schreurs@geo, unibe, ch) 2Institut Franfais du POtrole, P.O. B o x 311, F-92506 Rueil Malmaison, France Abstract: Experiments were performed to simulate deformation in zones of continental

transpression and transtension. Stratified models consisted of brittle analogue materials overlying a thin layer of viscous material. Oblique deformation was obtained by combining a basal, distributed strike-slip shear component with either transverse shortening (transpression) or transverse extension (transtension). In transpression experiments the imposed ratio of shear strain rate and shortening strain rate exerts an important control on initial fault evolution in the brittle layers of the model. In those experiments with a relatively high strain rate ratio (>3.6), subvertical, en echelon strike-slip faults develop first, striking at angles of 25-37 ~ to the shear direction. With increasing strain several convergent strike-slip fault zones form displaying positive flower structures. In low strain rate ratio experiments (_<2.7), gently dipping (30-45~ downward converging thrust faults accommodate initial failure. They bound pop-up structures that strike parallel to the shear direction. Increasing strain results in a fault pattern dominated by oblique-slip reverse faults. Partial partitioning of fault motion occurs at late stages of strain when strike-slip faults form within popup structures. The strike-slip faults merge at depth with confining oblique-slip reverse faults, have a curved shape in plan view and a dip direction which changes along strike. Fault patterns can be used as kinematic indicators. En echelon strike-slip faults initially accommodate deformation in a transtension experiment and strike at low angles (6-10 ~ to the shear direction. With increasing strain, normal faults form parallel to older strike-slip faults. They develop as a result of partitioning of fault motion and gravity failure. There is good agreement between experimental results and natural examples of continental transpressional and transtensional tectonics.

The relative motion between lithospheric plates is often oblique and major regional structures have been interpreted as the result of either transpression or transtension. E x a m p l e s of areas considered to have undergone transpression are the Alpine Fault zone of n o r t h e r n South Island in New Z e a l a n d (Norris et al. 1990), the Pyrenees (Roure et al. 1989), the Mongolian Western Altai (Cunningham et al. 1996) the East Anatolian fault zone in Turkey (Lyberis et al. 1992), the Venezuela Andes (Colletta et al. 1997), and central California (Mount & Suppe 1987). E x a m p l e s of t r a n s t e n s i o n include the Gulf of California (Withjack & Jamison 1986), the Malawi Rift (Chorowicz & Sorlien 1992) and the North A e g e a n Sea in Greece (Pavlides et al. 1990). Deformation in the upper parts of the crust is d o m i n a n t l y a c c o m m o d a t e d by faulting and is generally thought to be partly controlled by distributed flow of the underlying ductile parts of the crust and mantle (England 1989). Deformation of continental lithosphere is generally distributed over zones of at least several hundreds of kilometres in width (Molnar & Taponnier 1975;

McKenzie & Jackson 1986; England 1989). At a smaller scale, oblique deformation of competent sedimentary rocks may occur above a weak layer consisting of salt, evaporites or overpressured shales. In various tectonic settings world-wide, oblique deformation in the upper mantle or lower crust can apparently be resolved into nearly pure tangential and normal components of strain in the upper parts of the crust ('strain partitioning'; after Lettis & H a n s o n (1991)), and fault motion partitions into subparallel, coeval strike-slip and dip-slip faults. Oldow et al. (1990) concluded that the coexistence of such coeval structures requires a common detachment system in the middle to lower crust or upper mantle. Namson & Davis (1988) proposed that strain associated with oblique convergence b e t w e e n the Pacific and N o r t h American plate during late Cenozoic time may be resolved into c o m p o n e n t s n o r m a l and tangential to the plate margin. The tangential c o m p o n e n t is responsible for strike-slip disp l a c e m e n t s along the San A n d r e a s fault, whereas the normal component produces thrust

SCHREURS,G. & COLLETA,B. 1998. Analogue modelling of faulting in zones of continental transpression and transtension. In: HOLDSWORTH,R. E., STRACnAN,R. A. & DEWEY,J. E (eds) 1998. Continental Transpressional and Transtensional Tectonics. Geological Society, London, Special Publications, 135, 59-79.

59

60

G. SCHREURS & B. COLLETTA

faults and folds that strike nearly parallel to the San Andreas fault. In most previous experimental studies simulating oblique deformation, movement along a single basement strike-slip fault (basal velocity discontinuity) was combined with a transverse displacement of the longitudinal sidewalls of the model (e.g. Richard & Cobbold 1990). In such experiments, faults in a brittle overburden consisting of sand or clay were directly connected to the basement fault and limited to its immediate vicinity. A few studies simulated deformation over a larger zone by placing either a rubber sheet (Naylor et al. 1986) or a viscous material (Richard 1991; Richard et al. 1995) on top of the basement strike-slip fault. Oblique deformation experiments in which the load was transmitted from the sides were performed by Gapais et al. (1991). Our experimental set-up differs from previous ones in that the shear component of oblique deformation was distributed at the base of the model over its entire width. Such an experimental approach was chosen to simulate oblique deformation of a model driven by basal distributed flow. Experiments were performed in a normal gravity field and the distributed strikeslip shear movement was combined with either transverse shortening or transverse extension. The laboratory approach allowed us to vary the ratio of shear strain rate and shortening strain rate from one transpression experiment to another and to determine genetic relationships between imposed b o u n d a r y conditions and resulting structures. In our particular experimental set-up, brittle and viscous analogue materials were used. The viscous material distributed the imposed shear deformation homogeneously at the base of the model and at the same time acted as a horizontal decoupling material because of its contrasting rheological behaviour with respect to the overlying brittle material. At crustal scale, the viscous analogue material represents a detachment level in the middle to lower crust, whereas at basin scale it simulates a weak layer (e.g. salt, evaporites or overpressured shales) within a sedimentary sequence. The brittle analogue material overlying the viscous material simulates uppercrustal rocks (crustal scale) or c o m p e t e n t sediments (basin scale). In this paper we discuss the geometric and kinematic evolution of faulting in zones of transpression and transtension at basin and crustal scale, investigate to what extent fault motion becomes partitioned, make comparisons with natural examples, and propose structural criteria

for identification of natural zones of transpression and transtension.

Analogue materials and scaling Granular material (dry quartz sand and glass powder with an average grain size of c. 100 ~m) and a viscous material called polydimethylsiloxane (PDMS) were used in our experiments. Sand and glass powder deform according to the C o u l o m b slip criterion, and their angles of internal friction (about 30~ for sand and 39~ for glass powder) are approximately similar to those determined experimentally for upper-crustal rocks (40 ~for normal stresses <2 kbar and 31 ~ for normal stresses in the range 2-20 kbar; Byerlee 1978). Therefore, sand and glass powder are considered to be appropriate for simulating brittle deformation in the upper crust (Horsfield 1977; Byerlee 1978). PDMS has a density of 0.965 g cm -3 and a Newtonian viscosity of 5 • 104 Pa s at room temperatures and at strain rates below 3 • 10-3 s-1 (Weijermars 1986). It is considered to be a good analogue material to simulate viscous rheology of materials such as salt or rocks in the lower crust (Vendeville et al. 1987). To link an analogue model to its natural example, scaling parameters have to be applied. The question of whether or not a model is correctly 'scaled' is controversial and concerns what parameters should be scaled and what values should be used in scaling functions. Our models were scaled for lengths, viscosities and time by means of methods discussed by Hubbert (1937) and Ramberg (1981). At basin scale, scale ratios between models and natural examples are 2 • 10.5 for length (1 cm in the model represents 500 m in nature), 5 • 10-15 for viscosities (implying a viscosity of 1019 Pa s for evaporites) and 6.25 x 10-l~ for time (1 h of experiment represents 180 000 years in nature). At crustal scale, scale ratios are 2 • 10-6 for length (1 cm in the model represents 5 km in nature), 5 x 10-17 for viscosities (implying a viscosity of 1021 Pa s for the lower crust) and 6.25 • 10-11 for time (1 h of experiment corresponds to 1.8 Ma in nature). Although Hubbert (1937) stressed the importance of completely and properly scaled models, in practice it is unlikely that we can construct a model that is correctly scaled with respect to all parameters involved. Limitations are imposed by the fact that certain parameters are either not well known or difficult to model. In our experiments parameters such as temperature gradient with depth, effects of pore pressure and differential compaction were not considered. Furthermore, the use of granular materials neglects the effect of cohesive

ANALOGUE MODELLING OF FAULTS

61

Fig. 1. Experimental apparatus used for modelling transpression. (a) Perspective view; for ease of presentation transverse borders consisting of rubber sheets are not shown. (b) Base of model. strength in natural rocks, because it is practically zero in dry sand (Naylor et al. 1986). Scaling of grain size is also not considered, and Horsfield (1977) pointed out that fault zone width in granular material is dependent on grain size. Thus, in general we can only construct partially scaled models of natural systems. Nevertheless, such models can be very useful. They force us to explicitly consider and test which parameters may control the development of a structure and

which are only of minor importance. Moreover, modelling generates ideas and helps us to create hypotheses concerning the origin and development of structures in nature that should be tested or verified by field studies.

Experimental set-up for transpression The base of the experimental apparatus used for modelling transpression consisted of two

62

G. SCHREURS & B. COLLETTA

Fig. 2. Experimental apparatus used for modelling transtension. (a) Perspective view; extendable rubber sheets providing transverse confinement are not shown. (b) Base of model. plates (Fig. 1), one of which could be moved past the other by a geared motor drive. Some 60 plexiglass bars, each 5 mm wide, 1 cm high and 70 cm long, were stacked like cards between two parallel wooden bars that were attached to the base plates. The plexiglass bars were transversally confined by a thin wooden strip on either side. One end of each strip (A in Fig. lb) was attached to the moving base plate, whereas the other end was allowed to slip along a small pin (B in Fig. lb) attached to the fixed base plate. As one of the base plates was displaced, the

confined plexiglass bars slipped past one another and the initial rectangular configuration changed into a parallelogram, inducing strike-slip shear deformation in the overlying brittle-ductile analogue model. At the same time, transverse shortening was produced by displacement of one of the longitudinal sidewalls (C in Fig. lb) that overlie the array of plexiglass bars and their confining wooden strips. The geared motor drives inducing the transverse shortening were attached to the base plate. Movement of the longitudinal sidewall

ANALOGUE MODELLING OF FAULTS and base plate occurred at preset velocities by stepper motors with computer control.

Experimental set-up for transtension Analogue modelling of transtension was done with the experimental set-up shown in Fig. 2. Two base plates were overlain by an alternation of 21 plexiglass (each bar was 5 mm wide, 1 cm high and 70 cm long) and 20 foam bars (each bar was about 8.5 mm wide, 1 cm high and 70 cm long). Plexiglass and foam bars were stacked like cards between two parallel wooden bars that were fixed to the overlying longitudinal sidewalls of the model. Transverse confinement of the bars was provided by a thin wooden strip on either side. One end of each strip (B in Fig. 2b) was attached to the overlying longitudinal sidewall, whereas the other end (A in Fig. 2b) was allowed to slip along a small pin fixed to the opposite longitudinal sidewall. Before constructing the stratified brittleviscous analogue model, the assemblage of foam and plexiglass bars was shortened by 2 cm. In the shortened state, the width of each foam bar was reduced to c. 7.5 mm, whereas the width of the plexiglass bars remained unchanged. The model was then constructed on top of the shortened assemblage. As one of the base plates was displaced laterally, the overlying assemblage of plexiglass and foam bars changed into a parallelogram (Fig. 2b). At the same time, the extensional component of deformation was produced by displacing the longitudinal sidewall (C in Fig. 2b). The presence of foam bars alternating with plexiglass bars prevented localization of the extensional component of deformation near the moving longitudinal sidewalls. During transtension foam bars were 'decompressed' and deformation was distributed at the base of the model. The main difference between the initial experimental set-up for transtension and transpression is that in transtension both 'basement' (foam and plexiglass bars) and overburden (brittle-viscous analogue model) are affected by the transverse movement, whereas in transpression only the overburden undergoes transverse shortening.

Model set-up and procedure The base of each analogue model consisted of a thin (5 mm) layer of PDMS, which was placed directly on top of the bars. Previous experiments (Schreurs 1992, 1994) have shown that such a layer distributes the imposed strike-slip shear deformation evenly over the entire width of the model and prevents localization of deformation above discontinuities presented by

63

adjacent plexiglass (and foam) bars. Sand and glass powder were alternately poured on top of the PDMS to produce a 3-cm-thick stratified sand-glass powder cake. A square grid of coloured sand markers was then traced on the upper free surface. The long dimension of each model (70 cm) was parallel to the shear component of deformation. All models had an initial width of 26 cm and transverse borders consisted of rubber sheets (not shown in Figs. 1 and 2). In this paper four experiments simulating transpression and one experiment simulating transtension will be discussed. Values of parameters used are given in Table 1. From one transpression experiment to the next, the shear strain rate component was varied to investigate its influence on fault development in brittle layers. In all experiments the shortening strain rate was the same. Strain rate ratio (defined as shear strain rate divided by shortening strain rate) ranged between 1.8 and 7.3. Models were analysed by X-ray computerized tomography (CT) (Hounsfield 1973). On the basis of attenuation of X-rays by materials, this non-destructive technique generates crosssectional images through analogue models and allows a detailed analysis of their internal geometry and kinematic evolution (Mandl 1988; Colletta et al. 1991; Schreurs 1994)." Sequential cross-sectional slices can be used to compute sections in any direction. The different attenuation values of sand, glass powder and PDMS make it possible to image a stratified model very clearly. Faults in granular materials correspond to dilatant zones along which the arrangement of grains has been perturbed. These zones have lower densities and thus lower attenuations by X-rays compared with unfaulted domains. This difference in attenuation allows visualization of faults in X-ray CT images. Using a digital image processing program, X-ray CT images were enhanced by creating a pseudo-relief at the interface between different layers of the model and at the contact between faults and unfaulted domains.

Experimental results The distributed shear component of oblique deformation was arbitrarily chosen to be dextral in all experiments. With increasing deformation distributed grain flow of the brittle layers gives way to discrete faulting as the dominant mechanism of strain accommodation. Transpression experiments are described first in order of decreasing strain rate ratio and are followed by a description of the transtension experiment.

64

G. SCHREURS & B. COLLETTA

Table 1. Experimental parameters for transpression (experiments 1661, 1764, 1820 and 1770) and transtension

(experiment 2031) Experiment 1661 Transpression X Transtension Duration of experiment (min) 132 Velocity of base plate (cm/h) 8 Total movement of base plate (cm) 17.6 Maximum shear strain (~/) 0.62 Shear strain rate (10-5s-1) 7.8 Shortening or extension velocity (cm/h) 1 Final width of model (initial width 26 cm) 23.8 Max. shortening or extension (in %) 8.5 Shortening or extension strain rate (10-Ss-1) 1.07 Strain rate ratio 7.3

1764

1820

1770

x

x

x

170 4 11.3 0.40 3.9 1 23.2 10.9 1.07 3.6

200 3 10 0.35 2.9 1 22.7 12.8 1.07 2.7

240 2 8.0 0.28 1.9 1 22 15.3 1.07 1.8

2031 x 120 3 6 0.23 2.9 1 28 7.8 1.07 2.7

It should be noted that shear strain and shear strain rate for transpression experiments were calculated using the width of the array of Plexiglas bars (28.5 cm), which was slightly larger than the initial width of the overlying model (26 cm). Model widths were used to calculate shortening and shortening strain rate.

Transpression experiments Experiment 1661 (strain rate ratio 7.3); Fig. 3. At early stages of transpression two separate zones of en echelon dextral strike-slip faults (synthetic Riedel shears, R) form in the sand-glass powder cake (Fig. 3a and b). Surface strike of individual fault segments ranges between 24 ~ and 30 ~ with respect to the longitudinal borders (Fig. 3a). Early faults are vertical and extend down to the base of the brittle layers (Fig. 3g). Domains of positive vertical relief are created in the area where two adjacent en echelon faults overlap (Fig. 3b). As the tips of individual fault segments propagate sideways with increasing deformation their surface strike and at the same time their dip angle change laterally (e.g. fault X in Fig. 3c). They obtain a slight sigmoidal shape in plan view ('lazy' z-shape), have a small reverse separation at each fault tip, and a dip direction which changes along strike (see fault X in sections A, B and C of Fig. 3h). Characteristic of these socalled scissor faults is that if one follows the hanging wall along strike it changes into the footwall and vice versa. With increasing deformation lower-angle dextral faults striking at 15~ to 0 ~ (Ri in Fig. 3c, with lower surface strikes than the first formed synthetic faults) develop in

the overlap area between adjacent left-stepping faults and link with earlier formed faults. This results in major anastomosing fault zones defining shear lenses. Several sinistral faults striking at angles of 65-70 ~ (lower-angle antithetic faults, RI in Fig. 3c; see Schreurs 1994) develop once major dextral convergent strike-slip fault zones have formed. The final stage of the experiment (Fig. 3e and f) shows that deformation is accommodated on several steeply dipping oblique-slip fault zones with an overall surface strike of about 15 ~ Dextral strike-slip movement along the faults dominates over the reverse dip-slip movement (Fig. 3e and i). Late sinistral faults accommodate only minor displacement. Sinistral faults and unfaulted domains between them undergo slight clockwise rotations about vertical axes (Fig. 3e). Reverse faults forming at late stages are a boundary effect caused by 'scissoring' in the acute corners of the model (Fig. 3e and f).

Experiment1764 (strain rate ratio 3.6); Fig. 4. As in the previous experiment, dextral strike-slip faults (synthetic Riedel shears, R in Fig. 4a) form at early stages of transpression. Faults are again arranged in an en echelon pattern and are left-stepping. However, their surface strike

Fig. 3. Transpression experiment 1661. (a, c, e) Line drawings after photographs of surface illustrate fault evolution at three consecutive stages of distributed transpression; fine lines represent coloured sand markers (initially square grids); bold lines represent traces of visible faults. R, synthetic, dextral strike-slip fault (Riedel shear); R'I, lower-angle antithetic fault; R1, lower-angle synthetic fault. X marks location of a scissor fault. (g-i) Vertical sections showing fault evolution at three stages. Location of sections is indicated by A-D in corresponding surface drawings. (b, f) Surface photographs of Stage 1 and 3, respectively. (d) Schematic surface view and cross-sections of scissor fault. FW, Footwall; HW, hanging wall.

A N A L O G U E M O D E L L I N G OF FAULTS

65

66

G. SCHREURS & B. COLLETTA

(angle c~, see Fig. 3a) is now larger, varying between 28 ~ and 37 ~. With continuing deformation the steep strike-slip faults become oblique-slip displaying a component of reverse separation. They link up to form through-going dextral convergent strike-slip fault zones. Gently dipping reverse faults (e.g. fault Q in Fig. 4a and b) with dips of 30-50 ~ form in the central part of the model between dextral convergent fault zones. The surface strike of these reverse faults is at high angles to earlier formed faults. With increasing deformation, reverse faults form in the lower half of Fig. 4c, striking parallel to the longitudinal borders and converging downward to form a pop-up structure (see Fig. 11c, below). Dextral strike-slip faults develop in part simultaneously with the reverse faults. They strike at low angles (c. 20 ~ to the reverse faults and branch at depth with them (fault S in Fig. 4f). Continuing deformation results in an oblique-slip movement along the faults forming the pop-up structure, and simultaneously more dextral strike-slip faults form within the pop-up structure. Vertical sections illustrate how subvertical strike-slip faults and gently dipping oblique-slip reverse faults occur alongside in the same section (e.g. section B in Fig. 4d and f). Fault dip may vary considerably along strike. The surface photograph and horizontal section through the interior of the model (2 cm above its base) illustrate the complex fault pattern at the end of the experiment (Fig. 4g and h) and show how fault zone width decreases with depth.

Experiment 1820 (strain rate ratio 2.7); Fig. 5. Fault evolution in this experiment is markedly different from that in the previous two experiments. Transpression at early stages is now taken up by thrust faulting (Fig. 5a) and not by strike-slip faulting. The uplifted block between the thrust faults (dipping at 34-47 ~) verges in opposite directions and defines a pop-up structure parallel to the longitudinal sidewalls. As the back-directed thrust is hindered by the sidewall, the forward directed thrust of the pop-up structure is predominantly active with increasing deformation. Dextral strike-slip faults develop within the pop-up structure (Fig. 5b). They are subvertical near the free surface but curve at depth and branch with earlier formed faults. In front of the pop-up structure new faults develop in response to additional deformation (Fig. 5b). Partly, they represent a second pop-up structure and partly steeply dipping strike-slip faults. These structures interfere laterally and as a consequence fault dips vary considerably along strike: from 30 ~ to subvertical (see sections B - E

in Fig. 5d). Slip along the faults defining the popup structure becomes oblique (dextral reverse) with increasing strain. Dextral strike-slip faults (striking at 15-25 ~ form as a result within popup structures. The strike-slip faults branch at depth with confining oblique-slip reverse faults and change their dip direction along strike (Fig. 6). Domains within each of the two pop-up structures have undergone a slight clockwise rotation about vertical axes with respect to the nearly unfaulted domain between. The surface photograph and horizontal section (Fig. 5e and f) illustrate the merging of strike-slip faults with an oblique-slip reverse fault zone (left-hand side) as well as the second anastomosing fault zone (right-hand side), in which strike-slip faults merge with or are relayed laterally by oblique-slip reverse faults.

Experiment 1770 (strain rate ratio 1.8); Fig. 7. The evolution of this experiment is very similar to that of the previous one. At early stages of transpression a pop-up structure forms striking parallel to the longitudinal sidewalls. The downward converging reverse faults initially dip at about 35 ~. With increasing deformation they become dextral oblique-slip reverse faults. A second pop-up domain develops in front of the first one and also develops a strike-slip component. Dextral strike-slip faults (striking at about 25 ~ develop within the pop-up domains (Fig. 7a). They have a curved shape in plan view and coalesce with reverse faults. Some strike-slip faults form between pop-up structures. Vertical sections through the final stage of deformation clearly show the two pop-up domains. Late strike-slip faults within pop-up domains are subvertical near the upper surface but curve at depth and branch with the obliqueslip reverse faults. In the case where a steep strike-slip fault branches at depth with both forward and backward directed oblique-slip reverse faults of a pop-up structure, its dip direction changes along strike. The horizontal section shows a doubly plunging anticline (upper part of Fig. 7b): the hinge zone defined by bulging layers in the pop-up domain does not remain horizontal but changes its plunge laterally. The structure is partly bounded and partly cut by dextral strike-slip faults which coalesce with major oblique-slip reverse faults. Transtension experiments Experiment 203I (strain rate ratio: 2.7); Fig. 8. Dextral strike-slip faults appear at early stages

A N A L O G U E MODELLING OF FAULTS

Fig. 4. Transpression experiment 1764. (a, c, e) Line drawings after photographs of surface and (b, d, f) vertical sections showing fault evolution at three consecutive stages of distributed transpression. Faults P, Q and S are explained in the text. (g) Surface photograph of final stage and (h) horizontal section through interior of model showing complex fault pattern at end of experiment. Position of horizontal section is indicated in accompanying vertical sections (lines H-I and J-K in (h) and in (e).

67

68

G. S C H R E U R S & B. C O L L E T T A

Experiment 1820 tstrain rate ratio = 2.7)

,l crn/tl snortemng .

.

.

.

3cm/h movement of base plate

oblique-slip reverse

.

b. Stage 2: y -- 0.28 & shortening = 10 %

a. Stage 1: shear y -- 0.13 & shortening = 4.5 % H

partial strain subparallel st faults and ob] reverse faults

i i

BI

A 0 |

5cm i

C D/ K E ,lazy' z-shaped strike-slip fault confined between dextral oblique-slip reverse faults

C. Stage 3: y = 0.35 & shortening =12.8 %

d.

e. Stage 3: Surface view

Fig. 5. Transpression experiment 1820. (a-e) Line drawings after surface photographs document structural evolution at three consecutive stages of deformation. (d) Vertical sections of Stage 3 show partial partitioning of fault motion. (e) Surface photograph of Stage 3 with subparallel strike-slip faults and oblique-slip reverse faults. (f) Horizontal section through interior of model illustrating sigmoidal strike-slip faults coalescing with bounding oblique-slip reverse faults. Relative position of horizontal section is indicated by lines H - I and J - K in accompanying vertical sections.

ANALOGUE MODELLING OF FAULTS

69

horizontal section through the interior of the model show the parallel arrangement of normal and strike-slip faults at low angles to the longitudinal borders of the model.

Discussion

Fig. 6. Schematic sketch illustrating branching of strike-slip fault with bounding, downward converging oblique-slip faults. (Note that the strike-slip fault changes its dip direction along strike.)

of transtension (Fig. 8a). They are left-stepping and strike at low angles to the longitudinal sidewalls (6-11~ Fault dips vary between 80 ~ and 90 ~ (Fig. 8a). With continuing deformation several synthetic fault zones form that interfere with each other laterally. Near the right transverse border of the model sinistral strike-slip faults form with their surface strike at about 50 ~ to the longitudinal sidewalls (Fig. 8a). Normal faults develop between earlier formed major dextral strike-slip faults with increasing deformation (Fig. 8b). The normal faults dip between 60 ~ and 70 ~ and lack horizontal offset. Pure normal faults and vertical strike-slip faults strike parallel to one another and indicate partitioning of fault motion. Some of the initially vertical strike-slip faults display an oblique slip component with increasing strain. More sinistral, antithetic faults develop between parallel arranged major strike-slip faults, and with increasing deformation their surface strike is at progressively lower angles (decreasing to 40-36~ During deformation dextral strike-slip faults propagate laterally and interact with other dextral strike-slip faults that formed elsewhere within the model. In case overlap between major faults is right-stepping, because of lateral approach of two major strike-slip fault zones, small transtensional basins form in the overlap area (Fig. 8b). The surface photograph and the

According to Anderson (1951) shear stress is lacking at the Earth's surface, implying that two principal stresses have to be horizontal and the third vertical. Consequently, failure of homogeneous brittle material at a horizontal free surface should occur by reverse, normal or strike-slip faulting, but not by oblique-slip faulting. At depth, however, the orientation of principal stress axes need not necessarily be vertical or horizontal, and oblique-slip faulting might be possible. In our experiments initial oblique deformation in brittle materials is taken up by either nearly pure strike-slip faults or nearly pure dip-slip faults. This indicates that principal stresses throughout our model initially lie within subhorizontal and subvertical planes. This implies that the thin layer of PDMS at the base of the model is very effective in decoupling the sand from the plexiglass bars and reducing the basal drag exerted on the sand by the plexiglass bars. Thus, in our models the basal shear stresses are considered to be negligible. Therefore, Anderson's theory seems appropriate for initial faulting in our experiments, which occurs in response to a stress field in which one principal stress direction is vertical and the other two lie within a subhorizontal plane. Depending on the strain rate ratio, initial faults in as yet unfaulted granular material will be generated either as pure strike-slip faults or pure dip-slip faults. Once major faults have formed, however, the sand-glass powder cake consists of competent unfaulted material and incompetent dilatant fault zones. Additional deformation will then mostly be taken up by oblique-slip along favourably oriented pre-existing faults. Transpression

Displacement paths of surface marker particles were constructed using a computer program developed by Bons et al. (1993). Marker paths are illustrated for five consecutive stages in two transpression experiments (Fig. 9). In both cases the displacement paths are linear and the most obvious difference is the angle of obliquity to the bulk shear direction. This angle [3 is about 0-7 ~ in the high strain rate ratio experiment and 15-20 ~ in the low strain rate ratio experiment. Although the difference in angle is relatively minor, the difference in fault evolution between the two experiments is significant (see Figs 3 and

70

G. SCHREURS & B. COLLETTA

Fig. 7. Transpression experiment 1770. (a) Line drawing after surface photograph showing final stage of experiment. Fault style is dominated by oblique-slip reverse faults and steep, slightly curved strike-slip faults. (b-c) Horizontal sections at different levels through the interior of the model showing decreasing fault width with depth. Note also the doubly plunging anticline in (b). Position of horizontal sections is indicated in accompanying vertical section by lines H-I and J-K. (d) Vertical sections through final stage of experiment showing coalescence at depth of strike-slip faults with bounding faults of pop-up structure. Some steep strikeslip faults occur between the two pop-up structures. 5). Fault style in our experiments clearly depends on the applied strain rate ratio of distributed shear and transverse shortening. In experiments with relatively high ratios (7.3 and 3.6 in experiments 1661 and 1764, respectively) steep strike-slip faults dipping at 80-90 ~ develop at early stages of strain. At a strain rate ratio of 7.3, the mean value of the surface strike of these faults is at 27 ~ (Fig. 10c) and at 33 ~ for a ratio of 3.6 (Fig. 10d). Thus, obliquity of fault strike (angle o~in Fig. 10) increases with decreasing strain rate ratio. In distributed strike-slip shear experiments without a shortening component (Schreurs 1992) the surface strike of

early strike-slip faults is at still lower angles, with a mean value of 20 ~ (Fig. 10b). Based on initial fault orientation, it is inferred that the far-field principal stress axis (%), which lies in a horizontal plane at about 45 ~ to the shear direction for distributed shear (Schreurs 1992), rotates progressively counter-clockwise for decreasing strain rate ratios in transpression experiments. The en echelon arrangements of early strikeslip faults are important kinematic indicators: left-stepping faults in each zone reflect the dextral shear component of transpression. Likewise, right-stepping faults would indicate sinistral transpression. In high strain rate ratio

Fig. 8. Transtension experiment 2031. (a-e) Line drawings after photographs illustrating fault evolution at three consecutive stages. Fault pattern is dominated by subparallel trending strike-slip faults and normal faults. (d) Detail of Stage 3. (e) Surface photograph of Stage 3. (f) Horizontal section through part of the interior of the model showing faults subparallel to longitudinal borders of model. Relative position of horizontal section is indicated by lines H-I and J-K in accompanying vertical sections. (g) Vertical sections through Stage 3 showing subparallel strike-slip and normal faults.

A N A L O G U E MODELLING OF FAULTS

71

72

G. SCHREURS & B. COLLETTA

Experiment 1661 (strain rate ratio = 7.3) ~

~

,1[-

Experiment 1820 (strain rate ratio = 2.7)

__~ 11,

~

~-------~~~~~~---~:~=~

~

~

~

~

~-,,-~--_.~_-~-~_~-_.~...

boundary effects ("scissoring in acute comers of model)

Fig. 9. Displacement paths of surface marker particles shown for two distributed transpression experiments. [3 is angle between average trend of displacement path and longitudinal borders of model. Curved displacement paths in lower left-hand corner of (a) are related to scissoring effect in acute corner of model. experiments (1661 and 1764) the shortening component of transpression may cause laterally propagating strike-slip faults to swing around subparallel to the longitudinal borders of the model. This results in slightly sigmoidal-shaped (in plan view) scissor faults characterized by a small oblique-slip reverse component at each fault tip and a dip direction that changes along strike. The shape in plan view may also be used as a kinematic indicator, i.e. 'lazy' z-shape for dextral shear component and 'lazy' s-shape for sinistral shear component of transpression (see also Mandl 1988; Richard et al. 1995). In vertical sections approximately perpendicular to the strike of a single scissor fault, the hanging wall changes along strike to become the footwall. En echelon arranged strike-slip faults and sigmoidalshaped scissor faults occur also in distributed strike-slip shear experiments (Naylor et al. 1986; Mandl 1988; Schreurs 1992; Richard et al. 1995). Early faults determine to a large extent the subsequent fault pattern and fault evolution, because they are favourably oriented for reactivation. In cases where strike-slip faults develop initially, continuing transpression results in several major, anastomosing fault zones, consisting of steep oblique-slip faults along which the strike-slip component dominates. Positive flower structures are characteristic of such convergent strike-slip fault zones with upward flattening of fault dip. Between these major subvertical fault zones, sinistral strike-slip faults form in the high strain rate ratio experiment 1661. They are not conjugate R' shear (see Tchalenko 1970), but lower-angle (surface strike) antithetic shears (R]), which reflect local anticlockwise rotations of o3 (lying within a horizontal plane) into a direction subparallel to the overall strike of major fault zones. Major

subparallel fault zones act as deformation corridors and cause local stress field modifications. In experiment 1764, o.1 also undergoes an anticlockwise rotation and becomes subparallel to the trend of dextral convergent strike-slip fault zones. However, secondary strike-slip faults do not form between the major fault zones as they do in distributed strike-slip shear experiments (Schreurs 1994) and experiment 1661, but instead gently dipping reverse faults form that strike at high angles to the major fault zone trend (e.g. fault Q in Fig. 4c). This indicates local changes in orientation of the intermediate stress axis (o2) from vertical for strike-slip faulting to horizontal for reverse faulting. In low strain rate ratio experiments (experiments 1820 and 1770) with a ratio of 2.7 and 1.8, respectively) early stages of deformation are accommodated by gently dipping (30-45 ~) thrust faults and not by strike-slip faults. Increasing deformation leads to a fault pattern dominated by oblique-slip reverse faults. The oblique-slip component is thought to be the result of the maximum compressive stress axis, which is not perpendicular to the longitudinal walls, but at an acute angle. The orientation of early thrust faults is favourable for reactivation by strikeslip. The oblique-slip reverse faults verge in opposite directions and define pop-up structures. Subvertical strike-slip faults generally form at late stages of deformation either between or within pop-up structures. Coalescence patterns of strike-slip faults with their bounding oblique-slip faults are useful for making inferences about overall kinematics (Fig. 11). In map view 'lazy' z-shaped sigmoidal strike-slip faults within dextral oblique-slip pop-up structures indicate dextral transpression (Fig. 11a), whereas 'lazy' s-shaped sigmoidal

ANALOGUE MODELLING OF FAULTS "~0~

a.

73

strain rate ratio = 3.6

Distributed transtension

b.

0~ = 6-11 ~ mean value = 8 ~

Distributed strike-slip (Schreurs 1992) 0~= 17-23 ~ mean value = 20 ~

1

strain rate ratio = 7.3

~' strain rate ratio = 3.6

Distributed transpression C.

d. c~ =25-30 ~ mean value = 27 ~

Distributed transpression C~= 28-37 ~ mean value = 33 ~

Fig. 10. Surface strike of early faults in distributed shear and oblique deformation experiments. (a) Distributed transtension. (b) Distributed strike-slip shear. (e, d) Distributed transpression. strike-slip faults within sinistral oblique-slip pop-ups would point toward sinistral transpression (Fig. 11b). The close proximity of simultaneously active strike-slip faults and oblique-slip reverse faults points to abrupt lateral changes in orientation of principal stress axes, with 0-2 changing from horizontal to vertical. In low ratio experiments, minor block rotations occur about vertical axes. D o m a i n s confined by downward converging oblique-slip reverse faults undergo a relative clockwise rotation with respect to domains between pop-up structures. In very low ratio experiments, domainal rotations of adjacent fault-bounded blocks are less prevalent. Partial partitioning of fault displacment occurs at advanced stages of deformation, especially in low strain rate ratio experiments. Partitioning is partial, because fault displacement is not strictly separated into coeval, parallel striking strike-slip faults and thrust faults ('strain partitioning', after Lettis & Hanson (1991)) but rather into partially coeval, subparallel striking strike-slip faults and oblique-slip thrust or reverse faults.

Transtension Although only one transtension experiment was carried out, it was included in this contribution.

A description of the innovative experimental set-up and the first results might be of use for future experimental studies. It is emphasized that the results of only one experiment have to be viewed with caution. A t early stages of transtension left-stepping, synthetic strike-slip faults accommodate initial failure of the brittle layers. Their surface strike of 6-10 ~ (Fig. 10b) is lower than the strike of early strike-slip faults in distributed strike-slip shear experiments (Schreurs 1992) and is due to the extensional c o m p o n e n t of transtension. Antithetic, sinistral strike-slip faults develop between major dextral strike-slip fault zones. The surface strike of newly developed antithetic faults tends to diminish with increasing deformation (53-36 ~) and points toward local stress field modifications: the orientation of the main compressive stress axis rotates counterclockwise toward the strike of the early formed synthetic strike-slip faults. With continuing deformation normal faults develop between and parallel to earlier formed major synthetic strikeslip faults. Normal faults dip at angles between 60 ~ and 70 ~, and have no or only limited horizontal offset. Their development implies local changes in the orientation of the maximum principal stress axis from horizontal for strike-slip faulting to vertical for normal faulting. The

74

G. SCHREURS & B. COLLETTA V e r t i c a l sections

Plan view A

C

E

D

F

A ~

~____B

I

B

a. dextral transpression

A

C

E

D

F

-v

b. sinistral transpression Fig. 11. Idealized coalescence pattern of strike-slip faults and downward converging oblique-slip faults of popup structure in plan view and in vertical sections for (a) dextral transpression and (b) sinistral transpression. faults result from partitioning of fault motion and gravity failure.

Comparison with natural examples Despite the inherent idealizations of analogue models and the fact that they are only partially scaled, the structures in our models bear close resemblance to natural examples of distributed oblique deformation. The Confidence Hills in the southern Death Valley dextral strike-slip fault zone (eastern California) form a well-exposed positive flower structure (Fig. 12; Dooley & McClay 1996). It resembles the structures at early stages of high ratio transpression experiments (experiments 1661 and 1764, e.g. Fig. 3b) between two overlapping, left-stepping and steeply dipping oblique-slip faults. Both in the experiment and in the Confidence Hills, flower structures are formed by doubly plunging anticlines that trend subparallel to the bounding fault segments. Fold development in the Confidence Hills is considered to be aided by the presence of a basal salt deposit (corresponding to the basal PDMS in our experiments) and by buttressing and uplift

along and against the bounding oblique-slip reverse faults (Dooley & McClay 1996). Similarly, folds developed subparallel to major bounding oblique-slip faults in the Mecca Hills in southern California (Damte 1996), where the basal detachment horizon consists of shale. Recent structural studies of the Mongolian Western Altai (Cunningham et al. 1996) have shown that the current deformation regime of the area is transpressional with characteristic flower structures consisting of several active, dextral strike-slip master faults (>200 km long) and oblique-slip thrust faults with opposing vergence (Fig. 13). In plan and cross-section the structures of the Western Altai show strong similarities to advanced stages of our high ratio analogue models. Cunningham et al. (1996) also described the southern termination of a major dextral strikeslip fault in the Jargalant Range of the Western Altai (Fig. 14). Displacement along the fault system is accommodated by thrust faulting, oblique-slip faulting and uplift within the Jargalant Range. Cross-sectional geometry of the Range is that of an asymmetric flower structure. This geometry is similar to fault evolution from

ANALOGUE MODELLING OF FAULTS

75

Fig. 12. (a) Location of Confidence Hills within the southern Death Valley fault zone (SDVFZ) in eastern California and (b) 3D synoptic model of part of the Confidence Hills. Modified after Dooley & McClay (1996).

Fig. 13. Block diagram interpretation of part of Mongolian Western Altai, showing the High Altai, Sutai and Jargalant structural domains, each consisting of large-scale flower structures related to dextral transpressional strike-slip fault systems. Modified after Cunningham et al. (1996).

Fig. 14. Map showing location of Jargalant Range (Mongolian Western Altai). Dextral displacement along the Har Us Nuur fault system is accommodated by thrust faulting, oblique-slip faulting and uplift within the Jargalant Range (from Cunningharn et al. 1996).

76

G. SCHREURS & B. COLLETTA

Fig. 15. (a) Simplified structural map of the northwestern part of South America, showing the location of the Merida Andes and the dextral Bocono Fault in Venezuela (modified after Colletta et al. (1997). (b) Schematic crustal block diagram through the Merida Andes. Location of section is indicated in (a). stage 1 to 2 in the high ratio experiment 1661 (Fig. 3a-d), in which laterally propagating subvertical dextral strike-slip faults terminate by swinging around into oblique-slip reverse faults. A natural example of low ratio transpression experiments is possibly provided by the Merida Andes in Venezuela (Fig. 15a and b). This mountain belt is the result of Neogene dextral transpression between the Maracaibo plate and stable South America (Colletta et al. 1997). Oblique convergence is partitioned in the upper crust, with two conjugate northwest- and southeast-verging contractional structures, orthogonal to the plate boundary, and a major

intervening northeast-striking dextral strike-slip structure, the Bocono fault (Colletta et al. 1997). In the interpretation by Colletta et al., the steep Bocono fault is confined to the tectonic wedge and does not extend deeper into the crust than the sole thrust of the allochton. The structures in the transtension experiment are similar to those observed in the North Aegean Sea (Fig. 16), where dextral transtension has occurred since Late Miocene time (Lyberis 1985). Deformation occurs by a combination of en echelon arranged, E N E - W S W striking faults with normal character and subparallel striking steep faults with important

ANALOGUE MODELLING OF FAULTS

77

Fig. 16. Tectonic sketch map of North Aegean region (modified after Pavlides et al. (1990)). dextral strike-slip displacement (Pavlides et al. 1990). NW-SE striking faults have possibly originated as sinistral faults formed between major dextral fault zones and would correspond in that case to antithetic, sinistral strike-slip faults in the experiment which form after major dextral strike-slip faults have developed.

Conclusions Fault patterns at different stages of each experiment provide important information on overall kinematics, local stress field modifications, and partial partitioning of fault motion. The early fault style in transpression experiments clearly depends on the applied strain rate ratio. This ratio determines whether initial failure in the brittle layers is accommodated by steep strike-slip faults or by thrust faults. In high strain rate ratio experiments (_>3.6) steep strikeslip faults (dipping at 80-90 ~ formed early. Their en echelon arrangement can be used as an indicator of the overall sense of shear (i.e. a leftstepping pattern indicates dextral shear). The sigmoidal trace of strike-slip faults that laterally become oblique-slip reverse faults can also be used as kinematic indicator ('lazy' z-shape for dextral shear and 'lazy' s-shape for sinistral shear component). Both these kinematic indicators occur also in pure strike-slip fault systems (Naylor et al. 1986; Mandl 1988; Schreurs 1992; Richard et al. 1995). The difference with pure strike-slip fault systems is the obliquity of surface strike (with respect to the regional shear zone boundaries) of early strike-slip faults, which is larger in transpression experiments. Obliquity of surface strike in transpression

experiments increases as the strain rate ratio decreases. In low strain rate ratio experiments (<2.7) pairs of thrust faults (dipping at 30-45 ~) initially form parallel to the direction of shear. These faults have opposite vergence and bound pop-up structures. To a large extent, older faults determine the subsequent fault pattern and evolution. Where strike-slip faults develop initially during transpression, further deformation creates several major fault zones consisting of steep, obliqueslip faults along which the strike-slip component dominates. Positive flower structures with upward flattening of fault dip are characteristic of these convergent strike-slip fault zones. In experiments where gently dipping reverse faults initially accommodate transpression, an increase in strain leads to a fault pattern dominated by oblique-slip reverse faults (dextral component along the reverse faults indicates overall dextral shear component). Secondary faults forming between older major fault zones reflect local stress field perturbations that differ from the far-field stress system. Partial partitioning of fault displacement occurs at late stages of transpression, especially in low strain rate ratio experiments. Subvertical strike-slip faults generally form late between or within pop-up structures. The strikeslip faults strike subparallel to oblique-slip reverse faults and are simultaneously active. These strike-slip faults often merge at depth with oblique-slip reverse faults and generally have a curved fault trace and a dip direction which changes along strike. The fault pattern can be used as a kinematic indicator: 'lazy' zshaped strike-slip faults confined between

78

G. SCHREURS & B. COLLEq-TA

d e x t r a l o b l i q u e - s l i p r e v e r s e faults i n d i c a t e dextral transpression, w h e r e a s 'lazy' s-shaped strike-slip faults b e t w e e n sinistral oblique-slip reverse faults point to sinistral transpression. Vertical strike-slip faults a c c o m m o d a t e deform a t i o n of brittle layers at e a r l y stages of a transtension experiment. The en echelon a r r a n g e d faults strike at low angles (7-10") to the bulk shear direction (left-stepping faults indicate dextral shear c o m p o n e n t ) . W i t h c o n t i n u i n g d e f o r m a t i o n n o r m a l faults (dipping at 60-70 ~) form, w h i c h t r e n d parallel to o l d e r strike-slip faults a n d d e v e l o p as a result of partitioning of fault m o t i o n and gravity failure. T h e r e is g o o d a g r e e m e n t b e t w e e n our experim e n t s a n d n a t u r a l e x a m p l e s of c o n t i n e n t a l transpressional and transtensional tectonics. T h r e e - d i m e n s i o n a l imaging of a n a l o g u e m o d e l s m a y p r o v i d e constraints for g e o m e t r i c and kinematic i n t e r p r e t a t i o n s of c o m p l e x structures in natural zones of oblique d e f o r m a t i o n . J. Letouzey is thanked for providing the opportunity to carry out the experiments at the Institut Franqais du P6trole (Rueil Malmaison, France), M. Herwegh for help in constructing displacement paths, J.-M. Mengus for technical assistance, and the scanner group (M.-T. Bieber, F. Lamy and C. Schlitter) for its help in obtaining the X-ray CT images. The manuscript benefited from constructive reviews by T. Dooley and A. Sylvester. Financial support from the Swiss National Science Foundation is gratefully acknowledged.

References ANDERSON,E. M. 1951. The Dynamics of Faulting and

Dyke Formation with Applications to Britain. Oliver and Boyd, Edinburgh. BONS, P. D., JESSELL, M. W. & PASSCI4IER,C. W. 1993. The analysis of progressive deformation in rock analogues. Journal of Structural Geology, 15(3-5), 403-411. BYERLEE,J. 1978. Friction of rocks. Pure and Applied Geophysics, 116, 615-626. CI4OROWICZ,J. & SORLIEN,C. 1992. Oblique extensional tectonics in the Malawi Rift, Africa. Geological Society of America Bulletin, 104, 1015-1023. COLLETTA,B., BALE,P., BALLARD,J. E, LETOUZEY,J. & PINEOO, R. 1991. Computerized X-ray tomography analysis of sandbox models: examples of thinskinned thrust systems. Geology, 19, 1063-1067. , ROURE, E, DE TONI, B., LOUREIRO, D., PASSALACQUA, H. & Gou, Y. 1997. Tectonic inheritance, crustal architecture and contrasting structural styles in the Venezuela Andes. Tectonics, 16, 777-794. CUNNINGHAM, D., WINDLEY, B. E, DORJNAMJAA, D., BADAMGAROV,G. & SAANDAR,M. 1996. A structural transect across the Mongolian Western Altai: active transpressional mountain building in central Asia. Tectonics, 15, 142-156.

DAMTE, A. 1996. Styles of deformation in zones of oblique convergence: an example from the Mecca Hills, southern San Andreas Fault. PhD thesis, University of California, Santa Barbara. DOOLEY,T. P. & MCCLAu K. R. 1996. Strike-slip deformation in the Confidence Hills, southern Death Valley fault zone, eastern California, U.S.A. Journal of the Geological Society, London, 153, 375-387. ENGLAND,P. 1989. Large rates of rotation in continental lithosphere undergoing distributed deformation. In: KISSEL, C. & LAJ, C. (eds)

Paleomagnetic Rotations and Continental Deformation, Kluwer, Dordrecht, 157-164. GAPAIS, D., FIQUET, G. & COBBOLD, P. R. 1991. Slip system domains, 3. New insights in fault kinematics from plane-strain sandbox experiments. Tectonophysics, 188, 143-157. HORSFIELD,W. T. 1977. An experimental approach to basement-controlled faulting. Geologie en Mijnbouw, 56, 363-370. HOUNSFIELD, G. N. 1973. Computerized transverse axial scanning (tomography). British Journal of Radiology, 46, 1016--1022. HUBBERT, M. K. 1937. Theory of scale models as applied to the study of geologic structures. Geo-

logical Society of America Bulletin, 48,1459-1520. LETrIS, W. R. & HANSON,K. L. 1991. Crustal strain partitioning: implications for seismic-hazard assessment in western California. Geology, 19, 559-562. LVBERIS, N. 1985. G~odynamique du domain egOen depuis le Miocene sup~rieur. Th6se Doctorat d'6tat, Universit6 Paris VI. --, YuRuR, T., CHOROWlCZ, J., KASAPOGLU, E. & GUNDOGDU, N. 1992. The East Anatolian Fault: an oblique collisional belt. Tectonophysics, 204, 1-15. MANDL, G. 1988. Mechanics of Tectonic Faulting. Elsevier, Amsterdam. MCKENZIE, D. & JACKSON,J. 1986. A block model of distributed deformation by faulting. Journal of the Geological Society, London, 143, 349-353. MOLNAR, P. & TAPONNIER,P. 1975. Cenozoic tectonics of Asia: effects of a continental collision. Science, 198, 419-426. MOUNT, V. S. & SUePE, J. 1987. State of stress near the San Andreas fault: implications for wrench tectonics. Geology, 15, 1143-1146. NAMSON, J. S. & DAVIS,T. L. 1988. Seismically active fold and thrust belt in the San Joaquin Valley, central California. Geological Society of America Bulletin, 100, 257-273. 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 Geology, 8, 737-752. NORRIS, R. J., KOONS,R O. & COOPER,A. E 1990. The obliquely-convergent plate boundary in the South Island of New Zealand: implications for ancient collision zones. Journal of Structural Geology, 12, 715-725. OLDOW,J. S., BALLY,A. W. & Av~ LALLEMANT,H. G. 1990. Transpression, orogenic float, and lithospheric balance. Geology, 18, 991-994.

A N A L O G U E MODELLING OF FAULTS PAVLIDES,S., MOUNTRAKIS,D., KILIAS,A. & TRANOS, M. 1990. The role of strike-slip movements in the extensional area of Northern Aegean (Greece). A case of transtensional tectonics. Annales Tectonicae, 4, 196-211. RAMBERG, H. 1981. Gravity, Deformation and the Earth's Crust. Academic Press, New York. RICHARD, P. 1991. Experiments on faulting in a twolayer cover sequence overlying a reactivated basement fault with oblique (normal-wrench or reverse wrench) slip. Journal of Structural Geology, 13, 459--469. - & COBBOLD,P. 1990. Experimental insights into partitioning of fault motions in continental convergent wrench zones. Annales Tectonicae, 4, 35-44. - - , NAYLOR, M. A. & KOOPMAN,A. 1995. Experimental models of strike-slip tectonics. Petroleum Geoscience, 1, 71-80. ROURE, E, CHOUKROUNE, P. & ECORS PYRENEES TEAM 1989. ECORS deep seismic data and balanced cross sections: geometric constraints on the evolution of the Pyrenees. Tectonics, 8, 41-50. SCHREURS, G. 1992. Analogue modelling using X-ray

79

computed tomography analysis: experiments on distributed strike-slip shear deformation. Institut Fran~ais du P6trole. Rueil Malmaison, Report 39893.

1994. Experiments on strike-slip faulting and block rotation. Geology, 22, 567-570. TCHALENKO, J. S. 1970. Similarities between shear zones of different magnitudes. Geological Society of America Bulletin, 81,1625-1640. VENDEVILLE, g., COBBOLD, P. R., DAVY, P., CHOUKROUNE, P. & BRUN, J. R 1987. Physical models of extensional tectonics at various scales. In: COWARD,M. E, DEWEY, E & HANCOCK,P. L. (eds) Continental Extensional Tectonics, Geological Society, London, Special Publications, 28, 95-107. WEIJERMARS, R. 1986. Flow behaviour and physical chemistry of bouncing putties and related polymers in view of tectonic laboratory applications. Tectonophysics, 124, 325-258. WITHJACK,M. O. & JAMISON,W. R. 1986. Deformation produced by oblique rifting. Tectonophysics, 126, 99-124.

The structural response to evolving plate kinematics during transpression: evolution of the Lebanese restraining bend of the Dead Sea Transform R. W. H. B U T L E R 1, S. S P E N C E R 2 & H. M. G R I F F I T H S 1

1Department of Earth Sciences, The University of Leeds, Leeds LS2 9JT, UK 2Department of Geology, American University of Beirut, Beirut, Lebanon Abstract: Structural evolution along continental transform faults may be related to fault zone geometry and to regional variations in plate kinematics. Using a case study of the Lebanese sector of the Dead Sea Transform, the finite geometry of transpression at a restraining bend can be shown to have evolved in time. Relative structural chronologies, calibrated against dated landscape features such as lava-covered palaeosurfaces, coastal erosion surfaces and their incised drainage basins, are used to establish the timing of displacement activity on the major transcurrent faults. The early part of the region's structural history, up to late Miocene times, was controlled by the geometry of the through-going Yammouneh Fault. Transpression on this right-trending left-lateral structure was accommodated by strike-slip and distributed crustal shortening represented by the initial uplift of Mount Lebanon. For the past 6 Ma the principal active strand of the transform has been the Roum Fault. For much of this period it is presumed to have been a through-going fault which accommodated about 30 km left-lateral displacement. During the Quaternary, the fault zone has become strongly segmented. Although the location of active transcurrent faulting has migrated during the history of the transform, the major site of crustal shortening, the Mount Lebanon-Jabel Barouk structure, has remained broadly fixed. However, the rates of amplification of this structure and the coastal flexure appear to have varied. Continuing uplift of Mount Lebanon and local Plio-Quaternary folding suggest that the offshore continuation of the Roum Fault contains a rightwards, transpressive bend. We relate this multistage history for the Lebanese sector of the transform to an evolving plate tectonic setting: the rotation pole for relative plate motion between Africa and Arabia has migrated through time and the triple junction between the transform and the Tethyan destructive plate margin to the north has moved from onshore SE Turkey to now lie in the NE Mediterranean. Our case study illustrates the transient and evolving nature of deformation in continental restraining bends.

Continental transform faults are characterized by the interplay between strike-slip faulting and additional strains which accommodate crustal shortening or extension. Transpression has been used to explain major mountain ranges, transtension invoked to explain tectonic exhumation of deep crustal rocks and the origin of some sedimentary basins (e.g. Christie-Blick & Biddle 1985). These p h e n o m e n a may occur within the same broad zone of continental deformation, in some cases following in close succession at the same site (e.g. Crowell & Sylvester 1979). For ancient examples where the far-field plate motions are difficult to establish, it may also be difficult to assess the larger-scale controls on variations in transpression and transtension in space and time along a particular transform system. Many segments of the world's active plate boundary system may be inferred to have evolved kinematically, even over the past few million years. Consequently, to relate the tectonic response of a continental deformation zone

to evolving plate kinematics, it is useful to examine active systems that have accommodated displacements which, on the scale of relative plate motions globally, are low. H e r e we examine part of the D e a d Sea Transform, a continental plate boundary system that satisfies the n e e d for youthful tectonics and low displacements (e.g. Joffe & Garfunke11987). The transform contains one of the most obvious of the world's restraining bends, largely contained within Lebanon. It is this region that forms the focus of this study. It has been known for 40 years (Quennel11958) that the D e a d Sea System forms a left-lateral transform system linking the plate divergence in the Red Sea and Gulf of Suez in the south with plate convergence and continent-continent collision in the north. Extensive reviews of plate-scale geometry and the setting of the D e a d Sea Transform have been provided by Garfunkel et al. (1981), Q u e n n e l l (1984), H e m p t o n (1987), Girdler (1990) and, most thoroughly, by Joffe & Garfunkel (1987) and Westaway (1994).

BUTLER,R. W. H., SPENCER,S. & GRIFFITHS,H. M. 1998. The structural response to evolving plate kinematics during transpression: evolution of the Lebanese restraining bend of the Dead Sea Transform. In: HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. E (eds) 1998. Continental Transpressionaland Transtensional Tectonics. Geological Society, London, Special Publications, 135, 81-106.

81

82

R.W.H. BUTLER E T A L . 35 ~

30 ~

z5 o

East

Anatolian Fault

~"

I

~ld triple junction 1 6hab F a u l t

~ 3 5 ~

"1 ~.Mount Lebanon tronspressiv@ overlap Fault

Miocene pole further west

Pliocene pole

~

~11

..

b

~I

~

~

~ ~ _ _

Deed Sea

.i

".-. ,~.

i~i~i

plate 30 ~

Sinai Gulf of Aqobo frensfensive overlaps

--migrating ~

~

I.

7

frou.tens,on 35 ~

Fig. 1. (a) The active plate tectonics of the Dead Sea Transform, modified after Quennell (1984) and Girdler (1990). Their rotation pole for post-Miocene spreading of the Red Sea and relative motion between the Arabian and African plates places the southern part of the transform into weak transtension (Joffe& Garfunkel 1987), as indicated by active rifting in the Gulf of Aqaba. The Roum Fault forms a near-ideal small circle segment about the rotation pole and thus is a near-ideal transcurrent feature (Girdler 1990). The estimated rotation pole for seismicity along the Dead Sea Transform presented by Van Eck & Hofstetter (1990) is indicated, just north of the Nile Delta in SE Mediterranean. Small-circle arcs from this pole are indicated (dashed lines labelled r). Continuing coastal uplift and folding in Lebanon suggests transpression and a rightwards bend in the transform, presumably along the Levant continental margin west of Mount Lebanon. This area and the sea bed of the NE Mediterranean remain poorly known. Hence the position of the plate boundary triple junction between the Dead Sea Fault System and the Cyprus arc is uncertain. (b) The kinematic effects of the migration of a pole of rotation describing the relative motion on a transform fault, schematically illustrated for the Miocene-Recent history of the Dead Sea system (A-C in time). The resolved motion of Arabia (eastern side of transform) relative to Africa (western side) is indicated (arrows). For pole A the transform shows only transcurrent motion. Pole B generates transpression and transtension to the north and south sides, respectively (dark arrows). Pole C shows similar, but more intense, behaviour (lighter arrows).

G a r f u n k e l (1981) sited t h e t i m e - a v e r a g e d rotation pole for the D e a d Sea Transform during the P l i o - Q u a t e r n a r y at 32.8~ 22.6~ + 0.5 ~ just south of Crete (Fig. l a ) and suggested that the pre-Pliocene pole was located 4-5 ~ further west (Ionian Sea). Thus with time the rotation pole describing relative plate m o t i o n across the D e a d Sea Transform has m i g r a t e d to be closer to the fault. T h e c o n s e q u e n c e of a migration of rotation pole towards a transform fault is to tighten

the radius of curvature that describes the relative m o t i o n across the fault. T h e critical divide lies at the (radial) intersection of the linear projection linking the two rotation poles with the t r a n s f o r m fault. H e r e the o r i e n t a t i o n of t h e t a n g e n t for b o t h poles is identical. O n either side of this u n i q u e point on the transform fault the kinematics will change to either t r a n s t e n s i o n (relative plate divergence) or transtension (relative plate convergence). For a left-lateral fault

EVOLUTION OF THE LEBANESE RESTRAINING BEND with an approaching rotation pole to its west, points south of the radial intersection move into transtension whereas those to the north move into transpression (Fig. lb). Geometric changes in the structure of the Dead Sea Transform are well recorded for its southern and central parts, in Israel. Garfunkel (1981) described arrays of left-stepping transtensive relays in the Gulf of Aqaba (Fig. la) and further north. He estimated about 35--40 km of basin-opening of Plio-Quaternary age, indicative of a displacement rate of about 6 mm/a (Joffe & Garfunkel 1987). To the north of the Dead Sea itself, Heimann & Ron (1993) described the change from strike-slip to deformation with a strong compressive component on an array of structures associated with a strand of the transform (Almagor Fault). They ascribed the change to variations in the internal geometry of the transform during the Quaternary, presumably in response to evolving plate kinematics. There have, however, been few attempts to examine structural evolution in the Lebanese sector of the transform. This paper examines tectonic changes along the Lebanese area of the Dead Sea Transform system, linking structural evolution within the finite restraining bend to changes in plate kinematics that accompanied the migration of the Arabia-Africa rotation pole. Extensive folding and uplift appears to be associated with finite transpression; indeed, the Lebanese sector of the transform is one of the world's classic examples of a large-scale restraining bend. However, this finite description might be expected to be misleadingly simple given the variations in plate motion through time and the variations in transform behaviour evident further south (Garfunkel 1981). Our aims are to document the styles of deformation in the upper crust that characterize regional transpression on a major restraining bend, to show that this restraining bend has been bypassed so that parts are now experiencing transtension, and to relate these to changes in regional plate kinematics. We use quantitative landscape stratigraphy to date structural activity, thereby illustrating the time-scales of tectonic processes operating within continental transform systems. We examine the interplay between intra-transform fault activity and variations in relative plate vectors in controlling the switch from transpressive to transtensive behaviour.

Tectonic setting The Dead Sea Fault System forms part of the plate boundary network of the Mediterr a n e a n - A r a b i a n region linking sea-floor

83

spreading in the Red Sea with the destructive southern margin of the Tethyan collision belt (Fig. 1). The structure of the southern part of this fault system is relatively simple, with deformation focused along a narrow depression running up the Gulf of Aqaba, along the Dead Sea and along the Jordan valley (Quennel11958; Freund et al. 1970). A variety of pre-existing markers have been correlated across the fault system, including the continental margin across the Gulf of Aqaba (e.g. Girdler 1990), to establish bulk displacements (reviewed by Zak & Freund 1981; Quennell 1984; Hempton 1987; Joffe & Garfunke11987). Zak & Freund (1981), amongst others, estimated that the total left lateral displacement across the Dead Sea Fault System is to be 105 km, based on correlations across its southern segments. A swarm of diabase dykes dated at 22-18 Ma show the full offset (Eyal et al. 1981) and provide a maximum age for the onset of transform activity. The displacement apparently accumulated in two distinct periods (Quennell 1984; Hempton 1987; Joffe & Garfunkel 1987). The first of these occurred up to the late Miocene and accounted for 65 km displacement, at about 0.5 cm/a. The remaining 40 km of displacement has accumulated since the earliest Pliocene (> c. 4.5 Ma), at about i cm/a. The structure and kinematics of the Dead Sea Transform are more complex and obscure to the north. Ron & Eyal (1985), amongst many others (e.g. Quennell 1984), were able to recognize only about 25 km of left-lateral displacement in Lebanon and western Syria. Cenozoic deformation is distributed across a range of faults and folds from the Levant coast across to the Palmyrides of Syria. The timing and significance of these structures is highly controversional (e.g. Chaimov et al. 1990). To a large degree, the problems of identifying fault displacements have arisen because the surrounding rocks are dominantly sedimentary with originally subhorizontal stratigraphy; geometries which are not well disposed to charting strike-slip displacements. More subtle stratigraphic markers, such as lateral facies variations, have yet to be identified in the rather poorly studied, carbonate-dominated Mesozoic successions of Lebanon. Between 1975 and 1992, field geology was not a practical proposition in the country. Within Lebanon, the Dead Sea Transform is represented by some or all of an array of fault strands (e.g. Beydoun 1977; Arthaud et al. 1978; Walley 1988; Fig. 2). Most large-scale reviews of plate boundary continuity (e.g. Hempton 1987; Kempler & Garfunkel 1994) consider one of these structures, the Yammouneh Fault, to be

84

R.W.H. BUTLER ETAL.

•

Neoqene basalt

~

Neogene sedimenfs

l--]

older

:. '.~,i.;.: F."!:". ..-. ] : :..

strata

~~i)hs

Terbo,

." ," .

anticline

lie du

o i

I

7r._~poli

30 9

"v

I

"km

Joum~ Be, r u t

~Z

" :v :..-;... 9

9...-.;;:'. ,. ~.- .-.

~.~, . . { ::i.: :. :{ ...: . 5,do___~n

9 ". "" ' '". . . . . " , ,

DaMascus

.......

, . . . . .

"-.;"i"/_"

8, 9

.

":'

.......'..

"

9

-.,.

_ryrg tobotiy~

(,/

~:~::',7..':'.?' ,::;F.,,',I 9:'.: ; v,',', ..'- V..",.

l/

Anatolia

;...:%!v. :-::...;::.:,.vV.-..:.

fit

7::."

%

~it ts ire i

~...:..;.y,.

.:v.-...

" I XI :il sea "o~.::::v;::

~ ~ i

~/x

Sea

l~l

~o,,,e,.:-..v:,..l , ~ ....../~1 ~.~.v.
Palmyridos

A,-a,~,o

Arc,col ~ 0 0 km j

Fig. 2. Simplified map of the Lebanese area of the Dead Sea Transform (modified after Dubertret (1955) and Bartov (1994)) and (inset) its tectonic setting on the transform restraining bend. The map shows the major faults (RF, Rachaiya Fault; HF, Hasbaya Fault) and folds together with the sites of more detailed descriptions in other figures (appropriately numbered boxes) and in the text. the m a i n strand. This fault m a p s a l o n g a N N E - S S W trend ( D u b e r t r e t 1955), implying a

general restraining b e n d g e o m e t r y to the transform. M o r p h o l o g i c a l characteristics have b e e n

EVOLUTION OF THE LEBANESE RESTRAINING BEND documented by Garfunkel et al. (1981), who interpreted these in terms of an active tectonic landscape. They link the Yammouneh Fault northwards on the Ghab Fault in NW Syria to complete the connection with the southern margin of the Tethyan collision belt. Ophiolites in southern Turkey are offset by about 75 km across the Ghab Fault system (Freund et al. 1970), although Chaimov et al. (1990) suggested that the offset is just 20-30 km. A consequence of a composite Dead Sea-Yammouneh-Ghab system acting as the overall transform plate boundary is a transform-transform-transform triple junction with the left-lateral East Anatolian Fault in SE Turkey (e.g. Westaway 1994; Fig. 1). An alternative view of the modern plate boundary geometry is that the Dead Sea Fault System diverts offshore onto the Levantine continental margin. Girdler (1990) took the active transform out through SW Lebanon. Nur & BenAvraham (1978) suggested that some displacements pass offshore along the Carmel Fault to link out to the Levantine abyssal plain and hence to the Cypriot arc subduction zone. The coastal area north of the Carmel Fault is seismically active, with focal mechanisms suggesting leftlateral strike-slip (NW-SE) and broadly N-S extension (Hofstetter et aL 1996). In this case, a broad zone of transcurrent shearing represents the active Dead Sea Transform, linking into the Cypriot arc at a transform-transform-trench triple junction.

Tectonic structure of Lebanon and environs Cenozoic deformation in Lebanon is characterized by combinations of folding and dominantly left-lateral strike-slip faulting. Immediately to the south, the Dead Sea Transform lies within the Hula Valley (Fig. 2). Here the deformation zone is less than 10 km across and characterized by generally low-lying topography flanked by uplifts. Detailed mapping by Heimann & Ron (1987, 1993) showed the structure to consist of alternating transform overlaps with linking panels of deformation. These 'away-from-fault' deformations are best displayed by pressure ridges, rotated and uplifted blocks with subsidiary extensional faults - deformations typical of strike-slip relay-ramps (Peacock & Sanderson 1995). The transform is therefore represented by segmented active faults that alternate back and forth across the topographic depression. Given the large bulk displacements across the transform as a whole, Heimann & Ron (1993) suggested that fault segmentation is transitory.

85

Notwithstanding the incremental segmented nature of the Dead Sea Fault System in Israel, the structure of Lebanon is far more complex (e.g. Arthaud et al. 1978, Fig. 2). Here there are many strike-slip faults, generally oriented N-S or NNE-SSW, together with E - W cross-faults, and folds at the kilometre and 10 km scales (e.g. Beydoun 1977; Beydoun & Habib 1995). Walley (1988) termed the map-pattern 'braided' whereby the localized displacements from the south are distributed onto an array of five main transcurrent fault zones and a host of minor ones. These various deformation structures are reviewed in turn. Major transcurrent faults

Of the five major faults which outcrop in Lebanon (e.g. Beydoun 1977, Walley 1988), the Yammouneh is the most continuous, and is clearly evident on satellite images (e.g. Ambraseys & Barazangi 1989). Evidence for its continuity has been described by Garfunkel et al. (1981). They used its morphological characteristics, largely established from topographic maps, to infer that the fault is an active transcurrent structure. The southern portion of the fault coincides with the eastern slopes of the uplifted Jabel Barouk structure, forming the western edge of the southern Bekaa valley. To the north of the town of Chtaura (Fig. 2) the fault lies within the high ground of Mount Lebanon. Here it is characterized by a series of enclosed karstic sedimentary basins or poljes (Garfunkel et al. 1981). The Yammouneh Fault emerges from the northern edge of the mountains along Wadi Chadra (Fig. 2). Here it appears to offset the southern outcrop edge of the Horns Basalt by about 10 km. A similar offset is recognized on the north side of the Basalt, in NW Syria, and the Yammouneh is inferred to be continuous with the Ghab Fault (Garfunkel et al. 1981), linking northwards to the East Anatolian Fault (e.g. Hempton 1987; Joffe & Garfunkel 1987). Diverging from the southern end of the Yammouneh Fault near the Hula valley (Fig. 2), the Roum Fault has been mapped as a discontinuous array of left-stepping segments with associated basins. However, it has been difficult to trace for more than about 30-40 km to the north of Hula. There are no through-going structures mapped by Dubertret (1955). However, the region is characterized by flat-bedded Cretaceous carbonates and, as Dubertret's (1955) mapping was largely designed to show the distribution of lithostratigraphy, faults within individual geological units are generally not shown. Probable northward extensions to the Roum have been

86

R.W.H. BUTLER E T A L .

recognized in the Damour valley (Butler et al. 1997), just south of Beirut (Girdler (1990; his 'Ed Damour Fault') and within Beirut itself (Dubertret 1955). It has a clear geomorphological expression, bounding the coastal TyreNabatiy6 plateau to the west and the more dissected and uplifted ranges to the east (Sanlaville 1970). The Roum fault strands are associated with deflections in the courses of the Litani and Zahrani rivers (e.g. Walley 1988). Splaying to the east of the Yammouneh Fault from the Hula area, the Hasbaya and Rachaiya Faults cut the western flank of the Mount Hermon anticline (Fig. 2). The Hasbaya Fault runs broadly along the Jordan (Hasbani) valley north of Hula. A few kilometres to the east lies the Rachaiya Fault. Heimann et al. (1990) described small step-overs on associated subsidiary fault strands which develop minor basins. They estimated about 1 km of Quaternary horizontal movement on the Rachaiya Fault, based on these segment geometries. Dubertret's (1955) mapping shows neither the Rachaiya nor the Hasbaya Faults to be laterally continuous. If this is accurate, we infer that these structures have not accumulated more than a few kilometres total of transcurrent displacement. The Serghaya fault is the most eastern of the major faults of Lebanon, running along the length of the Mount Hermon-Anti-Lebanon anticline. In the Anti-Lebanon mountains the fault is very difficult to map because of the monotonous, shattered and deeply karstified upper Cretaceous carbonates through which it runs. In the south, where the Serghaya Fault converges with the main part of the Dead Sea Transform near Mount Hermon, there are indications of Plio-Pleistocene tectonic activity. Apart from the five major faults described above which outcrop in Lebanon, Khair et aL (1997) inferred the presence of a buried structure: their Mid-Bekaa Fault. This is inferred on the basis of a step in the buried basement morphology modelled on gravity data. The lack of surface expression makes it difficult to evaluate the significance of this structure as a late Cenozoic transcurrent fault. In summary, of the five major faults (six including the hypothesized Mid-Bekaa structure of Khair et al. (1997)), the only structures which can have accommodated a substantial proportion of the 105 km total offset of the Dead Sea Transform through Lebanon are the Roum, Serghaya and Yammouneh Faults. Of these, only the Yammouneh Fault appears to have the continuity and landscape characteristics consistent with slip-rates of 6 mm/a operating for several million years. However, to date, the

structural mapping of the Roum Fault has been inadequate to establish its continuity and hence its capability to accommodate significant (>10 km) displacement. Folds

A feature of the restraining bend on the Dead Sea Transform is the large-scale fold structures and associated topographic elevations, which are the greatest in the Levant (Fig. 2). The major anticline of the Mount Lebanon and Jabel Barouk ranges is sub-parallel to the trace of the Yammouneh Fault (e.g. Beydoun 1977; Hancock & Atiya 1979). The complementary syncline to the east contains the Bekaa valley, bounded to its east by the major anticlines of the AntiLebanon and Mount Hermon ranges. Within the Bekaa valley there are additional folds, lying exclusively to the east of the Yammouneh Fault, generally oriented oblique to its trace with NE-SW-trending hinges. Another oblique foldbelt is found in NW Lebanon, around the city of Tripoli. These complex deformations, which occur within Mesozoic-Cenozoic sediments, are possibly detached from deep-seated crustal transform faults along Triassic evaporites (e.g. Beydoun & Habib 1995). Beydoun (1977) and Khair et al. (1997) pointed out that the crust beneath the uplifted Mount Lebanon ranges is apparently of normal thickness. This observation has been used to infer that the Lebanese crust has not been shortened by Cenozoic tectonics (e.g. Ron 1987). However, the region was a site of substantial Mesozoic subsidence and extension (reviewed by Laws & Wilson 1997) with over 6000 m of sedimentary cover (Beydoun & Habib 1995). This implies a stretching factor of about 1.5, a value which estimates the pre-orogenic thickness of the crystalline crust at just 20 km. This is the thickness modelled by Khair et al. (1997) for the areas adjacent to the uplifted Mount Lebanon range. However, the present crustal thickness (excluding the cover) beneath Mount Lebanon is 23-25 km (Khair et al. 1997). Assuming homogeneous deformation beneath Mount Lebanon, area balancing implies about 8-10 km crustal shortening. Cross-faults and tectonic rotations

The oblique convergence across the Dead Sea Transform at the Lebanese restraining bend has been considered to be accommodated by block rotations and subsidiary faulting distributed across tens of kilometres off the main transcurrent faults. Following this proposal by Freund &

EVOLUTION OF THE LEBANESE RESTRAINING BEND Tarling (1979), Ron and co-workers (Ron 1987, Ron et al. 1990a,b) have identified tectonic rotations, counter-clockwise on the flanks of Mount Hermon and clockwise in Galilee (Fig. 2). Declination anomalies are recorded for Cretaceous volcanic rocks, compared with their predicted palaeo-pole positions for the Arabian sub-continent. Ron et al. (1990a) reported 69 ~ + 13 ~ counter-clockwise rotations from Hermon and correlated these with the 55.6~ + 10.4~ rotations of the same sense identified by Gregor et al. (1974) for Mount Lebanon. The rotations are apparently related to regenerating N W - S E right-lateral strike-slip faults which rotate to broadly E - W before locking (Ron et aL 1990a). These structures apparently accommodate about 50% E - W elongation, essentially forming zones of non-coaxial plane strain wherein the vertical thickness is conserved. Arrays of ENE-WSW-trending faults, generally assumed to accommodate right-lateral strike-slip (e.g. Ron 1987), are found on either side of the Roum Fault (Fig. 2). One array occupies a topographic low between Mount Lebanon and Jabel Barouk, running inland from Beirut into the Chouf hills (Fig. 2), dissecting strata which record counter-clockwise rotations (Gregor et al. 1974). Another suite of faults lies on the Tyre-Nabatiy6 plateau (Fig. 2), between the Roum Fault and the coast of Lebanon. This region has not been studied palaeomagnetically. In contrast to other parts of the restraining bend, the plateau contains few folds. The Cretaceous-Eocene stratigraphy is generally subhorizontal. Sparse field data confirm the assumption of strike-slip faulting (Nammour 1992) although map-patterns of the offset of stratigraphy suggest significant vertical slip components (e.g. Hancock & Atiya 1979). Summary

The L~banese restraining bend of the Dead Sea Transform contains a series of major transcurrent faults of which only the Yammouneh Fault is generally considered to be capable of accommodating many tens of kilometres of leftlateral transcurrent displacement. The restraining bend is characterized by distributed deformation adjacent to the Yammouneh Fault. This is most obviously expressed as longwavelength anticlines, generally aligned subparallel to the major fault. The Mount Lebanon anticline is probably the surface expression of crustal thickening, this deformation affecting previously extended Mesozoic sedimentary basins, thereby returning the crust to approximately its normal thickness. There are

87

additional minor folds with axes commonly oblique to the trend of the Yammouneh Fault. These presumably record distributed transcurrent shear. Other distributed shears are less obvious than the folds but may be more important as finite deformation mechanisms. Arrays of closely spaced, generally right-lateral, faults are spatially associated with strata with substantial declination anomalies. Ron et al. (1990a and others) related these to distributed block rotation within the restraining bend, locally accommodating net elongations of several tens of kilometres. The above summary is a simplified description of the finite structure of the restraining bend. Several workers (e.g. Westaway 1995) have attempted to explain the relationships between the orientation of the Yammouneh Fault, the total displacement on the Dead Sea Transform, the finite relative plate movement vector and the record of distributed longitudinal strains and tectonic rotations. As discussed above, however, the regional plate kinematics has varied during the 18 Ma history of the transform. Consequently, we suggest that an incremental analysis of tectonic evolution is required. We now review briefly the patterns of active tectonics as indicated by seismicity and then attempt to assess the deformation history, using geomorphology to date selected structures within the restraining bend.

Seismicity and active tectonics Analyses of seismicity in Lebanon have been greatly hindered by very poor instrumentation coverage. Local stations are restricted to Israel: there is only one station in Lebanon itself (at Bhannes, about 15 km east of Beirut). Thus precise hypocentre locations are possible only for large-magnitude events (Plassard & Kogol 1981). However, the seismicity recorded instrumentally shows an important pattern (Fig. 3). As recognized by Girdler (1990), the northern part of Lebanon is largely aseismic. However, the Levant has been prone to occasional very large earthquakes, not evident in the instrumental record of the past century (Plassard & Kogol 1981). Ambraseys & Barazangi (1989) suggested that the northern part of the Yammouneh Fault system has merely been in a period of seismic quiescence. They described clustering of devastating earthquakes, most notably in the eighteenth century, in the Bekaa valley. Although there were surface ruptures within the Bekaa valley, these have long since been obliterated. Consequently, the kinematics of these historical events is unknown.

88

R.W.H. BUTLER E T A L .

I.

discussed below, it is most unlikely that all this seismicity has occurred on major transcurrent faults such as the Yammouneh or Roum.

Geomorphology and dating structures

./

O'O O-

,o-O 9OLD. O 0

9

9 O0 9

f 'e

9 "' 9

go o I

9 9 ..o

,qaP

..l~.. o~ 0.0.

| 9

9 o."

9

9 " 99

9

9 Ilk

9o ~

.

9

9

9 8" o 9

99 9

.O

I 9

Fig. 3. Distribution of seismicity in the vicinity of the Lebanese restraining bend. Small dots, M 1 <2.0; large dots, Ml >2.0. RF, Roum Fault; YF, Yammouneh Fault. Detailed seismological work has been restricted to the south of the restraining bend, largely within Israel where instrumentation is present. Van Eck & Hofstetter (1990) showed seismicity for the period 1982-1989 to be clustered spatially along mapped segments of the transform. Focal mechanisms are largely consistent with N-S left-lateral faulting but include N-S extension and E - W compression adjacent to the intersection of the Carmel Fault with the main transform axis. Hofstetter and co-workers (Van Eck & Hofstetter 1990; Shapira & Hofstetter 1993) showed the main locus of larger magnitude earthquakes to coincide with the Roum Fault. Notwithstanding the technical problems in estimating event locations, micro-seismicity is spread in a broad swathe in southern Lebanon. Although there is a general N-S trend of earthquakes along the trace of the Roum Fault, there is a substantial amount of seismic energy distributed away from the main faults (Fig. 3). As

With all the structures described above it is critical to evaluate the timing of fault and fold activity, to establish tectonic evolution. The Lebanon area is ideal for using geomorphology to establish the timing of fault activity and uplift (Fig. 4). The coastal area contains an array of marine terraces now at elevations up to 400 m above modern sea level (Sanlaville 1970). The oldest recognized of these is of Miocene age (Nammour 1992). Late Miocene shallowmarine sands and clays together with local carbonates decorate a major plateau surface in SW Lebanon. This Tyre-Nabatiy6 plateau (Fig. 2) reaches elevations of c. 400 m. Its eastern margin is marked by conglomeratic fan deposits, presumably accumulated along the Miocene coastline (Nammour 1992). There are as yet no biostratigraphic data to date these sediments beyond being Miocene in age. However, they represent the greatest transgression across the area (e.g. Hirsch 1990) and the best candidate for this sea-level maximum is that which occurred in latest Tortonian times, c. 6.5 Ma. The Miocene marine terrace represented by the Tyre-Nabatiy6 plateau is dissected by a suite of deep canyons. Some of these have substantial offshore continuations (e.g. Goedicke 1972; Garfunkel & Almagor 1985; Hall et al. 1994), indicating that they were cut during a period of substantial sea-level lowstand. Goedicke (1972) mapped offshore continuations to the canyons of the modern Zahrani and the Nahr el Aouali. Garfunkel & Almagor (1985) traced the Litani canyon to depths in excess of 1000 m below modern sea level. Bathymetric surveys off the Levant coast (Hall et al. 1994) show possible offshore continuations to all the main drainage basins (discussed below with reference to Fig. lla). The most likely candidate for this incision episode occurred in late Messinian times (c. 5.5 Ma) during the Mediterranean-wide sea-level minimum. This chronology is consistent with the relative age of planation of the Tyre-Nabatiy6 plateau. Further along the Levant, Druckman et al. (1995) suggested that some incision occurred during the Oligocene. Uplift of the TyreNabatiyd plateau has continued through PlioQuaternary times as indicated by younger marine and alluvial terraces 9 Within the Bekaa-Jordan valley extensive studies of ancient fluvial and lacustrine deposits

EVOLUTION OF THE LEBANESE RESTRAINING BEND

89

fault WEST Messinion gorge

scarp

drainage basin M@ssinion gorg,

[

k~

/ ~

]

k~

s

poleeo lends ur f e t e

be sin

/

unconformity /

\

)

/

oo.rn gorge

( \ !

T

lava - filled palaeo-valley /

/

/

/.
appr~o/e~ Ikm 10 km

Fig. 4. The range of geomorphological features in Lebanon that may be used to establish deformation activity, illustrated on a schematic profile.

indicate that the Litani river originally drained into Hula and Galilee (e.g. Horowitz 1979). Thus it is possible to infer broad aspects of drainage basin evolution linking palaeo-valleys via landscape stratigraphy to various calibrated baselevels along the coast or within the main Dead Sea valley. Away from the coast, geomorphological surfaces may be dated using basalts of the G o l a n - H o m s system. This eruptive system has been reviewed extensively by Mor (1993). However, we will also present new K - A r ages from specific outcrops to establish clear absolute landscape chronologies, linking these to specific structural and stratigraphic relationships. Additional age relationships may be deduced relative to the continental deposits of the Bekaa valley, which include lacustrine silts and sands commonly interpreted to be of Miocene age (e.g. Dubertret 1955). There are several uses to which these datable geomorphological features may be put. The marine planation surfaces can be used to quantify the timing and rates of uplift. Offsets of rivers which dissect dated surfaces may be used to estimate fault slip-rates. However, our concerns are to establish fault activities and the timing of deformations within the Lebanese restraining bend. Our rationale is that deformation structures had surface expressions when active and thus geomorphological features which cut structures may be used to date the youngest possible age of the structure. Our first target is to address the age of the Yammouneh Fault and examine structural relationships in northern Lebanon.

The Y a m m o u n e h Fault in northern Lebanon The primary evidence for neotectonic activity along the Yammouneh Fault rests on its relationship to landscape (e.g. Garfunkel et al. 1981; Ambraseys & Barazangi 1989). Post Pliocene displacement has been estimated at c. 11 km, determined from the offset of the Hems Basalt on the northern slope of Mount Lebanon (e.g. Quennel11984). In obtaining this estimate there is a tacit assumption that the Hems Basalt originally formed a simple E - W outcrop belt and consequently that it erupted across a simple topographical surface. New field data suggest that this was not the case. A more complete description of these data has been published elsewhere (Butler et al. 1997; Butler & Spencer in press; Fig. 5). Lavas of the Hems Basalt dip gently away from the high ground of Mount Lebanon and unconformably overlie Cretaceous limestones on both sides of the Yammouneh Fault (Fig. 2). Critical relationships between these lavas and the fault are found in Wadi Chadra. Here a complete transect is preserved (Fig. 5b) from tilted but otherwise weakly deformed Cretaceous limestones on the east of the fault, with increasing fracture and brecciation into about 20 m of poorly consolidated carbonate gouge along the contact with the basalts (Fig. 6). This is a simple deformation gradient into a major fault: the carbonate gouge is presumably derived from the neighbouring Cretaceous limestones. The fault gouge contains no fragments of basalt. The

R . W . H . BUTLER ETAL.

90

..:..;v;..

. 9 ..'j

)

:~,..~.. 9

...., ,

..v ;;'i . J- " - ".~.~."

-V..'.:"

.. (...:.......';" .v..'v.....:""" 9

9

~

/)

I'

~ ". , .

.'V["

9 ".'.';...-'...-,'. "'.'.' V..', 9" V : ' . " V . ' " " : ' : " ' " " " "'J "'" :;" "" :: :":'.~ "'" """V "" """ o r o f i l e ..

ft

"

:iL ;-::...:i,..!.

9

". :-::'.i/".."i':W-.3":'.': .-:: :-:;..'.:...--:,

k....

,,'.b~


cb

..,

bert

/<) ,lie 0

\

~

Quaternary elastics

~Miocene ~ f (~

basalf

I

> modern streams show ~o o f f s e t ~

, ~

]older

oI

~

strata

5 1 km

IIII~~I ~

uncon

--C-~-~-7,-'.-, ~ 7P~-_~-"-.~

formi ty

)

Woo

-___

--400m

unstructured ~ o r b o n o t e gouge

b

t"

m

shattered lime s tone increosing

<

/71/ ,.x'..X,

bedded C r e t a c e o u s lime s fen e

__

cefaclosis

,,,

Fig. 5. (a) Sketch map (modified after Dubretret (1955)) of the northern part of Mount Lebanon, and profile (b) in Wadi Chadra, showing the relationship between the Yammouneh Fault and the Homs Basalt. Topographical contours are in metres 9The irregular outcrop pattern for the Homs Basalt may be explained by lavas infilling relatively low-lying topography, leaving the high-standing Jabal Akroum uncovered. The lavas are unfaulted. Sites for samples of basalt used for K-Ar geochronology are indicated on the map and section (7, 8, 9, 10, 12, 13 and 18b; see Table 1). (See Fig. 2 for location.)

basalt adjacent to the fault consists of a series of pillow lavas and hyaloclastites, overlain in turn by at least 200 m of subaerial flows. All units are essentially unfaulted, there is no sign of tectonic deformation or brecciation. Even those pillows

within a few metres of the carbonate fault gouge are undeformed. These relationships do not suggest a faulted contact. Rather, the basalts overlie unconformably the carbonates and the Y a m m o u n e h Fault. Palaeoslope estimates

EVOLUTION OF THE LEBANESE RESTRAINING BEND

91

Fig. 6. Unfaulted pillow lavas at the base of the Horns Basalt in Wadi Chadra, from the site of the profile in Fig. 5b, viewed looking NE. These basalts yielded geochronological samples R96-9 and R96-10. The outcrop of white rock immediateIy behind these pillow basalts (top right of photograph) is the unstructured carbonate fault gouge of the Yammouneh Fault. derived from pillow shapes indicate that the lavas banked against the outcrop of fault gouge. We conclude that the contact between the Cretaceous limestones and the Horns Basalt is a fault-scarp unconformity (Butler et al. 1997). The basalts have filled a topographical depression against the fault, and are absent from the high ground of Jabel Akroum to the east of the fault, hence the apparent offset in map pattern (Fig. 5a). The field relationships in Wadi Chadra indicate that at least the northern part of the Yammouneh Fault has accommodated no significant transcurrent displacement since the eruption of the Horns Basalt. Consequently, establishing the age of the palaeosurface across which the basalt erupted is critical. Existing studies (e.g. Gregor et al. 1974) used field relationships along the coastal plain, about 15 km west of Wadi Chadra, where the basalts interdigitate with Pliocene marine shales. However, recent studies of Levant basalt complexes indicate long and complex eruptive histories (Mor 1993). Consequently, we have dated the Homs Basalt at Wadi Chadra and elsewhere, using the K - A t method in Leeds. Our data are presented in Table 1. The locations of sites are identified in Fig. 5. We incorporated a simple stratigraphic test to establish the validity of our results, which the samples pass.

The pillow lavas at the base of the basalt pile in Wadi Chadra date at about 6.5 Ma, with the overlying subaerial flows yielding ages of about 6.2 Ma. Given the analytical uncertainties of these samples the age difference between these sites is not significant. However, the highest part of the lava pile yielded ages of 5.2 Ma. Thus all sites yield K - A r ages in accord with their relative stratigraphic position. The Horns Basalt at Wadi Chadra erupted over a period of 1-1.5 Ma, in Messinian times. Further evidence for the undeformed state of the Homs Basalt at Wadi Chadra comes from published palaeomagnetic data. Gregor et al. (1974) analysed the lower part of the pile of subaerial flows, recording both normal and reversed magnetic remanence directions with declinations within error of the expected late Miocene direction for the region. The upper flows, dated at 6.5-5.2 Ma by the K - A r method, give reversed directions which presumably record magnetic remanences from the early Gilbert, Chron C3r. The Horns Basalt also provides critical evidence for dating part of the uplift of the northern part of Mount Lebanon. To the west of the Yammouneh Fault the lavas unconformably overlie and seal stratigraphically a train of folds (the Jabel en Nsour fold belt in Fig. 5a). The lavas are not themselves folded but show a

92

R.W.H. BUTLER E T A L .

Table 1. Whole-rock K - A r ages from the Homs Basalt (see Fig. 5 for locations) Sample

%K

R96-7

1,013

R96-8

1,086

R96-9

0.671

R96-10

0,799

R96-12

1.607

R96-13 L95-18b

0.545 0.569

Volume of radiogenic 4~ (• 10.5 cm3g-1)

% radiogenic 4~

0.0211 0.0202 0.0238 0.0228 0.0154 0.0143 0.0209 0.0204 0.0411 0.0401 0.0118 0.0111

11.3 24.1 23.1 23.3 9.8 6.5 21.0 21.7 47.8 44.6 26.1 13.5

gentle (<5 ~ regional tilt towards the NNW. To the east of the fault, Jabel Akroum is incised by valleys which link onto the baselevel covered by the Horns Basalt, although the valleys are locally rejuvenated by incision which post-dates basalt eruption (Butler & Spencer in press). These relationships indicate that not only had the Yammouneh Fault accommodated all its left-lateral transcurrent displacement by the Messinian, the Mount Lebanon range had a strong morphological expression by this time.

The search for major transcurrent faulting in Lebanon during the Plio-Quaternary The Yammouneh Fault is only one of the major transcurrent structures in Lebanon, so the demonstration of its inactivity in Plio-Quaternary times does not imply that the plate boundary in Lebanon as a whole is blocked. On the basis of plate tectonic reconstructions (Joffe & Garfunkel 1987; Westaway 1994) there should be about 35 km of left-lateral displacement of Plio-Quaternary age, distributed onto one or several faults. We now consider the remaining faults in turn. The Mid-Bekaa Fault of Khair et al. (1997) has no clear surface expression and is unlikely to have accommodated major transcurrent displacements, at least since the deposition of Miocene continental deposits which now fill the Bekaa valley. It has been the site of historical seismicity (Ambraseys & Barazangi 1989), but there is no evidence that these seismogenic displacements were associated with transcurrent faulting. The analysis of gravity data by Khair et al. (1997) suggests the fault to have a substantial vertical displacement, perhaps by several kilometres, which has uplifted Mount Lebanon with

Age

Error

5.2

0.2

5.5

0.2

5.7

0.7

6.7

0.2

6.5

0.2

5.6 5.0

0.2 0.2

respect to the Bekaa. We tentatively interpret this structure to be a thrust. Further south in the Bekaa, the Hasbaya and Rachaiya faults broadly coincide with zones of diffuse seismicity (Fig. 3), as does the Yammouneh Fault in this region. Although offset Quaternary features (e.g. Mor 1987) and seismicity (Van Eck & Hofstetter 1990) indicate active transcurrent displacements, further north the active displacements on these structures appear to die out. Landscape features argue strongly against either the Hasbaya or Rachaiya Fault accommodating any substantial transcurrent displacement over the past few million years (Fig. 7). The geomorphology may be dated via a suite of valley-filling basalts, dated by Mor (1993, his Mechki Basalt) using the K - A r method at 2.23 +_ 0.1 Ma to 2.89 _+ 0.1 Ma. Our own data (Table 2) are in broad agreement with Mor's. These basalts overlie without offset the trace of the Hasbaya Fault (Fig. 7). Finite displacements on the fault are difficult to assess in the absence of any identified correlative markers. However, the age dates and field relationships indicate that at least the northern part of the Hasbaya Fault has not been an active transcurrent feature since at least mid-late Pliocene times. As with the Horns Basalt, Gregor et al.'s (1974) palaeomagnetic data for the Mechki Basalt show no evidence of tectonic rotations. The single site analysed from the flow shows the expected remanence direction for the mid-Pliocene. Gregor et al.'s site recorded a normal magnetic polarity, presumably from the Gauss (Chron C2An). Both the Rachaiya and Hasbaya Faults are marked by lines of poljes and valley deflections. These landscape features appear to link onto the old level of the Hasbani valley that contains the

EVOLUTION OF THE LEBANESE RESTRAINING BEND

93

Quaternary (lacustrine]

Mechki Basatt Neogeno ctastics old valley side

modernic~ciso~ drainage old vall~y e..~

basalt sample ',._., j Kaf@~ ~o

I 2 km

.Rachaiya Fig. 7. Sketch map showing field relationships between basalt lavas, geomorphology and the Hasbaya (HF) and Rachaiya Faults (RF). Location shown in Fig. 2. The pre-Miocene strata (chiefly late Jurassic-Eocene) are unornamented. Topographical contours are in metres. Table 2. Whole-rock K-Ar ages from the Mechki Basalt, (see Fig. 7for locations) Sample

%K

R96-2

0.958

R96-3

0.790

R96-4

0.806

R96-5

1.008

R96-6

1.062

Volume of radiogenic 4~ (• 10.5 cm3g-1)

% radiogenic 4~

0.0080 0.0076 0.0114 0.0117 0.0096 0.0094 0.0075 0.0082 0.0104 0.0083

16.3 20.2 38.6 35.7 14.5 18.3 24.8 34.9 30.4 26.1

Mechki Basalt, which seals the Hasbaya Fault. Consequently, these landscape features are Pliocene or older and hence their offset does not imply active transcurrent faulting. Walley (1988) considered these faults, away from the Hula area, to have been active only during the Miocene. The displacement activity of the Serghaya Fault is rather harder to establish than that of its neighbours to the east. Walley (1988) pointed out that its southern segments are overlain by part of Golan basalt edifice. The particular unit

Age

Error

2.1

0.5

3.8

0.1

3.0

0.1

2.0

0.1

2.5

0.2

was subsequently dated by Mar (1993, his Golan Formation) at 0.4-0.1 Ma. Further north there are several minor sedimentary basins, possibly associated with small relay ramps on the fault. In its type area on the Syria-Lebanon border, the Serghaya Fault is sealed by sands, muds and clays together with basalts. Dubertret & Vautrin (1945) dated these as Pliocene or older. Thus this northern segment of the Serghaya Fault has not been an active transcurrent structure for at least the past 5 Ma. Further precision awaits radiometric dating of the basalts.

94

R.W.H.

BUTLER

ETAL.

o~

oO E~

o# o,.~

o o

~9 ~:~

=

~.~

.g

.g

0

Cb

0

0

~..~ ~ e'~X ~,-~

aO! /

.~

EVOLUTION OF THE LEBANESE RESTRAINING BEND

95

The evidence of landscape stratigraphy from Lebanon strongly indicates that major faults to the east of the Yammouneh have not accommodated any significant proportion of the total plate displacement along the transform, certainly during the late Pliocene and Quaternary. This should perhaps be expected, as they trend to the NE, away from posible northern continuations of the transform boundary in NW Syria. They probably were important in transferring transform displacement into crustal shortening during the later stages of Palmyride deformation (e.g. Chaimov et al. 1990). It should be noted, however, that seismicity (Van Eck & Hofstetter 1990, Fig. 3) shows the southern parts of these faults to be active, a feature we shall address shortly.

Similar features were described by Nammour (1992) and considered to be of Miocene age. Consequently, we infer that the fault scarp on the Zrariy6 Fault is part of the preserved Miocene landscape and that it shows no evidence of more recent displacements. We suggest that the Tyre-Nabatiy6 plateau has remained essentially undeformed for the past 6.5 Ma. This conclusion is starkly different from that of Ron et al. (1990b) for northern Galilee. They suggested that the deformation is currently active. Certainly there is regional seismicity. We suspect that this is related to slip on the Carmel Fault together with compressional deformation related to the uplift of the plateau.

Tectonics and landscape evolution in SE Lebanon

The eastern edge of the Tyre-Nabatiy6 plateau is defined by a prominent scarp which locally coincides with the late Miocene shoreline (Nammour 1992) and the Roum Fault. This structure diverges northwards away from the Yammouneh in the north Hula area (Fig. 2) and its southern trace is obvious as an array of fault segments seen both in published geological maps (Dubertret 1955) and in satellite images (Girdler 1990). For reasons outlined above, this structure was not mapped by Dubertret (1955). However, we have identified major zones of fault breccia within the generally flat-lying Cretaceous limestones to the north of Dubertet's (1955) structures. Other splays are evident in air photographs and in 100 m-1 km offsets of streams. A probable northern continuation of the Roum Fault is well exposed as an array of steep faults in the Damour valley (Butler et al. 1997), where it separates gently dipping strata of the Tyre-Nabatiy6 plateau from the westdipping monocline of the Chouf mountains (Fig. 10). The fault zone in the Damour section consists of a 500 m wide zone within which the degree of brecciation and fracturing is highly heterogeneous. The main zones of fault gouge are poorly exposed but are probably a few tens of metres wide. Their map pattern implies that these fault zones are subvertical. They are linked by arrays of minor faults with generally strikeslip kinematics (Fig. 10c). Judging by the amount of deformation within the D a m o u r section, the fault zone here is clearly a major structure. The Damour river is currently only deflected by about 200 m (Fig. 10a). The area has a number of dry valleys and notched ridges indicative of a more highly disrupted landscape, so we consider the river offset to greatly underestimate displacement.

If no major transcurrent displacements can be accommodated to the east of the Yammouneh Fault in Plio-Quaternary times, then presumably they passed to its west. This area (Fig. 8) includes the Roum Fault. The rocks to the west of the R o u m Fault, cropping out on the Tyre-Nabatiy6 plateau, are generally fiat-lying and dissected by an array of N E - S W trending cross-faults. This structural domain forms the northern continuation of the belt of cross-faults discussed by Ron et al. (1990b) from northern Galilee. They relate the faulting to about 30~ clockwise tectonic rotations. Regrettably, there have been no suitable palaeomagnetic studies in SE Lebanon to detect rotations. However, the faults may be related to landscape evolution on the Tyre-Nabatiy6 plateau. Provided Ron et al.'s (1990b) model of concurrent rotation and faulting is accepted, dating the fault activity also constrains the timing of the rest of the deformation. The faults on the Tyre-Nabatiy6 plateau show substantial (>100 m) vertical offsets of Cretaceous-Eocene formations in places, although existing kinematic descriptions emphasize strike-slip (Nammour 1992). Despite these throws, the majority of the faults have no landscape expression. They appear to be planed off by the oldest marine terrace on the plateau, a Miocene landscape feature (Fig. 9a; Nammour 1992). Only the Zrariy6 Fault and its relays have visible fault scarps that apparently offset the upper terrace level. We investigated this fault (Fig. 9b) where it is marked by a 30 m scarp. Its base has a concave form, the cliff-base boulders containing Lithophaga. The scarp is clearly a relict sea cliff, now over 100 m above sea level.

T h e R o u m Fault Z o n e

96

R.W.H. BUTLER ETAL.

EVOLUTION OF THE LEBANESE RESTRAINING BEND

300m

~

,.,

/ @~ 9 ~.. .

~

.

Y

.

---2_-

Nj

97

:i : "

,~

..>..

9 ir

~

7/I

~ ~

, ,.-

9

"" "

",,\

M ~ 1 7 6 1 7 6 F,.

.oo;o0 ~ .A I _

j-

,,.,\ /

Hommooa Fm

cautr zooes

beddinq

j

-

,,,,

.

.

',,,, ,

/

/

C

foulted

/ cave

mfill

( 1/

~-a~~, y

+,oo

]

/:/ b Fig. 10. Sketch map (a), cross-section (b) and structural data (c: fault planes and striae) for the Roum Fault Zone at the Damour canyon. (See Fig. 2 for location.) Selected representative fault kinematic data are illustrated as 'pseudo-focal mechanisms' (see Butler et al. 1989). Much of the D a m o u r section exposes fault zones which cut and shatter palaeokarstic features (Fig. 9c). These include large volumes of compacted red muds and speleothems, the collapsed remnants of old cave systems. These must have existed w h e n this area lay above local watertable and consequently post-date the M i o c e n e marine terrace. Thus a substantial part of the d e f o r m a t i o n in the Darnour section post-dates the late M i o c e n e transgression and subsequent

regression. We conclude that at least these portions of the R o u m Fault post-date the strike-slip on the n o r t h e r n part of the Y a m m o u n e h .

Estimating post-Messinian offsets on the R o u m Fault O u r o b s e r v a t i o n s s u p p o r t t h e c o n t e n t i o n of Girdler (1990) that the R o u m Fault is a m a j o r zone of left-lateral d i s p l a c e m e n t that has b e e n

Fig. 9. Geomorphological features in SW Lebanon. (a) Upper marine planation surface of the Tyre-Nabatiy6 plateau, looking north across the Zahrani river valley. This surface here has an elevation of 350400 m and truncates all faults. It is inferred to be of late Miocene age (Nammour 1992). (b) Cliff line on the Zrariy6 Fault, interpreted as an un-faulted palaeo-seacliff, a pediment of conglomerate with Lithophaga and minor marine sediments, inferred to be Miocene in age. (e) Detail of palaeo-karst red soil involved within the Roum Fault Zone. Minor strike-slip faults cut this feature and the palaeo-karst is shattered adjacent to fault zones.

98

R.W.H. BUTLER E T A L . 20 km

/

/

IRUT

cofchm@nf

KELB

gorgo w,nd gap

"'""'"

....

Sidon

111 O

Present

(Okra)

b

15kin displacement

1 /

fC 30kin displacement

Fig. 11. Drainage basins and their offset across the Roum Fault. (a) Modern drainage basins and gorges across the Roum Fault. The sea-bed depths are after Hall et al. (1994). Z, Zahrani catchment. Quaternary basins are stippled. (b) A restoration of displacement on the Roum Fault which places the headwaters of the Zahrani river above the lower Litani canyon. In Pliocene and earlier times the Litani river flowed through what are now windgaps into the Hula valley. This restoration is not favoured as it offers no explanation of windgaps at the head of the modem Debb6 river. (e) The preferred restoration, which matches the headwaters of the Aouali river with the lower Litani canyon. The ancestral Zahrani headwaters may have fed the Debb6 valley.

active since the Miocene. Like the Yammouneh Fault, it is not a single fault but rather is represented by relays and strands which define a broad damage zone. Despite its seismogenic character, the Roum Fault is generally inferred to have accumulated little displacement and therefore to be a very young structure. Walley (1988) estimated displacements using river offsets, recording 9 km displacement of the Litani and 3.5 km of the Zahrani, and no mapscale offsets on the Aouali and Damour rivers. These offsets imply a dramatic displacement gradient on the Roum Fault Zone but are probably an artefact of the geomorphological analysis. The following discussion (Fig. 11) briefly reviews the results of a much larger geomorphological analysis that will be the subject of a forthcoming paper. The watershed running parallel to the R o u m Fault Z o n e between the Zahrani and Litani rivers contains a series of windgaps which betray the process of stream capture. Clearly, the modern courses of the rivers only record part of the displacement history of the fault. To establish a longer-term picture of fault displacement we now analyse the pattern of drainage basins,

windgaps and dry valleys across the Roum Fault (Fig. 11). The main period of canyon-cutting across the Tyre-Nabatiy6 plateau is likely to be late Messinian in age (c. 5.5 Ma). Consequently, these canyons offer excellent long-term markers of lateral offsets along the R o u m Fault Zone. Most critically, studies in the Hula valley and Galilee (Horowitz 1979) show that the ancestral Litani river flowed axially along the transform. As a consequence, the Litani gorge across the Tyre-Nabatiy6 plateau cannot have been cut by waters flowing from its current catchment. The next drainage basin to the north is that of the modern Zahrani river. If these headwaters lay directly above the lower Litani gorge during late Messinian times, the Roum Fault Zone must have accommodated at least 15 km left-lateral displacement (Fig. l l b ) . The progressive leftlateral offset of the Z a h r a n i headwaters is charted by an array of windgaps and dry gorges between the Z a h r a n i and Litani gorges. However, it seems unlikely to us that the deepest gorge in the Tyre-Nabatiy6 plateau with the most substantial offshore canyon should have developed from such a small catchment

EVOLUTION OF THE LEBANESE RESTRAINING BEND area as is provided by the headwaters of the Zahrani. If the Litani gorge was not cut by the headwaters of the Zahrani, presumably it was by those now lying further north along the eastern side of the Roum Fault (Fig. 11c). A 30 km displacement on the Roum Fault would bring the next catchment area in line with the Litani gorge, that of the Aouali river. The advantages of this restoration (Fig. 11c) are to align the Zahrani headwaters close to windgaps at the head of the Debb6 river. The Damour headwaters align with the Zahrani gorge and the Beirut river aligns with the lower gorge of the Aouali. We conclude that the Roum Fault Zone must have accommodated at least 15 km, probably more than 30 km, left-lateral displacement since the Messinian. Only some of the modern river courses record part of this displacement because the drainage system has been prone to stream capture. This process is evident in the presence of windgaps along the uplifted western flank of the Roum Fault and in the array of dry canyons which cross the Tyre-Nabatiy6 plateau.

Kinematic linkage between the Roum and Dead Sea Fault Systems The structural continuity between faults in southern Lebanon has been obscure (Girdler 1990). The southern part of the Roum Fault is marked by 7-10 km segments which define leftwards stepovers (Fig. 2). This geometry appears to be responsible for adjacent flank uplifts, concentrated on the west side of the fault zone, and associated topographical depressions (Fig. 8). The depressions contain Quaternary coarse clastic deposits, shed from the flanks. Only one sedimentary basin has developed, at Jarmaq (Fig. 8), largely because the low ground is drained by the Litani river and the flank uplift, which otherwise would pond sediment, is breached by the antecedent lower Litani gorge. The left-stepping segmentation on the Roum Fault defines an array of relay ramps (sensu Peacock & Sanderson 1995). The pattern of uplifts and basins associated with this system appears to continue to the northernmost strand of the main Dead Sea Fault System near Metula (Figs 2 and 8). Heimann & Ron (1987) described active faulting on the eastern side of the Hula Valley (Azaz Fault), relaying transtensionally onto the southernmost part of the Hasabaya Fault. This in turn appears to relay across onto the southern part of the Yammouneh. The leftstepping relays generate basins, such as the Quaternary E1 Marj depression (Fig. 8). Rightward bends on the faults, perhaps inherited from

99

when they operated as a linked system with the main Yammouneh Fault, generate minor pressure ridges (e.g. Jebel Hamamiss, Heimann & Ron 1987). Thus the modern transfer of displacement from the Dead Sea Fault System onto the Roum is accomplished by an array of relay ramps, parts of which accommodate transtension and basin formation whereas other parts are associated with transpression and uplift. The segmented nature of the modern Roum Fault and others linking across the the main Dead Sea Fault System in the Hula Valley raises important issues of displacement compatibility and kinematic viability. As active fault segment lengths within the relay ramp zone are between 5 and 10 kin, even displacements on fault strands of 1-2 km would generate very high strains and rotations in the connecting wallrocks. Neither are observed; the Plio-Quaternary basalts record no palaeomagnetic rotations (Ron 1987). This fault array cannot accommodate the much larger displacements estimated for the Roum Fault on the basis of geomorphology or predicted from plate tectonics. We conclude therefore that fault segmentation here is transitory and overprints a through-going, fully linked fault system. It should be noted that transient segment activity must also be a feature of other parts of the Dead Sea Fault System.

The uplift of Mount Lebanon The ages of the major folds are rather less controversial than those of the principal transcurrent faults in Lebanon. The Jabel en Nsour fold belt (Figs 2 and 5) are truncated by the unconformity beneath the Homs Basalt. The Bekaa valley existed as a sedimentary basin during the Miocene, with the accumulation of continental fluvial and lacustrine deposits of this age within the regional syncline. These sediments onlap unconformably the flanks of the Mount Lebanon uplift, particularly to the north of Chtaura (Fig. 2). They also form the fills to valleys emanating from the flanks of the range (Butler & Spencer in press). On the coastal side of the range, transgressive Miocene carbonates, presumably of late Tortonian-Early Messinian age, unconformably overlie discordant Cretaceous strata, notably near Jouni6 (Fig. 2). The ancestral Jabel Barouk was also uplifted at the time, shedding sediment onto the marine terrace now represented by the Tyre-Nabatiy6 plateau (Nammour 1992). That the major uplifts have continued to amplify is indicated by the tilts in Miocene strata away from Mount Lebanon in the Bekaa velley and at Jouni6 (Fig. 2). New folds have grown

100

R.W.H. BUTLER E T A L .

during the Plio-Quaternary. These include the Jabel Turbol anticline and the Kousba fold belt near Tripoli (Fig. 2). These structures have offshore equivalents, for example, the folds at the Ile du Palmier (Beydoun & Habib 1995). Amplification of the coastal monocline is indicated by the array of uplifted marine terraces (Sanlaville 1970; Nammour 1992) throughout coastal Lebanon. That uplift is continuing is indicated by a prominent notch level at about 1 m above sea level between Tyre and Tripoli (Fig. 2). Clearly, folding has occurred throughout the late Miocene-Recent history of the restraining bend. It thus accompanied displacement on the Roum and Yammouneh Faults. They remain the clearest indications of transpression within the Lebanese restraining bend. As discussed above, the major folds are oriented sub-parallel to the Yammouneh and associated faults and therefore show so-called 'in-line' behaviour. Smaller folds, with wavelengths of 1-5 km, can occur in en echelon arrays. Hancock & Atiya (1979) proposed that Lebanese folds generally initiated with E - W trending hinge lines and have subsequently been rotated counter-clockwise during distributed wrench shearing. In contrast, Westaway (1995) suggested that, of all the folds, at least the major anticlines initiated 'in-line', as a result of fault-normal compression. To test between these two models we have analysed the orientation of minor structures with respect to fold hinges on a range of folds. Our rationale was that if folds amplified as foldhinges rotated then minor structures should transect the fold hinge (for models, see Woodcock & Schubert 1994). The range of minor structures within the folds of Lebanon has been described by Hancock & Atiya (1979). Here we present data from part of the major in-line fold of Jabel Barouk (Fig. 12), although the same general patterns have been replicated at every one of the in-line folds so far investigated. The Niha syncline contains a core of upper Cretaceous carbonates between two uplifted tracts of Jurassic carbonates. The larger of the uplifts contains the topographical crest of Jabel Barouk and trends parallel to the Yammouneh Fault. To the north and south of our study area, published maps (e.g. Dubertret 1955) show the Jabel Barouk structure to be an anticline. However, in our study area there is no evidence of an eastern limb to this fold. We will examine this aspect shortly. Minor structures within the Cretaceous carbonates consist of bedding-orthogonal joint sets, bedding sub-parallel stylolites and minor veins. Measurements of these structures, together with bedding measurements to obtain estimates of

v

91

u trot- E o c o ~ o

~ ) ~ J

//

Lwr -M;d C r o t

22

Jurossic

k

L2

Fig. 12. Simplified geological map of the Niha syncline area, part of the in-line fold belt to the Yammouneh Fault in south-central Lebanon (see Fig. 2 for location). The cross-section (X-Y) is a semi-schematic representation of fold geometries adjacent to the fault, showing the Jabel Barouk halfanticline, presumed amplified from the edge of the Mesozoic carbonates on the west side of the fault. The stereonets show poles to bedding (upper one, n = 291) and intersections between bedding and joints or extension veins (lower one, n = 622). fold-hinge orientation, were taken on four transects across the Niha syncline. Poles to bedding define a simple girdle with an average hinge that is essentially horizontal, trending 200 ~. This trend is consistent with the map-pattern (Fig. 12). The lines of intersection between joints or veins and bedding planes show a complex pattern but define a point maximum parallel to the average hinge line to the syncline. These data indicate no sign of transection. We infer therefore that the history of incremental strains (represented by the minor structures) and the finite strain (represented by the major fold) were essential coaxial.

EVOLUTION OF THE LEBANESE RESTRAINING BEND

Relationship between 'in-line' folds and displacement on the Y a m m o u n e h Fault Deformation during transpression is likely to have been partitioned into strike-slip on the major faults, initially the Yammouneh, and fault-normal compression generating folds. The Jabel Barouk structure developed parallel to the Yammouneh Fault. There are sound mechanical reasons for this. Classic studies of buckle folds (e.g. Biot 1961) show that competent layers will preferentially develop fold hinges parallel to their edges, so, if the Yammouneh Fault cut the competent Mesozoic carbonates, the cut itself would have seeded the orientation of the subsequent fold hinge. However, the fold could amplify differentially along the edge of the fault. Half-folds, as shown in our study area, are a common result of such processes (Biot 1961), with fold amplification enhanced because of a locally reduced mechanical resistance to bending. We infer that in this sector the Yammouneh Fault accommodated not only regional left-lateral displacements but also local differential uplift. Certainly, on the eastern flank of Jabel Barouk, there are numerous outcrops of fault breccia that are consistent with our kinematic model. It also provides an explanation for apparently young landscape features along this segment of the fault (Garfunkel et aL 1981) and the current seismicity (Fig. 3) without invoking Plio-Quaternary strike-slip on it. The southern part of the Yammouneh Fault may therefore be behaving as a steep reverse fault whereas the transcurrent displacement is currently accommodated by the Roum Fault Zone.

101

strata (chiefly upper Cretaceous basalts), and their apparently associated strike-slip fault arrays discussed by Ron et aL (1990a). The notion that subsidiary strike-slip faulting and rotation occurred throughout the evolution of the restraining bend may be assessed using the data of Gregor et aL (1974; see Ron et al. 1990a), in conjunction with the structure of their sites. These workers identified counter-clockwise rotations of 55.6~ (+10.4 ~ of Mesozoic volcanic rocks with respect to their expected orientation. Ron et al. (1990a) argued that these rotations require multiple active strike-slip faults to generate these rotations. In flat-lying strata it may be difficult to recognize such faults. However, the sites studied by Gregor et aL (1974) are adjacent to steeply dipping strata. The best example comes from the Laklouk area of Mount Lebanon (Figs 2 and 13). Although the palaeomagnetic sites generally lie

Sonnin~limestone I 3 km I i underlying Cretoceous U~ ~ JUFOSSiC

~"

expected U ~j ~ /direction

~~

Relationship between folds, cross-faults and vertical rotations The conclusion reached above on the basis of minor structures, that the major in-line folds grew coaxially, is in conflict with the notion that the region around the major transcurrent faults such as the Yammouneh experienced a general leftlateral shearing. The general shearing model (e.g. Ron 1987; Ron et aL 1990a) suggests that rotations about vertical axes accompanied by rightlateral slip on subsidiary faults accommodate part of the oblique slip during transpression. In some cases this style of deformation is clear. The 11~ counter-clockwise rotation of Plio-Pleistocene basalts in the Korazim block (described by Heimann & Ron (1993); location shown in Fig. 2) clearly developed synchronously with transcurrent shearing along the major strike-slip faults of the Dead Sea Transform. It is much harder to establish the timing of the larger-scale rotations, based on palaeomagnetic data from Mesozoic

Xt I i -

v=,

.L../

I

t

I

Fill. 13. Simplified geological map and profile through the Qartaba 'horst', showing Gregor et al.'s (1974) palaeomagnetic declinations for Late Jurassic (stippled arrows) and early Late Cretaceous (white arrows) sites. It should be noted that the fold structure of the horst is not disrupted by transverse (E-W) faults. (See Fig. 2 for location.)

102

R.W.H. BUTLER E T A L .

i~"

transform fault zone

}: ~.-~norrow

I! l}I

migrating triple junction ? triple junction r relocate s ~ ~ II

f" /,.:: oblique ~. ,,:~. folds ~ ~ Yommouneh links /: ~ *~;. "" across dlstribufed

transform

"in-line" folds

,

narrow

fault zone

t! A ) !4i''':(~ II~'"'"

/~

ifii

g

,r

distributed trans,o.~

,,

__....,~At',

Yom,ouoeh

..: ;.:,. ,"....".::.:!

tron.form'/

~." ',

/I ~

iocote~ along

9

1\$,1:~~

.one

~

~;

C.

b. 50kin

,,

f

(\~r~,

9~ I

taulting~

through-going ~ f~oumFault / "

/ YA~[ I. . . #./Ji~2s aoanaonea [{l'--transform I .

Jarmoq

I

5kin El Marj basin

Levantine

/-\x d

~. \ I I

,ronstensioo on ~armo, ~ao/,

Nf

~'o I.~ ~"~:~ ~::::!;i!ii :~:basin ;::::i:; ~:::~1

e.

.........

~I

Fig. 14. A model for the kinematic evolution of the Lebanese restraining bend on the Dead Sea Transform, seen in map view. The Palmyrides fold belt (inverted sedimentary basins) is not illustrated but may have been important during the early stages of transform activity. Arabia v. Africa motion is indicated for various stages (white outlined arrows, see also Fig. 1). (a) illustrates a possible early stage of development (c. 18-15 Ma) where the region contains a distributed left-lateral transcurrent shear zone, before nucleation of a throughgoing Yammouneh Fault; (b) shows the main stage of transform activity with the restraining bend represented by the left-lateral Yammouneh Fault and fault sub-parallel folds (Mount Lebanon, Mount Herman-Anti Lebanon); (c) shows the situation with the change in plate convergence vector, causing the restraining bend geometry to geometrically harden and be bypassed by the Roum Fault; (d) shows the modern situation with the Roum Fault becoming segmented with individual fault strands, particularly in the south; (e) shows the resultant geometry of transtension at leftward stepovers but transpression on right-trending portions of faults (fault names in the Hula area after Heimann & Ran (1987)). in gently dipping strata, these sites lie close to a major fold structure, the Qartaba 'horst' (Fig. 13). The eastern limb of the 'horst' is sub-vertical with a N N E - S S W strike. It retains this strike alongside all the Laklouk sites and is not cut by any map-scale faults. Thus, if the palaeomagnetic rotations were accommodated by subsidiary strike-slip faulting, this deformation must entirely precede the development of the Qartaba structure. The rotations therefore formed early in the local structural chronology. As at Laklouk, other sites studied by Gregor et al. (1974) lie adjacent to steeply dipping strata

which define part of the fold structure of Mount L e b a n o n or its southern continuation in the Chouf mountains (Fig. 2). These limbs are only locally cut by cross-faults. The regional extent of fold limbs essentially prohibits any large-scale cross-fault development after folding. Thus if the cross-faults and palaeomagnetically determined rotations are part of the Dead Sea Transform history in the Lebanese restraining bend they must represent a very early stage and have not operated since the Miocene. Gregor et al. (1974) suggested that the rotations affected the whole region and represent large-scale plate

EVOLUTION OF THE LEBANESE RESTRAINING BEND processes entirely preceding the initiation of the transform.

Linking restraining bend evolution to plate tectonics Many existing attempts to link structural evolution within the Lebanese restraining bend to plate kinematics on the Dead Sea Transform have assumed that the array of structures seen in the bend were active contemporaneously. By linking landscape evolution to tectonic structures a picture emerges of evolving deformation. Here we review our findings of structural evolution, schematically illustrated in Fig. 14. In our structural model, the earliest structures within the Lebanese sector of the transform are the oblique folds, cross-faults and tectonic rotations that collectively represent a broad zone of distributed left-lateral shear (Fig. 14a). These field structures form the earliest part of the relative structural chronology. However, the initiation of transform tectonics remains poorly understood, not least because of the difficulties in isolating these effects from those associated with Palmyride deformation (e.g. Chaimov et al. 1990) in the late Paleogene. As noted by Quennell (1984) and others, Palmyride deformation and the D e a d Sea Transform may be kinematically linked, but such an appraisal lies outside the scope of this discussion. However, for Hancock & Atiya (1979), much of the displacement on the Yammouneh post-dates the oblique folds and crossfaults of the Lebanese area. After a period of distributed strain, our model shows the Yammouneh Fault cutting through the Lebanese area to provide a 'hard-linked' through-going transcurrent structure linking the Ghab Fault with the southern part of the Dead Sea Transform (Fig. 14b). This must have occurred in Miocene times, presumably while the Arabia-Africa rotation pole lay far to the west. The position of this pole is critical in determining the amount of transpression across the Yammouneh Fault. The transpression is manifest by coeval transcurrent displacement on the Yammouneh Fault together with increasing uplift and amplification of the Mount Lebanon and related in-line folds. These folds initiated parallel to the main faults and thus record nearideal partitioning of transpressive strain. During the late Miocene the Arabia-Africa rotation pole migrated much closer towards the transform. The effect was to increase the angle of convergence across the Yammouneh Fault. For the restraining bend to remain active, the shortening and amplification rate of in-line folds

103

would have had to increase greatly. Rather, the transform fault zone evolved, with transcurrent displacements transferred off the Yammouneh and onto the Roum Fault system (Fig. 14c). The effect of this reorganization in the plate boundary further north would have been dramatic, with the triple junction between the transform and the Tethyan collision zone migrating from SE Turkey into the NE Mediterranean. There are few studies as yet of structural evolution in this region to confirm our prediction. Nevertheless, the Lebanese sector of the transform remained generally transpressive, as indicated by continued deformation and uplift of Mount Lebanon. Presumably, right-ward steps on the transform existed offshore. In part, the differential uplift of Mounts Lebanon and Hermon relative to the Bekaa was accommodated by steep reverse movements reactivating segments of the abandoned transcurrent faults. This may still be continuing, particularly adjacent to the southern Bekaa area, where there is local seismicity (Van Eck & Hofstetter 1990). For much of the Pliocene the Roum Fault system presumably acted as a through-going transform which accommodated the great bulk of the total 6 mm/a displacement between the Arabian and African plates. It was geometrically well suited to do this, forming a small-circle segment about the Pliocene rotation pole (Quennell 1984; Girdler 1990). However, the focal mechanism studies of Van Eck & Hofstetter (1990; Fig. 1) indicate that the modern rotation pole has moved much closer to the transform. Thus the trend of the Roum Fault system no longer forms a small circle about the active rotation pole and would be strongly in transpression. It appears to have modified, however, into an array of distinct fault segments (Fig. 14d and e). That these segments are generally only a few kilometre long implies that the fault geometries have not accommodated much displacement. We infer that the segments are very young, perhaps only a few thousand years old. In their present configuration (Fig. 14e) the segments step rightwards and are associated with local transtension. This is indicated by an array of relay-ramp basins (Fig. 14e; in the sense of Peacock & Sanderson 1995). However, the western flanks of the fault segments are strongly uplifted and internally folded. Our observations at the Beaufort ridge (Fig. 14e), together with those by Heimann & Ron (1987) along the Shehumit Fault, suggest that these fault flanks are nucleating in-line folds. Thus the individual fault segments are undergoing transpression, presumably because they step leftwards across the active direction of relative plate motion. This

104

R.W.H. BUTLER E T A L .

interpretation is consistent with the modern location of the rotation pole identified in Fig. 1.

Discussion The finite description of the Lebanese restraining bend with its multiple transcurrent faults, cross-faults, tectonic rotations and folds, belies a more complex tectonic history where the different structures record different deformation histories. In this sense, our conclusions amplify those of Heimann & Ron (1993), who described transient fault segmentation. Such variations in fault segmentation may be expected without necessarily invoking any change in plate kinematics. However, the Dead Sea Transform has experienced such variations in long-range kinematics and they are recorded in its Lebanese segment. Although the region displays the finite geometry of a restraining bend, this geometry reflects only part of the tectonic development and transform history. We have been able to build up a view of incremental structural evolution through landscape stratigraphy. Many of the landforms, such as fault scarps, minor offsets of streams and basins at relay ramps, have remained little modified by surface erosional processes long after the cessation of local tectonic activity. Clearly, in the semi-arid, karst-prone environment of Lebanon, many landforms are highly stable (Butler & Spencer in press). Consequently, it is difficult to use satellite images to build a picture of active faults. Similarly, existing geological maps, constructed before modern developments in tectonic geology and without regard to tracing faults, provide incomplete pictures of fault continuity. The following conclusions, based on a reinterpretation of these types of data but enhanced with our own field observations and dating, may require further modification as the structural geology of Lebanon is mapped in its own right. By piecing together local structural evolution within the restraining bend, it is possible to relate the crustal and local lithospheric response to evolving plate kinematics. As a consequence of the migration of the pole of rotation towards the Dead Sea Transform, new fault segments were activated, bypassing the restraining bend and placing part of the crust from transpression into ideal transcurrent or even locally transtensional tectonics. Thus small parts of the fault system, particularly those at the southern end of the restraining bend, have experienced a rather complex geological history over the past 18 Ma, associated with fault array evolution in a rather low-displacement plate boundary system.

Perhaps such complexities should be expected for most continental transforms. A migrating rotation pole represents one of the simplest examples of plate kinematic evolution. The history of, for example, the East Anatolian Transform (Fig. 1) is likely to have been far more complex, with migrating triple junctions and the related, local changes in fault kinematics this process implies. There is a well-established discrepancy between the amount of crustal shortening in the restraining bend compared with its bulk transcurrent displacement (e.g. Hancock & Atiya 1979; Quennell 1984). However, these analyses were made with the Pliocene rotation pole and the total transform displacement of 105 km. We suggest that this discrepancy has been overestimated because the restraining bend geometry (Fig. 14b) was active for only 50-60% of the total displacement on the transform. Furthermore, the Miocene rotation pole implies a much lower angle of relative plate convergence across the restraining bend. Finally, we point out that there are many unsolved problems within the Dead Sea Transform. Although we have presented a structural history for the restraining bend, it is qualitative. Each stage of the history requires quantification of the angular velocities of plate motion about the particular pole and this compared with resolved magnitudes of transcurrent offset, transpression and the observed strains in the field. An incremental approach is required to achieve these goals, with the search for more, datable, landscape features to provide a calibrated tectono-morphological stratigraphy. We thank Z. Beydoun, C. Walley and K. Khair of the American University of Beirut for discussions on Lebanese tectonics, together with M. Casey for discussions on folding. The original draft of this contribution has been improved greatly thanks to the forthright reviews of G. Roberts and two anonymous referees, although the opinions presented here remain the authors'. H. M. G. was supported by a NERC research studentship. R. W. H. B. was supported by a Nuffield Foundation Research Fellowship. K-Ar analyses were performed in Leeds by P. Guise and D. Rex.

References AMBRASEYS,N. N. & BARAZANGI,M. 1989. The 1759 earthquake in the Bekaa valley: implications for earthquake hazard assessment in the eastern Mediterranean region. Journal of Geophysical Research, 94, 4007-4013. ARTHAUD,E, DUBERTRET,L., MASSON,P. ~; MERCIER, J. L. 1978. I~tats de constraintes successifs posteoc6nes les long des failles libano-syriennes au nord de Beyrouth. 6e ROunion Annales Science de

EVOLUTION OF THE LEBANESE RESTRAINING BEND

la Terre, Orsay (en depot d la Societ~ Gdologique de France). BARTOV, Y. 1994. Geological photomap of Israel and adjacent areas at 1:750 000, 2nd edition. Geological Survey of Israel, Jerusalem. BEYDOUN, Z. R. 1977. The Levantine countries: the geology of Syria and Lebanon (maritime regions). In: NAIRN,A. E. M., KAYES,W. H. & STI-mLI,E G. (eds) The Ocean Basins and Margins, 4A, the Eastern Mediterranean. Plenum, New York, 319-353. & HABIB, J. G. 1995. Lebanon revisited: new insights into Triassic hydrocarbon prospects. Journal of Petroleum Geology, 18, 75-90. BLOT,M. A. 1961. Theory of folding of stratified viscoelastic media and its implications in tectonics and orogenesis. Geological Society of America Bulletin, 72, 1595-1620. BUTLER, R. W. H. & SPENCER,S. (in press). Landscape evolution and the preservation of tectonic landforms along the northern Yammouneh Fault, Lebanon. In: WHALLEu W. B., SMITH, B., WIDDOWSON, M. & SUMMERFIELD, M. (eds) Uplift, Erosion and Stability: Geological and Geomorphological Perspectives on Landscape Evolution. Geological Society, London, Special Publication. - - , PRIOR, D. J. & KNIPE, R. J. 1989. Neotectonics of the Nanga Parbat syntaxis, Pakistan, and crustal stacking in the northwest Himalayas. Earth and Planetary Science Letters, 94, 329-343. & GRIFFITHS,H. M. 1997. Transcurrent fault activity on the Dead Sea Transform in Lebanon and its implications for plate tectonics and seismic hazard. Journal of the Geological Society, London, 153, 757-760. CHAIMOV,T. A., BARAZANGI,M., AL-SAAS, D., SAWAF, T. & GEBRAN,A. 1990. Crustal shortening in the Palmyride Fold Belt, Syria, and implications for movement along the Dead Sea Fault System. Tectonics, 9, 1369-1386. CHRISTIE-BLICK,N. & BIDDLE,K. T. (eds) 1985. Strikeslip Deformation, Basin Formation and Sedimentation. Society of Economic Palaeontologists and Mineralogists, Special Publications, 37. CROWELL, J. C. & SYLVESTER,A. G. (eds) 1979. Tectonics of the Junction between the San Andreas Fault System and the Salton Trough, South-eastern California. University of California, Santa Barbara. DRUCKMAN,Y., BUCHBINDER,B., MARTINOTTI, G. M. SIMON TOY, R. & AHARON, P. 1995. The buried Afiq canyon (eastern Mediterranean, Israel): a case study of a Tertiary submarine canyon exposed in Late Messinian times. Marine Geology, 123, 167-185. DUBERTRET, L. 1955. Carte gdologique du Liban au 1/200 000, avec notice explicative. Minist~re des Travaux Publics, Beirut. -& VAUTRIN, M. H. 1945. Carte gOologique du Liban au 1/50 000, feuille Rachaiya-Nord, avec notice explicative. Minist6re des Travaux Publics, Beirut. EYAL, M., EYAL, Y., BARTOV,Y. & STEINITZ,G. 1981. The tectonic development of the western margin -

-

,

-

-

105

of the Gulf of Elat, Aqaba, rift. Tectonophysics, 80, 39-66. FREUND,R. & TARLING,D. H. 1970. Preliminary Mesozoic palaeomagnetic results from Israel and inferences for microplate structure in Lebanon. Tectonophysics, 60, 189-205. ~ - , GARFUNKEL,Z., ZAK, I., GOLDBERG, M., WEISBROD, T. & DERIN, B. 1970. The shear along the Dead Sea rift. Philosophical Transactions of the Royal Society of London, Series A, 267, 107-130. GARFUNKEL, Z. 1981. Internal structure of the Dead Sea leaky transform (rift) in relation to plate kinematics. Tectonophysics, 80, 81-108. & ALMAGOR, G. 1985. Geology and structure of the continental margin off northern Israel and the adjacent part of the Levantine basin. Marine Geology, 62, 105-131. , ZAK, I. & FREUND,R. 1981. Active faulting in the Dead Sea Rift. Tectonophysics, 80, 1-26. GIRDLER, R. W. 1990. The Dead Sea transform fault system. Tectonophysics, 180, 1-13. & SOUTHREN,T. C. 1987. Structure and evolution of the northern Red Sea. Nature, 330, 716-721. GOEDICKE, T. R. 1972. Submarine canyons of the central continental shelf of Lebanon. In: Stanley, D.J. (ed.) The Mediterranean Sea: a Natural Sedimentation Laboratory. Dowden, Hutchinson & Ross, Stroudsburg, PA, 655-670. GREGOR, C. B., MERTZMAN, S., NAIRN, A. E. M. & NEGENDANK, J. 1974. The palaeomagnetism of some Mesozoic and Cenozoic volcanic rocks from the Lebanon. Tectonophysics, 21, 375-395. HALL, J. K., UDINSTEV,G. B. & ODINOKOV,Y. Y. 1994. The bottom relief of the Levantine Sea. In: KRACHENINNIKOV,V. A. & HALL, J. K. (eds) Geological Structure of the North Eastern Mediterranean (Cruise 5 of Research Vessel 'Akademik Nikolaj Strakhov'). Historical Productions Hall, Jerusalem, 5-32. HANCOCK, P. L. & ATIYA, M. S. 1979. Tectonic significance of mesofracture systems associated with the Lebanese segment of the Dead Sea transform fault. Journal of Structural Geology, 1, 143-153. HEIMANN,A. & RON, H. 1987. Young faults in the Hula pull-apart basin, central Dead Sea transform. Tectonophysics, 141, 117-124. & -1993. Geometric changes of plate boundaries along part of the northern Dead Sea Transform: geochronologic and palaeomagnetic evidence. Tectonics, 12, 477--491. --, EYAL, M. & EYAL, Y. 1990. The evolution of Barhta rhomb-shaped graben, Mount Hermon, Dead Sea Transform. Tectonophysics, 180, 101-110. HEMr'rON, M. R. 1987. Constraints on Arabian plate motion and extension history of the Red Sea. Tectonics, 6, 687-705. HIRSCH, F. 1990. Apercu de l'histoire phan6rozoique d'Israel. Journal of" African Earth Sciences, 11, 177-196. HOFSTETTER, A., VAN ECK, Y. & SHAPIRA, A. 1996. Seismic activity along fault branches of the Dead Sea-Jordan Transform System: the Carmel-Tirtza fault system. Tectonophysics, 267, 317-330. -

-

106

R.W.H. BUTLER ETAL.

HOROWITZ, A. 1979. The Quaternary of Israel. Academic Press, New York. JOFFE, S. & GARFUNKEL, Z. 1987. Plate kinematics of the circum Red Sea - - a re-evaluation. Tectonophysics, 141, 5-22. KEMPLER, D. & GARFUNKEL, Z. 1994. Structure and kinematics in the northeastern Mediterranean: a study of an irregular plate boundary. Tectonophysics, 234, 19-32. KHAIR, K., TSOKAS, G. N. & SAWAF,Y. 1997. Crustal structure of the northern Levant region: multiple source Werner deconvolution estimates for Bouguer gravity anomalies. Geophysical Journal International, 128, 605-616. LAWS, E. D. & WILSON, M. 1997. Tectonics and magmatism associated with Mesozoic passive continental margin development in the Middle East. Journal of the Geological Society, London, 154, 459-464. MOR, D. 1987. Geological Map: Har Odem. Geological Survey of Israel. 1993. A time-table for the Levant Volcanic Province, according to K - A t dating in the Golan Heights, Israel. Journal of African Earth Science, 16, 223-234. NAMMOUR, M. T. 1992. Plages a Turrides Mioc6nes dans le Liban-Sud, leurs rapports avec celles du Quaternaire, avec les glacis-cones et les talwegs fossiles, dans le sublittoral libanais. Hannon (R~vue Libanaise de Gdographie), 21, 7-50. NUR, A. & BEN-AVRAHAM, Z. 1978. The eastern Mediterranean and the Levant: tectonics of continental collision. Tectonophysics, 46, 297-311. PEACOCK, D. C. P. & SANDERSON,D. J. 1995. Strike-slip relay ramps. Journal of Structural Geology, 17, 1351-1360. PLASSARD, J. & KOGOL, B. 1981. Seismicitd du Liban. Conseil National de la Recherche Scientifique, Beirut. QUENNELL, A. M. 1958. The structural and geomorphologic evolution of the Dead Sea Rift. Quar-

terly Journal of the Geological Society, London, -

-

114, 1-24. 1984. The western Arabian rift system. In: DKXON,

J. E. & ROBERTSON,A. H. E (eds) The Geological Evolution of the Eastern Mediterranean. Geological Society of London, Special Publications, 17, 775-788. Roy, H. 1987. Deformation along the Yammuneh, the restraining bend of the Dead Sea transform: palaeomagnetic data and kinematic implications. Tectonics, 6, 653-666. -& EYAL,Y. 1985. Intraplate deformation by block rotation and mesostructures along the Dead Sea Transform, northern Israel. Tectonics, 4, 85-105. - - , NUR, A. & EYAL, Y. 1990a. Multiple strike-slip fault sets: a case study from the Dead Sea Transform. Tectonics, 9, 1421-1432. - - & HOFSTETrER, A. 1990b. Late Cenozoic and Recent strike-slip tectonics in Mt. Carmel, Northern Israel. Annales Tectonicae, 4, 70-80. SANLAVILLE,19. 1970. l~tude gOomorphique de la r~gion littorale du Liban. Publications de l'Universit6 Libanaise, Beyrouth. SHAPmA, A. & HOVSTETrER, A. 1993. Source parameters and scaling relationships of earthquakes in Israel. Tectonophysics, 217, 217-226. VAN ECK, T. & HOFSTEaq'ER,A. 1990. Fault geometry and spatial clustering of microearthquakes along the Dead Sea-Jordan rift fault zone. Tectonophysics, 180, 15-27. WALLEY, C. D. 1988. A braided strike-slip model for the northern continuation of the Dead Sea Fault and its implications for Levantine tectonics. Tectonophysics, 145, 63-72. WESTAWAY, R. 1994. Present-day kinematics of the Middle East and eastern Mediterranean. Journal of Geophysical Research, 99, 12071-12090. 1995. Deformation around stepovers in strikeslip faults. Journal of Structural Geology, 17, 831-846. WOODCOCK, N. H. & SCHtmERT, C. 1994. Continental strike-slip tectonics. In: HANCOCK,P. L. (ed.) Continental Tectonics. Pergamon Press, Oxford, 251-263. ZAK, I. & FREUND, R. 1981. Asymmetry and basin migration in the Dead Sea Rift. Tectonophysics, 80, 27-38. -

-

Tectonic evolution of the Northern Salinian Block, California, USA: Paleogene to Recent shortening in a transform fault-bounded continental fragment ENRICO TAVARNELLI

Centro di Geodinamica, Universitd della Basilicata, 85100, Potenzal Italy

Abstract: The complex structural setting of the western margin of North America is interpreted to result from oblique convergence of the North American and Pacific plates, accommodated by both right-lateral slip along the San Andreas fault and shortening east and west of it. Strike-slip movements along the San Andreas fault led to the detachment of a continental fragment (the Salinian Block) from the North American margin during early Miocene, and its translation northwestwards for over 300 km. Structural analysis in the Northern Salinian Block, west of the San Andreas fault, reveals a NE-SW-directed shortening accommodated by NW-SE-trending folds and thrusts with a dip-slip kinematic character. The stratigraphic record of progressively younger unconformities affected by both folds and thrusts, as well as the overprinting relationships among these contractional structures, enables us to clarify the tectonic evolution of the region since Paleocene time. The Paleocene to Recent history of the Salinian Block was dominated by strike-slip along the San Andreas fault, and by shortening perpendicular to it. The partitioning between strikeslip and dip-slip movements appears to be controlled by a pre-existing tectonic feature. The results from structural analysis along the Salinian Block are integrated into a deformation model for the western margin of North America, providing additional constraints on the timing of deformation and helping to clarify the role of strain-partitioning processes in the obliquely convergent California margin. The general parallelism between map- and outcrop-scale structures, known among geologists as Pumpelly's rule (see Price & Cosgrove 1990), renders mesoscopic analysis important for defining the kinematic history of deformed areas. A l t h o u g h outcrop-scale observations have long proven successful in inferring kinematic models (Wilson 1961; Cosgrove 1980), a limit of geodynamic reconstructions solely based on minor structures is that the latter are usually overprinted by later deformations, thus making tectonic inferences problematic. A well-defined geodynamic framework determined by means of independent methods, instead, makes structural analysis a reliable tool for relating detailed field information to larger-scale features. The tectonic setting of the NE circum-Pacific subduction system is characterized by the juxtaposition of the North American plate, largely made of continental lithosphere, to the Pacific, Juan de Fuca and Cocos plates, almost entirely made of oceanic lithosphere. Well-defined seafloor magnetic isochrons, and the presence of two presumably fixed reference frames, the Hawaii and Yellowstone hot-spots, constrain both relative and absolute plate motions (Atwater 1970, 1989; Stock & Molnar 1988; Doglioni & Harabaglia 1996). In central-northern California, the Pacific and North American plates are separated by a continental fragment,

the Salinian Block, where the record of Paleogene to Recent deformations is preserved along superb coastal exposures. These elements, in combination, make central-northern California an unusually favourable setting where mesoscopic data can tentatively be linked to plate-tectonic reconstructions. This paper aims to unravel the deformation history of the Northern Salinian Block, bounded by the active San Andreas fault system to the east, and by the Pacific plate to the west. The results of a structural analysis provide the basis for discussing the role of strike-slip partitioning processes in the tectonic evolution of the North American ocean-continent transform margin.

Geological setting The juxtaposition of the North American and Pacific plates occurs along a belt of anastomosing NW-SE-trending faults, the San Andreas fault system, which extends from the Gulf of California to Cape Mendocino for over 2500 km. Together with its northern continuation, i.e. the Queen Charlotte fault zone, the San Andreas fault system transfers extension produced at the East Pacific Rise to convergence and subduction in the Aleutian Trench with a dominantly rightlateral strike-slip, thus defining a major o c e a n - c o n t i n e n t transform margin (Wilson

TAVARNELLI,E. 1998. Tectonic evolution of the Northern Salian Block, California, USA: Paleogene to recent shortening in a transform fault-bounded continental fragment. In: HOLDSWORTH,R. E., STRACHAN, R. A.& DEWEY,J. E (eds) 1998. Continental Transpressionaland Transtensional Tectonics. Geological Society, London, Special Publications, 135, 107-118.

107

108

E. TAVARNELLI

1965; Atwater 1970). The inception of the transform regime is generally referred to the Late Oligocene (29-26 Ma: Stock & Molnar 1988; Atwater 1989). A diffuse seismicity associated with the San Andreas fault system, outlined by destructive historical and recent earthquakes (Hill et al. 1990), indicates that this regime is still tectonically active. In spite of a general agreement on the role of the San Andreas fault as a major tectonic boundary, controversy exists on the interpretation of its continuation at depth. Wilson (1965), Atwater (1970) and Namson & Davis (1988), among many others, considered it as a deep-rooted fault cutting the entire lithosphere, i.e. the stacked Pacific and North American plates, whereas Jones et al. (1994) and Holbrook et al. (1996) interpreted it as a shallow feature confined to the North American plate and emanating from a 20-25 km deep, gently NE-dipping d6collement. The tectonic history of California before activation of the San Andreas fault system was dominated by the Andean-type convergence with subduction of the Pacific plate under the North American continent, which had been initiated during Late Mesozoic time (Dickinson 1981). Continued subduction gave rise to an arc-trench system, with development of a frontal accretionary prism (the Franciscan Complex), a forearc basin (the seat of deposition of the Great Valley Sequence) floored by oceanic crust (i.e. the future Coast Ranges Ophiolites), and an igneous complex (the Sierra Nevada Magmatic Arc). Subduction-related shortening was probably accompanied by rightlateral slip along an ancestor lineament precursor of the San Andreas fault (the Proto-San Andreas fault: Nilsen 1981; Page 1990), which could have accommodated up to 300 km displacement in pre-Eocene time (Graham & Dickinson 1978). The present-day San Andreas fault in central-northern California separates continental North America from a dominantly granitic fragment, the Salinian Block (Fig. 1). This fragment is thought to represent the former southern continuation of the Sierra Nevada batholith (Graham 1978; Ross 1983), or an exotic terrane accreted to North America during subduction of the Farallon plate (Vedder et al. 1983). Since early Miocene time, the Salinian Block has been detached from the North American margin and has been translated 300 km northwestwards along the San Andreas fault to its present position in western California (Graham & Dickinson 1978; Blake et al. 1984). Four smaller fragments are recognized within the Salinian Block: the Southeastern, Central, Western and Northern

Fig. 1. Sketch map of the Northern Salinian Block, showing location of study areas.

Blocks (Ross 1983). This study will focus on the structural evolution of the Northern Salinian Block, well exposed along the central-northwestern coast of California (Fig. 1). Rocks of the region consist of Lower Cretaceous, mainly granitic and metamorphic basement, unconformably overlain by Upper Cretaceous to Recent sediments. Granitic rocks crop out extensively in the Montara Mountains, in the Point Reyes Peninsula, at Bodega Head, and, together with other igneous and metamorphic assemblages, are believed to underlie the sedimentary sequences of Point Arena and Pigeon Point (Fig. 2). The Upper Cretaceous to Recent mainly marine sequences show abrupt lateral stratigraphic differences over little horizontal distance, which reflect deposition in local, restricted basins, thus suggesting a complex history of continued synsedimentary deformation (Dickinson et al. 1979). Highly simplified

NORTHERN SALINIAN BLOCK, CALIFORNIA

109

Fig. 2. Simplified composite stratigraphic columns of the outcrop locations of the Northern Salinian Block, based on Nilsen et al. (1981), Graham et al. (1989), and references therein. Arrows indicate the deformation episodes documented in this paper.

composite stratigraphic columns of the exposed sediments are summarized in Fig. 2.

General structure Since inception of the transform regime, the deformation history of central-northern California was characterized by dominant rightlateral strike-slip along the San Andreas and attendant San Gregorio, Pilarcitos, Hayward and Calaveras faults (Johnson & Normark 1974; Graham & Dickinson 1978; McLaughlin et al. 1996). Right-lateral movements were accompanied by coeval SW-NE-directed shortening, mainly expressed by NW-SE-trending folds and thrusts, which are abundant east of the San Andreas fault (Aydin & Page 1984; Namson & Davis 1988; Bloch et al. 1993; Jones et al. 1994,

and references therein). By contrast, the internal architecture of the Northern Salinian Block, west of the San Andreas fault, is poorly constrained and structural information is relatively rare, with important exceptions (Galloway 1977; Coppersmith & Griggs 1978; Joyce 1981; Duane Gibson 1983; Wiley & Moore 1983). This section outlines the character, geometry and orientation of mesoscopic (outcrop-scale) and macroscopic (map-scale) contractional structures, i.e. folds and thrusts, which have affected the Northern Salinian Block since the (?)Late Cretaceous-Paleocene interval. Investigated anticlines and synclines range in both wavelength and amplitude from kilometres (i.e. first-order folds) to metres (i.e. third-order folds, according to the definition by Nickelsen (1963)). Deformation occurred under brittle to

110

E. TAVARNELLI

semi-brittle conditions at shallow crustal levels. Two separate tectonostratigraphic terranes are recognized in the Northern Salinian Block (Fig. 3): the La H o n d a Block, between the San Andreas, Pilarcitos and San Gregorio faults, and the Pigeon Point Block, west of the San Gregorio fault.

Pigeon Point Block The structure of the Pigeon Point Block is characterized by both folds and thrusts. Firstorder folds range from open to isoclinal and their wavelengths vary from a few hundred metres to several kilometres. The major mapscale fold is represented by a syncline which affects the Upper Cretaceous Pigeon Point Formation and whose hinge line intersects the coast near Bolsa Point (Fig. 3). This fold trends 305 ~ and plunges 17 ~ towards the NW (Fig. 4a). Second- and third-order folds, parasitic to the Bolsa Point syncline, trend 308 ~ and plunge 19~ towards the NW (Fig. 4a). The Oligocene-Miocene Vaqueros Fm lies unconformably over the Pigeon Point Fm, post-dating the Bolsa

Point syncline, which probably developed during the Paleocene-Eocene interval (see also Joyce 1981). Remnants of the Vaqueros Fm crop out between Franklin Point and Point Afio Nuevo and at Pescadero Beach (Fig. 3), dipping homoclinally 30 ~ towards the N N E (Duane Gibson 1983), and could represent a portion of a WNW-trending map-scale fold. The Vaqueros Fm is unconformably overlain by the Late Pliocene Purisima Fm, which is affected by a first-order anticline-syncline pair whose hinge lines intersect the coast at San Gregorio Beach and Pomponio Beach, respectively. These folds of probable late Pliocene-(?)Pleistocene age, trend NW-SE, subparaUel to the Tertiary Bolsa Point syncline, and plunge gently towards the NW (Fig. 4b). The effects of reverse faulting are particularly evident near Point Afio Nuevo, where mudstones and cherts of the upper Miocene Monterey Fm are thrust southwestwards over Quaternary marine and continental deposits. This fault, known as the Point Afio Nuevo thrust (Coppersmith & Griggs 1978), strikes N W - S E and dips 35~ towards the NE. Mechanical striae and grooves along the main thrust and attendant faults, a weak cleavage in the fault rock and rare synthetic (R) Riedel shears indicate a dip-slip kinematic character. Displacements are towards the SW with a 41 ~ mean slip direction (Fig. 5a). The fault is parallel to bedding in the hanging wall, and makes angles of 20-35 ~ with beds in the footwall, thus defining a hanging wall flat-footwall ramp geometry (Fig. 6a): this pattern indicates a minimum displacement of about 8 m. Other mesoseopic reverse faults, along which the Upper Cretaceous Pigeon Point Fm is thrust over Pleistocene terrace deposits, cropout near Pescadero Beach (Fig. 3). These faults strike N W - S E and dip 20-45 ~ towards both the SW and NE (Fig. 5b). Striae indicate a dip-slip kinematic character and define a 43 ~ mean slip direction (Fig. 5b).

La Honda Block Structural analysis in the northern part of the La Honda Block was carried out in the Montara Mountains, the Point Reyes Peninsula, Bodega Head and Point Arena areas (Fig. 1).

Montara Mountains Fig. 3. Sketch map of the structure of Pigeon Point Block, west of the San Gregorio fault (see inset in Fig. 1); asterisks indicate the outcrop location of described structures.

The main structure recognized in the Montara Mountains is a first-order, asymmetrical NNEverging anticline that is well exposed in the cliffs between Shelter Cove and Devil's Slide (Fig.

NORTHERN SALINIAN BLOCK, CALIFORNIA

~1

b

PigeonPoint

San GregorioPomponio Beach

111

~

9

Devil'sSlides

x<,/ • -~

/

,%

9

xx

9 , 9149 9

9

128 Data

d

axial trend 305 o

MOSSBeach

/,x ,,,, -U

~

n 176 Data

o - p o l e s to b e d d i n g

meanf o l d

41 Data

e

axial trend 316 o

x

x

257 Data

f

Point Reyes

~149

~..~/~mean

told

axial trend 307 o

PointArena

\

,

axial tren~t 304 e

114 Data

,~ - p o l e s to c l e a v a g e

axial trend 313 o

• - minor fold axes

~

l

~

:

g

108 Data

O t-

Id 318 o

map-scale fold axes

Fig. 4. Orientation of major and minor folds in the Northern Salinian Block (equal-area projection, lower hemisphere). The dashed line represents the local trend of the San Andreas fault. Filled and open circles in (b) refer to the Pescadero Beach syncline and the San Gregorio Beach anticline, respectively (see Fig. 3 for location). Filled and open circles in (d) refer to the northern and southern synclines at Moss Beach, respectively (see Fig. 7b for location).

7a). This structure affects the Cretaceous Montara Granodiorite and the unconformably overlying Paleocene Point San Pedro Fm; the fold trends 307 ~ and plunges 21 ~ towards the NW (Fig. 4c). Second- and third-order folds, parasitic to the major anticline, are well exposed along the California Highway 1 at Devil's Slide (Fig. 6b): their mean trend is 304 ~ (Fig. 4c). This structure is post-dated by an unconformably overlying Pleistocene marine terrace (Fig. 7c). Further to the south, at Moss Beach, a superb exposure in the Fitzgerald Marine Reserve allows for accurate definition of the local 3D structural geometry (Fig. 7b). The stratigraphy, and depositional and deformation history of the area were accurately defined by Wiley & Moore (1983), and further structural details are reported here. The Upper Pliocene Purisima Fm u n c o n f o r m a b l y overlies the Cretaceous

Montara Granodiorite, and is affected by two second-order asymmetrical synclines separated by a NE-verging anticline (Fig. 7b). The trends of these structures range between 302 ~ and 307 ~, with plunges up to 24 ~ towards the NW (Fig. 4d). These structures are post-dated by a Pleistocene marine terrace, which unconformably overlies both the granodiorite and the Purisima Fm (Fig. 7b). Impressive structural relationships are observed along the northern cliffs of the nearby Montara Point, where mesoscopic reverse faults determine local stratigraphic repetitions, carrying the Cretaceous granodiorite over the Pleistocene marine deposits (Fig. 6c). These faults, which trend N W - S E to W N W - E S E and dip generally 35 ~ towards NE, also affect Holocene deposits. Mechanical striae indicate a dip-slip kinematic character, with displacement of

112

E. TAVARNELLI

a

Pt. Afio Nuevo

b

Pescadero Beach

C

Montara Beach

.y 61 Data

d

mean slip dkection 41 o

Point Reyes

~ 14 Data

e

f

Bodega Head

~

-

striae on thrust faults

direction 47 0

l

i

'I~,I~L~'\

\

~_ planes

24 Data

9- poles to S surfaces

direction 37 o

Point Arena

al~e~tloon 21 Data

9- R Riedel shears

p

7 Data

t 66 Data

s

direction 43 o

mean slip direction 56 o

9- poles to C surfaces

Fig. 5. Orientation and kinematics of mesoscopic thrusts and related shearing fabrics showing a dominantly dip-slip kinematic character (equal-area projection, lower hemisphere). The dashed line represents the local trend of the San Andreas fault.

NORTHERN SALINIAN BLOCK, CALIFORNIA

113

Fig. 7. Geological setting of the Montara Mountains (see Fig. 1 for location). (a) Simplified tectonic sketch map. (b) Detailed map of the Fitzgerald Marine Reserve at Moss Beach (partly based on mapping by Wiley & Moore (1983)) at low tide (see inset in (a) for location). (e) Simplified cross-section (trace A-A' of (a)) showing unconformities of Paleogene (1), Pliocene (2) and Pleistocene (3) ages offset by progressively younger contractional structures. Kgr, Montara Granite (Cretaceous); Pal, Point San Pedro Fm (Paleocene); P1, Purisima Fm (Late Pliocene); Q, marine terrace deposits (Pleistocene and Holocene).

Fig. 6. Mesoscopic reverse faults and related shearing fabrics from outcrops described in the paper (redrawn from photographs).

114

E. TAVARNELLI

hanging wall blocks generally towards the SW (Fig. 5c).

Point Reyes Peninsula The Point Reyes Peninsula is the only area where continuous outcrop along coastal exposures made it possible to reconstruct the general structure. This was done with mapanalysis techniques, using the 1 : 48 000 map by Galloway (1977). The main structure is the Horse Ranch syncline, a broad first-order fold which affects the C r e t a c e o u s - L a t e Miocene sequence and underlying rocks. This fold, which approximately trends 140 ~ (Fig. 8a), is postdated by the unconformably overlying Late Pliocene Drake's Bay Fm (Fig. 8b). Another important structure is the Point Reyes syncline, an open first-order fold which affects both the U p p e r Pliocene Purisima Fm and the basal unconformity that separates it from the underlying Miocene sequence (Fig. 8a and b). This structure trends 313 ~, and plunges 7~ towards the NW (Fig. 4e). Other mapped folds at Drake's Beach are a second-order anticline and syncline pair, parasitic to the Point Reyes syncline. Their hinge line trends range from 135 ~ to 146 ~ (Fig. 8a). Minor, asymmetrical folds parasitic to the Horse Ranch syncline are common in the upper Miocene Monterey Fm. Their average trend is 315 ~ (Fig. 4e). Mesoscopic thrust faults are abundant in the northernmost Peninsula. A reverse fault, along which the Early Miocene Laird Sandstone is overridden by Cretaceous granitic basement rocks, is exposed at Keohe Beach (Fig. 6d). This structure, and related minor thrusts, strike W N W - E S E and dips 70 ~ towards the NNE.

Mechanical striae on the fault surfaces indicate a dip-slip kinematic character, and displacement towards the SSW with a mean 47 ~ slip direction (Fig. 5d).

Bodega Head The Bodega H e a d Peninsula represents the northernmost exposure of the Salinian granitic basement rocks along the California coast, and is separated from the mainland by the 2-3 km wide San Andreas fault zone. The granitic basement, consisting of quartz diorite and diorite, is intruded by numerous pegmatite dykes (Ross 1983). These intrusive rocks are unconformably overlain by Quaternary marine terrace deposits, which were tilted towards the N E presumably as a result of slip along the San Andreas fault (Koenig 1963). At Horseshoe Cove, a subvertical NW-SE-trending fault juxtaposes the basement rocks, in the west, against Quaternary sediments, in the east. No kinematic indicators were found along the fault surface, which probably represents a strand of the San Andreas fault. Evidence for recent shortening is found at Mussel Point, where small reverse faults truncate the unconformity which separates the marine terrace deposits from the underlying basement rocks. These faults strike N W - S E and dip 25-46 ~ towards either the SW or NE. Mechanical striae on the fault surfaces indicate a dip-slip character, and a mean 44 ~ slip direction (Fig. 5e). Other mesoscopic thrust faults occur at Windmill Beach, where the pegmatite dykes are offset with small (3-10 cm) displacements. These faults strike N W - S E and dip 20 ~ towards the SW (Fig. 5e).

Fig. 8. Geological setting of the Point Reyes Peninsula (see inset in Fig. i for location). (a) Highly simplified hinge-point tie-line map. (b) Simplified cross-section showing unconformities of Paleogene (1), Miocene (2) and Pliocene (3) ages overprinted by progressively younger structures (modified after Galloway (1977)).

NORTHERN SALINIAN BLOCK, CALIFORNIA Point Arena

The northernmost portion west of the San Andreas fault, between Fort Ross and Point Arena, exposes three sedimentary sequences separated by two major unconformities (Fig. 2), which are believed to be underlain by the Salinian basement (McCulloch 1989). The oldest unit, represented by Upper Cretaceous to middle Eocene rocks deposited in the forearc Gualala basin (Loomis & Ingle 1994), is affected by tight to isoclinal, NW-SE-trending small folds. This sequence is unconformably overlain by an Oligo-Miocene unit which was deposited in the rapidly subsiding Point Arena basin. Siliceous strata of the Point Arena Fm are affected by gentle to tight, NW-SE-trending first-order folds (Fig. 4f). These folds are truncated by an erosional surface which represents the base of a Pleistocene marine terrace. This unconformity is truncated by three small thrust faults along which the Miocene Point Arena Fm overrides the Pleistocene terrace (Fig. 6e). The thrust contact is marked by a 1 m thick fault gouge with well-developed shearing fabrics (Fig. 60 . Mechanical striations along these structures indicate a dominant dip-slip character, with SWdirected thrust kinematics (Fig. 50. Holocene terrace deposits post-date thrusting (Fig. 6e).

Tectonic significance of Paleogene to Recent shortening The data presented above indicate that the Northern Salinian Block has experienced shortening since the (?)Paleocene, i.e. the proposed age for the Bolsa Point Syncline (Fig. 3), and that this deformation continued up to the Recent, as documented by faulted Holocene deposits at Montara Point (Fig. 6c). The recognized folds range in size from mesoscopic (outcrop-scale) to macroscopic (map-scale), the former being often parasitic to the latter; this suggests that contractional structures of different scales are geometrically and kinematically linked, and that the small strains inferred from minor structures reflect larger strains produced by plate-tectonicscale features. Palaeomagnetic data show that folds and thrusts, although disrupted by the Neogene San Gregorio and Pilarcitos rightlateral strike-slip faults, did not significantly change orientation (Duane Gibson 1983). These combined lines of evidence can be used to test the existing plate-tectonic models for the western margin of continental North America. The NW-SE trend of Paleogene to Recent folds and thrusts, and the inferred mean SW-NE shortening direction, poses the question of how

115

these structures relate to coeval right-lateral slip along the San Andreas and attendant faults. Sylvester & Smith (1976) and Aydin & Page (1984), among many others, ascribed compressional structures to wrench tectonics induced by the Neogene transform regime. Although this model accounts for contractional structures trending oblique to strike-slip faults (Wilcox et al. 1973), conventional wrench tectonics fails to explain the marked parallelism between the observed fold-thrust structures and the San Andreas fault (Figs 4 and 5). Geodetic data show that the Pacific plate is at present moving toward N37~ with respect to North America, whereas the San Andreas fault south of the Mendocino triple junction strikes N41~ (Jordan & Minster 1988; Gripp & Gordon 1990; Doglioni & Harabaglia 1996); this discrepancy results in slightly oblique convergence, i.e. wrench-dominated transpression, between the Pacific and North American plates (Tikoff & Teyssier 1994). Geodynamic reconstructions for the past 5 Ma indicate that the relative motions of the Pacific and North American plates were also mainly characterized by right-lateral strike-slip accompanied by a minor component of convergence (Atwater 1989). As we go further back in time, the kinematic character of the Pacific-North America plate boundary becomes controversial, and the San Andreas fault is alternatively interpreted as a transpressional (Engebretson et al. 1985; Stock & Molnar 1988), or a transtensional (Atwater 1989; Bohannon & Parson 1995) feature. The reasons for these discrepancies originate in the different methods used to determine relative and absolute plate motions (Atwater 1989). Based on the present work, the reconstruction by Stock & Molnar (1988) for the 29-5 Ma interval, where the Pacific-North America relative movements are transpressional, is preferred because it accounts for the kinematic evidence in the Northern Salinian Block. The coexistence of strike-slip and contractional deformations is best explained by strikeslip partitioning (sensu Teyssier et al. 1995) of oblique convergence at the North AmericaPacific plate boundaries, where oblique movements produce a major component of rightlateral strike-slip parallel to the San Andreas fault, and a minor component of shortening perpendicular to it. A kinematic model for the western margin of California is illustrated in Fig. 9. As first proposed by Oldow et al. (1990), displacement compatibility suggests that thrusts and related folds, as well as strike-slip faults, share a common d6collement. This is broadly consistent with the hypothesis of Jones et al.

116

E. TAVARNELLI

(1994) for an active, low-angle subduction of the Pacific under the North American plate, and the consequent inference of a shallow San Andreas fault. The model closely recalls those by Jones et al. (1994) and Doglioni & Harabaglia (1996), yet differs as it incorporates structural evidence from the Northern Salinian Block. The model is also similar to those proposed for other transpressional orogens, such as the Caledonian belt of Greenland (Holdsworth & Strachan 1991) and Norway (Northrup & Burchfiel 1996), where deformation occurred at deeper crustal levels under dominantly ductile conditions, thus suggesting that P - T variations do not significantly influence kinematic strain-partitioning. An interesting question concerns the factors influencing the proposed strike-slip partitioning processes. According to Zoback et al. (1987), the San Andreas fault is a mechanically weak structure and the stress field in central California is mainly controlled by fault properties. Tikoff & Teyssier (1994), on the other hand, have argued that lithospheric-scale deformation is primarily controlled by boundary conditions imposed by plate motion, and not by fault properties. Integrated field and geophysical evidence is consistent with both views. In addition to these influences, tectonic heritage can play an important role in strain-partitioning processes. It is widely agreed that, unlike oceanic lithosphere, continental crust accommodates much of its deformation by reactivation of pre-existing structures which can effectively act as loci for stress concentration during superimposed deformations (Schedl & Wiltschko 1987; Holdsworth et al. 1997 and references therein). Stratigraphic

r,lxlgfloi ,G'

Fig. 9. Kinematic deformation model for western California, where contractional and strike-slip deformation result from strike-slip partitioning processes related to oblique subduction of the Pacific plate under North America.

documentation of a Proto-San Andreas fault, subparallel to the present-day transform boundary, and active in the Late Cretaceous-Paleocene interval (Nilsen 1981; Page 1990), makes this feature an excellent candidate as an important mechanical anisotropy responsible for strike-slip partitioning during inception of wrench-dominated transpression. These elements suggest that the Paleocene to Recent evolution of the Northern Salinian Block and the inferred strain-partitioning along the San Andreas fault were controlled by a combination of relative plate motion, intrinsic fault properties and fault reactivation processes.

Concluding remarks The tectonic evolution of the Northern Salinian Block was dominated by SW-NE-directed shortening expressed by NW-SE-trending folds and thrusts. Structural data provide the analytical background to a kinematic deformation model, where coeval contractional and strike-slip deformations result from strike-slip partitioning of oblique convergent movements at the Pacific-North American plate boundary. In combination with other factors, such as boundary conditions imposed by plate motion and/or intrinsic fault properties, strain-partitioning processes could have been controlled by a major preexisting tectonic feature, the Proto-San Andreas fault, active since Late Cretaceous time. Paleogene to Recent strike-slip deformations and coeval SW-NE-directed shortening support the hypothesis of a shallow San Andreas fault branching from a gently NE-dipping d6collement. This view contrasts with the general notion of a conventional transform fault affecting the whole lithosphere, i.e. the stacked Pacific and North American plates. This study illustrates the effectiveness of mesoscopic analysis as a tool to ultimately test existing regional tectonic models. This work was carried out while I was in receipt of a NATO-CNR Advanced Fellowship at the Department of Geology and Geophysics of the University of California, Berkeley, USA. ! am indebted to W. Alvarez for introducing me to the fascinating geology of the Salinian Block. D. Jones, P. Claeys, D. Karner, A. Clymer and B. Fouke are also gratefully acknowledged for stimulating discussions in the field. Suggestions and comments by B. Tikoff, C. Teyssier and B. Holdsworth, along with critical reviews by Ian Alsop and an anonymous referee, greatly improved the original manuscript.

References ATWATER,T. 1970. Implications of plate tectonics for the Cenozoic tectonic evolution of Western North

NORTHERN SALINIAN BLOCK, CALIFORNIA America. Geological Society of America Bulletin, 81, 3513-3536. 1989. Plate tectonic history of the northeast Pacific and western North America. In: The Geology of North America, Volume N. Geological Society of America, Boulder, CO, 21-72. AYDIN, A. & PAGE, B. M. 1984. Diverse PlioceneQuaternary tectonics in a transform environment, San Francisco Bay region, California. Geological Society of America Bulletin, 95, 1303-1317. BLAKE, M. C., HOWELL, D. G. & JAYKO, A. S. 1984. Tectonostratigraphic Terranes of the San Francisco Bay region. In: BLAKE,M. C. (ed.) Franciscan Geology of Northern California. Pacific Section, Society of Economic. Paleontologists and Mineralogists, Pacific Coast Paleogeography Symposium, 43, 5-22. BLOCH, R. B., VON HEUNE, R., HART, P. E. & WENTWORTH, C. M. 1993. Style and magnitude of tectonic shortening normal to the San Andreas fault across Pyramid Hills and Kettleman Hills South Dome, California. Geological Society of America Bulletin, 105, 464-478. BOHANNON, R. G. & PARSON,Z. 1995. Tectonic implications of post-30 Ma Pacific and North American relative plate motions. Geological Society of America Bulletin, 107, 937-959. COPPERSMITH,K. J. & GRIGGS,G. B. 1978. Morphology, recent activity and seismicity of the San Gregorio fault zone. In: SILVER,E. A. & NORMARK,W. R. (eds) San Gregorio-Hosgri Fault Zone, California. California Division of Mines and Geology Special Report 137, 33-42. COSGROVE, J. W. 1980. The tectonic implications of some small-scale structures in the Mona Complex of Holy Isle, North Wales. Journal of Structural Geology, 2, 383-396. DICKINSON,W. R. 1981. Plate tectonics and the Continental Margin of California. In: ERNST,W. G. (ed.)

The Geotectonic Development of California (Rubey Volume 1). Prentice-Hall, Englewood Cliffs, N J, 1-28. INGERSOLL,R. V. & GRAHAM,S. A. 1979. Palaeogene sediment dispersal and palaeotectonics in Northern California. Geological Society of America Bulletin, 90,1458-1528. DOGLIONI, C. &; HARABAGLIA,P. 1996. The kinematic paradox of the San Andreas Fault. Terra Nova, 8, 525-531. DUANE GIBSON, J. 1983. Tectonic setting and deformation associated with the San Gragorio fault system, San Francisco Peninsula, California. In: ANDERSEN,D. W. & RYMER,M. J. (eds) Tectonics

--,

and Sedimentation along Faults of the San Andreas System. Pacific Section, Society of Economic Paleontologists and Mineralogists, 4, 67-77. ENGEBRETSON,D. C., Cox, A. & GORDON, R. G. 1985. Relative motions between oceanic and continental plates in the Pacific Basin. Geological Society of America, Special Paper, 206, 1-59. GALLOWAY,A. J. 1977. Geology of the Point Reyes Peninsula, Marin County, California. California Division of Mines and Geology Bulletin, 202, 1-72.

117

GRAHAM, S. A. 1978. Role of Salinian Block in evolution of San Andreas fault system, California.

Bulletin, American Association of Petroleum Geologists, 62, 2214-2231. --

& DICKINSON,W. R. 1978. Evidence for 115 kilometers of right slip on the San Gregorio-Hosgri Fault trend. Science, 199, 179-181. - - , STANLEY,R. G., BENT,J. V. & CARTER,J. B. 1989. Oligocene and Miocene palaeogeography of central California and displacement along the San Andreas fault. Geological Society of America Bulletin, 101, 711-730. GRIPP, A. E. & GORDON, R. G. 1990. Current plate velocities relative to the hot-spots incorporating the NUVEL-1 global plate motion model Geophysical Research Letters, 17, 1109-1112. HILL, D. P., EATON, J. P. & JONES, L. M. 1990. Seismicity, 1980-86. In: WALLACE,R. E. (ed.) The San Andreas Fault System, California. US Geological Survey Professional Paper, 1515. HOLBROOK, W. S., BROCHER,T. M., TEN BRINK, U. & HOLE, J. A. 1996. Crustal structure of a transform plate boundary: San Francisco Bay and the central California continental margin. Journal of Geophysical Research, 101, 22311-22334. HOLDSWORTH, R. E. & STRACHAN,R. A. 1991. Interlinked system of ductile strike slip and thrusting formed by Caledonian sinistral transpression in northeastern Greenland. Geology, 19, 510-513. --, BUTLER, C. A. & ROBERTS, A. M. 1997. The recognition of reactivation during continental deformation. Journal of the Geological Society, London, 154, 73-78. JOHNSON, J. D. & NORMARK,W. R. 1974. Neogene tectonic evolution of the Salinian Block, West-Central California. Geology, 2, 11-14. JONES, D. L., GRAYMER,R., WANG, C., McEVILLY,T. V. & LOMAX,A. 1994. Neogene transpressive evolution of the California Coast Ranges. Tectonics, 13, 561-574. JORDAN,T. H. & MINSTER,J. B. 1988. Measuring crustal deformation in the American West. Scientific American, 256, 48-58. JOYCE,J. M. 1981. A deformational history of the Pigeon Point Formation. In: FRIZZELL,V. (ed.) Upper Cre-

taceous and Palaeocene Turbidites, Central California Coast. Pacific Section, Society of Economic Paleontologists and Mineralogists, 2, 57-59. KOENIG, J. B. 1963. The geologic setting of Bodega Head. California Division of Mines and Geology Bulletin, 16, 1-15. LOOMIS, K. B. & INGLE, J. C. 1994. Subsidence and uplift of the Late Cretaceous-Cenozoic margin of California: new evidence from the Gualala and Point Arena Basins. Geological Society of America Bulletin, 106, 915-931. MCCULLOCH, D. S. 1989. Evolution of the offshore central California margin. In: WINTERER, E. L., HUSSONG, D. M. & DECKER, R. W. (eds) The Eastern Pacific Ocean and Hawaii. The Geology of North America, Volume N. Geological Society of America, Boulder, CO, 43%469. MCLAUGHLIN, R. J., SLITER, W. V., SORG, D. H., RUSSELL, P. C. • SARNA-WOJCICKI,A. M. 1996.

118

E. TAVARNELLI

Large-scale right-slip displacement on the East San Francisco Bay Region fault systm, California: implications for location of late Miocene to Pliocene Pacific plate boundary. Tectonics, 15, 1-18. NAMSON, J. S. &; DAVIS,Y. L. 1988. Seismically active fold and thrust belt in the San Joaquin Valley, central California. Geological Society of America Bulletin, 100, 257-273. NICKELSEN, R. E 1963. Fold patterns and continuous deformation mechanisms of the central Pennsylvania folded Appalachians. In: Pittsburgh Geo-

logical Society Guidebook to Tectonics and Cambro-Ordovician Stratigraphy, Central Appalachians of Pennsylvania, 13-29. NILSEN, Y. H. 1981. Late Cretaceous geology of California and the problem of the Proto-San Andreas fault. In: HOWELL, D. G. & McDOUGALL, K. A. (eds) Mesozoic Palaeogeography of the Western United States. Pacific Section, Society of Economic Paleontologists and Mineralogists, Pacific Coast Paleogeography Symposium, 2, 559-573. - - , CLIVrON,H. E. & HOWELL, D. G. 1981. Introduction to Upper Cretaceous and Palaeocene turbidites of Central California. In: FRIZZELL, V. (ed.) Upper Cretaceous and Palaeocene Turbidites, Central California Coast. Pacific Section, Society of Economic Paleontologists and Mineralogists, 2, 1-4. NORTHRUP, C. J. & BURCHFIEL, B. C. 1996. Orogenparallel transport and vertical partitioning of strain during oblique collision, Efjorden, north Norway. Journal of Structural Geology, 18, 1231-1244. OLDOW, J. S., BALLY,A. W. & AVE LALLEMANT,H. G. 1990. Transpression, orogenic float, and lithospheric balance. Geology, 18, 991-994. PAGE, B. M. 1990. Evolution and complexities of the transform system in California, U.S.A. Annales Tectonicae, special issue, supplement to 4, 53-69. PRICE, N. J. & COSOROVE,J. W. 1990. Analysis of Geological Structures. Cambridge University Press. Ross, D. C. 1983. The Salinian Block - - a structurally displaced granitic block in the California Coast Ranges. Geological Society of America, Memoirs, 159, 255-264.

SCHEDL, A. & WILTSCHKO,D. V. 1987. Possible effects of pre-existing basement topography on thrust fault ramping. Journal of Structural Geology, 9, 1029-1037. STOCK,J. & MOLNAR,E 1988. Uncertainties and implications of the Late Cretaceous and Tertiary position of North America relative to the Farallon, Kula and Pacific Plates. Tectonics, 7, 1339-1384. SYLVESTER,A. G. • SMITH, S. S. 1976. Tectonic transpression and basement-controlled deformation in San Andreas fault zone, Salton Trough, California. Bulletin, American Association of Petroleum Geologists, 30, 2081-2102. TEXSSlER, C.,TIKO~, B. & MARKLEY,M. 1995. Oblique plate motion and continental tectonics. Geology, 23, 447-450. TIKOFF,B. & TEYSSIER,C. 1994. Strain modeling of displacement-field partitioning in transpressional orogens. Journal of Structural Geology, 16, 1575-1588. VEDDER, J. G., HOWELL, D. G. & MCLEAN, H. 1983. Stratigraphy, sedimentation and tectonic accretion of exotic terranes, Southern Coast Ranges, California. Memoirs, American Association of Petroleum Geologists, 34, 471-496. WILCOX, R. E., HARDING, Y. P. & SEELY, D. R. 1973. Basic wrench tectonics. Bulletin, American Association of Petroleum Geologists, 57, 74-96. WILEY, T. J. & MOORE, E. J. 1983. Pliocene shallowwater sediment gravity flows at Moss Beach, San Mateo County, California. In: LARUE, D. K. & STEEL, R. J. (eds) Tectonics and Sedimentation along Faults of the San Andreas System. Pacific Section, Society of Economic Palaeontologists and Mineralogists, 29-43. WILSON, G. 1961. The tectonic significance of smallscale structures and their importance to the geologist in the field. Annales de la SocidtO GOologique de Belgique, 84, 423-548. WILSON, J. T. 1965. A new class of faults and their bearing on continental drift. Nature, 207, 343-347. ZOBACK,M. D., ZOBACK,M. L., MOUNT,V. S. et al. 1987. New evidence on the state of stress of the San Andreas fault system. Science, 238, 1105-1111.

Contractional and extensional structures in the transpressive 'Big Bend' of the San Andreas fault, southern California DEREK

RUST

Neotectonics Research Centre, Brunel University, Uxbridge UB8 3PH, U K (e-mail." derek, rust@bruneL ac. uk)

Abstract: The San Andreas Big Bend is well known as a restraining bend producing transpressional conditions. However, field mapping in the southern part of the bend reveals a very wide (10 km) zone of active San Andreas faulting displaying a marked structural asymmetry on either side of the conspicuous central main fault trace, a result unexpected in the context of theoretical and model studies. The southern half of the fault zone is dominated by contraction and uplift in a strike-slip duplex, whereas the adjacent northern half displays a much greater number of probably older faults, particularly a succession of apparent Pshears as well as left-lateral probable Rl-shears, producing extension and subsidence. It is proposed that these two halves of the fault zone represent mirror images of each other as a result of their positions relative to the most restraining section of the Big Bend, the shortening in the southern half being initiated as this section is approached, and the extension and complex faulting history of the northern half occurring as this terrain moves away from the most restraining section. These asymmetrical structural relationships, as well as their history of development, may be more generally applicable to major strike-slip fault zones in transpressional and, by analogy, transtensional situations. Clearly, the two halves of such fault zones are, structurally, remarkably discrete.

The well known 'Big Bend' in the San Andreas fault is regarded as the classic example of a restraining bend producing transpressive conditions along a strike-slip fault (Crowell 1974; Fig. 1). Similar conditions have been obtained in model studies of strike-slip faulting where a component of convergence is introduced, and generally result in a symmetrical pattern of compressional faults; flower structures being a notable example (Lowell 1972; Wilcox et al. 1973; Bartlett et al. 1981; Naylor et al. 1986). However, field documentation of such structures on the San Andreas fault, such as the work of Sylvester & Smith (1976) in the Mecca Hills exposures of the fault near the Salton Sea, is limited by the small scale and incompleteness of fault zone maps of the San Andreas system (Moore & Byerlee 1991), as well as the difficulty in distinguishing between active faults related to the prevailing stress regime and inactive faults which may be inherited from past conditions (Rust 1993). Detailed, large-scale mapping studies of well-preserved ground ruptures carried out shortly after major strike-slip faulting events in arid regions (Clark 1968; T c h a l e n k o & A m b r a s e y s 1970) are few in number and do not include restraining bends approaching the size of the Big Bend. Moreover, such single events do not rupture all the active faults making up the fault zone, thus diminishing the applicability of the work to the longer-term

faulting behaviour of the zone as a whole. Against this background, the aim of the present study is to analyse the results of large-scale field mapping of active faults which make up the San Andreas fault zone in the southern part of the Big Bend. This mapping records active extensional as well as contractional structures, and a marked asymmetry in structural style between the two halves of the fault zone.

The Big Bend and transpressional tectonics The overall strike of the San Andreas transform for most of its length is approximately 040 ~. This is maintained for some 500 km from the Mendocino triple junction in the north as far as the northern margin of the Transverse Ranges in southern California, where the strike changes to about 080 ~ in the Big Bend reach before curving back to a strike of about 065 ~ as the fault continues southwards for some 300 km towards the head of the Gulf of California (Moore & Byerlee 1991; Fig. 1). Such a structure in a right-lateral system operates as a restraining bend, and this appears to be supported by the pattern of historical ground rupturing events on the fault, as originally noted by Allen (1968). The origin of the Big Bend has been attributed to the opening of the Gulf of California associated with the East Pacific Rise stepping inboard of the Mexican

RUST.D. 1998. Contractional and extensional structures in the transpressive 'Big Bend' of the San Andreas fault, southern California. In: HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. E (eds) 1998. Continental Transpressionaland Transtensional Tectonics. Geological Society, London, Special Publications, 135, 119-126.

119

120

D. RUST

Fig. 1. Location map showing the mapped segment of the San Andreas fault zone (Fig. 2) in the southern part of the large restraining bend ('Big Bend') in the San Andreas fault.

margin about 4.5 Ma ago, causing Baja California to rift away from mainland Mexico (Crowell 1974, 1979), and to interactions between the San Andreas and major left-lateral regional faults (Garlock and Big Pine) in the Transverse Ranges (Davis & Burchfiel 1973; Bohannon & Howell 1982; Fig. 1). May et al. (1993) and Westaway (1994) have provided more recent analyses of these structures and events in producing the Big Bend.

structural style of the active subsidiary faulting making up the opposing halves of the fault zone on either side of the conspicuous main fault trace (Fig. 3), this being equivalent to the principal displacement zone of theoretical studies as summarized by Sylvester (1988). These contrasts can be appreciated by reference to the fault map shown in Fig. 2, and will be highlighted during the following consideration of the two halves of the fault zone.

Fault mapping and fault zone asymmetry

Fault zone south of the main fault trace

Field mapping was carried out at a scale of 1 : 12 000 with the help of panchromatic, colour and infra-red aerial photographs, including a low-sun series, ranging from 1:6000 to 1:20 000 scale. This work documented a zone of active faulting (Rust 1993) up to 10 km wide (Fig. 2), significantly wider than the maximum fault zone width of 2.3 km in the southern Big Bend recorded by the 1:24 000 photo-interpreted fault maps available to Moore & Byerlee (1991) A further observation was the contrasting

As shown by the fault map (Fig. 2), this half of the fault zone has significantly fewer active faults than the adjacent part of the fault zone on the other side of the main fault trace. The most prominent structure present is a fault-bounded ridge trending sub-parallel to the main fault trace (Figs 3 and 4). The mapping (Fig. 2) shows that uplift of the ridge is principally accomplished along a fault which, near the main fault trace, is initially vertical and striking approximately E-W, then curves progressively southwards with

CONTRACTIONAL AND EXTENSIONAL STRUCTURES

a p

~

~

t

\ ~ , ,

PDz

:

FI

/'

..

~

..

/

9

9.'

--':.

.........

""

.....

..".-"

.'" /'-".'/ .'..;"" ~ --..=_.' ' _ _ _ X ~ - - ~ - ~ - - - - - - - - - - - . - - - - - -

~,--~_.~_

rr:..;;--

."

,

.'..;/

.- . . . . . . . .

_ ~

-'-~'~ f

[

........ - . . . . ~ -

/_.-.:,

'"9

: ....

-T.~'-

'.

,:

/;' / .... ... ~t. ....... 1 / ...

I-

....

,.-

....

.

: .:

," " , ..

,"-"" ,~,

9

,,

/

/,,.__!_

......

~-.----"" --" ~ " ..1~ ~ - - - - - - - - " ~ - ' ~ " ~ " - " - - ' ~ ~

~

..-%.

__~

~-/Y~

/

,

."_-,-r--~_,.

#

~

.

/

.

- - _ ' , ...... ,. /. .,~,

,i ''~

,

-'"

.' -9

/

2.

9

-

I

-.-'_-~ _cz

/ ,K_~-J~-", ,,- .," / ;~....~ ,, ,.

..

.'"

,_'."

,,

:

4

","

.,

2:- .........

1-

/

,,,,-._

,

i ..

121

.z"

..,,

....

,

.-~

..

"--~" ~ " - ' ~ ' ~ -

u----'----~----~ ~

~'/~:'

~

O -

xq

: .......... ~ ......~

, ./~-.r

",,z

.... " ' "

......

.......

u_ D

7, A

_. ~ Km

,

_ _ - - ------._.~,,

Fig. 2. Map of active faults comprising the San Andreas fault zone in the southern part of the Big Bend. (See Figs i and 3 for location.) The map is bisected by the main fault trace (principal displacement zone) which, together with the flanking subsidiary faults, makes up a zone of active faulting up to 10 km wide. Conventional symbols are used to show thrust faulting, sense of strike-slip movement and relatively upthrown and downthrown sides of faults with significant components of dip-slip movement. Outlined areas in the southern half of the fault zone represent remnants of a prominent fault-disrupted erosion surface developed on ridge crests, and an arrow labelled '4' identifies the viewpoint for the low-altitude oblique aerial photograph in Fig. 4. The high-altitude oblique aerial photograph in Fig. 3 was taken from a viewpoint to the west of the map. Inset diagram shows Riedel shears (R, R 1 and P) and the principal displacement zone (PDZ), orientated to correspond to the main San Andreas fault trace os the map, predicted by theoretical studies of right-lateral simple shear (modified from Woodcock & Schubert (1994)) 9(b is the angle of internal friction, commonly about 30~ (Tchalenko & Ambraseys 1970), of the deformed material.

decreasing dip and ultimately defines the western end of the most uplifted section of the ridge as a thrust dipping generally away from the main fault trace (Fig. 4). The vertical component of offset on this fault is recorded by a summit erosion surface preserved along the full length of the ridge crest and is c. 260 m. This planar surface (Fig. 4), developed across the Mesozoic granitic basement complex rocks making up the ridge, is underlain by a thick clay-rich regolith which supports a distinctive vegetation assemblage of relatively luxuriant grasses with scattered mature oaks at higher elevations. This assemblage contrasts sharply with the brush vegetation on the steep and rapidly eroding slopes bordering the surface, allowing easy recognition on infra-red aerial photographs in particular. The age of the surface is u n k n o w n but, if the suggestion (Crowel11979) is correct that this part of the San Andreas system was initiated after the opening of the Gulf of California, and that San Andreas displacement was previously accommodated along the neighbouring San Gabriel fault to the

south (Fig. 1), the surface probably pre-dates faulting in this area. The most uplifted section of the ridge is bisected by a prominent cuvilinear right obliqueslip fault also following an approximate E - W strike (Fig. 2). The component of reverse slip on this fault produces a vertical offset in the erosion surface of about 120 m, as shown in Fig. 4, and the fault then continues westwards to terminate against another prominent active fault which defines the southern margin of the ridge and has a strike which may reflect an R-shear affinity (Fig. 2). At its eastern end this fault displays right-lateral displacement in accommodating eastwards upper-plate movement on a prominent thrust striking at a high angle to the main fault trace and marking the eastern end of the ridge (Fig. 2). By contrast, at the western end of this section of the ridge, where upper-plate thrusting is westwards, the fault exhibits leftlateral displacement. Between these two opposing displacement directions dip-slip movement occurs on this fault as the ridge is uplifted.

122

D. RUST

Fig. 3. High-altitude oblique aerial photograph overlooking the junction of the Garlock and San Andreas faults which define the delta-shaped tip of the westernmost Mojave Desert (Fig. 1). View looking ENE close to the strike of the Garlock fault, which forms the linear mountain front between the desert and the mountains on the left of the view. The San Andreas fault strikes from near the lower left corner of the photograph towards the mid-upper right margin of the view (between the two arrowheads at the margin of the photograph). The linear mountain ridge being actively uplifted within a contractional strike-slip duplex (Fig. 2), and described in the text, can be seen to the right (south) of the San Andreas fault in the right middle ground of the photograph. The curving fault which progressively takes on a dominant thrust component as it defines the western end of the most uplifted section of this ridge (Figs 2 and 4) can be recognized, as can the viewpoint for Fig. 4, approximately 180~ from the present viewing direction and at a much lower altitude. Together, these faults can be r e g a r d e d as forming a strike-slip d u p l e x structure with the faultb o u n d e d ridge comprising a series of horses (Woodcock & Fischer 1986). Moreover, adjacent to the northwest side of this feature, additional curvilinear faults, also exhibiting a progressively increasing reverse c o m p o n e n t of slip, define a continuation of the ridge (Figs 2 4 ) . A large landslide, originating near the crest of this part of

the ridge, has m o v e d downslope across one of these additional curvilinear faults and, at the foot of the ridge, across the adjacent main fault trace. This landslide produces a time line docum e n t i n g an approximate 1 : 3 ratio of slip on the subsidiary fault in relation to the m a i n fault trace (Rust 1993), establishing the activity and importance of the curvilinear faults in this part of the duplex. These faults probably terminate against

CONTRACTIONAL AND EXTENSIONAL STRUCTURES

123

Fig. 4. Low-altitude aerial photograph looking WSW along strike of the curvilinear oblique-slip fault, identified at the bottom margin of the photograph by an arrowhead labelled '1', which bisects the most prominent part of the mountain ridge uplifted within a contractional strike-slip duplex in the southern half of the San Andreas fault zone. The crestal areas of the ridge preserve an erosion surface, recognizable by its distinctive smooth grassy slopes in comparison with the steep brush-covered flanks of the ridge, which records a dip-slip component of offset on this fault of c. 120 m. (See Fig. 2 for map of the duplex and erosion surface, and the camera viewpoint.) Farther to the right (north) in the photograph the foot of this most uplifted section of the ridge is defined by another active fault which curves away from the viewer. As it curves away the dip of this initially vertical fault progressively lessens, becoming a thrust where it marks the western end of this ridge section. Beyond this the ridge continues, and the smooth grass-covered erosion surface capping it can be detected (particularly with reference to Figs 2 and 3). This offset surface records a vertical component of displacement on the thrust of about 260 m. Similar curving faults associated with the duplex bound the northern slopes (near the right-hand edge of the photograph) of this continuation of the ridge. The main San Andreas fault trace itself (Fig. 2), which is located between the two marginal arrowheads each labelled '2', is marked here by prominent scarps in the apron of alluvial debris shed from the actively uplifting ridge.

a continuation of the linear fault which forms the southern boundary of the ridge (Fig. 2), although continuity of surface outcrop is disrupted by downslope movements. Overall, the duplex appears to be actively propagating nortwestwards into the Big Bend (Fig. 1). The remainder of this half of the fault zone is dominated by a fault some 8 k m in length which juxtaposes basement complex rocks and very recent fanglomerate deposits (Rust 1993). This fault, which also disrupts the erosion surface, displays active displacement in a synthetic sense, and a strike which suggests an R-shear affinity (Fig. 2).

Fault z o n e north o f the main fault trace The most apparent structural contrast here is the much greater n u m b e r of faults in comparison with the opposing half of the fault z o n e described above (Fig. 2). In addition, the strike of the faults (Fig. 5) suggests that this half of the fault zone is dominated by P-shears, although left-lateral faults with probable Rl-shear affinity represent a significant subset in the distribution. The best example of this subset is a fault some 5 km long which occurs near the eastern margin of the mapped area, diverging from the strike of the main fault trace by 40-45 ~ (Fig. 2). Faults

124

D. RUST canyons are alluviated. These contrasts can be appreciated on the high-altitude oblique aerial photograph, which also allows a comparison of the width and elevation of the bedrock ridges that border each side of the linear trough following the main fault trace (Fig. 3). These bedrock ridges are the result of subsidiary faulting activity and effectively document the width of the fault zone in the area studied. This bordering strip of bedrock is particularly noticeable on the north side of the fault zone in Fig. 3, and can be seen to progressively diminish in width and elevation in a southeasterly direction away from the camera viewpoint and out of the Big Bend (Fig. 1).

020o

1 km

Fig. 5. Rose diagram illustrating the strike and length of subsidiary faults in the north half of the fault zone shown in Fig. 2. The vertical line trending 020~is perpendicular to the overall strike of the main fault trace in Fig. 2. Consequently, although the main fault trace itself is not plotted, the strike of subsidiary faults in relation to it can be seen as angular variations from a horizontal line on the diagram. Faults which continue beyond the boundaries of Fig. 2 are excluded, as are inferred faults fully concealed beneath alluvium. An overall strike for curved faults was obtained by concentrating on reaches where the fault is clearly defined; curvature near fault terminations was disregarded. The plot clearly indicates that this half of the fault zone is dominated by faults with a P-shear orientation, although a significant subset of the distribution is made up of left-lateral faults with probable Rl-shear affinity striking at 160-165 ~ in the lower quadrant of the plot (Fig. 2). with possible R-shear affinity are almost absent. The apparent P-shears are generally arrayed in a releasing sense of stepover or jog (Segall & Pollard 1980; Sibson 1985), as exemplified by a well-developed pull-apart basin in one particularly prominent stepover. Overall, the north half of the fault zone is undergoing crustal extension and subsidence, in marked contrast to the pronounced contraction and uplift in the adjacent southern half of the fault zone. This observation is also supported by sedimentation and elevation patterns: the high-elevation south half of the fault zone exhibits marked erosion and canyon incision, wheres the relatively low-elevation north half is dominated by deposition and the

Discussion The structures described above demonstrate a marked asymmetry in the San Andreas fault zone in the southern part of the Big Bend, a result unexpected in the context of the broadly symmetrical patterns produced in model studies of convergent strike-slip faulting. The southern half of the fault zone is characterized by contraction and uplift, and the overall strike of the curvilinear compressional faults corresponds to that predicted for reverse faulting under transpressional conditions (Sanderson & Marchini 1984). The small n u m b e r of faults, and the apparent importance of R-shears (which form early in model experiments of strike-slip fault development), suggests that this part of the fault zone is being actively d e f o r m e d by newly formed faults created as the bedrock terrain bordering the San Andreas enters the curved reach and moves northwestwards towards the most restraining part of the Big Bend. This notion is consistent with evidence for a contractional strike-slip duplex structure actively propagating in the same direction. It is noteworthy that thrusting in the duplex is directed partly towards the main fault trace, in disagreement with flower structure predictions, and it may be that the propagating fault will progressively assume increasing importance in accommodating San Andreas displacement in the Big Bend. U n d e r this proposal, strain partitioning between the main fault trace and the bordering subsidiary faults will vary in both time and space as a particular part of the bedrock terrain adjacent to the San Andreas fault moves through the Big Bend. The adjoining northern half of the fault zone displays a significantly different structural style, particularly in the number, apparent affinity and overall extensional character of the faults. Pshears are not inconsistent with transpressional

CONTRACTIONAL AND EXTENSIONAL STRUCTURES settings, as pointed out by Tikoff & Teyssier (1992) in their study of pluton emplacement within dilational jogs between P-shears under convergent strike-slip conditions. Similarly, R 1shears are predicted as late-developing faults by theoretical (Sanderson & Marchini 1984) and model (Naylor et al. 1986) studies of transpressional conditions. However, these studies indicate clockwise rotation of the long axis of the horizontal strain ellipse under transpression, and consequently the prominent left-lateral subsidiary fault with probable Rl-shear affinity discussed above (which exhibits a broad sigmoidal curvature consistent with clockwise rotation) would be expected to diverge from the main fault trace at a much higher angle (up to 90~ than the observed 40-45 ~ Nevertheless, a number of faults with apparent P-shear affinity abut this particular left-lateral fault (Fig. 2), and it seems possible that this fault formed after this part of the fault zone had moved through the most restraining part of the Big Bend farther to the northwest. That is to say, it may have formed as the intensity of transpressive conditions was relaxing. These considerations, together with the essential absence of faults with possible R-shear affinity, suggest that the northern half of the fault zone is more 'evolved' in terms of strikeslip faulting development than the adjacent part of the zone to the south. It may be that this structural arrangement provides the counterpart to that proposed for the southern part of the zone, in that the bedrock terrain bordering the San Andreas fault to the north is leaving the most restraining part of the Big Bend, effectively in mirror image to the terrain bordering to the south which is entering the Big Bend. In this scenario, early formed structures within a wide but relatively simple belt of faulting, such as now displayed in the southern half of the zone, would become progressively overprinted and replaced by a succession of later-forming active structures such as P-shears and Ra-shears, and increasing contraction and uplift associated with approach to the restraining bend would be replaced by extension and subsidence as the terrain moves away from the bend. This picture is also likely to involve abandonment and reactivation of early formed structures and produce a wide, highly faulted and complex terrain such as now makes up the northern half of the fault zone in the area studied (Fig. 2). Because it shows a significantly wider zone of faulting in the Big Bend in comparison with the rest of the San Andreas, the fault map compilation carried out by Moore & Byerlee (1991) is also consistent with this evolutionary picture of fault zone development.

125

Condu~on Points arising from this study which may have general applicability in explaining geological relationships in similar settings are as follows: (1) Structural asymmetry of major strike-slip fault zones in transpressional and, by analogy, possibly in transtensional, situations is to be expected, despite the generally symmetrical structural patterns predicted by theoretical and model studies. (2) The zone of faulting in these situations is likely to evolve through space and time in a succession of structures, and involve abandonment and reactivation of early formed structures. (3) Although it is essential to consider the entire fault zone in assessing the activity level of major faults (Rust 1993), this study indicates that the two halves of the zone are, structurally, remarkably discrete. (4)The work highlights the need for more detailed fieldbased studies to supplement and control the results of theoretical and model studies. Fieldwork on which this paper is based was carried out under US Geological Survey Contracts 14-08-0001-G395 and 14-08-001-18200.

References ALLEN, C. R. 1968. The tectonic environments of seis-

mically active and inactive areas along the San Andreas fault system. In: DICKINSON,W. R. & GRANTZ,A. (eds) Proceedings of the Conference on Geologic Problems of the San Andreas Fault System. Stanford University Publications in the Geological Sciences, 11, 70-82. BARTLETT,W. L., FRIEDMAN,M. & LOGAN,J. M. 1981.

Experimental folding and faulting of rocks under confining pressure, Part IX. Wrench faults in limestone layers. Tectonics, 79, 255-277. BOHANNON, R. G. & HOWELL, D. G. 1982. Kinematic

evolution of the junction of the San Andreas, Big Pine, and Garlock faults, California. Geology, 10, 358-363. CLARK, M. M. 1968. Surface Rupture along the Coyote Creek Fault, the Borrego Mountain Earthquake of April 9, 1968. US Geological Survey Professional Paper, 787, 55-86. CROWELL,J. C. 1974. Origin of Late Cenozoic basins in southern California. In: DICKINSON,W. R. (ed.) Tectonics and Sedimentation. Society of Economic Paleontologists and Mineralogists Special Publication, 22, 190-204. -1979. The San Andreas fault system through time. Journal of the Geological Society, London, 136, 293-302. DAVIS, G. A. & BURCHFIEL,B. C. 1973. Garlock fault -

an intracontinental transform structure, southern California. Geological Society of America Bulletin, 84, 1407-1422. LOWELL,J. D. 1972. Spitsbergen Tertiary orogenic belt and the Spitsbergen fracture zone. Geological Society of America Bulletin, 83, 3091-3102.

126

D. RUST

MAY, S. R., EHMAN, K. D., GRAY, G. G. & CROWELL,J. C. 1993. A new angle on the tectonic evolution of the Ridge basin, a 'strike-slip' basin in southern California. Geological Society of America Bulletin, 105, 1357-1372. MOORE, D. E. & BYERLEE, J. D. 1991. Comparative geometry of the San Andreas fault, California, and laboratory fault zones. Geological Society of America Bulletin, 103, 762-774. NAYLOR, M. A., MANDL, G. t~ SIJPESTEIJN, C. H. K. 1986. Fault geometries in basement-induced wrench faulting under different initial stress states. Journal of Structural Geology, 8, 737-752. RUST, D. J. 1993. Recognition and assessment of faults within active strike-slip fault zones: a case study from the San Andreas fault in southern California. In: STEWART,I., VrrA-FtNZL C. & OWEN, L. (eds) Neotectonics and Active Faulting. Zeitschrift fiir Geomorphologie, Supplement-Band, 94, 207-222. SANDERSON, D. J. t~ MARCHINI,W. R. D. 1984. Transpression. Journal of Structural Geology, 6, 449-458. SEGALL, P. & POLLARD,D. D. 1980. Mechanics of discontinuous faults. Journal of Geophysical Research, 85, 4337-4350. SmsoN, R. H. 1985. Stopping of earthquake ruptures at dilational fault jogs. Nature, 316, 248-251.

SYLVESTER, A. G. 1988. Strike-slip faults. Geological Society of America Bulletin, 100, 1666-1703. -& SMrrH, R. R. 1976. Tectonic transpression and basement-controlled deformation in the San Andreas fault zone, Salton Trough, California.

Bulletin, American Association of Petroleum Geologists, 60, 2081-2102. TCHALENKO,J. S. & AMBRASEYS,N. N. 1970. Structural analysis of the Dasht-e-Bayaz (Iran) earthquake fractures. Geological Society of America Bulletin, 81, 41~50. TIKOFF, B. & TEYSSIER, C. 1992. Crustal-scale, en echelon P-shear tensional bridges: a possible solution to the batholithic room problem. Geology, 20, 927-930. WESTAWAY,R. 1994. Deformation around stepovers in strike-slip fault zones. Journal of Structural Geology, 17(6), 831-846. WILCOX, R. E., HARDING, Z. P. & SEELY, D. R. 1973. Basic wrench tectonics. Bulletin, American Association of Petroleum Geologists, 57, 74--96. WOODCOCK, N. H. & FISCHER, M. 1986. Strike-slip duplexes. Journal of Structural Geology, 8, 725-735. SCHUBERT,C. 1994. Continental strike-slip tectonics. In: HANCOCK, P. L. (ed.) Continental Deformation. Pergamon Press, Oxford, 251-263.

Salar Grande pull-apart basin, Atacama Fault System, northern Chile JURRIAAN

REIJS* & KEN MCCLAY

Fault Dynamics Project, Geology Department, Royal Holloway, University o f London, Egham TW20 OEX, UK *Present address: EPT-SG, Shell International Exploration & Production B. V., Research & Technology Services, PO Box 60, 2280 A B Rijswijk, The Netherlands

Abstract:The Salar Grande basin is a 'poly-history'strike-slip basin along the north-trending Atacama Fault System in northern Chile (21~ 70~ It exhibits a range of structures that indicate a complex evolution with multiple deformation styles. This paper presents a new interpretation of Salar Grande as a strike-slip pull-apart basin with three stages of evolution. In the first stage (Oligocene to Early Miocene), the basin opened as a sinistral strikeslip pull-apart structure at a left-stepping releasing jog of the Atacama Fault System. At the end of this stage, the basin had a rhomboidal geometry with dimensions of about 35 km length and 10 km width, bounded by a series of linked oblique-extensional basin-sidewall faults. The western sidewall faults linked to the northern fault strand of the Atacama Fault System whereas the eastern sidewall faults linked to the southern strand. The basin interior was a series of tilted fault blocks forming half-grabens. During the second stage (Oligocene to Early Miocene) a NNW-SSE basin short-cut fault developed, dissecting the basin with 6 km of sinistral offset. In the third stage (Pleistocene-Recent), NNW-SSE strike-slip faults were dextrally reactivated with displacements of up to 1 km. The understanding of field examples such as Salar Grande provides new insights into the geometries and evolution of complex multi-phase strike-slip terranes. The geometries recognized in Salar Grande compare well with analogue modelling studies of strike-slip fault systems. Salar Grande is located at approximately 21~ 70~ along the northern end of the Salar del Carmen segment of the Atacama Fault System in the Cordillera de la Costa of northern Chile (Fig. 1). The ranges of the Cordillera de la Costa form an elevated plateau, separated from the Pacific shore on the west by the 900 m high scarp of the Coastal Scarp Fault System (Fig. 1). Salar Grande is a Recent saltpan (salar), c. 265 km 2 in area, with a N-S elongate shape. It is 50 km long and 5-8 km wide, and is bounded by major N-S strike-slip faults. The basin fill consists of a minimum of 162 m of massive rock salt (Ericksen 1993) and the total basin depth is unknown. This paper presents a case study that aims to provide new insights into the evolution and geometry of complex multi-phase strike-slip fault systems. The Salar Grande strike-slip basin is an example of a pull-apart basin that has undergone a long and complex, polyphase strike-slip history. Field studies, L A N D S A T TM and aerial photograph analyses permit a new model for the structural evolution of the Salar Grande basin to be postulated.

Regional geology Salar Grande is located in the Cordillera de la Costa in northern Chile, which is mainly

composed of Lower Jurassic to Lower Cretaceous arc-related rocks (Fig. 1). Three other arc systems were also developed in northern Chile (Fig. 1), related to subduction episodes of the A n d e a n Orogenic Cycle (Coira et al. 1982; Schreuber & Reutter 1992). From west to east these are: the mid-Cretaceous arc in the Longitudinal Valley (Fig. 1), the Late Cretaceous to Paleogene arc in the Chilean Precordillera, and the Early Miocene-Recent arc in the western Cordillera (Schreuber & Reutter 1992). In the Cordillera de la Costa, Jurassic La Negra Formation andesites and coastal batholith granodiorites document the first subduction episode (Schreuber & Reutter 1992). In the Eocene, the Incaic orogenic phase uplifted the Cordillera de la Costa on a deepseated thrust system (Buddin et al. 1993). The Cordillera de la Costa was subsequently levelled in the Oligocene-Early Miocene to form a peneplain (A. Tomlinson, pers. comm. 1997). The subsequent Quechua orogenic phase (Buddin et al. 1993) occurred at the M i o c e n e - P l i o c e n e b o u n d a r y to form the Coastal Scarp Fault System (CSFS), the present western boundary of the plateau (Santanach et al. 1996). The oldest strike-slip fault system in the Cordillera de la Costa is the Atacama Fault System (AFS), which developed in the Early

REIJS,J. & MCCLAY,K. 1998. Salar Grande pull-apart basin, Atacama Fault System, northern Chile

In: HOLDSWORTH,R. E., STRACHAN,R. A.& DEWEY,J. E (eds) 1998. Continental Transpressionaland Transtensional Tectonics. Geological Society, London, Special Publications, 135, 127-141.

127

128

J. REIJS & K. McCLAY

Fig. 1. (a) Regional map of the Atacama Fault System (modified after Schreuber & Andriessen (1990)). (b) Schematic regional fault map indicating the main traces of the northern portion of the Atacama Fault System (i.e. Salar del Carmen segment), interpreted from LANDSAT TM image 001/75. Cretaceous as a ductile, sinistral, trench-linked, strike-slip fault system (Taylor et al. 1996). The steeply dipping north-trending AFS is more than 1000 km long, stretching from Salar Grande (20~ to north of La Serena (29~ and

accommodates the oblique subduction of the Nazca plate at the South A m e r i c a n active margin (Naranjo et al. 1984; Thiele & Pincheira 1987; Reutter & Schreuber 1988; Schreuber & Andriessen 1990; Brown et al. 1991, 1993;

SALAR GRANDE PULL-APART BASIN Schreuber & Reutter 1992). The AFS consists of three arcuate, concave to the west, fault segments (Fig. 1), from south to north: the E1 Salado, Paposo and the Salar del Carmen segments, respectively (e.g. Schreuber & Andriessen 1990; Brown et al. 1991, 1993). The Mesozoic AFS

4~ dating shows that sinistral ductile motion on the AFS was Early Cretaceous in age and associated with granitic plutonism between 132 and 126 Ma, and that motion may have continued until 106 Ma (Randall et al. 1996; Dallmeyer et al. 1996; Taylor et al. 1996). Herr6 (1987b) reported a sinistral ductile displacement of 34 km between 144 and 131 Ma ( K - A r dating). Transitions from ductile to brittle sinistral strike-slip have occurred in space and time around plutonic complexes (Brown et al. 1993). They occur at a given distance from the pluton when faults cut into cooler country rock rather than hot plutons, and later within the intrusive rocks when they cooled down during the abandonment of the Lower Cretaceous magmatic arc (Brown et al. 1993). In the E1 Salado segment, this brittle deformation is expressed by large sidewall rip-out structures, some of which control mineralization (Brown etal. 1993). K - A r dating of mineralized samples associated with the brittle strike-slip phase indicate a mid-Cretaceous age (Arabasz 1971; Zentilli 1974; Colley et al. 1990). Northwest-trending, brittle, sinistral strikeslip strands of the AFS have displacements of up to 70 km and, in places, offset the main northtrending strands of the AFS (Randall et al. 1996). These faults play an important role in recent discussions about the origin of large, mid-Late Cretaceous, clockwise palaeomagnetic rotations of the area (e.g. Randall et al. 1996). The Cenozoic AFS

The Cenozoic AFS is described in the literature as a strike-slip fault with a minor displacement (St Armand & Allen 1960; Herv6 1987a; Armijo & Thiele 1990; Dewey & Lamb 1992; Buddin et al. 1993). The Cenozoic movement history of the AFS is controversial, with reports of both Recent dextral (Herv6 1987a; Dewey & Lamb 1992; Buddin et al. 1993) and sinistral movements (St Armand & Allen 1960; Armijo & Thiele 1990). Plate reconstructions show a dextral component to oblique plate subduction during the

129

Tertiary (Pardo-Casas & Molnar 1987; Buddin et aL 1993). Tomlinson et al. (1993) reported a dominance of sinistral strike-slip fault systems in the Precordillera in Tertiary time, when subduction is known to have had a dextral component. This sinistral slip in an apparent dextral regime may have been the result of a reduced rate of plate convergence together with a possible reduced tectonic coupling, as Reutter et al. (1996) suggested for Tertiary sinistral motions along the Precordillera Fault System (PFS). In situ stress measurements around Salar Grande show that the present regime favours dextral shear (Sch~ifer & Dannapfel 1994). A magnitude 7.5 earthquake in Taltal in 1966 indicates that at least parts of the AFS are still active (Brown et al. 1991).

Geology of the Salar Grande area Figure 2 shows the stratigraphy of the Salar Grande area, and Fig. 3 the distribution of the main stratigraphic units. In the Early to Middle Jurassic, andesites of the La Negra Formation were erupted with an interval of limestone deposits in the Pliensbachian (see section below), and followed by intrusion of the coastal batholith in the Late Jurassic-Early Cretaceous. Continental red-beds were deposited in the Early Cretaceous. The Oligocene to Pleistocene Soledad Formation of rock salt fills the Salar Grande basin, and alluvium was deposited from Miocene to Quaternary. La Negra Formation

This formation consists of Lower to Middle Jurassic brownish red, grey to dark grey porphyritic pyroxene andesites and basaltic andesites (Maksaev & Marinovic 1980; Skarmeta & Marinovic 1981). The total minimum thickness has been estimated at 3800 m, including marine sedimentary intercalations (Maksaev & Marinovic 1980; Skarmeta & Marinovic 1981). Ages for the La Negra Formation range from Sinemurian to Kimmeridgian (Naranjo 1978; Rogers 1983b; Rogers & Hawkesworth 1989; Dallmeyer et al. 1996). Pliensbachian limestones

Reefal limestones are found on N-S-trending ridges or down-faulted blocks on the northwestern margin of Salar Grande (Figs 2 and 3). They are grey-green to brownish, and rarely purple, and occur in 10-20 cm thick beds. They are fossiliferous, with a shallow marine fauna of

130

J. REIJS & K. M c C L A Y

Miocene to Quaternary alluvium Gravels and sands reactivation: P/eistocene to Recent Soledad Formation Pure halite with locally thin anhydrite crust and sandstone at the base pull-apart & basin dissection: Oligocene to Early Miocene Continental red-beds Red sand and siltstones with conglomerate intervals Minimum thickness estimated 550 m

Coastal batholith granodiorites and diorites La Negra Formation Total thickness = 3800 m

Pliensbachian limestone intercalation Fossiliferous limestone with sand and siltstone intercalations Shallow marine fauna of bivalves, Cideroidea, corals (Isastrea) and ammonites (Uptonia jamesom) .

Andesites and basaltic andesites with limestone and sand intercalations

Alterations of chrysocolla, limonite and hematite

[

I Continental Red-beds Granodiorite

~

Pliensbachian Limestone

~

La Negra Formation

SALAR GRANDE PULL-APART BASIN

131

Fig. 3. Geological map of the Salar Grande area (see Fig. 1 for location) compiled from field data, LANDSAT TM and aerial photograph interpretations and previously published mapping by Skarmeta & Marinovic (1981). Borehole data from Ericksen (1993).

Fig. 2. Schematic stratigraphy of the Salar Grande area (not to scale). Compiled from field data, LANDSAT TM and aerial photograph interpretations and Skarmeta & Marinovic (1981) and Scanlan (1992).

132

J. REIJS & K. McCLAY

ammonites, bivalves, echinoid spines (Cideroidea) and corals (Isastrea). The limestones are interbedded with thin layers of silt and mudstone and poorly consolidated, green, glauconitic, medium- to coarse-grained sandstones. The sandstones contain feldspars and black lithic fragments probably derived from the La Negra Formation. A minimum thickness of 650 m has been estimated for the Pliensbachian limestone sequence. An ammonite specimen identified as Uptonia jamesoni (J. Wright, pers. comm., 1996) was found, constraining this unit to an early Pliensbachian age (Harland et al. 1990). No definitive stratigraphic relationships between the limestones and the La Negra Formation are exposed in the area. However, La Negra Formation rocks have been dated as being both older and younger than Pliensbachian (Naranjo 1978; Rogers 1983b; Rogers & Hawkesworth 1989; Dallmeyer et al. 1996) and therefore the limestones are best interpreted as an intercalation within the La Negra Formation.

Coastal batholith This unit was intruded into the La Negra Formation in the western part of the northern Cordillera de la Costa (Figs 2 and 3) and consists mainly of diorite to granodiorite plutons (Maksaev & Marinovic 1980; Skarmeta & Marinovic 1981). The coastal batholith has been dated as Late Jurassic-Early Cretaceous, based on 4~ dates of 132-106 Ma (Dallmeyer et al. 1996), Rb-Sr whole-rock dates of 158-154 Ma (Rogers 1983a; Rogers & Hawkesworth 1989) and a K - A r date of 123 Ma (Alfaro 1972).

Continental red-beds Northwest of Salar Grande, c. 550 m of continental red-beds are preserved in fault-bounded blocks (Figs 2 and 3). These fluvial sedimentary rocks consist of interbedded, well-sorted and sub-rounded, fine-grained sandstones to coarsegrained siltstones, poorly sorted immature sandstones, conglomerates and debris flow deposits. Some lithic fragments have been identified as andesite and granodiorite derived from the La Negra Formation and the coastal batholith. The age of this unit is therefore post-batholith (Late Jurassic-Early Cretaceous), with a possible age of Early Cretaceous (based on correlations with the Atajafia Formation; Scanlan (1992) and references therein).

Soledad Formation The present-day Salar Grande basin is filled with the Soledad Formation (Figs 2 and 3) which includes rock salt, gypsum and anhydrite deposits (Bobenrieth 1979; Maksaev & Marinovic 1980; Skarmeta & Marinovic 1981). Three drill holes have shown that the rock salt continues to depths of as much as 162 m, and is underlain by dry gravels and alluvium (Stoertz & Ericksen 1974; Ericksen & Salas 1989; Ericksen 1993). The age of this formation may range from Oligocene(?) to Pleistocene (Jensen et al. 1995; Santanach, et al. 1996).

Miocene to Quaternary alluvium Alluvium up to 80 m thick is found throughout the area. It consists of gravels and sand derived from erosion of surrounding units. This unit was assigned to the Quaternary by Skarmeta & Marinovic (1981) and to the Plio-Quaternary by Buddin et al. (1993). However, SERNAG E O M I N work indicates that much of the alluvium is intercalated with Lower to Middle Miocene ashes (A. Tomlinson, pets. comm., 1997), making part of this unit Lower-Middle Miocene in age.

Structure The Salar Grande area is still tectonically active. In the alluvium, several recent ground ruptures are found in the form of small scarps, in-line sag basins (up to 3 m deep) and faultoverstep or fault-bend features such as push-up swells (up to 1 m of uplift) and a small-scale pull-apart basin (40 cm deep). The ground surface of the salar shows that the theologically mobile salt of the Soledad Formation is characterized by scarps (up to 10 m), sag basins (up to 3 m deep) and push-up swells (up to 10 m uplift), indicating very recent deformation.

Geometry o f the Salar Grande Basin Salar Grande has an elongate shape of 50 km length and 5-8 km width with the long axis in a north-south direction (Fig. 3). The southern strand of the Atacama Fault System bounds the basin on the east and the northern strand finks to the western sidewall faults (Fig. 4). The basin bounding faults are linked oblique-extensional basin-sidewall faults. The basin interior is composed of a series of mainly west-tilting fault blocks, forming half-grabens (Fig. 5).

SALAR GRANDE PULL-APART BASIN

133

Fig. 4. Fault map of the Salar Grande area (see Fig. 1 for location). The surface of Salar Grande varies in altitude from 642 to 750 m, which is c. 950 m below some of the surrounding peaks, both west and east of the basin. This, together with a minimum of 162 m of the basin fill from borehole data (Fig. 3;

Ericksen 1993), means that the basin floor (top La Negra Formation) must have subsided at least 1.1 km below the surrounding peaks after the Oligocene peneplainization (A. Tomlinson, pers. comm. 1997) of the coastal plateau.

134

J. REIJS & K. McCLAY

A

A'

km

SGF

1

(~ (~

Salar

- -

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

-._ "~"~"~"| --

.

.

.

.

.

0 ~__~_

.

.

.

.

.

.

.

.

i

k

\\

B

70000'

\

\

J

Km

~,,,.

\

\\

.

~ .

SGF \ ~

-..

""~

\

~~~.r

'~

B' ~

"~ "" -... / /

:::::~':::~':::.

~::7.:::'C.~'.:~"

~

70600' "-..., "--

/ ~

/I

~

~

"~ ~

~

70600' \

\

~]

++

?

? ?

I

I

\ (~

E

~

~

\

(~) SalarGrande

?

SGF (~

r

I

~_~2~;~'~j,v~,~.~. ++

++

+ + ~..~--~_X--_"F_~3~]~SS"~2r'~-'r-~_t"~;J"/~X_

Quaternary alluvium

70000'

..~,,,,~-

~,r

,,-

- , ) 1

~

_

~

_

~

_

~

_

~

_ ~

from viewer dextral reactivation (2~x~ towards & away from viewer movement during sinistral phase

Coastal batholith La Negra Formation

-

9

i

sinistral movement

Continental red-beds F-~

-

7

@@ towards & away

Soledad Formation ~ E ~ Pliensbachian limestone

i

I

\W

\ \

++

~

I \

\

,

-%----%~

I~+

:

C'

~ ~

\\

21 I Pacific 1 1 Ocean

~

] ~ ) /

D km

1 km H=V

(~

~-SalarGran~e~

\

~

//

?

\

~

/

/ /

km

~

/"

"-C

LI i_

i

SalarGrande 9; ~ - . . . . . . . . . . . . . . .

1t

J

;..-..-. ;. ;. ;..-..-..-..-. ;[-.~1~...~....~'Ll \

1t

.

movement during Jv dextral phase

A B c

D

Fig. 5. East-west cross-sections through the Salar Grande strike-slip basin.

Faults cut through the La Negra andesite, coastal batholith and continental red-beds, indicating that at least part of the displacement postdates the Early Cretaceous. Santanach et al.

(1996) reported evidence from the sedimentary record in the Rio Loa canyon (Fig. 1) for an Oligocene to Early Miocene deformation event which formed the Salar Grande basin.

SALAR GRANDE PULL-APART BASIN

Salar Grande Fault The Salar Grande basin is cross-cut by the Salar Grande Fault (hereafter SGF; Fig. 4). In the salar, the SGF consists of active fault traces that produce monoclines (up to 10 m in height), sag basins of several metres in depth and push-up swells at overstepping fault strands. The most prominent structure is a c. 10 m high scarpbounded uplift zone of c. 14 km 2 in the basin fill along a left bend of the SGF (Figs 3 and 4). In cross-section (Fig. 5), this uplift zone is interpreted as a pop-up along a dextral restraining bend of the SGF. Stream cuts into the alluvium are dextrally offset by c. 300 m in the northern part of the SGF. Along the western border fault of Salar Grande a sidewall rip-out with sinistral asymmetry (Swanson 1989) has formed west of the SGF (Figs 4 and 7, below). A lineament recognized on LANDSAT TM imagery can be interpreted as an old trace of the SGF, which continues in a SSE direction from the southeastern corner of Salar Grande, instead of bending to the south (Figs 4 and 5). This suggests that the SGF was previously a continuous S S E - N N W fault. From maps and the L A N D S A T TM image (Figs 3 and 4), the portion of the Salar Grande basin south of the SGF appears sinistrally offset with respect to the northern block. Reconstruction of possible movements of the early SGF with the LANDSAT TM image (see Fig. 8, below) yields a perfect match of the northern and southern block and three N-S faults, when assuming 6 km sinistral displacement. This reconstruction supports the interpretation that the basin was dissected by this SSE-NNW fault. The SGF dies out in a west-bending horsetail splay of (oblique)-extensional faults (Figs 3 and 4) northwest of Salar Grande, which also reflect a sinistral displacement history.

Eastern normal faults and Cerro Chuculay Fault East of Salar Grande, a number of prominent E-W-striking extensional faults can be identified (Figs 3 and 4). These faults dip towards north with scarps of up to 300 m and produce south-tilted domino-style fault blocks in their hanging walls. These scarps have formed in the Miocene to Quaternary alluvial cover and are therefore Tertiary, probably post-Miocene in age. Parallel and adjacent to these faults granodiorite dykes are found, suggesting there was an earlier N-S extensional event during the Late Jurassic-Early Cretaceous coastal plutonism.

135

The Cerro Chuculay Fault displaces both the eastern normal faults and the Salar de Llamara Border Fault dextrally by c. 1 km (Figs 4 and 6). The Cerro Chuculay Fault is parallel to the active trace of the SGF and probably formed at the same time. This overprinting relationship shows that the eastern normal faults were active before the Recent dextral NNW-SSE faulting phase.

Rio Seco Fault The Rio Seco Fault was identified on LANDSAT TM imagery as a sinistral strike-slip fault with an offset of 750 m (Fig. 7). This fault most probably formed directly after basin dissection, cutting off the original western border faults.

Discussion Figure 9 summarizes the complex fault history of the Salar Grande strike-slip basin as determined in this study. The first phase of sinistral strikeslip faulting formed the Salar Grande rhomboidal pull-apart basin c. 35 km long and c. 10 km wide in a left step in the Atacama Fault System. Vertical offset of the coastal peneplain records a subsidence of the basin floor of at least 1.1 km since the Oligocene. Santanach et al. (1996) found evidence for an Oligocene-Early Miocene formation of the Salar Grande basin. The minimum amount of strike-slip displacement necessary to create a pull-apart of the dimensions of Salar Grande has been estimated at 3 km. The basin interior consists of tilted fault blocks forming flat-bottomed half-grabens. From the westward tilting of most of the fault blocks, it is concluded that western border faults were dominant in basin formation. Tilted fault blocks and half graben geometries have also been recognized in analogue modelling studies of pull-apart basins (Dooley 1994; McClay & Dooley 1995; Dooley & McClay 1997). During basin formation the basin was largely starved or under-filled. The basin fill consists of at least 162 m of Soledad Formation which is partly syntectonic. As deformation progressed, a basin short-cut fault formed that connected the offset principal displacement zones of the AFS. Analogue modelling studies demonstrate that basin shortcut faults develop typically at a late stage of pullapart formation (Dooley 1994; McClay & Dooley 1995; Dooley & McClay 1997). However, because of a change in tectonic regime, the short-cut fault changed its trace and

136

J. REIJS & K. McCLAY

Fig. 6. LANDSAT TM image of the ranges bordering Salar Grande in the east. The NNW-SSE-striking Cerro Chuculay Fault clearly offsets E-W-striking normal faults dextrally over 1 km. completely cross-cut the AFS and continued beyond the pull-apart basin. During this stage, displacement along the AFS was concentrated on the SGF offsetting the basin by 6 km (Fig. 8). The strike-slip movement along the dissected western border fault of Salar Grande was subsequently 'short cut' by the Rio Seco fault. Along this fault lithology markers appear to be sinistrally offset by 750 m (Fig. 7).

Oligocene-Miocene sinistral movements along the AFS are poorly documented. However, there are reports of a sinistral strikeslip movement phase along the parallel Precordillera Fault System (PFS) around Calama (Fig. 1) during the Early Miocene (25-17 Ma; May et al. 1996) and from 32 Ma to at least 24 Ma (Reutter et al. 1996). Jensen et al. (1995) reported Upper Eocene to Middle Miocene

SALAR GRANDE PULL-APART BASIN

137

Fig. 7. LANDSAT TM image of the ranges bordering Salar Grande in the west. The N-S-striking Rio Seco Fault shows sinistral offsets over c. 750 m of contacts. The offset markers on the LANDSAT TM image signify lithological boundaries, separating zones with different reflection colours within the La Negra Formation. East of the Rio Seco Fault, the Salar Grande fault produced a side-wall rip-out with sinistral asymmetry. formation of the Quillagua Trough, a continental pull-apart basin in the Longitudinal Valley (21~ formed as a result of local transtension between the AFS and PFS. Plate vector reconstructions for the subducting Nazca plate at the north Chilean active margin, however, seem to suggest a stress regime that favoured dextral faulting along N-S faults in the last 50 Ma (Pardo-Casas & Molnar 1987; Buddin et al. 1993). These reconstructions also show a decrease in the convergence rate and obliquity in the Oligocene-Early Miocene. This

time coincides with the initiation of the Salar Grande basin and the Quillagua Trough (Jensen et al. 1995), and with the reports of sinistral strike-slip motion on the PFS around Calama (Reutter et al. 1996; May et al. 1996). Sinistral strike-slip at this time may have resulted from this reduced convergence rate and angle, combined with a possible reduction in tectonic coupling (Reutter et al. 1996). After the Pleistocene, N N W - S S E faults were dextrally reactivated (SGF) and new N N W - S S E dextral faults formed (e.g. Cerro Chuculay

138

J. REIJS & K. McCLAY

Fig. 8. (a) LANDSAT TM reconstruction of the structural geometry of Salar Grande at Early Cretaceous times. (b) LANDSAT TM image with interpretation of Plio-Pleistocene to Recent structural geometry.

Fault). This Recent deformation produced active pressure ridges and sag basins in the rock salt of the Oligocene-Pleistocene Soledad Formation, and pull-aparts, scarps and stream offsets in the Miocene to Quaternary alluvium. Dextral strike-slip offsets of 300 m along the SGF and 1 km along the Cerro Chuculay Fault are recorded. In situ stress measurements in the Salar Grande area (Schiller & Dannapfe11994) reveal a current stress situation that favours dextral strike-slip movements along NNW-SSE faults. Plate reconstructions for the Miocene to Recent show an increased dextral component and an increased convergence rate of the subducting slab at the Chilean active margin (Pardo-Casas & Molnar 1987; Buddin et al. 1993). These changes may have triggered the switch from sinistral to dextral slip along the AFS. Using terminology of Nilsen & Sylvester (1995), the Salar Grande strike-slip basin can be classified as a 'poly-history' basin which has recorded pulses of varying structural styles.

Conclusions The Salar Grande poly-history strike-slip basin records three stages of deformation in different tectonic regimes. During the first stage (Oligocene-Early Miocene), at least 3 km of sinistral strike-slip along the Atacama Fault System opened the 35 km long and 10 km wide Salar Grande pull-apart basin. The basin consist of segmented, mostly west-tilted fault blocks that form flat-bottomed half-grabens. In the second stage of evolution (Early Miocene-?) of the Salar Grande pull-apart, a basin short-cut fault was formed (the Salar Grande Fault), across the overstepping fault strands. This NNW-SSE basin short-cut fault subsequently offset the Salar Grande basin over 6 km. In turn, the Rio Seco Fault formed to short cut the shear along the offset western basin border fault and has recorded c. 750 m of sinistral displacement. In the third stage (Pleistocene to Recent), the

SALAR GRANDE PULL-APART BASIN

139

Structural Evolution of the Salar Grande Strike-Slip Basin A

Oligocene-Early Miocene

Stage h Opening pull-apart basin I

3 km minimum sinistral strike-slip 1.1 km minimum subsidence

Development of basin short-cut fault

\

B

\ Early Miocene

Stage Ih Basin dissection Change of trace of basin short-cut fault -6 km sinistral strike-slip along the basin short-cut fault dissecting the basin into northern and southern fault blocks Rio Seco Fault active

C

Pleistocene to Recent

Stage II1: Dextral reactivation

-300 m dextral strike-slip Formation of pressure ridge in rock salt Cerro Chuculay Fault active

Fig. 9. Schematic summary of structural evolution of the Salar Grande poly-history strike-slip basin from the Oligocene-Early Miocene to Recent. N N W - S S E Salar Grande Fault and parallel Cerro Chuculay fault were dextrally reactivated with over 0.3 and 1 km of displacement, respectively. Changes in the subduction vector of the Nazca plate, timing and kinematics of AFS and PFS deformation, and in situ stress measurements all compare well with the proposed model of multistage fault evolution in the Salar Grande basin. However, it is difficult to link d e f o r m a t i o n events directly to plate tectonics, as rheological differences and topography may also have been important. The geometries of the basin interior of Salar Grande are similar to the features identified in analogue modelling studies of pullapart basins.

The Salar Grande basin is a good example of a poly-history strike-slip basin, which shows that along a trench-linked strike-slip fault system, with a long and complex history, early-formed pull-apart basin geometries may be difficult to recognize. The characteristics unravelled for this example may lead to the identification and understanding of new pull-apart basins in complex multi-phase strike-slip terranes. This research was supported by the Fault Dynamics Project (sponsored by Arco British Limited, Brasoil, UK Ltd, BP Exploration, Conoco (UK) Limited, Mobil North Sea Limited, and Sun Oil Britain). K. McC. also acknowledges funding from Arco British Limited. The authors express their gratitude to RTZ Mining and Exploration, for all their financial and

140

J. REIJS & K. McCLAY

logistic support of the field expeditions. The staff of RTZ's Antofagasta office were of tremendous help. In particular, we express our gratitude to D. Andrews for endorsing the project; A. Pope and L. Pope for their help and hospitality; F. Quijano for helpful discussions; and V. Ramos and D. Salinas-Hernandez for their assistance in the field. The manuscript has benefited greatly from reviews and suggestions by A. Tomlinson, J. Grocott and N. Mufioz. Fault Dynamics Publication No. 71.

References ALFARO, G. 1972. Situaci6n de la miner'a en la zona de Tocapilla. Instituto de Investigaciones Geologfcas, Santiago. ARABASZ, W. J., JR 1971. Geological and geophysical studies of the Atacama fault zone in northern Chile. PhD thesis, California Institute of Technology, Pasadena. ARMIJO, R. & THIELE, R. 1990. Active faulting in northern Chile: ramp stacking and lateral decoupling along a subduction plate boundary? Earth and Planetary Science Letters, 98, 40-61. BOBENRIETH, L. 1979. Geologia de los cuadrdngulos Cerro Desamparado y Cerro Soledad. Proyecto Cordillera de la Costa Tocopilla-Rio Loa. Instituto de Investigaciones Geol6gicas, Santiago. BROWN, M., DALLMEYER,R. D., DIAZ, E & GROCOTr, J. 1991. Displacement history and tectonic significance of the Atacama fault system (El Salado Segment), N Chile: 4~ mineral age constraints. Eos Transactions, American Geophysical Union, 72, 263. --, DIAZ, E & GROCOTr, J. 1993. Displacement history of the Atacama fault system 25~176 northern Chile. Geological Society of America Bulletin, 105, 1165-1174. BUDDIN, Z. S., STIMPSON,I. G. • WILLIAMS,G. D. 1993. North Chilean forearc tectonics and Cenozoic plate kinematics. Tectonophysics, 220, 193-203. COIRA, B., DAVIDSON, J., MPODOZIS, C. & RAMOS, V. 1982. Tectonic and magmatic evolution of the Andes of Northern Argentina and Chile. EarthScience Reviews, 18, 303-332. COLLEY, H., TREOLAR, P. J. & DIAZ, E 1990. Gold-silver mineralisation in the E1 Salvador region, northern Chile. In: KEAY,R. RAMSEY,R. & GROVES, D. (eds) The Geology of Gold Deposits: the Perspective in 1988, Economic Geology Monograph, 6, 208-217. DALLMEYER, R. D., BROWN, M., GROCOTr, J., TAYLOR, G. K. & TREOLAR, P. J. 1996. Mesozoic magmatic and tectonic events within the Andean plate boundary zone, 26~176 North Chile: constraints from 4~ mineral ages. Journal of Geology, 104, 19-40. DEWEY, J. E & LAMB,S. H. 1992. Active tectonics of the Andes. Tectonophysics, 205, 79-96. DOOLEY, T. P. 1994. Geometries and kinematics of strike-slip fault systems: insights from physical modelling and field studies. PhD thesis, University of London.

--

& MCCLAY,K. R. 1997. Analog modelling of pullapart basins. Bulletin, American Association of Petroleum Geologists, 81, 1804-1826. ERICKSEN, G.E. 1993. Upper Tertiary and Quaternary continental saline deposits in the Central Andean region. In: KIRKHAM, R. V., SINCLAIR W. D., THORPE, R. I. & DUKE, J. M. (eds) Mineral Deposit Modelling, Geological Society of Canada, Special Paper, 40, 89-102. -& SALAS, R. 1989. Geology and resources of Salars in the Central Andes. In: ERICKSEN,G. E., CAt'AS PINOCHET,M. T. & REINEMUND,J. A. (eds) Geology of the Andes and its' Relation to Hydrocarbon and Mineral Resources. Circum-Pacific Council for Energy and Mineral Resources, Earth Science Series, 11, 151-164. HARLAND,W. B., ARMSTRONG,R. L., COX,A. V., CRAIG, L. E., SMITH,A. G. & SMITH, D. G. 1990. A Geologic Time Scale 1989. Cambridge University Press, Cambridge. HERVE, M. A. 1987a. Movemiento normal de la falle Paposo, Zona de Falla Atacama en el Mioceno al norte de Paposo (24~ Chile. Revista Geol6gica de Chile, 31, 31-36. 1987b. Moviemento sinistral en el Cretficico Inferior de la Zona de Falla Atacama al norte de Paposo (24~ Chile. Revista Geol6gica de Chile, 31, 37-42. JENSEN, A., DORR, M. J., GOTZE, H. J., KIEFER, E., IBBEKEN, H, & WILKE, H. 1995. Subsidence and sedimentation of a forearc-hosted, continental pull-apart basin: the Quillagua Trough between 21~ and 21~ northern Chile. In: A. SAEz (ed.) Abstracts G L O P A L S - I A S Meeting, Antofagasta, 12-18 November, Chile, 5-6. MAKSAEV, V. & MARINOVIC, N. 1980. Cuadr~ingulos Cerro de la Mica, Quillagua, Cerro Posada y Oficina Prosperidada, Regi6n de Antofagasta. Carta Geologica de Chile. Instituto de Investigaciones Geol6gicas, 45-48. MAC, G., HARTLEY,A. J. & STUART,E M. 1996. Oligocene-Recent sedimentary and tectonic evolution of the Calama basin, N. Chilean forearc. Third International Symposium on Andean Geodynamics, Saint-Malo (France), 17-19 September 1996, 435-437. MCCLAY, K. R. & DOOLEY, T. R 1995. Analogue modelling of pull-apart basins. Geology, 23, 711-714. NARANJO, J. A. 1978. Zona interior de la Cordillera de la Costa entre los 26000 ' y 26 ~ region de Atacama. Carta Geologica de Chile, escala 1:100 000. - - , HERVI2, M. A., PRIETO, X. & MUNZINGA E 1984. Actividad Cretficica de la Falla al este de Chafiaral: milonitizaci6n y plutonismo. Comunicaciones Departemento de Geologia, Universidad de Chile, 34, 57-66. NILSEN, T. H. & SYLVESTER,A. H. 1995. Strike-slip basins. In: BUSBY, C. J. & INGERSOLL,R. V. (eds) Tectonics of Sedimentary Basins. Blackwell, Oxford, 425-457. PARDO-CAsAS,E & MOLNAR,P. 1987. Relative motion of the Nazca (Farallon) and South American

SALAR G R A N D E PULL-APART BASIN plates since late Cretaceous time. Tectonics, 6, 233-248. RANDALL, D. E., TAYLOR, G. K. & GROCOTT, J. 1996. Major crustal rotations in the Andean margin: paleomagnetic results from the Coastal Cordillera of northern Chile. Journal of Geophysical Research, 101(B7), 15783-15798. REUTrER, K. J. 8r SCHREUBER, E. 1988. Relation between tectonics and magmatism in the Andes of Northern Chile and the adjacent areas between 21 ~ and 25~ V Congreso Geol6gico Chileno, Tomo I, A345-A363. , & CHONG,C. 1996. The Precordilleran fault system of Chuquicamata, Northern Chile: evidence for reversals along arc-parallel strike-slip faults. Tectonophysics, 258, 213-228. ROGERS, G 1983a. A geochemical traverse across the north Chilean Andes. PhD thesis, Open University. -1983b. The petrogenesis of the 'La Negra' Formation, N Chile. Eos Transactions, American Geophysical Union, 64, 329. & HAWKESWORTH,C. J. 1989. A geochemical traverse across the north Chilean Andes; evidence for crust generation from the mantle wedge. Earth and Planetary Science Letters, 91(3--4), 271-285. SANTANACH,P., SERRAT,D., S/i~EZ,A. CABRERA,L. & CHONO, G. 1996. Tertiary extensional faulting at the lower Loa valley, Northern Chile (21-22~ Third International Symposium on Andean Geodynamics, Saint-Malo (France), 17-19 September 1996, 489--492. SCANLAN,P. M. 1992. Tectonic rotations and Mesozoic basin evolution in the Central Andes. PhD thesis, University of Birmingham. SCHAFER,K. H. • DANNAPFEL,M. 1994. State of in situ stress in Northern Chile and North-western Argentina. In: REUTTER, K. J., SCHREUBER,E. & WIGGER, P. J. (eds) Tectonics of the Southern Central Andes: Structure and Evolution of an Active Continental Margin. Springer-Verlag, Berlin, 103-110. -

-

141

SCHREUBER,E. & ANDRIESSEN,P. A. M. 1990. The kinematic and geodynamic significance of the Atacama fault zone, northern Chile. Journal of Structural Geology, 12(2), 243-257. & REUTTER,K. J. 1992. Magmatic arc tectonics in the central Andes between 21 ~ and 25~ Tectonophysics, 205, 127-140. SKARMETA,J. & MARINOVlC,V. 1981. Hoja Quillagua, Regi6n de Antofagasta. Carta Geologica de Chile. Instituto de Investigaciones Geol6gicas, 51. ST ARMAND,P. & ALLEN, C. R. 1960. Strike-slip faulting in northern Chile (abstract). Geological Society of America Bulletin, 71, 1965. STOERTZ, G. E. & ER~CKSEN,G. E. 1974. Geology of Salars in Northern Chile. US Geological Survey Professional Paper, 811. SWANSON, M. T. 1989. Sidewall ripouts in strike-slip faults. Journal of Structural Geology, 11(8), 933-948. TAYLOR, G., RANDALL,D. & GROCOTT, J. 1996. Paleomagnetism, strike-slip systems and crustal rotation in the region 25-27~ of northern Chile. Third International Symposium on Andean Geodynamics, Saint-Malo (France) 17-19 September 1996, 509-512. THIELE, R. dc PINCHEIRA,M. 1987. Tectonica transpresiva y movemiento de desgarre en el segmento sur de la Zona de Falla Atacama, Chile. Revista Geol6gica de Chile, 31, 77-94. TOMLINSON, A., MPODOZIS, C., CORNEJO, P. & RAMIREZ, C. 1993. Structural geology of the Sierra Castillo-Agua Amarga fault system, Precordillera of Chile, E1 Salvador-Portrerillos. Second international Symposium on Andean Geodynamics, ORSTROM and University of -

-

Oxford, Oxford. ZENTILLI, M. 1974. Geological evolution and metaUogenetic relationships in the Andes of northern Chile between 26 ~ and 29 ~ south. PhD thesis, Queens University, Kingston, Canada.

Strike-slip partitioned transpression of the San Andreas fault system: a lithospheric-scale approach CHRISTIAN

TEYSSIER 1 & BASIL TIKOFF 2

1Department o f Geology and Geophysics, University o f Minnesota, Minneapolis, M N 55455, USA (e-maik teyssier@maroon, tc. umn. edu) 2Department o f Geology and Geophysics, Rice University, Houston, T X 77005, USA (e-mail." btikoff@geophysics, rice.edu) Abstract: Physical and numerical experiments on transpression indicate that strike-slip partitioning is facilitated by low angle of convergence, such as occurs on the modern San Andreas fault system. Cross-sections across the San Andreas fault overestimate the amount of margin-normal contraction, because they do not include the effect of horizontal stretching caused by the wrench component of deformation. When this correction is made assuming a strike-slip partitioned transpression model, contraction estimated from cross-sections is consistent with both the normal and tangential movements imposed by plate motion: there is no San Andreas discrepancy. Seismic anisotropy (shear-wave splitting) and teleseismic analyses suggest that the upper mantle in the San Andreas region is thick and largely undeformed to the SW of the San Andreas fault, but thin and intensely sheared in a zone 50-100 km wide beneath the NE region of the San Andreas fault system. Seismic anisotropy is best explained by a transpression zone of vertical foliation (and possibly horizontal lineation) which penetrates the asthenospheric mantle. Using a lithospheric-scale approach, the transfer of the displacement field from the penetratively deformed mantle to the strike-slip partitioned upper crust necessitates zones of accommodation in the mid- to lower crust, corresponding to the fiatlying detachments that underlie the San Andreas fault system. The mechanics of the San Andreas fault are best considered in connection with the strength of the various lithospheric layers involved in this system.

Strike-slip partitioning is commonly observed at zones of oblique convergence or divergence of tithospheric plates. Our definition of strike-slip partitioning is as follows: deformation is accomm o d a t e d simultaneously by strike-slip fault systems or shear zones and contractional or extensional structures (Fig. 1). Sumatra (e.g. McCaffrey 1992; Tikoff & Teyssier 1994), the San Andreas fault system (e.g. Mount & Suppe 1987; Sylvester 1988), Turkey (Jackson 1992), and New Zealand (Cashman et al. 1992) all display strike-slip partitioning. Although more difficult to recognize, the same behaviour is observed in ancient orogenic belts (Vauchez & Nicolas 1991) including the Archaean Canadian Shield (Hudleston et al. 1988), the Variscan Iberian-Armorican arc (Dias & Ribeiro 1994), the Mesozoic N o r t h A m e r i c a n Cordillera (Oldow et al. 1989), the Tertiary European Alps (Ratschbacher 1986), and Spitsbergen (Teyssier et al. 1995a). In this paper, we investigate two kinematic implications of strike-slip partitioning in the upper crust of the San Andreas fault (SAF) system, one of the best documented strike-slip

partitioned systems: (1) deformation of the San Andreas borderlands; and (2) deformation of the lithospheric layers that underlie the strikeslip partitioned SAF system. (1) Upper-crustal deformation is characterized by strike-slip partitioning, where c. 73% of the tangential relative plate motion occurs on the SAF; this difference has been referred to as the 'San Andreas discrepancy'. Cross-sections across central California (Namson & Davis 1990; Bloch et al. 1993) are now sufficiently constrained by surface geology, seismic sections and borehole data to estimate convergence rates and amounts across the SAF. These cross-sections do not integrate the amount of horizontal extension in a direction perpendicular to the contraction direction (an effect present in all transpressional zones that contain a wrench component (Jamison 1991; Krantz 1995)) and overestimate the true amount of plate convergence across the SAF system. Using these crosssections and a strike-slip partitioning model, the amounts of normal and tangential movements are calculated; these rates match plate motion estimates (e.g. DeMets et al. 1990). Therefore,

TEYSSIER,C. & TIKOFF,B. 1998. Strike-slip partitioned transpression of the San Andreas fault system: a lithospheric-scale approach. In: HOLDSWORTH,R. E., STRACHAN,R. A.8r DEWEY,J. E (eds) 1998. Continental Transpressionaland Transtensional Tectonics. Geological Society, London, Special Publications, 135, 143-158.

143

144

C. TEYSSIER & B. TIKOFF

Fig. 1. Strike-slip partitioned transpression, as simulated by the physical experiments of Richard & Cobbold (1990). A discrete offset of rigid plates occurs at the lowest levels. Deformation is distributed within the middle silicone layer. Both strike-slip and thrust faults occur within the uppermost sand layer. there is no 'San Andreas discrepancy'; motion is simply accommodated in the wrench borderlands of the San Andreas system. (2) How does strike-slip partitioning of the upper crust extend with depth? Lowercrustal and upper-mantle d e f o r m a t i o n is investigated using published results of seismic sections and seismic anisotropy interpreted from shear-wave splitting. These studies suggest a pervasive zone of deformation below the SAF system characterized by vertical foliation and probably horizontal lineation. Therefore, the upper-crustal strike-slip partitioning of the SAF system is coupled to a pervasively shearing upper mantle through a complexly deforming ductile mid- to lower crust. This ductile layer transfers shearing of the lithospheric and asthenospheric mantle below central California (e.g. Molnar 1992) to the upper crust, resulting in the development of the flat-lying detachment zones recognized in seismic sections. Relative plate motion in California is responsible for a broad zone of deformation encompassing the borderlands of the SAF and going down to depths of over 100 km, into the asthenosphere. The broad zone of deformation in the mantle imposes the dominant boundary condition at the base of the SAF system. Transpression deformation in continental regions may be dominated by this basal displacement field which expresses itself differently in the various rheological layers of the lithosphere, culminating in severe strike-slip partitioning of the upper crust.

Kinematics In this paper, we consider the case of transpression defined by Sanderson & Marchini (1984), where the boundary conditions include a simple shear component accommodating lateral translation along the deforming zone, a pure shear component allows horizontal shortening across the deformation zone accommodated by vertical elongation that results in the upward motion of the Earth's flee surface, and there is no volume change. Fossen & Tikoff (1993) and Tikoff & Teyssier (1994) defined two types of homogeneous transpression, based on the orientation of the infinitesimal strain axes: wrench dominated and pure-shear dominated. In wrench-dominated transpression, 81 and s3 (infinitesimal contraction direction) lie in the horizontal plane and oblique to the vertical shear plane, and &2is vertical. In pure-shear-dominated transpression, &2 and &3lie in the horizontal plane and are oblique to the shear plane, and &l is vertical. Thus, the infinitesimal contraction direction (/~3) always lies in the horizontal plane, and its orientation relative to the zone boundaries is used to define different cases of transpression. Wrench-dominated transpression corresponds to angles of convergence 0~ < oL< 20 ~ (45 ~< k3 <55~ and pureshear-dominated transpression corresponds to 20~ < a < 90~ ( 550 < &3<900) 9As noted by Tikoff & Teyssier (1994), the angle between the direction of infinitesimal contraction and relative movement of plates is 45 ~ for wrenching, 0~ for normal convergence, and intermediate for

LITHOSPHERIC-SCALE TRANSPRESSION

145

Fig. 2. Top: map views of transpression experiments, based on set-up of Withjack & Jamison (1986) and used by Tikoff & Peterson (1998), showing the difference between relative motion (arrow) and infinitesimal extension and contraction (crosses, short axis is contraction). Pre-stretched rubber sheet provides homogeneous deformation at base of model. Bottom: parallel and straight flow lines typify homogeneous transpressional deformation, for any angle of convergence. transpressional cases (Fig. 2). Therefore, the contraction direction deduced from young (not rotated) folds and thrusts, or from borehole breakouts or first-motion tensors derived from seismic events, should always be at an angle to the plate motion vector, if transpression is the appropriate model for oblique plate convergence. There are two possible orientations of the finite strain axes for transpressional deformations, y and z lie in the horizontal plane, and x is vertical, for all cases of pure-shear-dominated transpression and some cases of wrench-dominated transpression when the angle of convergence is close to 20 ~ or strain is high (Tikoff & Greene 1997). x and z lie in the horizontal plane, and y is vertical for some cases of wrench-dominated transpression. Wrench-dominated transpression leads to a fundamental misorientation between infinitesimal and finite strain axes. Tikoff & Teyssier (1994) proposed that this misorientation may be at the origin of strike-slip partitioning in the upper crust. Strike-slip faults that form early in wrench-dominated transpression settings are not well oriented to accommodate the threedimensional finite strain which results from contraction across the deformation zone. N e w structures (e.g. thrust faults, folds) must form to accommodate this contraction. In other cases, strike-slip partitioning may be related to the

intrinsic properties of the crust or lithosphere (e.g. weak zone in magmatic arc settings, de Saint Blanquat et al. this volume). The observation that orogenic belts commonly partition a component of the transcurrent motion onto a strike-slip fault (Fig. 1; Fitch 1972; Vauchez & Nicolas 1991), requires a more detailed model. Tikoff & Teyssier (1994) created a model of strike-slip partitioned transpression in an attempt to quantify this type of deformation. In this model, some of the transcurrent motion imposed by relative plate motion is accommodated on discrete faults oriented parallel to the transpression zone boundary. The remaining portion of the transcurrent deformation and all of the normal plate motion component are accommodated in the areas adjacent to the strike-slip faults (wrench borderlands; Jamison 1991). Thus, the relative motion is accommodated by discrete faults and broad zones of transpressional deformation. The strain model predicts a relationship between the percentage of slip partitioned on strike-slip faults (called strike-slip efficiency), the angle of relative plate motion, and the orientation of infinitesimal strain axes (correlated with stresses). By comparing analytical models with tectonic examples (Sumatra, central California, New Zealand), Tikoff & Teyssier (1994) and Teyssier et al. (1995b) tested the hypothesis that strike-slip partitioned systems can be explained in terms of bulk kinematics.

146

C. TEYSSIER & B. TIKOFF

Margin-parallel extension and overestimation of the normal component of plate motion Determining the exact relation between structures within a deformation zone and the motion of the deformation-zone boundaries is not straightforward. This is particularly true for transpressional zones, because of the threedimensional nature of deformation (Koons &

Henderson 1995). Let us consider a wrench zone, which may or may not be strike-slip partitioned (Fig. 3a). The simple shear zone boundaries are sliding past each other, maintaining a constant width of the zone, yet en echelon folds are formed in competent layers (Odonne & Vialon 1983; Tikoff & Peterson 1998). Calculation of shortening from folds does not correlate with the motion of the deformation-zone boundaries. The only case where the palinspastic reconstruction of folded regions gives information on the motion

Fig. 3. Diagrammatic illustration of horizontal extension in transpression (partly after Twiss & Moores 1992), particularly addressing the effect of strike-slip partitioning. (a) Partitioned wrenching, in which apparent contraction across the deforming zone is exactly matched by horizontal extension. (b)-(d) Increasing amounts of strike-slip partitioning decreases the horizontal extension in the wrench borderlands. True marginperpendicular contraction is only recorded in the case of complete strike-slip partitioning.

LITHOSPHERIC-SCALE TRANSPRESSION

147

1

ca

0.9 0.8

= ._o

0.7

o-

0.6

=~ ~

0.5

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

0.4 O ""

"NO r 0.3 ( ~Norme~Caht ,~'~"=~"*"'ed dominated x 0.2 =

0. I

........... ~............~

0

10

20

30

~

40

50

~

60

'

70

80

90

Angle of convergence (alpha) Fig. 4. Graph demonstrating the relationship between horizontal contraction and horizontal extension in transpression, as a function of plate motion. Horizontal extension is caused by a wrench component of deformation and results in overestimation of contraction for transpressional settings. For simple shear deformation (a = 0~ all horizontal contraction is matched by horizontal extension and no true contraction occurs. The border between wrench-dominated and pure-shear-dominated transpression is noted by a switch from more horizontal to more vertical infinitesimal stretching.

of the deformation-zone boundaries is when convergence is strictly normal. In a transpression zone, folds develop because of the combined effect of simple shear and pure shear (Fig. 3b and c; Jamison 1991; Fossen & Tikoff 1993; Krantz 1995). Measuring the shortening based on the cross-sectional balancing of a folded region does not give information on the exact motion of the deformation-zone boundaries. It is necessary to evaluate not only the horizontal shortening, but also the horizontal extension, to make inferences on the boundary conditions, and determine whether the deformation zone is a wrench zone or a transpression zone. In principle, an alternative would be to measure the vertical thickening, but this is commonly difficult because there are few structural markers recording elongation in this direction. Horizontal extension caused by shear deformation is particularly well studied in regions of en echelon folding. As indicated by Jamison (1991), all folds in transpression zones should accommodate a certain amount of extension parallel to the fold hinges. Hinge-parallel extension increases with decreasing convergence angle (because the simple shear component increases), and is particularly acute for wrenchdominated transpression (eL <20~ for which the long axis of finite strain remains parallel to the fold hinge up to relatively high strain. Typically, this hinge-parallel extension is accommodated by normal faults or conjugate strike-slip faults (Jamison 1991). In a zone adjacent to the SAF

(Durmid Hills), Bargmann (1991) recognized boudinage of a traceable ash layer and was able to quantify hinge-parallel extension. Faulting accommodating horizontal extension is also well documented in transtension experiments (Withjack & Jamison 1986). To determine the amount of boundary-normal convergence across a transpression zone, one must subtract the apparent contraction matched by hinge-parallel extension which results from the simple shear component. Several ways of calculating this horizontal extension have been proposed: direct measurement (Btirgmann 1991), fold geometry (Jamison 1991), and regional strain (Krantz 1995). If the angle of convergence and the amount of contraction are known, a simple graph can be constructed to plot the relative ratio of infinitesimal contraction v. extension in the horizontal plane for transpression (Fig. 4). The margin-normal convergence across a transpressional zone is the percentage of observed contraction not accommodated by horizontal extension. For wrench deformation (o~ = 0~ all of the horizontal contraction is exactly compensated by horizontal extension (plane strain), and no margin-normal convergence occurs (Fig. 3a). For head-on convergence (o~ = 90~ the normal contraction matches normal convergence (Fig. 3d). For intermediate cases (Fig. 3b and c), for example a = 45 ~ and contraction of 10 km, the amount of horizontal extension to horizontal contraction is 0.18, and thus 82% or 8.2 km of the observed contraction

148

C. TEYSSIER & B. TIKOFF

is actual margin-parallel contraction (Fig. 4). Once the angle of plate motion and the marginnormal contraction are known, the amount of wrenching is easily calculated from geometry. We demonstrate below that this kinematic consideration allows us to dismiss the concept of a 'San Andreas fault discrepancy' that has permeated the literature: the missing strike-slip component is contained in the borderlands of the San Andreas system. It should be noted that Fig. 4 is only valid for transpression in the sense of Sanderson & Marchini (1984) or type B transpression of Fossen & Tikoff (this volume). The relation between boundary-normal contraction and horizontal extension changes if deformation includes margin-parallel stretching (Fossen & Tikoff this volume).

Physical and numerical experiments The few published physical and numerical experiments which attempted to simulate threedimensional deformation give us significant insight into the parameters causing strike-slip partitioning. The physical experiments of Richard & Cobbold (1990) produced a series of faults in sand during oblique convergence. Two experimental configurations were used. In the first, sand directly overlies a sheet of Mylar and no strike-slip motion occurs. Deformation is accommodated by oblique-slip movement on the faults, which is a typical response of sand in oblique convergence (e.g. Lamons 1991). In the second configuration, a layer of sand overlies a ductile layer of silicone, all of which rests on rigid plates. In this case, coeval strike-slip and thrust faults develop in the layer of sand. Strikeslip partitioning preferentially occurs with low angles of convergence. As both types of faults occur in the same material, the partitioning of deformation cannot be explained by intrinsic fault behaviour or fault weakness. The mechanical role of the silicone layer is debatable, but this layer undoubtedly serves to distribute the deformation imposed by the velocity discontinuity at the boundary between the two underlying moving plates. Molnar (1992) reinterpreted the results of these experiments and proposed that the silicon layer couples the moving plates at the base of the experiments with the sand. These experiments have been recreated using three-dimensional finite-element modelling, by Braun & Beaumont (1995). These workers showed that, for a low angle of convergence (r < 0.25, where r is the ratio of normal to transverse components of movement; e.g. oL <18.4~ both thrusts and vertical strike-slip faults form

simultaneously, and strike-slip partitioning occurs. For higher angles of convergence, oblique-slip faults form to accommodate the imposed motion, and no strike-slip partitioning occurs. The physical experiments of Richard & Cobbold (1990) and the numerical modelling of Braun & Beaumont (1995), which examined the temporal evolution of oblique convergent fault systems, agree with the idea derived from kinematic analysis that low angles of convergence favour strike-slip partitioning (e.g. wrenchdominated transpression). The effect of a 'soft' basal layer was investigated by recent experiments conducted at the University of Minnesota. These experiments were performed to study the process of folding in transpression (Peterson & Tikoff 1995; Tikoff & Peterson 1998). The experimental apparatus was designed directly after the machine utilized by Withjack & Jamison (1986), to simulate the kinematics of transpression (Fig. 3). Peterson & Tikoff (1995) constructed a three-layer model: a rubber sheet at the base, which acted to homogenize deformation, an overlying layer of Newtonian silicone (RD-20; Rh6ne-Poulenc), and a competent layer of combined plasticine-silicone on top. All angles of convergence (oL --- 0~ 15~ 25~ 45~ 60~ and 90~ indicated that: (1) the folds initiate perpendicular to the infinitesimal contraction direction; and (2) the fold hinges rotate parallel to the long axis of the horizontal finite strain ellipse during deformation. The average trend of folds lies near the expected value for the horizontal finite strain axes predicted from a transpression model, supporting the notion that hinges parallel the finite strain axis rather than rotate as a material line (e.g. Treagus & Treagus 1992). Thus, in the case of en echelon folds, it is possible to infer a relation between structures within a deformation zone and the motion of the deformation-zone boundaries.

Importance of basal boundary condition and justification of transpressional kinematics in oblique convergence The exact rheological nature of the lower basal layer, particularly with respect to upper-crustal deformation, is critical. As discussed above, the experiments of Richard & Cobbold (1990) only exhibited strike-slip partitioning in the presence of the silicone layer. The mechanical role of this layer was inferred to either decouple (Richard & Cobbold 1990) or couple (Molnar 1992) the sand from the underlying plates. A simple experiment was conducted at the University of Minnesota to investigate this effect. Sand overlay silicone, which both overlay a basal rubber sheet. During

LITHOSPHERIC-SCALE TRANSPRESSION the experiment, the sand layer thickened and sheared, as noted by rotation of grid lines placed on top of the surface. Thus, if the sand and silicone layers are analogous to the Richard & Cobbold (1990) experiments, the layers must be at least partially coupled. Therefore, we support the notion of Molnar (1992) that strike-slip partitioning requires coupling. Much discussion by Sanderson & Marchini (1984), Schwerdtner (1989) and Robin & Cruden (1994), has centred around the exact kinematics associated with transpressional deformation. These workers typically considered that the transpressional 'walls' provide the boundary conditions to which transpressional deformation responds. If this is the case, the 'walls' must have very particular mechanical properties (e.g. Schwerdtner 1989) to resolve the problems of strain compatibility. We propose an alternative, or 'bottoms-up', view to transpressional systems (Fig. 5). Rather than considering a deforming zone between two colliding blocks, orogenic transpression is better viewed in the context of flow lines from below, driven by deformation of the strong subcontinental mantle. A basal layer (middle or lower crust) acts to couple and control deformation in the overlying layer (upper crust), as occurred in the folding experiments. Outside the deforming zone, the upper and lower crust are attached and do not move with respect to each other (Fig. 5). Above the deforming lower crust, the upper crust must also deform. This upper-crustal deformation is commonly dominated by strikeslip partitioning (Fig. 1), but this behaviour is not required (e.g. South Island of New Zealand; Norris et al. 1990; Teyssier et al. 1995b). If strikeslip partitioning occurs, detachments must exist between the sides of the strike-slip fault, as well as between the upper and lower crust, to accommodate the difference in deformation style between these two crustal levels. This view potentially explains how the fractured upper crust responds to 'bulk' transpressional tectonics: the upper crust deforms above a flow field imposed from below (McKenzie & Jackson 1983; Molnar 1992). Flow lines in transpressional deformation are always straight and parallel lines (e.g. Fossen et al. 1994). The observations of gradients of simple shear movement are consistent with a model of heterogeneous transpression (Tikoff & Greene 1997). Gradients in the wrench component of deformation cause the orientation and magnitude of flow lines to change in zones normal to the margin (Fig. 6), but do not result in major margin-parallel or vertical strain compatibility problems. The essential role of the basal boundary condition

149

Fig. 5. Two contrasting views of transpression: (a) deformation between two rigidly converging plates; (b) and (c) 'bottoms-up' view in which transpressional deformation is driven by basal movement, with top and bottom views, respectively. The latter may be more appropriate for orogenicscale transpression. explains why transpression, which is a rather simple type of three-dimensional deformation, appears to work so well to model the orientation of infinitesimal and finite structures in a number of orogens developed in oblique convergence. Experimental modelling of lithospheric layers has shown that the ductile layer (mid- to lower crust) is capable of forming 'detachments' (very strong gradients of deformation) while remaining

150

C. TEYSSIER & B. TIKOFF

Fig. 6. Flow lines in heterogeneous transpression, viewed from fixed (right-hand) side. Large arrow: vector of movement; small arrows: vector of domainal flow; dashed lines: vector of contractional component. Flow lines are straight and parallel within domains. Each domain contains uniform contraction but a varying amount of wrench deformation, and thus the orientation of flow vectors varies. in bulk mechanical continuity (i.e. coupled). The existence of strike-slip partitioning at the surface, but distributed deformation at the base of the experiment (e.g. Richard & Cobbold 1990) implies that some horizontal shearing must occur within the basal layer. Thus, the basal layer couples lithospheric layers, but also provides a medium that accommodates the steep strain gradients. In nature, if the lithospheric mantle deforms over a wide zone and strike-slip partitioning occurs in the upper crust, it follows that the midto lower crust must contain structures that accommodate the differential displacements. One possible field example of this behaviour has been documented by Pavlis & Sisson (1995), in a Tertiary forearc system (Chugach complex, Alaska). They found a pervasive sub-horizontal foliation (D2), which they interpreted as the detachment between the lower and upper crust. Coeval contractional deformation, evidenced by vertical foliation, was inferred to have occurred above and below the detachment zone. Another example of a sub-horizontal, mid- to lowercrustal detachment occurs in northeastern Greenland (Holdsworth & Strachan 1991; Strachan et al. 1992). This high-strain zone records low angle foliation, prominent stretching lineations, and sheath-fold geometries consistent

with orogen-parallel movement. This zone is infered to have formed early during the Caledonian orogeny (Stage 1), in reponse to oblique sinistral convergence (Strachan et al. 1992).

Application to central California Upper-crustal kinematics

The well-documented geology in central California provides an opportunity to determine the true offset across a distributed wrench zone. A small reorientation of relative plate motions at 3.4-3.9 Ma is hypothesized to have caused a switch from a transcurrent-transtensional to a transpressional setting in central California (Harbert 1991). The present angle of convergence between the North American and Pacific plates is c. 5~ at the latitude of central California (DeMets et al. 1990), and the rate of 48.0 m m / a is resolvable into transcurrent (47.8 mm/a) and normal (4.2 mm/a) components. Geodetic surveys indicate that the SAF accommodates only 70-75% (35 mm/a) of the strikeslip motion predicted from the plate motion (e.g. Rymer et aL 1984; Prescott & Yu 1986). The exact amount of the discrepancy depends on whether the Basin and Range extension is added to the North America-Pacific relative motion (e.g.

LITHOSPHERIC-SCALE TRANSPRESSION Minster & Jordan 1978). If the Basin and Range extension is considered, the San Andreas discrepancy is resolvable into transcurrent (4 mm/a) and normal (7 mm/a) components ('modified' discrepancy of DeMets et al. (1990)). However, we consider that the relative plate motion should match deformation in central California and thus consider the discrepancy as consisting of transcurrent (47.8 mm/a) and normal (4.8 mm/a) components. Some slip on other margin-parallel strike-slip fault zones, including the Hosgri and Rinconada (e.g. Graham & Dickinson 1978; Sedlock & Hamilton 1991), and the eastern California shear zone (Dokka & Travis 1990), may partially alleviate the transcurrent discrepancy. The remaining discrepancies, both tangential and normal, are potentially recorded in distributed wrench deformation in western California (Jamison 1991). Central California is a strike-slip partitioned transpressional system (Tikoff & Teyssier 1994), in which the SAF accommodates a major portion of the transcurrent movement provided by the plate motion. However, distributed dextral deformation occurs in the borderland regions of the fault. In these areas, en echelon faults are commonly observed (e.g. Diblee 1976; Harding 1976) flanking the major strike-slip faults. As summarized by Jamison (1991), these en echelon folds have an orientation of c. 20 ~ to the adjacent major faults and moderate limbdips of 20-40 ~. These systems have been particularly well studied with field mapping, drill cores, and seismic sections because of petroleum reserves; for example, by Namson & Davis (1988, 1990) and Bloch et al. (1993). Based on their work, and that of Namson & Davis (1988), Bloch et al. (1993) were able to estimate a contraction of 33.1 km across the SAF system. The models are constructed assuming fault-bend folding and a d6collement at 11-14 km depth. For the purposes of our calculation we use 3.65 Ma as the initiation of transpression, which results in a margin-normal convergence rate of 9.2 mm/a. It is critical to realize that although the geological data are of very good quality, the contractional rate is not within the error bars of either plate motion prediction of 4.2 mm/a or even the estimates that account for Basin and Range extension (7 mm/a). In any case, the amount of contraction appears too high. There are several possibilities to solve this problem. As indicated by Jamison (1991), the assumption of fault-bend folding (e.g. Namson & Davis 1988) tends to maximize the amount of contraction. Alternatively, deformation may have initiated before 3.9 Ma. However, even these problems do not account for the large discrepancy

151

between observed contraction and plate motion. The major oversight of these analyses is the failure to take into account horizontal extension into the models of normal convergence. Following the analysis of Jamison (1991) and Krantz (1995), we suggest that the contraction across the SAF system is overestimated by not considering the hinge-parallel stretching. This effect is easily envisioned with pure wrenching (Fig. 4), which exhibits large amounts of contraction. However, this contraction is exactly counterbalanced by horizontal extension, which results in the required 'bulk' plane strain deformation. As the angle of convergence increases, the amount of horizontal extension decreases (Fig. 4). Jamison (1991) noted that deformation in the wrench borderlands deviates significantly from pure contraction. According to geodetic surveys, the San Andreas fault accommodates c. 73% of the tangential plate motion. This 73 % strike-slip partitioning for a 5~ convergence angle predicts that the borderlands are characterized by an infinitesimal contraction direction of 55 ~ (Fig. 3; Teyssier et al. 1995b). This modelling also implies that the borderlands are characterized by a convergence angle of 20~ For this angle of convergence (Fig. 4), 50% of the contraction in the horizontal plane is compensated by extension in the horizontal plane. The rest of the motion is actual contraction perpendicular to the fault. Thus, using the contraction estimate of 33.1 km (Bloch et al. 1993), only 16.6 km of actual offset occurs perpendicular to the fault. This provides a normal contraction rate of 4.6 mm/a, close to the 4 mm/a predicted by plate motion. The same analysis suggests 45.6 km of wrench motion accommodated in the en echelon structures of central California. This provides a tangential motion rate of 12.7 mm/a, compared with a discrepancy of 12.8 mm/a. Within the error of the analyses, transpressional deformation in central California is shown to accommodate plate motion. Thus, the San Andreas discrepancy is compensated by distributed deformation in central California. The strike-slip partitioned model (Tikoff & Teyssier 1994), geodetic information (e.g. Prescott & Yu 1986), plate motion (DeMets et al. 1990) and cross-section balancing (e.g. Namson & Davis 1988) indicate a coherent dataset that describes deformation in central California. The normal component of deformation is matched by plate motion. If considered in terms of rates, c. 27% of the lateral motion within the SAF system is accommodated by pervasive wrenching in central California. In terms of total offset (45.6 km), from an estimated offset of 330 km, 14% of

152

C. TEYSSIER & B. TIKOFF

Fig. 7. Map of the Pacific-North America plate boundary at the latitude of California, showing the directions of fast polarization of SKS waves (after 0zalaybey & Savage 1995). In the vicinity of the San Andreas fault, two superimposed anisotropic layers are defined; the direction of fast polarization of the shallower layer follows faithfully the trend of the San Andreas fault. Length of line proportional to delay time as shown. the tangential plate motion is accommodated by wrenching. In all cases, a major percentage of strike-slip movement is accommodated by pervasive wrenching, which is expected from a partially strike-slip partitioned transpression zone and should not be overlooked.

Mantle kinematics In the last decade, much progress has been made to study the structure of the subcontinental mantle using the principles of seismic anisotropy. The record of over 300 stations distributed on all major continents, on both active and ancient orogenic zones, indicates that, in general, mantle anisotropy is coherent with crustal deformation (Silver 1996), suggesting that the anisotropy is acquired during lithospheric deformation. On the other hand, laboratory experiments have shown

that the lithospheric mantle is the strongest layer of the lithosphere; therefore, it is crucial to consider the behaviour of the subcontinental mantle in orogenic zones, because this part of the lithosphere dominates the mechanics of the system on the scale of plates, and probably controls the large-scale features of orogenesis. Although there are several possible sources of anisotropy, the dominant contribution appears to be the preferred orientation of mantle minerals (Nicolas & Christensen 1987), acquired during deformation and giving the bulk anisotropy of elastic properties of mantle rocks. One popular technique for recording mantle anisotropy is shear-wave (SKS) splitting (see Silver (1996) for a review of the technique, assumptions, and interpretations). In the San Andreas region, Savage & Silver (1993) and Ozalaybey & Savage (1995)

LITHOSPHERIC-SCALE TRANSPRESSION interpreted shear-wave splitting data (Fig. 7) in terms of two superimposed anisotropic layers. The nature of the lower layer is debatable (Ozalaybey & Savage 1995); its E - W fast polarization direction has been interpreted to be due to either strained asthenosphere or a fossil anisotropy preserved in the subducted old Farallon plate. The upper layer, recognized only in the vicinity of the SAF, contains a fast polarization direction exactly parallel to the surface trace of the fault. This layer is estimated to be 50-100 km wide and 115-125 km thick. The thickness estimation is highly dependent on the degree of anisotropy modelled to explain the delay time. Given the strong probability that the mantle fabric is characterized by a vertical foliation and horizontal lineation, the optimal orientation for shear-wave splitting, the estimate of 4% anisotropy used by Ozalaybey & Savage (1995) is conservative, which means that the anisotropic layer may be somewhat thinner. However, it is unlikely that this layer is as thin as the lithosphere proposed by Zandt (1981) and Furlong et al. (1989) who considered, from teleseismic data, that the asthenosphere reaches shallow levels (45 km) below the SAF system. Therefore we envision that the anisotropic layer measured by SKS splitting consists of the currently thickening lithosphere plus the underlying asthenospheric upwelling directly below the San Andreas system. If simple shear is assumed, the amount of shear strain necessary to produce a fast direction very closely parallel to the wrench zone boundary in the mantle is on the order of -y -~ 5 (A. Tommasi, pers. comm.) or even ~ ~ 2 according to the shear experiments of Zhang & Karato (1995). Given the few hundred kilometres of wrench displacement accommodated by the Pacific-North America plate boundary over a zone only 50-100 km wide in the late Cenozoic, the parallelism of the fast anisotropy direction and the SAF system is easily accounted for. The localization of the shallow anisotropy layer below the San Andreas system strongly suggests that the anisotropy is produced by active shearing along this plate boundary and is not a fossil anisotropy. Progressive cooling of the lithosphere over the last 20-30 Ma, as the Mendocino triple junction moved NW, also suggests that mantle fabric was acquired in recent times. Another important result from the analysis by Ozalaybey & Savage (1995) is that the anisotropy zone in the upper layer is well expressed on the NE side of the SAF over a width of 50-100 km (Helffrich et al. 1994), but is not developed on the SW side of the fault. This spatial distribution of anisotropy directly correlates with lithospheric

153

thickness. Zandt (1981) and Furlong et al. (1989) have demonstrated that the lithosphere is thinner on the NE side of the San Andreas fault. Along the same lines, Liu & Furlong (1992) proposed that the late Cenozoic volcanic belt, off-centred with respect to the SAF system, lies above the zone of mantle upwelling to the NE of the fault. Therefore, a mantle shear zone has concentrated in the thin and hot lithosphere, rather than in the thick lithospheric segment of the Pacific plate. R e l a t i o n b e t w e e n m a n t l e a n d crustal kinematics

We propose a geological view of the structure of the lithosphere in the SAF region that integrates crustal kinematics and mantle flow (Figs 8 and 9). Based on recent studies of seismic anisotropy beneath California (Ozalaybey & Savage 1995), a new picture of mantle structure emerges. The layer of anisotropy is thick (probably over 100 km) whereas the lithospheric mantle is less than half that thickness. Therefore, we propose that vertical foliation and horizontal lineation in the lithospheric shear zone permeate the asthenospheric mantle down to c. 100 km. The main result of this model is that a major mantle shear zone underlies the SAF system, offset to the NE relative to the trace of the SAF. The abrupt change in mantle properties across the SAF suggests that the SW side moves north as a relatively rigid mantle block. At shallow levels, the crust is characterized by strike-slip faults and fold-and-thrust belts developed on both sides of the SAF (e.g. Namson & Davis 1988). Seismic reflection and refraction profiling (Saleeby 1986; Fuis & Clowes 1993) has suggested the existence of a mid-crustal detachment zone under the SAF system, leading some researchers to suggest that the SAF is only a shallow, upper-crustal feature (Jones et al. 1994). In addition, Late Cenozoic crustal evolution in California has probably produced partial melting of the mid- to lower crust which gave rise to felsic volcanism (Liu & Furlong 1992). This ductile lower crust has probably accommodated gradients in the displacement fields between the lithospheric mantle and the upper crust, allowing extreme strike-slip partitioning to take place in the upper crust, above a pervasively deforming mantle. The orientation of lineations in the detachment zones (Fig. 9) may be rather complex, depending on the degree of strike-slip partitioning and possible rotation of upper-crustal blocks. This concept of pervasive lower-crustal detachments is similar to that of orogenic float developed by Oldow et al. (1990). In our model, the displacement field in the

154

C. TEYSSIER & B. TIKOFF

Fig. 8. Lithospheric cross-section of the San Andreas fault system in central California, oriented SW-NE. Upper-crustal structure from Jones et al. (1994). In contrast to those workers, we propose that the San Andreas fault is rooted in the mantle. The mid-crustal reflectors are detachment zones that transfer motion in the penetratively deforming mantle to the upper crust dominated by strike-slip partitioning. The width and depth of the mantle shear zone are interpreted from seismic anisotropy.

crustal layers is driven by the displacement field at the base of the system, determined by the mantle shear zone. Although the mantle shear zone is developed only to the N E of the SAF, a typical fold-andthrust belt developed on the SW side of the fault in the upper crust (e.g. Namson & Davis 1990). The symmetry of displacement field relative to the SAF in the upper crust contrasts with the asymmetry of mantle deformation. This situation can only be resolved by the presence of flat detachments within the lower crust which act to transfer the displacement field from the mantle to the upper crust. We propose that the major active lithospheric plate boundary is the mantle shear zone interpreted by Ozalaybey & Savage (1995) from shear-wave splitting. This shear zone is probably diffuse (50-100 km wide) and its relation to the upper crust is unclear because of the presence of

subhorizontal detachments that accommodate the motion of upper-crustal blocks relative to the mantle. The observation that the SAF lies directly above the SW margin of the mantle shear zone suggests that the SAF is rooted in the mantle. The SW boundary of the shear zone is probably the location of a major displacement gradient within the mantle, which would localize the S A F in the upper crust. This effect is observed in physical experiments in which faults localize above velocity discontinuities (i.e. line discontinuity at base of experiment) even if an intermediate ductile layer (silicone) is present (e.g. Richard & Cobbold, 1990). Alternatively, the upper-crustal SAF may have moved relative to the mantle during its development. Whether the lithospheric mantle has moved northeastward relative to the upper crust and the S A F is debatable ( H a d l e y & Kanamori 1977; Furlong et al. 1989; Furlong

LITHOSPHERIC-SCALE TRANSPRESSION

155

Fig. 9. Schematic representation of different levels of lithospheric deformation in the San Andreas fault system. The upper crust responds by strike-slip partitioning, the lower crust by heterogeneous flow, and the upper mantle by penetrative shearing. The shear zone in the upper mantle drives the system. The mid- to lower crust undergoes complex shearing to allow transfer of motion from the upper mantle to the strike-slip partitioned upper crust; in particular, flat-lying detachments are expected to develop, in which the lineation's orientation is very sensitive to the local strike-slip partitioning, in a manner similar to that proposed by Oldow et al. (1990) for the orogenic float concept. The upper and lower crust are coupled by the bulk flow in the deforming zone, although the lower crust is capable of producing high strain shear zones, presumably subhorizontal, which locally detach the two layers. 1993; Ozalaybey & Savage 1995). The fold-andthrust belt developed within and to the SW of the Salinian block on the western side of the SAF during the Tertiary requires a d6collement system (Tavernelli this volume). In our view, the motion of the SAF relative to the mantle is dependent on the dynamics of the mantle shear zone, which may vary with time. Large gradients in shear strain rates within the mantle are likely to influence not only the location of the SAF, but also the activation of other strike-slip faults such as the Hayward and Calaveras faults (Furlong 1993), and possibly as far as the eastern California shear zone (Dokka & Travis 1990; Pez-

zopane & Weldon 1993), as long as the lower crust transfers mantle displacements.

Conclusions Strike-slip partitioning is u n d e r s t o o d in the context of different lithospheric layers. Deformation of the upper crust depends on the basally imposed motion of lower lithospheric layers. Thus, upper-crustal transpressional deformation is better considered as a result of relative motion from below, rather than deformation between rigidly deforming blocks. Cross-sections across the San Andreas fault

156

C. TEYSSIER & B. TIKOFF

system (constrained by surface geology, seismic sections, and borehole data (e.g. Bloch et al. 1993)) overestimate (9.2 mm/a) the amount of contraction compared with plate motions (4.2 ram/a). This overestimation is an artefact of cross-section balancing that does not integrate the amount of horizontal extension (perpendicular to the contraction direction) caused by the wrench c o m p o n e n t of deformation (e.g. Jamison 1991). W h e n this effect is included, using a strike-slip partitioned transpression model, the normal contraction and tangential motion rates become 4.6 mm/a and 12.7 mm/a, respectively. Adding these results to the observed 35 mm/a of transcurrent motion on the San Andreas fault from geodetic surveys (e.g. Prescott & Yu 1986), there is no longer a discrepancy of movement. D e f o r m a t i o n in central California, to a first approximation, matches what is expected of a transpressional system driven by relative plate motion, and the San Andreas discrepancy does not exist. Strike-slip partitioning in the upper crust results from shearing and offset in the upper mantle. Shear-wave splitting analysis suggests the existence of a pervasive shear zone in the upper mantle (Ozalaybey & Savage 1995). The same technique also suggests that the NE side of the SAF is highly anisotropic, whereas the SW side does not show that anisotropy. This result is consistent with the idea developed by Furlong (1993), based on teleseismic studies, that a wedge of asthenosphere exists at shallow levels on the NE side of the SAF; the thin lithosphere participates in localizing deformation in this zone. This mantle deformation is coupled to the brittle upper crust via heterogeneous deformation in the ductile crust. The mid- to lower crust also allows the existence of flat detachments where high strain accumulates to allow the translation of rigid, upper-crustal blocks over the shearing mantle. We emphasize that an upper-crustal structure such as the SAF cannot be understood without consideration of its borderlands and underlying lower-crustal and mantle layers. Future research on obliquely convergent systems should focus on documenting deformation in the poorly known lower crust, the critical layer that allows motion to be transferred from the lithospheric plates to the upper crust. This work was supported by US National Science Foundation Grant EAR-9607018. We thank G. Barruol, M. Mattauer, A. Tommasi, and A. Vauchez for fruitful discussions. C. T. acknowledges support from a Bush Foundation grant for partial funding of a sabbatical leave to the Laboratoire de Tectonophysique, Universit6 Montpellier II, and support from this university.

References

BLOCH,R. B., HUENE,R.V., HART,P. E. & WENTWORTH, C.M. 1993. Style and magnitude of tectonic shortening normal to the San Andreas fault across the Pyramid Hills and Kettleman Hill South Dome, California. Geological Society of America Bulletin, 105, 464-478. BRAUN,J. & BEAUMONT,C. 1995. Three dimensional numerical experiments of strain partitioning at oblique plate boundaries: implications for contrasting tectonic styles in California and South Island, New Zealand. Journal of Geophysical Research, 100, 18059-18074. BI3RGMANN,R. 1991. Transpression along the Southern San Andreas fault, Durmid Hill, California. Tectonics, 10, 1152-1163. CASHMAN, S. M., KELSEY, H. M., ERDMAN, C. F., CUaSEN,H. N. C. & BERRYMAN,K. R. 1992. Strain partitioning between structural domains in the forearc of the Hikurangi subduction zone, New Zealand. Tectonics, 11, 242-257. DEMETS, C., GORDON,R. G., ARGUS,D. E & STEIN,S. 1990. Current plate motions. Geophysical Journal International, 101, 425-478. DE SAINT BLANQUAT,M., TIKOFF, B., TEYSSIER, C. VIGNERESSE,J. L. 1998. Transpressional kinematics and magmatic arcs. This volume. DIAS, R. & RIBEIRO,A. 1994. Constriction in a transpressive regime: an example in the Iberian branch of the Ibero-Armorican arc. Journal of Structural Geology, 16, 1543-1554. DIBLEE, T. W. 1976. The Rinconda and Related Faults in the Southern Coast Ranges, California, and their Tectonic Significance. US Geological Survey Professional Paper, 981. DOKKA, R. K. &: TRAVIS, C. J. 1990. Late Cenozoic strike-slip faults of the central Mojave desert, California. Tectonics, 9, 311-340. FITCH, J. T. 1972. Plate convergence, transcurrent faults, and internal deformation adjacent to Southeast Asia and the Western Pacific. Journal of Geophysical Research, 77, 4432-4460. FOSSEN,H. & TIKOFF,B. 1993. The deformation matrix for simultaneous pure shear, simple shear, and volume change, and its application to transpression/trantension tectonics. Journal of Structural Geology, 15, 413-425. - & -1998. Extended models of transpression and transtension, and application to tectonic settings. This volume. & TEYSSIER, C. 1994. Strain modeling of transpressional and transtensional deformation. Norsk Geologisk Tidsskrift, 74, 134-145. FuIs, G. S. & CLOWES,W. D. 1993. Comparison of deep structure along three transects of the western North American continental margin. Tectonics, 12, 1420-1435. FURLONG,K. P. 1993. Thermal-theologic evolution of the upper mantle and the development of the San Andreas fault system. Tectonophysics, 223, 149-164. - - , HUGO,W. D. & ZANDT,G. 1989. Geometry and evolution of the San Andreas Fault. Journal of Geophysical Research, 94, 3100-3110.

LITHOSPHERIC-SCALE TRANSPRESSION GRAHAM, S. A. & DICKINSON,W. R. 1978. Evidence for 115 kilometers of right slip on the San Gregorio-Hosgri fault trend. Science, 199, 179-181. HADLEY, D. & KANAMORI,H. 1977. Seismic structure of the Transverse Ranges, California. Geological Society of America Bulletin, 88, 1469-1478. HARBERT,W. 1991. Late Neogene relative motions of the Pacific and North American plates. Tectonics, 1 0 , 1-16. HARDING, T. P. 1976. Tectonic significance and hydrocarbon trapping consequence of sequential folding synchronous with San Andreas faulting, San Joaquin Valley, California. Bulletin, American Association of Petroleum Geologists, 60, 356-378. HELVFRICH,G., SILVER,R G. & GIVEN,H. 1994. Shearwave splitting variation over short spatial scales on continents. Geophysical Journal International, 119, 561-573. HOLDSWORTH, R. E. ~1; STRACHAN,R. A. 1991. Interlinked system of ductile strike-slip and thrusting formed by Caledonian sinistral transpression in northeastern Greenland. Geology, 19, 510-513. HUDLESTON, P. J., SCHULTZ-ELA,D. & SOUTHWICK,D. L. 1988. Transpression in an Archean greenstone belt, northern Minnesota. Canadian Journal of Earth Sciences, 25, 1060-1068. JACKSON, J. 1992. Partitioning of strike-slip and convergent motion between Eurasia and Arabia in Eastern Turkey and the Caucasus. Journal of Geophysical Research, 97, 12471-12479. JAMISON,W. R. 1991. Kinematics of compressional fold development in convergent wrench terranes. Tectonophysics, 190, 209-232. JONES, D. L., GRAYMER,R., WANG, C., McEVILLY,T. V. & LOMAX,A. 1994. Neogene transpressive evolution of the California Coast Ranges. Tectonics, 13, 561-574. KOONS, R O. & HENDERSON, C. M. 1995. Geodetic analysis of model oblique collision and comparison to the Southern Alps of New Zealand. New Zealand Journal of Geology and Geophysics, 38, 545-552. KRANTZ, R. W. 1995. The transpressional strain model applied to strike-slip, oblique-convergent and oblique-divergent deformation. Journal of Structural Geology, 17, 1125-1137. LAMONS,R. C. 1991. Oblique convergence in modeling experiments and in the field. PhD thesis, University of Minnesota, Minneapolis. LIu, M. & FURLONG,K. P. 1992. Cenozoic volcanism in the California Coast Ranges: numerical solutions. Journal of Geophysical Research, 97, 4941-4951. MCCAFFREY, R. 1992. Oblique plate convergence, slip vectors, and forearc deformation. Journal of Geophysical Research, 97, 8905-8915. MCKENZIE, D. & JACKSON,J. 1983. The relationship between strain rates, crustal thickening, paleomagnetism, finite strain and fault movements within a deforming zone. Earth and Planetary Science Letters, 65, 182-202. MINSTER,J. B. &JORDAN,T. H. 1978. Present-day plate motions. Journal of Geophysical Research, 83, 5331-5354. MOLNAR, P. 1992. Brace-Goetze strength-profiles, the partitioning of strike-slip and thrust faulting at

157

zones of oblique convergence, and the stress-heat flow paradox of the San Andreas fault. In: EVANS, B. & WONG,T.-F. (eds) Fault Mechanics and Transport Properties of Rocks. Academic Press, New York, 435-459. MOUNT, V. S. & SUPPE, J. 1987. State of stress near the San Andreas fault: implications for wrench tectonics. Geology, 15, 1143-1146. NAMSON, J. S. & DAVIS,T. L. 1988. Seismically active fold and thrust belt in the San Joaquin Valley, central California. Geological Society of America Bulletin, 100, 257-273. - & -1990. Late Cenozoic fold and thrust belt of the southern Coast Ranges and Santa Maria basin, California. Bulletin, American Association of Petroleum Geologists, 74, 467-492. NICOLAS,A. & CHRISTENSEN,N. I. 1987. Formation of anisotropy in upper mantle peridotites - A review.

In: Crystalline Plasticity and Solid State Flow in Metamorphic Rocks. John Wiley, New-York. NORRIS, R. J., KOONS,P. 0. & COOPER,A. E 1990. The obliquely-convergent plate boundary in the South Island of New Zealand. Journal of Structural Geology, 12, 715-725. ODONNE, E & VIALON, P. 1983. Analogue models of folds above a wrench fault. Tectonophysics, 9 9 , 31-46. OLDOW, J. S., BALLY,A. W., AVE LALLEMANT,H. G. & LEEMAN,W. P. 1989. Phanerozoic evolution of the North American Cordillera: United States and Canada. In: BALLY,A. W. & PALMER,A. R. (eds) The Geology of North America: an Overview. Geological Society of America, Boulder, CO, 139-232. , & - 1990. Transpression, orogenic float, and lithospheric balance. Geology, 18, 991-994. (~)ZALAYBEY, S. ~r SAVAGE, M. K. 1995. Shear-wave splitting beneath western United States in relation to plate tectonics. Journal of Geophysical Research, 100, 18135-18149. PAVLIS,T. L. & SISSON,V. B. 1995. Structural history of the Chugach metamorphic complex in the Tana River region, eastern Alaska: a record of Eocene ridge subduction. Geological Society of America Bulletin, 107, 1333-1355. PETERSON, K. & TIKOFF, B. 1995. An experimental approach to folding during transpression. Geo-

logical Society of America, Abstracts with Programs, A220-221. PEZZOPANE, S. K. & WELDON,R. J. 1993. Tectonic role of active faulting in central Oregon. Tectonics, 12, 1140-1169. PRESCOTT,W. H. & Yu, S.-B. 1986. Geodetic measurement of horizontal deformation in the northern San Francisco Bay region, California. Journal of GeophysPcal Research, 91, 7475-7484. RATSCHBACHER,L. 1986. Kinematics of austro-alpine cover nappes: changing translation path due to transpression. Tectonophysics, 125, 335-356. RICHARD, P. & COBBOLD, P. 1990. Experimental insights into partitioning of fault motions in continental convergent wrench zones. Annales Tectonicae, IV, 35-44. ROBIN, R-Y .E & CRUDEN,A. R. 1994. Strain and vorticity patterns in ideally ductile transpression zones. Journal of Structural Geology, 16, 447-466.

158

C. TEYSSIER & B. TIKOFF

RYMER, M. J., LISOWSKI,M. & BUR_FORD,R. O. 1984. Structural explanation for low creep rates on the San Andreas fault near Monarch Peak, Central California. Bulletin of the Seismological Society of America, 74, 925-931. SALEEBY,J. B. 1986. C-2: Central California offshore to

Colorado Plateau, Centennial Continent~Ocean Transect, 10, 1:500 000. Geological Society of America, Boulder, CO. SANDERSON, D. & MARCHINI, R. D. 1984. Transpression. Journal of Structural Geology, 6, 449-458. SAVAGE, M. K. & SILVER, P. G. 1993. Mantle deformation and tectonics: constraints from seismic anisotropy in western United States. Physics of the Earth and Planetary Interiors, 78, 207-227. SCHWERDTNER, W. M. 1989. The solid-body tilt of deformed paleohorizontal planes: application to an Archean transpression zone. Journal of Structural Geology, 11, 1021-1027. SEDLOCK, R. L. & HAMILTON,D. H. 1991. Late Cenozoic tectonic evolution of southwestern California. Journal of Geophysical Research, 96, 2325-2351. SILVER,P. G. 1996. Seismic anisotropy beneath the continents: probing the depths of geology. Annual Review of Earth and Planetary Sciences, 24, 385.432. STRACHAN,R. A., HOLDSWORTH,R. E., FRIDERICHSEN, J. D. & JEr'SEN, H. E 1992. Regional Caledonian structure within an oblique convergence zone, Dronning Louise Land, NE Greenland. Journal of the Geological Society, London, 149, 359-371. SYLVESTER, A. 1988. Strike-slip faults. Geological Society of America Bulletin, 100, 1666-1703. TAVERNELLI,E. 1998. Tectonic evolution of the Northern Salinian Block, California, USA: Paleogene to Recent shortening in a transform faultbounded continental fragment. This volume.

TEYSSIER, C., KLEINSPEHN,K. L. & PERHING, J. 1995a. Analysis of fault populations in western Spitsbergen: implications for deformation partitioning along transform margins. Geological Society of America Bulletin, 107, 68-82. - - , TIKOFF,B. & MARKLEY,M. 1995b. Oblique plate motion and continental tectonics. Geology, 23, 447-450. TIKOFF,B. & GREENE, D. 1997. Stretching lineations in transpressional shear zones. Journal of Structural Geology, 19, 29-40. & PETERSON,K. 1998. A n experimental approach to folding during transpression. Journal of Structural Geology, in press. & TEYSSIER, C. 1994. Strain modeling of displacement-field partitioning in transpressional orogens. Journal of Structural Geology, 16, 1575-1588. TREAGUS,J. E. & TREAOUS,S. H. 1992. Transected folds and transpression: how are they related? Journal of Structural Geology, 14, 361-367. TwIss, R. J. & MOORES, E. M. 1992. Structural Geology. W. H. Freeman, New York. VAUCHEZ,A. & NICOLAS,A. 1991. Mountain building: strike-parallel displacements and mantle anisotropy. Tectonophysics, 185, 183-201. WITHJACK,M. O. & JAMISON,W. R. 1986. Deformation produced by oblique rifting. Tectonophysics, 126, 9%124. ZANDT, G. 1981. Seismic images of the deep structure of the San Andreas fault system, central Coast Ranges, California. Journal of Geophysical Research, 99, 9651-61. ZHANG, S. & KARATO, S.-I. 1995. Lattice preferred orientation of olivine aggregates deformed in simple shear. Nature, 375, 774-777. -

-

Exhumation of UHP rocks by transtension in the Western Gneiss Region, Scandinavian Caledonides MAARTEN

KRABBENDAM

& J O H N F. D E W E Y

Department of Earth Sciences, University of Oxford, Oxford OX1 3PR, UK *Present address: Department of Earth Sciences, Monash University, Clayton, VIC 3168, Australia (e-maik [email protected], edu.au)

Abstract: In the Western Gneiss Region (WGR) in the Scandinavian Caledonides, Scandian eclogites (P = 16 to >28 kbar) occur in a large area of reworked Proterozoic gneisses, structurally below a series of Scandian nappes. The top-to-the-west, extensional Nordfjord-Sogn Detachment (NSD) separates the WGR from allochthonous units, which include several late-orogenic Devonian basins. The allochthon has not experienced Scandian high-pressure (HP) metamorphism. Below the NSD, the WGR is intensely deformed under late-orogenic amphibolite-facies conditions. This deformation is bulk-constrictional, as indicated by a linear feldspar fabric within augen gneisses and tight to isoclinal, lineationparallel folds within layered gneisses. In a later stage, the NSD, the WGR and the Devonian basins were folded by east-west trending folds, coeval with continuing movement along detachments. To explain these features we propose a transtensional model for the late-orogenic evolution of the WGR. Transtension in West Norway had a sinistral sense and was partially partitioned with increasing transtensional angle towards the NE-SW trending MOreTrOndelag Fault Zone in the NW. During transtension, there is a strong tendency for rejuvenation of detachments, because detachments fold and may lock as they move. In the WGR, the younger Hornelen Detachment developed above the older NSD. Transtension was the principal exhumation mechanism of the HP and ultra-high-pressure (UHP) rocks in the WGR and involved oblique plate divergence of Laurentia and Baltica during the Early Devonian. Exhumation of high-pressure (HP) rocks can be achieved by a variety of mechanisms, including erosion, buoyancy, extrusion, corner flow and extension, with extension favoured for regional HP terranes (Platt 1993). Extensional exhumation has been proposed for many HP terranes, both in subduction and continent-continent collision settings. In most cases this extensional exhumation is thought to have occurred during plate convergence (Platt 1993). Broadly speaking, two mechanisms of extensional exhumation in a convergent setting have been proposed. First, exhumation by underplating and extension during convergence, in which continuous underplating at depth is accompanied by extension at the top of a wedge-shaped orogen or accretionary prism (Platt 1986). This has been proposed for the Franciscan Complex, the Alps, and the Himalayas (Platt 1986; R o y d e n & Burchfie11987; Wheeler 1991). Second, exhumation by extensional collapse, in which removal of the lithospheric root triggers a more sudden extension in the upper part of the orogen (Dewey 1988; Platt & England 1993). This has been proposed for the Tibetan Plateau, the Basin and Range, and the Betic Cordillera (Dewey 1988; England & Houseman 1988; Platt

& Vissers 1989). The driving forces of extension in both cases are body forces, so that either underplating or removal of the lithospheric root results in the vertical stress exceeding the horizontal stress of convergence, thus bringing the internal part of an orogenic system into extension. It is, however, clear from the geological record that convergence cannot continue forever, and that plate motions may change direction or even reverse relatively rapidly (Dewey 1988). In this paper we present an example from the Scandinavian Caledonides, where Devonian exhumation of HP and ultrahigh-pressure (UHP) rocks was achieved by transtension driven by sinistral oblique plate divergence of the Laurentian and Baltic Plates.

Geological setting of the Western Gneiss Region (WGR) The Western Gneiss Region (WGR) in West Norway is a large basement window within the Scandinavian Caledonides (Fig. 1). The W G R contains p r e d o m i n a n t l y granodioritic and granitic gneisses, most of which are heterogeneously layered, some with augen textures.

KRABBENDAM,M. & DEWEY,J. E 1998. Exhumation of UHP rocks by transtension in the Western Gneiss

159 Region, Scandinavian Caledconides. In: HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. F. (eds) 1988. Continental Transpressionaland TranstensionalTectonics.Geological Society, London, Special Publications, 135, 159-181.

160

M. K R A B B E N D A M & J. F. D E W E Y

EXHUMATION OF UHP ROCKS BY TRANSTENSION Minor amounts of HP granulite, anorthosite, quartzite, marble, ultrabasic inclusions and eclogite pods occur within the W G R (Bryhni 1966; Lappin 1966, Bryhni & Sturt 1985) indicating both igneous and sedimentary protoliths. The W G R gneisses are of Proterozoic origin, reworked by the continent-continent collision of Baltica and Laurentia during the Scandian Phase (Late Silurian) of the Caledonian orogeny (Kullerud et al. 1986). The Caledonian reworking of the W G R increases to the west (Dietler et al. 1985). The W G R is a huge metamorphic core complex covering some 35 000 km 2. In the south and east, it is overlain by a series of nappes of great extent (Gee 1978; Bryhni & Sturt 1985). The direction of mid to late Silurian thrusting was to the east and SE but the basal thrust ('Jotun Detachment') was reactivated by top-tothe-west extension (Fossen & Holst 1995; Milnes et aL 1997). The L~erdal-Gjende Fault Zone (LGFZ) is a top-to-the-west extensional structure and cuts the basal thrust - Jotun Detachment (Dietler et al. 1985; Milnes et al. 1997). Norton (1987) also identified the extensional RCragen Detachment farther to the east, underlying the small RCragen Basin of Early Devonian age (Steel et al. 1985), although Gee et al. (1994) argued that the extent and displacement along the RCragen Detachment is rather limited. In the north, the W G R is cut by the M0re-TrCndelag Fault Zone (MTFZ), a 300 km long strike-slip fault with a long and complicated history. During the Devonian, the MTFZ had a sinistral shear sense; Mesozoic reactivation had a dextral shear sense (Grcnlie & Roberts 1989; Torsvik et al. 1989; S6ranne 1992; GrCnlie et al. 1994). A series of small but deep Old Red Sandstone sedimentary basins of Late Silurian to Middle Devonian age occur along the MTFZ (Steel et al. 1985). The continuation of the W G R north of the MTFZ, usually referred to as the Vestranden Gneiss Complex (VGC), contains eclogite only rarely (Bryhni & Andr6asson 1985; Gilotti & Hull 1993). In the western part of the W G R (Fig. 2), eclogite-bearing gneisses are separated from allochthonous units by the Nordfjord-Sogn Detachment (NSD) (Norton 1987; Andersen & Jamtveit 1990). The allochthonous units contain

161

Precambrian gneisses, ?Late Precambrian metasediments, Ordovician ophiolite and island-arc fragments, Silurian sediments and melanges, and minor late Caledonian intrusions (Brekke & Solberg 1987; Dunning & Pedersen 1988; Furnes et al. 1989; Andersen et al. 1990). Some of these allochthonous units in the west have been correlated with nappes east of the WGR; for example, the Dalsfjord suite is correlated with the vast Jotun Nappe (e.g. Brekke & Solberg 1987). The allochthonous rocks do not contain eclogites or other HP rocks. The metamorphic grade is generally greenschist facies; in the northern part it reached amphibolite-facies conditions (Cuthbert 1991; Chauvet et al. 1992; Dewey et al. 1993; Dransfield 1994; Wilks & Cuthbert 1994). Thus, a very significant metamorphic break occurs across the NSD. Four Devonian basins occur within the hanging wall above the NSD, the biggest of which (the Hornelen Basin) is underlain by a second detachment, the Hornelen Detachment (Dewey et aL 1993). The Devonian sediments are weakly metamorphosed (sub-greenschist facies) and locally possess a cleavage in finegrained beds (Roberts 1983; Bryhni & Andr6asson 1985; Torsvik et al. 1986; Bee et al. 1989). Devonian sediments are thought to have been deposited while the basins moved westward over the detachments (Hossack 1984; S6ranne & S~guret 1987; S6guret et al. 1989). Clast provenance studies in the Hornelen Basin indicate that sediments were derived from the allochthonous units; not a single clast could be correlated with the W G R (Steel & Gloppen 1980; Cuthbert 1991). This implies that the W G R was not at the surface in Devonian times and that erosion did not unroof the WGR. Also, this allows a minimum estimate of westward displacement along the NSD of 12 km for the Kvamshesten area and 38 km for the Hornelen area. Scandian contractional structures are preserved in the Allochthons (Hartz et al. 1994; Osmundsen & Andersen 1994) and in the eastern part of the W G R (Dietler et al. 1985; Fossen 1992). In the western part of the WGR, late-orogenic extensional deformation structures are dominant and are the subject of this paper.

Fig. 1. Tectonostratigraphic map of Western Norway, showing swing in lineation and fold axis direction. Structural data after Chauvet & S6ranne (1989, 1994), Fossen & Rykkelid (1992), S6ranne (1992), Gilotti & Hull (1993), Fossen & Holst (1995) and the authors' observations. Isotherms of eclogite metamorphism after Griffin et al. (1985), UHP localities after Smith (1995) and Wain (1997). EDZ, Einarsdalen D6collement Zone; HD, HCybakken Detachment; Ho, Hornelen Basin; H~, H~steinen Basin; Kv, Kvamshesten Basin; LGFZ, L~erdal-Gjende Fault Zone; MTFZ, M0re-TrCndelag Fault Zone; NSD, Nordfjord-Sogn Detachment; R, RCragen Basin; RD, RCragen Detachment; So, Solund Basin; VGC, Vestranden Gneiss Complex.

162

M. KRABBENDAM & J. F. DEWEY

Fig. 2. Tectonostratigraphic map of western part of the WGR. Positions of Figs 4, 5, 6, 8 and 9 are indicated. EF, Eikefjord Fault; FF, FCrdefjorden; G, GrCndalen. Modified from Andersen & Jamtveit (1990), Dewey et aL (1993) and Andersen et al. (1994). Isotherms of eclogite metamorphism after Griffin et al. (1985); UHP localities after Smith (1995) and Wain (1997).

EXHUMATION OF UHP ROCKS BY TRANSTENSION

163

amphibolite-facies metamorphism (e.g. Heinrich 1982; Mcrk 1985; Jamtveit 1987), aided by strong deformation during late-orogenic extension (Dewey et al. 1993; Krabbendam & Wain 1997). The UHP province, although defined by a limited number of localities, has the same NE-SW strike as the Caledonian HP isotherms and is positioned in the most coastal areas (Figs 1 and 2), with the Totland eclogite pod (Wain 1997) as the most southern locality and the FjCrtoft diamond locality, NNE of Alestmd, as the most NE locality discovered so far (Dobrzhinetskaya et al. 1995). As yet, the relationships of the UHP province with the rest of the WGR are not clear (Wain 1997). The gneisses immediately surrounding the UHP eclogites are similar to the gneisses elsewhere, suggesting an in situ origin for the UHP eclogites. The pervasive amphibolite-facies metamorphism and associated high strain have obliterated any original contacts between the UHP and the HP terrane (Krabbendam & Wain 1997), so that if the UHP province is 'foreign' (e.g. Smith 1995), any juxtaposition of UHP rocks against HP rocks must have occurred before the late-orogenic amphibolite-facies metamorphism.

Caledonian P - T - t evolution The eclogite pods within the W G R indicate pressures of 15-24 kbar (Wain 1997) with a UHP province of coesite-bearing eclogites indicating pressures in excess of 28 kbar (Smith 1984, 1995; Smith & Lappin 1989; Wain 1997) centred around Stadlandet (Fig. 2). Micro-diamonds have been recorded in a gneiss from Fjcrtoft (Dobrzhinetskaya et al. 1995). Caledonian eclogite-facies isotherms strike NE-SW (Figs 1 and 2) with peak temperatures increasing to the NW (Griffin et al. 1985). It is commonly assumed that the peak metamorphic pressures increase in the same way. This is corroborated by peak pressures in the Sunnfjord area (P < 14-18 kbar, Griffin et al. 1985; Andersen & Jamtveit 1990; Chauvet et al. 1992) and in the Nordfjord area (P 20-24 kbar with _ 28 kbar in the UHP terrane (Fig. 3a) (Smith 1995; Wain 1997). In contrast, the gneisses surrounding the eclogites show predominantly amphibolite-facies metamorphism (Bryhni 1966; Chauvet et al. 1992; Dewey et al. 1993; Dransfield 1994). The 'normal HP' eclogites were metamorphosed in situ, the surrounding felsic gneisses having been retrogressed during the late-orogenic pervasive

T,b 0

0 I

I

200 [

I

400 I

I

600 I

I

800 I

I

1000

Ordovician

I

Silurian

440

430

I

I

SSOC

//

Early Devonian 420

I

sst

--,

\

mus.9

\

......

1 .... ,,,sionyet ]

] 80 q

410

I

Ringerike

I

ORS

390 I

basins,

380

370

I

I

time,Ma )

W. Norway

1 lmm/yr

hb,

cooling "g T

~

eclogitefacies ~(. shortening~ / ~ /

l

400

Mid. Dev LateDevonian

amphibolite

facies

..'~" ~ ' ~

~ m i ~

\ eclOgitefacies extension

100

l

. , TRANSPRESSION?

P, kbar

a

depth, km

g"

COLLAPSE? TRANSTENSlON?

depth, km

b

Fig. 3. (a) P - T - t diagram for the Sunnfjord and Nordfjord areas in the WGR. (b) Depth-time evolution of the WGR. Approximate exhumation rates are indicated. Data after Griffin & Brueckner (1980, 1985), Bjcrlykke (1983), Smith (1984, 1995), Gebauer et al. (1985), Griffin et al. (1985), Lux (1985), Steel et al. (1985), Kullerud et aL (1986), Cotkin et al. (1988), Dunning & Pedersen (1988), Chauvet & Dallmeyer (1992), Chauvet et al. (1992), Koenemann (1993), Wilks & Cuthbert (1994), Krogh & Carswell (1995), Boundy et al. (1996) and Wain (1997).

164

M. KRABBENDAM & J. F. DEWEY

The timing of the orogenic evolution of the Scandinavian Caledonides is fairly well constrained, and indicates a relatively short event for the W G R (Fig. 3b). The youngest ophiolite age of 443 + 3 Ma is obtained in the Solund Stavfjord Ophiolite Complex (Dunning & Pedersen 1988) and indicates that full continent-continent collision must have occurred afterwards. Peak (U)HP metamorphism in the W G R is dated at 420-400 Ma (Griffin & Brueckner 1980, 1985; Gebauer et al. 1985; M~rk & Mearns 1986), and occurred during the Scandian continent-continent collision. Earlier ages of U H P metamorphism, dated at c. 1700 Ma, are restricted to eclogites enclosed within garnet peridotite bodies, which probably represent subcontinental mantle material incorporated into the W G R before or during the Scandian orogeny (Jamtveit et al. 1987). In the WGR, hornblende cooling ages (closing T ~ 500~ cluster around 410-395 Ma and muscovite cooling ages (closing T ~ 350~ cluster around 405-385 Ma (Lux 1985; Chauvet & Dallmeyer 1992; Boundy et al. 1996). Mineral cooling ages in the allochthonous areas are older (450-420 Ma, Berry et al. 1994; Boundy et al. 1996) so that the NSD also represents a geochronological break. Cooling ages in the W G R suggest very rapid cooling and exhumation, and indicate that by 390-380 Ma the W G R had been exhumed significantly and that the continental crust had returned to (almost) normal thickness (Fig. 3b). Thus, the Scandian continent-continent collision in Norway was a very short-lived event as the build-up a n d the destruction of the orogen were achieved in about 40 Ma. As a result, rates of burial and exhumation for the deepest rocks were fast, about 6-9 mm/a for the exhumation from c. 30 kbar to c. 10 kbar. The later part of the exhumation, from 410 to 375 Ma, slowed progressively to less than 0.2 mm/a (Wilks & Cuthbert 1994). Exhumation, therefore, occurred directly after peak pressure at the late stage of the orogenic cycle (Fig. 3b). This evolution is in contrast to many other HP terranes in the world, which were exhumed at an early stage of the orogenic evolution and are commonly overprinted by a later HT event and subsequent thrusting. This is exemplified by the allochthonous HP terranes in the Scandinavian Caledonides such as the Seve Nappe, Bergen Arcs and Tromsr Nappe, which overlie units with lower peak pressure (Roberts & Gee 1985; Van Roermond 1985; Boundy et al. 1997). HP metamorphism in the Seve Nappe has been dated at 505 Ma (MCrk et al. 1988) and in the Bergen Arcs at c. 460 Ma (Boundy et al. 1997),

i.e. well before the Scandian Phase, and cooling ages of both Seve Nappes and Bergen Arcs predate the eclogite metamorphism in the W G R (Dallmeyer & Gee 1986; Dallmeyer 1990; Boundy et al. 1996, 1997). The burial and exhumation of the Seve Unit were not the same events as the burial and exhumation of the W G R and occurred in a different tectonic setting (MCrk et al. 1988); this also applies to the Bergen Arcs (Boundy et al. 1997). In this paper we will focus on the exhumation of U H P rocks in the WGR, rather than on the exhumation of the allochthonous HP terranes.

Late-orogenic extension The NSD and its associated mylonite zones are the most conspicuous structures related to lateorogenic extension in the Caledonides and are among the most spectacular extensional detachments in the world in representing a geochronological and metamorphic break of at least 15 kbar, and are thus responsible for a vast amount of crustal excision-attenuation. In many places, eclogite occurs as close as 1 km below the Devonian sediments (Fig. 4) and, in the Nordfjord area, the Totland U H P pod (Wain 1997) occurs less than 8 km from the NSD. The NSD is folded along east-west trending, upright folds (Norton 1987; Chauvet & S6ranne 1994), with allochthonous units preserved in the synforms (Fig. 2). South of the H~steinen Basin, the NSD has a relatively simple geometry, with a single detachment separating the W G R in the Lower Plate from the allochthonous units in the Upper Plate (Andersen & Jamtveit 1990). North of the H~steinen Basin, the NSD bifurcates into the lower NSD and a complex system of other detachment faults, including the Eikefjord Fault, with the Hornelen Detachment as the highest detachment (Figs 2 and 4). It is convenient to separate the tectonostratigraphy into the Lower Plate (WGR) below the NSD, a Middle Plate (Middle Allochthon) between the NSD and the Hornelen Detachment, and the Upper Plate (Allochthon and Devonian) above the Hornelen Detachment (Dewey et al. 1993). To regard the whole of the Middle Plate as a single mylonitic detachment zone (Wilks & Cuthbert 1994) is too simple, because it contains both high- and lowstrain zones. In the central part of the WGR, the dominant late-orogenic extension is east-west and is responsible for the exhumation of HP rocks. Extensional structures show top-to-the-west sense of shear (at greenschist to amphibolite facies) close to the NSD (Andersen & Jamtveit

EXHUMATION OF UHP ROCKS BY TRANSTENSION

165

W

E Mid-Lower Allochthon

minor extension in Ul:g)er AIIochthon

/

A

brittle HD

H o , ~ , Basin

i

, 3 0 km ,

,

~

mylonite complex

~

constriction in W G R

_..---~-).~___

semibrittle N S D A'

Fig. 4. Section A-A'. East-west schematic cross-section through the WGR, Allochthons and Hornelen Basin, showing 'pure-shear' and 'simple shear' in two dimensions. A plunge culmination occurs close to the Nordfjord-Sogn Detachment. NSD, Nordfjord-Sogn Detachment; HD, Hornelen Detachment. Black dots represent schematically the position of eclogites. (See Fig. 2 for location.) 1990; Dewey et al. 1993; Wilks & Cuthbert 1994). Andersen & Jamtveit (1990) showed that below the NSD deformation (at amphibolite facies) becomes progressively more coaxial (Fig. 4). Whereas in an east-west cross-section this still holds true, we will show in this paper, that, in three dimensions, the coaxial extension is constrictional, with east-west extension and vertical thinning accompanied by north-south contraction.

Constrictional structures at Standal A detailed structural study has been made along the well-exposed coast of Standal, SW of the Hgtsteinen Basin. At Standal, the NSD (or as it appears on the N G U (Norges Geologiske UndersOkelse) maps, the Standal Fault) dips north, whereas beneath the northern margin of the Kvamshesten Basin, the NSD dips south (Figs 2 and 5). The described section occurs in the core of an east-west trending anticline. We do not use $1 and $2 terminology, because the structures are developed in rocks with a very complicated Precambrian history; instead Sa, La, Sb, etc., is used. In general, Sa structures are amphibolite-facies-grade structures (biotite, muscovite, amphiboles, epidote-zoisite) whereas Sb are lower-grade structures that deform the amphibolite-facies fabrics. All these structures are related to late-orogenic extension.

Augen gneiss with linear fabric (La) The lowest structural level exposed at Standal is represented by granodioritic to granitic augen gneiss (Figs 5 and 6). Monomineralic K-feldspar and plagioclase augen and polymineralic quartzo-feldspathic augen occur in a fine-

grained, quartz-rich matrix with biotite, muscovite and minor epidote. The augen are strongly prolate, defining an almost perfect Lfabric (La) (Fig. 7). K-Feldspar and polymineralic augen 1-2 cm across are up to 20 cm long when seen along the lineation (aspect ratio 1:1-2:5-10). Plagioclase augen are smaller and show a higher aspect ratio (1 : 1-2:5-20), probably resulting from higher ductility during deformation. The lineation is defined also by quartz rods and micas. When viewed normal to the lineation, preferred orientations are absent or extremely poorly defined (Fig. 7a). Micas, when observed normal to the lineation, may be randomly orientated, or display a poorly defined folding, or wrap the feldspathic augen in circular fashion. These textures are indicative of an almost purely constrictional strain regime, approaching K = ~. Within the augen gneiss, strongly prolate bodies of dark green biotite-bearing amphibolite occur; these are up to i m across and at least 10 m long. The contacts of these bodies with the augen gneiss are intensely folded with fold axes parallel to the well-developed stretching lineation but with axial surfaces in all directions parallel to the lineation. Axial surface fabrics are very poorly developed. These structures were probably formed by contraction in all directions normal to the stretching lineation, i.e. by constriction.

Layered gneiss with abundant lineationparallel folds (Fa) Above the augen gneiss, finely layered gneiss occurs (Fig. 6). The layering is distinct and, typically, consists of 10-20% quartzo-feldspathic layers, 10-40% mafic layers, rich in amphiboles,

166

M. KRABBENDAM & J. F. DEWEY

Fig. 5. Structural map of the Standal area, indicating different strain regimes. (See Fig. 2 for location.)

biotite and epidote, and 50-80% intermediate layers, containing feldspar, quartz and biotite. This domain is characterized by an abundance of folds, axis-parallel to the lineation. The Fa folds are tight to isoclinal on a centimetre to metre scale and have amphibolite-facies assemblages as axial surface fabrics (Sa: biotite and amphibole). These folds are not sheath folds, because no curvature of hinge lines is present. This domain contains strongly prolate bodies of amphibolite and anorthosite with intensely folded contacts, identical to the amphibolite bodies within the augen gneiss. Occasional layers of augen gneiss included within the layered gneiss show the same linear fabric as the larger occurrence of augen gneiss described above. On the basis of the similarity of the shape of the amphibolite bodies and the occurrence of La-fabric augen gneiss within the layered gneiss, it is concluded that the layered gneiss was subjected to the same constrictional strain as the Lafabric augen gneiss, with contraction in the y and z direction causing the intense lineation-parallel folding. The strong pre-existing anisotropy in the layered gneisses precluded the formation of

unequivocal L-tectonites. Lineation-parallel folds are not diagnostic for constriction, but they can be explained by it and we believe that the lineation-paraUel folds formed in a constrictional strain field.

Layered mylonitic gneiss with planar fabric (Sa) Lithologically, this zone is identical to the folded layered gneiss but is characterized by the absence of tight folding and the presence of structures indicative of strong non-coaxial strain. In feldspar-rich layers, asymmetrical feldspar augen (or clasts) are common and show a top-to-thewest sense of shear. Extensional crenulation cleavages with the same sense of shear are abundant, in particular in more mafic layers. Occasional layers of almost pure quartz show a very strong ribbon texture. Amphibolite and anorthosite bodies occur as sheets rather than as prolate bodies. The anorthosite sheets (from 15 cm to tens of metres thick) show pinch-and-swell structures (Fig. 6). Anorthosite may show a mylonitic fabric but a centimetre-spaced fracture

EXHUMATION OF UHP ROCKS BY TRANSTENSION I" / ~

167

S

\\~

"

/

arnDh'b~ greenschist-facies~ ~ N

~

/

----. ~\~

.v.v.. . . . .

mylonites; S>L~

[~ . ~

layered gneiss; L>S

~

~ ~

layered gneiss; S = LI ~ ' ~

/

p

o

r/

~ ~ ~ ~ a

~-~ ~

augen gneiss; L>>S ~ ' - - - -

~

~

~

0 d

o

l

a

t

~ -

n

s

o

h

r

t

a

p

h

o

e

s

/

.~--.

t

9-o .\

-

Fsp-augen . \\ " r , v / ~'

"

bundantfolding II L"

e

amphibolite bodies

d

i

3~__~ ~1_ i~,, /:,37/

~~~/-'~'~///

e

/

/

pinch & swell anorthosite sheets

Fig. 6. Schematic block diagram of part of the Nordfjord-Sogn Detachment, Standal. View towards the SE. Not to scale. cleavage is dominant in places. Top-to-the-west extensional crenulation cleavages are abundant within the anorthosite sheets.

Mylonite, chlorite breccia and cataclasite The zone of mylonitic gneiss gradually passes into a zone characterized by both ductile and brittle structures directly below the Standal Fault (NSD) (Figs 5 and 6). This zone consists of dark grey-green to black, fine-grained semipelitic to quartzo-feldspathic schistose greenschist-facies mylonite to ultramylonite with occasional small (millimetre to centimetre) augen of quartz and feldspar and/or a finely spaced lamination defined by <1 mm wide quartz-rich layers. Cataclastic structures overprint the mylonitic to ultramylonitic structures and consist of dense networks of late slip planes coated with chlorite, domains of extremely fine-grained fault gouge and thin cataclastic layers with very dark, glassy (?pseudotachylitic) margins. Broken fragments within the cataclastic layers contain mylonitic foliation, now rotated by subsequent cataclasis. Such cataclastic overprinting of mylonite and ultramylonite is typical of classic extensional detachments (Lister & Davis 1989).

The N S D - Standal Fault The brittle Standal Fault is not exposed near Standal but its trace can be constrained to within 5-20 m. It is probably very similar in appearance to the Dalsfjord Fault, the section of the NSD

which is well exposed when traced around the Kvamshesten Basin (Torsvik et aL 1992). The fault trace at Standal appears to be steeper than the mylonitic fabric below. This small angle (about 10-20 ~) implies that the brittle faulting occurred later and cross-cut the more ductile shearing. Torsvik et al. (1992) suggested that the brittle NSD was reactivated during the Permian-Mesozoic, based on palaeomagnetic dating of fault rocks.

Structures in the allochthonous Upper Plate The allochthonous Upper Plate at Standal consists of dark grey calcareous mica-quartz schist, with occasional marble bands, feldspathic bands and quartz sweats. Top-to-the-west shear sense indicators include: widely spaced (5-10 cm) sets of extensional crenulation cleavage, asymmetrical wrapping of quartz porphyroclasts by the micaceous foliation and asymmetrical quartz fibre growth in the pressure shadows of pyrite crystals. This extension occurred in the semibrittle regime. No true mylonitic texture is developed, feldspathic bands are not sheared, and quartz occurs as porphyroclasts rather than behaving in a ductile manner. Therefore, the conditions of deformation were much cooler than below the NSD.

S u m m a r y o f structures at Standal At Standal, linear augen gneisses (La) are overlain by layered gneisses with abundant

168

M. KRABBENDAM & J. F. DEWEY W G R are layered, and the tight lineation-parallel folding as described above is found throughout most of the western part of the W G R and also in the Middle Plate.

Late east-west trending folds (Db) East-west trending Fb folds, overprinting the amphibolite-facies structures described above, occur throughout the W G R gneisses but also affect the NSD, the Middle Allochthon, the Devonian basins and the Hornelen Detachment (Fig. 8). The folds overprint or fold the earlier Fa structures. Axial surface fabrics are generally poorly developed and can be seen only in the hinge zones of close to tight Fb folds in the more micaceous lithologies where a crenulation cleavage (Sb) is developed, overprinting an earlier mica fabric (Sa). It has been claimed that these Fb folds are always parallel to the La mineral lineation (Chauvet & S6ranne 1994). Although statistically this is correct, in many places the La mineral lineation can be seen to be folded around Fb folds, in particular north of Nordfjord, attesting to the overprinting of Fb on Fa.

Fig. 7. L-Fabric in augen gneiss, indicating constriction, Standal. (a) View parallel to L: virtually no fabric elements are present in this section. View to the east, pencil is 13 cm long. (b) View normal to L: a strong L-fabric of K-feldspar augen and plagioclase. Top to the north, hammer is 40 cm long. lineation-parallel folds (Fa) (Figs 5 and 6). These structures are interpreted as having formed by constrictional strain during amphibolite-facies conditions. This constrictional domain is overlain by mylonitic gneiss, mylonites and ultramylonites, interpreted as strongly non-coaxial during conditions varying from amphibolite to greenschist facies (Db). The discrete, brittle Standal Fault cross-cuts the mylonites at a small angle. Semi-brittle extensional structures occur in the Upper Plate.

Constrictional fabrics at other localities Linear or strongly prolate augen fabrics are also observed in W G R augen gneisses along Nordfjord and in Middle Allochthon augen gneisses north of Hyen; all augen gneisses observed by us show constrictional fabrics. L > S fabrics have also been reported from the VGC (Gilotti & Hull 1993) and S6ranne (1992) reported prolate mylonitic fabrics close below the HCybakken Detachment, just north of the M T F Z (Fig. 1). However, augen gneisses make up only a minor part of the WGR. The bulk of the gneisses in the

East-west trending folds (Fb) within the WGR The folds within the W G R gneisses are generally 1-100 m scale, open to close folds that fold the penetrative amphibolite-facies fabric (Sa) and gneissic layering and overprint the lineation-parallel (Fa) folds described above. Fb fold axes are subparallel to but, in places, fold the regional amphibolite-facies lineation La. Most of the Fb folds are upright but many recumbent folds also occur, in particular on either side of Nordfjord. Commonly recumbent and upright sets occur very close together, without displaying overprinting relationships. The intensity of Fb folding varies greatly; along F0rdefjorden and Nordfjord, the folding led to a north-south shortening of the order of 30-50%. Elsewhere, as on Stadlandet, such folds are open or absent (Fig. 8a). Further north, close to the MTFZ, the fold intensity increases again (S6ranne 1992; Robinson 1995). A detailed down-plunge section has been constructed from excellent and continuous roadcuts between Naustdal and Fcrde (Fig. 8b). Here, two sets of later folds occur, one set dips to the north, the other to the south. No overprinting relationships have been found. Restoration of these folds indicates a n o r t h - s o u t h shortening of 12-29%. Such folding may be expected if contraction was occurring by constriction normal to the fold axes.

EXHUMATION OF UHP ROCKS BY TRANSTENSION

169

N - S SHORTENING N

Slsdllar~iet

0

< 15%

} ,,'- 1~-...-50%,'~,11., NOrdflord ~ , "-e~

~20% ~,~ .....

26~ n

ere

~

H&stelnen

Kvsmhesten

~ ===~~ ~ . ~ . . ~ - - ~

8olund

I soon~f|erd 8 -, ~ I

25km

_

B

-

~

DevonianBasins

~

Western GneissRegion

~7

UFfPeclogites

--

AIIochthonous

a

~ i ~

.

"

N-S shortening: 12-29%

~//,~., L~~176

b

Fig. 8. (a) Section B-B'. North-south cross-section across the WGR, showing approximate amount of north-south shortening during east-west extension. Partly after Kildal (1970) and Osmundsen (1996). North of Hornelen Basin: own observations. (b) Down-plunge projection normal to the stretching lineation (090/30 to E) near Naustdal, showing late folding, coeval with extension.

East-west trending folds in the Middle Allochthon

East-west folding of the Devonian sediments

On the eastern side of the Hornelen Basin, the steep Hyenfjord offers a good opportunity to study east-west trending folding in the Middle Plate between the Hornelen Detachment and the NSD. The Allochthon rocks here consist of gneisses (including augen gneiss), quartzite, anorthosite and minor calc-silicates. Despite its relatively high structural position the Middle Plate has experienced much extension under amphibolite-facies conditions (Wilks & Cuthbert 1994). As a result, many of the extensional structures are very similar to those in the underlying WGR: mylonitic gneiss with top-to-thewest shear sense, a strong east-west lineation, locally intense amphibolite-grade lineationparallel folding and an L >> S fabric in the sporadic augen gneisses. A detailed cross-section along the new road along the eastern shore of Hyenfjord has been made (Fig. 9); outcrops on the steep cliffs on the western shore have been studied with the aid of binoculars from a high viewpoint on the opposite fjord wall. The later folds (Fb) are very well developed along Hyenfjord. Generally, they are close folds with wavelength varying from 10 cm to 10 m and are parasitic to larger open to close folds with a wavelength of 1-4 km. Upright folds are dominant but, locally, 1-10 m scale recumbent folds occur. Total north-south shortening is estimated to be a minimum of 18%. Along Hyenfjord, the Fb folds are cut by the Hornelen Detachment; this can be clearly observed in the field below the GrCndalen Syncline (Fig. 9).

East-west folding also affects the Devonian basins, especially the Hornelen and Kvamshesten Basins. The Kvamshesten Basin and the NSD are folded in a single upright, kilometrescale syncline about the same size as the basin itself. North-south shortening across the Kvamshesten Basin is estimated to be 11-25% (Osmundsen 1996). In the Hornelen Basin, several kilometre-scale, open to close folds occur with most folding occurring close to the north and south margins, for instance the Grcndalen syncline in the south (Fig. 9). Bedding is locally vertical or overturned, as along the northern margin. The plunge of the folds is 20-30 ~ towards the east together with the overall east-dipping Devonian strata. The total north-south shortening of the Hornelen Basin along the Hyenfjord section is estimated to be about 12%. The GrCndalen Syncline and the folds near Skjerdal are cut by the Hornelen Detachment, indicating that north-south shortening occurred before the latest stages of top-tothe-west displacement (Fig. 9).

East-west folding of the Detachments The NSD is folded in east-west folds with wavelength of about 10 km, the Devonian basins generally occurring in the synclines (Figs 2 and 8). The total north-south shortening of the NSD as measured from Hornelen to Solund is c. 20-25 % (Fig. 8). The Hornelen Detachment, however, is an almost flat structure, with less than 2% north-south shortening along Hyenfjord (Fig. 9).

170

M. KRABBENDAM & J. F. DEWEY

S

Hyen

N

Nordfjord

Gronclilen Syncline .,, ~ : ~

o~ o o : f=o oo ~ Hornelen ~ ~176 o 0~.~ ~176176 ~o~ o0 ~ o~=_;.~0~ o0 o0 o~ ~ o o 00 ~

Basin

o-,., -

C

Middle AIIochthonous mylonites

Linear a u g e n gneiss

2%

C w

with top to w e s t shear sense

Fig. 9. Section C-C'. South-north cross-section of eastern end of the Hornelen Basin, showing different amounts of north-south shortening of the Hornelen Basin, Hornelen Detachment (HD) and Middle-Lower Allochthons. Structural data in the Allochthons based on detailed fieldwork; structural data from Devonian and the HD based on binocular observation and reconnaissance fieldwork. Southern part of section partly after Bryhni & Lutro (1989). The northern part of the section as drawn is extrapolated from outcrops farther west, where the northern contact of the Hornelen Basin is exposed. (See Fig. 2 for location.)

This suggests that the Hornelen Detachment is a younger structure than the NSD and the mylonites in both the W G R and the Middle Plate.

Relative age o f north-south shortening The relative age of the north-south shortening resulting from the east-west trending folds (Fb) has been somewhat controversial and has been connected with the elusive Solundian-Svalbardian Orogeny (e.g. Torsvik et al. 1988). The relationship between north-south shortening and Devonian deposition is, as yet, also unclear. Bryhni & Skjerlie (1975) argued for basin formation during north-south shortening for the Kvamshesten Basin. This was further supported by Chauvet & S6ranne (1994), who claimed that all the east-west folds of the Devonian rocks and the detachment were syn-depositional, based on different types of unconformities as interpreted from aerial photographs of the Kvamshesten and Hornelen Basins (S6ranne et al. 1989). Fieldwork by Osmundsen (1996) has not confirmed the existence of the unconformities in the Kvamshesten Basin. Osmundsen (1996) argued that most of the north-south contraction in the Kvamhesten Basin is late- to post-depositional. On the other hand, reconnaissance fieldwork near the southern margin of the Hornelen Basin in Grcndalen strongly suggests that the complementary anticline south of the Grcndalen Syncline contains an angular unconformity on its north limb and is syn-depositional. The Hyenfjord cross-section (Fig. 9), however, displays the relevant field relationships,

confirming the mapping of Bryhni & Lutro (1989). The Hornelen Detachment cuts the Grcndalen Syncline and other folds in the Devonian rocks near Skjerdal. Regardless of whether these folds are syn-depositional (which is not of particular relevance here), they clearly pre-date substantial movement along the Hornelen Detachment. The south bounding fault of the Hornelen Basin cuts the Hornelen Detachment (Bryhni & Lutro 1989) and is, locally, steeper than bedding and, therefore, post-dates both deposition and extensional movement along the Hornelen Detachment. Significantly, the north-south shortening by Fb folds of the gneissic layering (>18%) exceeds the Fb north-south shortening of the Hornelen Basin (c. 12%), which in turn exceeds the Fb north-south shortening of the Hornelen Detachment (c. 2%) (Fig. 9). This suggests the following evolution for the Hyenfjord area: (1) folding of gneissic layering and mylonites, including the underlying NSD; (2) deposition of Devonian basin, at that moment positioned farther east; (3) folding of Devonian basin; (4) emplacement of Devonian basin along the Hornelen Detachment; (5) further folding of the Hornelen Detachment, the Hornelen Basin and the underlying gneisses (about 2%). This evolution strongly suggests folding, possibly progressive, before and during Devonian sedimentation and during top-to-the-west movement along the Hornelen Detachment. Progressive folding, before and after faulting along the NSD, was also suggested by Torsvik et al. (1986) for the Kvamshesten Basin. We suggest that north-south shortening affecting the Devonian basins is related in time to the regional east-west

EXHUMATION OF UHP ROCKS BY TRANSTENSION

171

extension, indicating bulk constriction. In summary, north-south contraction was synchronous with vertical thinning and east-west extension for a considerable period during the late-orogenic evolution of the WGR, indicating late-orogenic bulk constriction.

excision or attenuation from the Hornelen Basin to the U H P province on Stadtlandet is at least 80 km, this implies that at least 80% of the exhumation was achieved by bulk constriction and only 20% by excision along the NSD (see also Krabbendam & Wain 1997).

Orientation of lineations and fold axes

Interpretation and discussion

Lineations and fold axes, derived from various sources, show a regional swing in the plunge direction within the W G R (Fig. 1). To the south, the lineations are W N W - E S E plunging, in the central part the plunge direction is east-west, whereas close to the MTFZ the lineations and fold axes are NE-SW, subparallel to the MTFZ. This swing in orientation affects both Fa and Fb structures. Fb orientations also swing in areas where close or tight Fa structures are absent, so that the influence of pre-existing corrugations of Fa on the orientation of Fb can be regarded as insignificant.

To exhume the U H P and HP rocks in the W G R in Norway, some 100 km of overburden must have been removed from a restricted area (the U H P province), and some 50-60 km of overburden over a large area (>10 000 km2). The widespread occurrence of extensional structures, the very significant metamorphic break across the NSD, the scarcity of Devonian sediments in the foreland regions of the Scandinavian Caledonides and the presence of intramontane Devonian basins above extensional detachments indicate that the W G R was exhumed mainly by extensional tectonics with only a minor contribution by erosion. To explain the constrictional fabrics, the syn-extensional north-south shortening and the Devonian strike-slip movement along the MTFZ we present below a model of sinistral transtensional exhumation in the WGR.

Strain estimation In the absence of stratigraphic or other displacement markers, strain is hard to estimate in the WGR. In the western part of the WGR, vertical shortening has been estimated by Dewey et al. (1993) to be about 0.2 (80%) on the basis of flattening of foliation around eclogite pods and shortening of granite and quartz veins in Sunnfjord, and by assuming (from eclogite metamorphic assemblages) an original crustal thickness of c. 150 km that has been reduced to the present c. 30 km. The linear feldspar augen at Standal have aspect ratios of 1 : 1 - 2 : 5 - 1 0 , indicating minimum vertical shortening of 0.1-0.2 (80-90%), concomitant with similar horizontal north-south shortening. As described, the north-south shortening responsible for the Fb structures varies between <5% and 50%. The bulk finite strain ellipse has values of about x : y : z -- 10:0.5:0.2 to 5.3:0.95:0.2. The minimum displacement along the NSD is constrained by the distance between the eastern end of the Devonian basins and the points where the unconformities on the western side meet the NSD. This minimum displacement is 12 km for the Kvamshesten Basin and 38 km for the Hornelen Basin (Fig. 2). It is unlikely that the Upper Plate originated from (much) further east than the present position of the Jotun Detachment, thus providing a maximum displacement of about 100 km. If the detachments originated with the same dip as they have today (c. 10 o), this would mean a vertical displacement of minimum 6.5 km and maximum 17 km. As the

B o u n d a r y - c o m p a t i b l e transtension: s o m e basic characteristics It is sufficient, here, to briefly state the most important characteristics of transtension and indicate some structural features. Transtension is regarded in this paper as volume constant, homogeneous and boundary compatible. Transtension is a combination of plane strain extension and non-coaxial strain (strike-slip) (Fig. 10). The relative amounts of the plane s~rain extension component and the strike-slip component can be expressed by the transtensional angle, which is the angle between the transport direction and the coaxial extension component. The transtensional angle increases with increasing strike-slip component. The K value in a Flinn diagram is directly related to the transtensional angle. Pure constriction (x >y -- z, K = oo) occurs at a transtensional angle of 70.5~ at all other transtensional angles 'general constriction' occurs (x > 1 > y > z, 1 < K < ~). At angles smaller than 70.5 ~ vertical shortening exceeds horizontal shortening (z is vertical) whereas, at angles greater than 70.5 ~, horizontal shortening exceeds vertical shortening (y is vertical) (Fig. 10). In sinistral transtension, the transport direction (TD) is anticlockwise to the finite stretching direction (x), which is in turn

172

M. KRABBENDAM & J. F. DEWEY ~.

transport

(180-t8)>70.5~

8

Fig. 10. Volume-constant, homogeneous, boundary-compatible transtension. Transtensional angle is (180 -13). (a) Transtensional angle >70.5~ horizontal shortening exceeds vertical shortening and z is horizontal. (b) Transtensional angle <70.5~ vertical shortening exceeds horizontal shortening and y is horizontal. z Av

transport direction 1

J~

~

~

__

/ ~j~ .~ ~~" ~~ ~

constrictional shzone ear

Cz ~ Cz ~ (exhumation)

Fig. 11. Strain and displacement indicators of sinistral transtension. Indicators include: exhumation by bulk vertical shortening (Cz ~ Cz'); L>>S fabrics parallel to x; boudinage and pinch and swell structures with long axes parallel to x; recumbent folding normal to z; upright folding normal to y; late dykes and normal faulting normal to xi; folding of detachment parallel to the transport direction (TD); faulting and block rotation along axes normal to x. anticlockwise to the i n s t a n t a n e o u s stretching direction (xi) (Figs 10 and 11). T h e following strain a n d d i s p l a c e m e n t indicators are to be

expected to be p r o d u c e d by transtension (Fig. 11). B u l k transtension leads to vertical shortening and h e n c e to exhumation. Prolate boudins

EXHUMATION OF UHP ROCKS BY TRANSTENSION 9

9

detachment A

9

r

9

9

m

173 9

Y\

.

C1

\r a

/~m~

detachment A'

C2

b detachment A" (inactive),=

Fig. 12. Schematic crustal sections normal to finite stretching direction (x), indicating crustal thinning (C1 to C3), vertical and horizontal shortening normal to x and rejuvenation of detachment. Transtensional angle <70.5~ (a) Development of detachment A. (b) Detachment A folds. (c) Detachment A locks; a new detachment (B) develops.

and L>>S fabrics form parallel to x (finite stretching direction). Pinch-and-swell structures may develop at low transtensional angles parallel to the xy-plane with long axes parallel to x. Both recumbent and upright folding are expected with axes parallel to x. Late dykes and normal faulting may occur normal to xi. Block rotation associated with faulting will have axes of rotation normal to x. A crustal cross-section normal to the finite stretching direction x (Fig. 12) shows that during bulk transtension the crustal thickness can be reduced dramatically, coeval with bulk horizontal shortening normal to x. If a low-angle extensional detachment zone develops (vorticity partitioning in the xz plane) it must fold and extend as it moves (Fig. 12). There is a strong tendency for the detachment to lock itself, not only as it increases its surface area per unit volume as a result of folding but also because of the vertical rotation of the finite stretching direction with progressing deformation, so that subsequent displacement has to

move over, rather than along, the hinges. Rejuvenation of detachments is, therefore, to be expected (Fig. 12).

Partially partitioned, sinistral transtension in the W G R To explain the structures described in this paper, the following exhumation model for the W G R is proposed (Figs 13 and 14). Transtension in the W G R was sinistral with the strike-slip component parallel to the MTFZ (NE-SW). The transtension was partially partitioned, with an increase in transtensional angle towards the NW. The increasing transtensional angle is indicated by the regional swing in the mineral lineations and fold axes' azimuths from N W - S E close to the Jotun Detachment, W N W - E S E around Solund, east-west in the study area to W S W - E N E to S W - N E close to the MTFZ (Fig. 1). Variable transtensional angles are indicated also by the variable intensity of the late east-west folds as described above (Fig. 8).

174

M. KRABBENDAM & J. F. DEWEY

Fig. 13. Schematic model of partially partitioned sinistral transtension in West Norway. (See text for explanation.) JD, Jotun Detachment; HD, HCybakken Detachment; LGFZ, L~erdal-Gjende Fault Zone; MTFZ, MCre-TrCndelag Fault Zone; NSD, Nordfjord-Sogn Detachment; R, R0ragen Basin; RD, RCragen Detachment; VGC, Vestranden Gneiss Complex; WGR, Western Gneiss Region.

EXHUMATION OF UHP ROCKS BY TRANSTENSION The transtensional angle was close to zero (near plane strain extension) near the present east edge of the Jotun Nappe (Fig. 13), as indicated by the approximate down-dip, top-to-theNW movement along the Jotun Detachment and the Lerdal-Gjende Fault Zone (Milnes et al. 1988, 1997; Fossen & Holst 1995) and by limited lineation-parallel folds in this area, suggesting limited horizontal shortening normal to the extension direction. In the study area around the Devonian basins, the transtensional angle is c. 40-50 ~, compatible with the values of the strain estimation of bulk-constriction. Towards the MTFZ, the transtensional angle approaches 90~ (near strike-slip) towards the MTFZ (Fig. 11) supported by intense, upright folding and linear fabrics, subparallel to the MTFZ, in this area (see below). In an early, hot stage and/or deeper part of the orogen, transtension was taken up by relatively homogeneous constriction, producing linear fabrics in augen gneisses and the finely spaced gneissic layering and the tight, lineation-parallel folding (Fa) of this layering. In a later, cooler, shallower part of the orogen, strong vorticity partitioning in the xz-plane took place, with the development of the distinct extensional NSD in the upper crust. The NSD folded while it was moving to the west, and rejuvenation of the detachment took place to the north of the H~steinen Basin, thus developing the Hornelen Detachment. In this late stage, north-south partitioning of the transtension was possibly more pronounced with the development of distinct NE-SW trending strike-slip faults, such as the sinistral MTFZ and subsidiary sinistral strikeslip fault zones. It was in the cooler parts of the orogen that conspicuous mylonite zones (both underlying the NSD and elsewhere in the WGR) were developed, although the bulk of the extension was taken up by the more homogeneously distributed strain of the earlier, hotter phase. The syn-tectonic low-grade metamorphism of the Devonian basins above the W G R can be explained by the juxtaposition of these basins on top of warm lower crust along the NSD and the Hornelen Detachment; the cleavage formation within the basins was caused by the north-south shortening accompanying transtension. Within the Hitra and Smr Devonian basins, positioned along the MTFZ, syn-metamorphic (sub-greenschist-facies) structures include upright and recumbent ENE-WSW trending folds, with associated cleavages (Bee et al. 1989). Locally, two cleavages are developed, with the younger cleavage anticlockwise to the older cleavage (Bee et al. 1989). These structures are compatible with partitioned transtension near the MTFZ, although the structural evolution in

175

this area may be more complex because of the suggested docking and rotation of the Smr terrane (Torsvik et al. 1989). Transtension north o f the M T F Z : the Vestranden Gneiss C o m p l e x

North of the MTFZ, a similar partitioned transtension took place with the transtensional angle increasing south-eastward towards the MTFZ (Fig. 13). Close to the MTFZ, the Vestranden Gneiss Complex (VGC, Fig. 1) is characterized by upright NE-SW trending folds, (sub)-parallel to the lineation with rare new axial surface fabrics (Gilotti & Hull 1993). The gneisses show NE-SW trending, amphibolitefacies L-S and L > S fabrics. Numerous smallscale steep shear zones show ductile sinistral strike-slip movement (Gilotti & Hull 1993). Constrictional fabrics also occur below the Hcybakken Detachment (Fig.l, S6ranne 1992). All these structures are very similar to those observed in the western part of the WGR. In the NW of the VGC, the Roan Window (Fig. 1) contains HP granulite (T -- 870~ P = 14 kbar (Johannson & M611er 1986; MOiler 1988), with peak metamorphism dated at 432 _+ 6 Ma (Dallmeyer et al. 1992), i.e. early Scandian. The Einardsdalen D6collement Zone (EDZ) separates the high-grade Roan Window in the footwall from a lower-grade hanging wall with deformation structures developed at granulite facies but continuous fabric formation at lower T and P (Gilotti & Hull 1993). Although interpreted as a thrust by Gilotti & Hull (1993), the above features are more easily explained by extension rather than thrusting (Lister & Davis 1989; Wheeler & Butler 1994). The shear sense of the Einardsdalen D6collement is to the NW (Gilotti & Hull 1993), which is more compatible with the regional extension direction than the direction of overthrusting, which is to the SE (Fossen & Rykkelid 1992). We suggest, therefore, that the E D Z is a top-to-the-NW extensional detachment and not a thrust, and that the structures observed by Gilotti & Hull (1993) can be better explained by partitioned, sinistral transtension, and reject their transpressional interpretation. In the VGC area, close to the MTFZ, the transtensional angle was high, resulting in strongly constrictional strain (Fig. 13). Towards the NW, the transtensional angle decreases and is low in the area around the Roan Window, dominated by top-to-the-NW, near-plane strain extension. Thus the partitioning from low to high transtensional angle towards the MTFZ was more or less symmetrical, occurring on either side of the MTFZ.

176

M. KRABBENDAM & J. E DEWEY

increasing strike slip component (increasing transtensionalangle) ~o~e ~t

1

2

~ -

~~-I

_

I x - - - ~ t "-, _~_L~'.~.,..~\ "

~

orthogonal stretching component

~

Fig. 14. Schematicdiagramof partially partitioned transtension(volumeconstant,boundarycompatible). Partitioning of transtensJona]angleonly; orthogonalstretchingis kept constant. Transtension and oblique plate divergence between Laurentia and Baltica Of interest is whether the transtension in the WGR is associated with oblique, sinistral plate divergence between Laurentia and Baltica. Fossen (1992) argued that late-orogenic extension was accompanied by plate divergence, based on the very widespread evidence of lateorogenic extension in SW Norway and on the systematic overprint of contractional structures by extensional structures along the Jotun Detachment. The syn- or ?late-orogenic Ringerike Group sediments in the Oslo area are thrust towards the east (Bjcrlykke 1983); this thrusting, however, is not balanced by the amount of extension in the WGR and Central Norway, again suggesting plate divergence (Fossen 1992, 1993). Oblique plate divergence is supported also by the reinterpretation of the Caledonian sole thrust in central East Greenland as a top-to-theeast extensional detachment (Hartz & Andresen 1995), which separates Precambrian gneisses, reworked during the Caledonian, from Late Proterozoic and Carnbro-Ordovician meta-sediments. Above the detachment a small but deep Devonian basin occurs, bounded to the west by the Western Fault Zone, which is interpreted by some workers (Larsen & Bengaard 1991) as sinistral during the Devonian. It appears, therefore, that many of the late-orogenic, Devonian deformational features in the East Greenland Caledonides can be explained by sinistral transtension, which was strongly partitioned between extensional detachments and strikeslip faults. This suggests that the late-orogenic evolution of the Caledonides was relatively symmetrical. This study shows that the plate divergence

between Baltica and Laurentia during the Devonian must have been sinistrally oblique as the north-south shortening structures and the east-west constrictional fabrics in the WGR are not compatible with either orthogonal plate divergence or with convergent orogenic collapse. We suggest that Devonian plate divergence in Scandinavia was highly oblique and that transtension was the most important factor in the exhumation of the UHP rocks in Norway. Silurian convergence and collision of Laurentia and Baltica was also sinistrally oblique (Soper et al. 1992), as suggested by sinistral transpression in the Greenland Caledonides (Holdsworth & Strachan 1991), SE Scandian thrusting in the Scandinavian Caledonides (Fossen 1992) and large-scale sinistral movements between Laurentia and Avalonia in the British Caledonides (Dewey & Shackleton 1984; Hutton 1987; Hutton & McErlean 1991). We propose that this sinistral oblique convergence in the Silurian was followed by sinistral oblique divergence in the Devonian. Thus, the change in plate motion vectors between Baltica and Laurentia was less than 180~ possibly not exceeding 45~. The structures in the WGR, presented in this paper, were formed mainly under amphibolitefacies conditions, relatively late in the orogenic evolution. There is a gap in the structural record of about 5-10 Ma, between the formation of eclogite-facies structures (Andersen et al. 1991, 1994; Dewey et al. 1993) and the formation of the late-orogenic amphibolite-facies structures described in this paper (Fig. 3b). During this time it is possible that some form of orogenic collapse may have taken place, quickly followed by the transtension, for which the evidence is presented here. Whether such a collapse was caused by removal of the thermal boundary

EXHUMATION OF UHP ROCKS BY TRANSTENSION layer (Dewey 1988; Platt & England 1993) or by slab break-off (von Blanckenburg & Davies 1995) is uncertain, although some kind of such process is very likely to have taken place. Although the best argued examples of orogenic collapse happen to occur in plate convergent regimes (the Betic Cordillera (Platt & Vissers 1989) and the Himalayas (Dewey 1988; England & Houseman 1988), there is no reason to assume that orogenic collapse and convective removal of the lithospheric root cannot be accompanied by (or even result in) plate divergence.

Oblique plate divergence as an effective exhumation m e c h a n i s m This study shows that oblique plate divergence, expressed by bulk transtension over a wide area, can be a very effective mechanism to exhume a large (U)HP terrane. Various criteria can be put forward to recognize this mechanism in other collision belts: (1) Late-orogenic extension must exceed late-orogenic shortening. (2) Lateorogenic extension is likely to affect the entire crustal thickness. (3) Constriction, represented either by bulk-constrictional fabrics or by folded detachments, and other strain features characteristic of transtension are expected if plate divergence is oblique. (4) Age and structural position of HP units. HP m e t a m o r p h i s m is expected to be peak orogenic and exhumation is expected to be late orogenic and not followed by more convergence. As a result, HP units are likely to be the lowest tectonostratigraphic unit, with HP metamorphism relatively late in the orogenic evolution. If such criteria are not met, HP units are likely to be exhumed by other mechanisms, such as underplating and extension, erosion, extrusion or orogenic collapse during plate convergence. Subduction roll-back, which may lead to an extensional regime where many of the above criteria are met, evidently needs a subducting oceanic crust relatively close to the exhumed HP terrane. The observation of Platt (1993) that exhumation usually occurs while convergence is active, may be biased, as this observation is mainly based on studies of active or young mountain belts. We feel that although the statement may be true for many belts, its generalization is not appropriate. In other words, in 100-200 Ma, the Himalayas and the Alps may well look very much like the WGR, with the largest HP unit (currently still residing at depth) being exhumed by plate divergence, with some tiny, early orogenic HP units occurring at higher structural levels, exhumed while convergence was still active. In this context, it is worth noting

177

that most of the large exposed (U)HP terranes in the world are no longer in convergence.

Conclusions Transtension was the d o m i n a n t e x h u m a t i o n mechanism for (U)HP rocks in the WGR. This study shows that transtension is a powerful exhumation mechanism. Transtension in the W G R was partially partitioned with increasing transtensional angle towards the M T F Z and requires active oblique plate divergence between Laurentia and Baltica. A change in plate motion between Laurentia and Baltica from sinistral oblique convergence to sinistral oblique divergence took place at the end of the Silurian-Early Devonian. It remains uncertain whether this change in plate motion was associated with or even triggered by TBL removal or slab break-off. T. B. Andersen is thanked for many discussions and for taking M. K. up to critical exposures around the Hornelen Basin by helicopter. A. Wain is thanked for many discussions on the metamorphic evolution of the WGR. B. Sturt and an anonymous reviewer are thanked for comments on the manuscript. NERC funding for M. K. is acknowledged.

References ANDERSEN, T. B. & JAMTVEIT,B. 1990. Uplift of deep

crust during orogenic extensional collapse: a model based on field studies in the Sogn-Sunnfjord region of Western Norway. Tectonics, 9, 1097-1111. , JAMTVEIT,B., DEWEY,J. F. & SWENSSON,E. 1991. Subduction and eduction of continental crust: major mechanisms during continent-continent collision and orogenic extensional collapse, a model based on the south Norwegian Caledonides. Terra Nova, 3, 303-310. --, OSMUNDSEN, P. T. & JOLIVET, L. 1994. Deep crustal fabrics and a model for extensional collapse of the southwest Norwegian Caledonides. Journal of Structural Geology, 16, 1191-1203. --, SKJERLIE, K. P. & FURNES, H. 1990. The Sunnfjord Melange, evidence of Silurian ophiolite accretion in the West Norwegian Caledonides. Journal of the Geological Society, London, 147, 5%68. BERRY, H. N., Lux, D. R., ANDRESEN,A. & ANDERSEN, T. B. 1994. Argon 40-39 dating of rapidly uplifted high pressure rocks during late-orogenic extension in southwestern Norway. Geological Society of America, Abstracts' with Programs, 25 6, A-477. BJf~RLYKKE, K. 1983. Subsidence and tectonics in late Precambrian and Palaeozoic sedimentary basins in Southern Norway. Norges Geologiske UndersOkelse Bulletin, 380, 159-172.

178

M. K R A B B E N D A M & J. F. DEWEY

BOUNDY, T. M., ESSENE, E. J., HALL, C. M., AusTRHEIM, H. & HALLIDAY, A. N. 1996. Rapid exhumation of lower crust during continent-continent collision and late extension: evidence from 40Ar/39Ar incremental heating of hornblendes and muscovites, Caledonian orogen, western Norway. Geological Society of America Bulletin, 108, 1425-1437. - - , MEZGER, K. & ESSENE, E. J. 1997. Temporal and tectonic evolution of the granulite-eclogite association from the Bergen Arcs, western Norway. Lithos, 39, 159-178. BOE, R., ATAKAN,K. & STURT,B. A. 1989. The style of deformation in the Devonian rocks on Hitra and Sm01a, Central Norway. Norges Geologiske Undersokelse Bulletin, 414, 1-19. BREKKE, H. & SOLBERG, P. O. 1987. The geology of Atl0y, Sunnfjord, Western Norway. Norges Geologiske UndersOkelse Bulletin, 410, 73-94. BRYHNI, I. 1966. Reconnaissance studies of gneisses, ultrabasites, eclogites and anorthosites in Outer Nordfjord, Western Norway. Norges Geologiske UndersOkelse Bulletin, 241, 1-68. & ANDRI2ASSON, P. G. 1985. Metamorphism in the Scandinavian Caledonides. In: GEE, D. G. & STURT, B. A. (eds) The Caledonide OrogenScandinavia and Related Areas. John Wiley, New York, 763-782. & SKJERLIE, F. J. 1975. Syndepositonal tectonism in the Kvamshesten district (Old Red Sandstone), western Norway. Geological Magazine, 112, 593-600. -& LUTRO, O. 1989. Naustdal. Berggrunnsgeologisk kart 1218 III, scale 1 : 50 000. Forelcpig utgave Norges geologiske unders0kelse. & STURT, B. A. 1985. Caledonides of southwestern Norway. In: GEE, D. G. & STURT, B.A. (eds) The Caledonide Orogen - Scandinavia and Related Areas. John Wiley, New York, 89-107. CHAUVET, A. & DALLMEYER, R. D. 1992. 4~ mineral dates related to Devonian extension in the southwestern Scandinavian Caledonides. Tectonophysics, 210, 155-177. & St~RANNE,M. 1989. Microtectonic evidence of Devonian extensional westward shearing in SW Norway. In: GAYER, R. A. (ed.) The Caledonide Geology of Scandinavia, Graham & Trotman, London, 245-254. & -1994. Extension-parallel folding in the Scandinavian Caledonides: implications for lateorogenic processes. Tectonophysics, 238, 31-54. , KIENAST, J. R., PINARDON, J. L. & BRUNEL, M. 1992. Petrological constraints and PT path of Devonian collapse tectonics within the Scandian mountain belt (Western Gneiss Region, Norway). Journal of the Geological Society, London, 149, 383-400. COTKIN, S. J., VALLEY, J. W. & ESSENE, E. J. 1988. Petrology of a margarite-bearing metaanorthosite from Seljeneset, Nordfjord, western Norway: implications for the P-T history of the Western Gneiss Region during Caledonian uplift. Lithos, 21, 117-128. CUTHBERT, S. J. 1991. Evolution of the Devonian -

-

-

-

-

-

-

-

-

-

Hornelen Basin, Western Norway: new constraints from petrological studies of metamorphic clasts. In: MORTON, A. C., TODD, S. P. & HAUGHTON, P. D. W. (eds) Developments in, Sedimentary Provenance Studies. Geological Society, London, Special Publications, 57, 343-360. DALLMEYER, R. D. 1990.4~ mineral age record of a polyorogenic evolution within the Seve and Kr nappes, Tr0ndelag, Norway. Tectonophysics, 179, 199-226. & GEE, D. G. 1986.4~ mineral dates from retrogressed eclogites within the Baltoscandian miogeocline: implications for a polyphase Caledonian orogenic evolution. Geological Society of America Bulletin, 97, 26-34. , JOHANSSON, L. & MOLLER, C. 1992. Chronology of Caledonian HP-granulite-facies metamorphism, uplift and deformation within northern parts of the Western Gneiss Region. Geological Society of America Bulletin, 104, 444-455. DEWEY, J. F. 1988. Extensional collapse of orogens. Tectonics, 7, 1123-1139. & SHACKLETON,R .M. 1984. A model for the evolution of the Grampian tract in the early Caledonides and Appalachians. Nature, 312, 115-121. , RYAN, P. D. & ANDERSEN, T. B. 1993. Orogenic uplift and collapse, crustal thickness, fabrics and metamorphic phase changes: the role of eclogites. In: ALABASTER, H .M., HARRIS, N. B. W. & NEARY, C. R. (eds), Magmatic Processes and Plate Tectonics, Geological Society, London, Special Publications, 76, 325-343. DIETLER, T. N., KOESTLER, A. G. & MILNES, A. G. 1985. A preliminary structural profile through the Western Gneiss Complex, Sognefjord, southwestern Norway. Norsk Geologisk Tidsskrift, 65, 233-235. D OBRZHINETSKAYA,L. F., EIDE, E. A., LARSEN,R. B., ET AL. 1995. Microdiamond in high-grade metamorphic rocks of the Western Gneiss region, Norway. Geology, 23, 597-600. DRANSFIELD, M. W. 1994. Extensional exhumation of high-grade metamorphic rocks in Western Norway and the Zanskar Himalayas. PhD thesis, University of Oxford. DUNNING, G. R. & PEDERSEN, R. g. 1988. U/Pb dating of ophiolites and arc-related plutons of the Norwegian Caledonides: implications for the development of Iapetus. Contributions to Mineralogy and Petrology, 98,13-23. ENGLAND, P. C. & HOUSEMAN, G. 1988. The mechanics of the Tibetan Plateau. Philosophical Transactions of the Royal Society of London Series A, 326, 301-320. FOSSEN, H. 1992. The role of extensional tectonics in the Caledonides of south Norway. Journal of Structural Geology, 14, 1033-1046. 1993. The role of extensional tectonics in the Caledonides of south Norway: Reply. Journal of Structural Geology, 15, 1381-1383. -& HOLST, T. B. 1995. Northwest-verging folds and the northwestward movement of the Caledonian Jotun Nappe, Norway. Journal of Structural Geology, 17, 3-15. -

-

-

-

E X H U M A T I O N OF U H P ROCKS BY TRANSTENSION & RYKKELID, E. 1992. Postcollisional extension of the Caledonide orogen in Scandinavia: structural expressions and tectonic significance. Geology, 20, 737-740. FURNES, n . PEDERSEN, P. B., SUNDVOLL, B., TYSSELAND, M. & TYMYR, O. 1989. The age, petrography, geochemistry and tectonic setting of the late Caledonian GhsCy Intrusion, west Norway. Norsk Geologisk Tidsskrifi, 69, 273-289. GEBAUER, D., LAPPIN, M. A., GRUNENFELDER, M. & WYTTENBACI-I, A. 1985. The age and origin of some Norwegian eclogites: a U - P b zircon and R E E study. Chemical Geology, 52, 227-247. GEE, D. G. 1978. Nappe displacement in the Scandinavian Caledonides. Tectonophysics, 47, 393-419. , LOBKOWICZ, M. & SINGH, S. 1994. Late Caledonian extension in the Scandinavian Caledonides - the RCragen Detachment revisited. Tectonophysics, 231, 139-155. GILOTTI, J. A. & HULL, J. M. 1993. Kinematic stratification in the hinterland of the central Scandinavian Caledonides. Journal of Structural Geology, 15, 629-646. GRIFFIN, W. L. & BRUECKNER,H. K. 1980. Caledonian Sm-Nd ages and a crustal origin for Norwegian eclogites. Nature, 285, 319-321. & -1985. REE, Rb-Sr and Sm-Nd studies of Norwegian eclogites. Chemical Geology, 52, 249-271. , AUSTRHEIM, H. BRASTAD, K., ET AL. 1985. High-pressure metamorphism in the Scandinavian Caledonides. In: GEE, D. G. & STURT, B. A. (eds) The Caledonide Orogen - Scandinavia and Related Areas. John Wiley, New York, 783-802. GRONLIE, A. & ROBERTS, D. 1989. Resurgent strikeslip duplex development along the Hitra-Smr and Verran Faults, MCre-Trondelag Fault Zone, Central Norway. Journal of Structural Geology, 11, 295-305. , NAESNER, C. W., NAESER, N. D., MITCHELL, J. G., STURT, B. A. & INESON, P. R. 1994. Fissiontrack and K - A r dating of tectonic activity in a transect across the MCre-TrCndelag Fault Zone, central Norway. Norsk Geologisk Tidsskrift, 74, 24-34. HARTZ, E. & ANDRESEN, A. 1995. Caledonian sole thrust of central East Greenland: a crustal-scale Devonian extensional detachment? Geology, 23, 637-640. & ANDERSEN, T. B. 1994. Structural observations adjacent to a large scale extensional detachment zone in the hinterland of the Norwegian Caledonides. Tectonophysics, 231, 123-137. HEINRICH, C. A. 1982. Kyanite-eclogite to amphibolite facies evolution of hydrous mafic and pelitic rocks, Adula Nappe, Central Alps. Contributions to Mineralogy and Petrology, 81, 30-38. HOLDSWORTH, R. E. & STRACHAN,R. A. 1991. Interlinked system of ductile strike-slip and thrusting formed by Caledonian sinistral transpression in NE Greenland. Geology, 19, 510-513. HOSSACK, J. R. 1984. The geometry of listric growth faults in the Devonain basins of Sunnfjord, W. -

-

-

-

,

-

-

179

Norway. Journal of the Geological Society, London, 141, 629-637. HurroN, D. H. W. 1987. Strike-slip terranes and a model for the evolution of the British and Irish Caledonides. Geological Magazine, 124, 405-425. & MCERLEAN, M. 1991. Silurian and Early Devonian sinistral deformation of the Ratagain Granite, Scotland; constraints on the age of Caledonian movements on the Great Glen fault system. Journal of the Geological Society,

London, 148, 1-4. JAMTVEIT,B. 1987. Metamorphic evolution of the Eiksundal eclogite complex, Western Norway, and some tectonic implications. Contributions to Mineralogy and Petrology, 95, 82-99. , CARSWELL, D. A. & MEARNS, E. W. 1987. Chronology of the high-pressure metamorphism of Norwegian garnet peridotites/pyroxenites. Journal of Metamorphic Geology, 9, 125-139. JOHANSSON, L. d~;MOLLER, C. 1986. Formation of sapphirine during retrogression of a basic high-pressure granulite, Roan, Western Gneiss Region, Norway. Contributions to Mineralogy and Petrology, 94, 29-41. KILDAL, E. S. 1970. Geologisk kart over Norge, berggrunnskart. Mhloy, 1:250 000. Norges Geologiske UndersCkelse. KOENEMANN, F. H. 1993. Tectonics of the Scandian Orogeny and the Western Gneiss Region in southern Norway. Geologische Rundschau, 82, 696-717. KRABBENDAM, M. &; WAIN, A. 1997. Late-orogenic structures, differential retrogression and structural position of HP and U H P rocks in the Nordfjord Stadtlandet area, Western Gneiss Region. Norges Geologiske UndersOkelse Bulletin, 432, 127-139. KROC;H, E. J. & CARSWELL,D. A. 1995. HP and U H P eclogites and garnet peridotites in the Scandinavian Caledonides. In: COLEMAN,R. C. & WANG, I. (eds) Ultra High Pressure Metamorphism. Cambridge University Press. KULLERUD, L., TORUDBAKKEN, B. O. & ILEBEKK, S. 1986. A compilation of radiometric age determinations from the Western Gneiss Region, South Norway. Norges Geologiske UndersOkelse Bulletin, 406, 17-42. LAPPIN, M. A. 1966. The field relationships of basic and ultrabasic masses in the basal gneiss complex of Stadtlandet and Almklovdalen, Nordfjord, southwestern Norway. Norsk Geologisk Tidsskrift, 46, 439-496. LARSEN, P. H. & BENGAARD, H. J. 1991. Devonian basin initiation in East Greenland; a result of sinistral wrench faulting and Caledonian extensional collapse. Journal of the Geological Society, London, 148, 355-368. LISTER, G. S. & DAVIS, G. A. 1989. The origin of metamorphic core complexes and detachment faults formed during Tertiary continental extension in the northern Colorado River region, U.S.A. Journal of Structural Geology, 11, 65-94. Lux, D. R. 1985. K/Ar ages from the Basal Gneiss Region, Stadlandet area, Western Norway. Norsk Geologisk Tidsskrift, 65, 227-286.

180

M. K R A B B E N D A M & J. F. DEWEY

MILNES, A. G., DIETLER, T. N. & KOESTLER, A. G. 1988. The Sognefjord northshore log - a 25 km depth section through Caledonized basement in western Norway. In: KRISTOFFERSEN, Y. (ed.) Progress in Studies of the lithosphere in Norway. Norges Geologiske UndersCkelse Special Publication, 3, 114-121. - - , WENNBERG, O. P., SKAR, • & KOESTLER,A. G. 1997. Contraction, extension and timing in the South Norwegian Caledonides: the Sognefjord transect. In: BURG, J. P & FORD, M. (eds) Orogeny through Time. Geological Society, London, Special Publications, 121, 123-148. MOLLER, C. 1988. Geology and metamorphic evolution of the Roan area, Vestranden, Western Gneiss Region, Central Norwegian Caledonides. Norges Geologiske UndersOkelse Bulletin, 413, 1-32. MORK, M. B. E. 1985. A gabbro to eclogite transition on FlemsCy, SunnmCre, western Norway. Chemical Geology, 50, 283-310. & MEARNS, E. W. 1986. Sm-Nd isotopic systematics of a gabbro-eclogite transition. Lithos, 19, 255-267. - - . , KULLERUD, K. & STABEL, A. 1988. Sm-Nd dating of Seve eclogites, Norrbotten, Sweden evidence for early Caledonian (505 Ma) subduction. Contributions to Mineralogy and Petrology, 99, 344-351. NORTON, M. G. 1987. The Nordfjord-Sogn Detachment, W. Norway. Norsk Geologisk Tidsskrift, 67, 93-106. OSMUNDSEN, P. T. 1996. Late-orogenic structural geology and Devonian basin formation in Western Norway: a study from the hanging wall of the Nordfjord-Sogn Detachment in the Sunnf]ord Region. Doctor Scientiarum thesis, University of Oslo. & ANDERSEN, T. B. 1994. Caledonian compressional and late-orogenic extensional deformation in the Staveneset area, Sunnfjord, Western Norway. Journal of Structural Geology, 16, 1385-1401. PLATr, J. P. 1986. Dynamics of orogenic wedges and the uplift of high-pressure metamorphic rocks. Geological Society of America Bulletin, 97, 1037-1053. 1993. Exhumation of high-pressure rocks: a review of concepts and processes. Terra Nova, 5, 119-133. -& ENGLAND, P. C. 1993. Convective removal of lithosphere beneath mountain belts: thermal and mechanical consequences. American Journal of Science, 293, 307-336. & VISSERS,R. L. M. 1989. Extensional collapse of thickened continental lithosphere: a working hypothesis for the Alboran Sea and Gibraltar Arc. Geology, 17, 540-543. ROBERTS, D. 1983. Devonian tectonic deformation in the Norwegian Caledonides and its regional perspectives. Norges Geologiske UndersOkelse Bulletin, 380, 85-96. ROBINSON, P. 1995. Extension of Trollheimen tectonostratigraphic sequence in deep synclines near

-

-

-

-

-

-

Molde and Brattvfig, Western Gneiss Region, southern Norway. Norsk Geologisk Tidsskrifl, 75, 181-198. ROYDEN, L. & BURCnFIEL, B. C. 1987. Thin-skinned extension within the convergent Himalayan region: gravitational collapse of a Miocene topographic front. In: COWARD, M. P., DEWEY, J. F. & HANCOCK, P. L. (eds) Continental Extensional Tectonics. Geological Society, London, Special Publications, 28, 611-619. SEGURET, M., S~;RANNE,M., CHAUVET,A. & BRUNEL, M. 1989. Collapse basin: a new type of extensional sedimentary basin from the Devonian of Norway. Geology, 17, 127-130. SERANNE, M. 1992. Late Palaeozoic kinematics of the MCre-Trr Fault Zone and adjacent areas, central Norway. Norsk Geologisk Tidsskrift, 72, 141-158. -& SEGURET, M. 1987. The Devonian basins of western Norway: tectonics and kinematics of an extending crust. In: COWARD, M. P., DEWEY, J. F. & HANCOCK, P. L. (eds) Continental Extensional Tectonics. Geological Society, London, Special Publications, 28, 537-548. - - , CHAUVET,A. ,St;GURET, M. & BRUNEL,M. 1989. Tectonics of the Devonian collapse-basins of Western Norway. Bulletin de la Soci~tO GOologique de France, 8, 489-499. SMITH, D. C. 1984. Coesite in clinopyroxene in the Caledonides and its implication for geodynamics. Nature, 310, 641-644. 1995. Microcoesites and microdiamonds in Norway: a brief review. In: COLEMAN, R. C. & WANG, I. (eds) Ultra High Pressure Metamorphism. Cambridge University Press. & LAPPIN, M. A. 1989. Coesite in the Straumen kyanite-eclogite pod, Norway. Terra Nova, 1, 47-56. SOPER, N. J., STRACHAN, R. A., HOLDSWORTH, R. E., GAYER, R. A. & GREILING, R. O. 1992. Sinistral transpression and the Silurian closure of Iapetus. Journal of the Geological Society, London, 149, 871-880. STEEL, R. J. & GLOPPEN, T. G. 1980. Late Caledonian (Devonian) basin formation, western Norway: signs of strike-slip tectonics during infilling. In: BALLANCE, P. E & READING, H. G. (eds) Sedimentation in Oblique-slip Mobile Zones. Special Publication of the International Association of Sedimentologists, 4, 79-104. - - , SIEDLECKA,A. & ROBERTS,D. 1985. The Old Red Sandstone basins of Norway and their deformation: a review. In: GEE, D. G. & STURr, B. A. (eds) The Caledonide Orogen - Scandinavia and Related Areas. John Wiley, New York, 293-316. TORSVIK, T. H., STURT,B. A., RAMSAY,D. M., BERING, D. & FLUGE, P. R. 1988. Paleomagnetism, magnetic fabrics and the structural style of the Hornelen Old Red Sandstone, Western Norway. Journal of the Geological Society, London, 145, 413-430. - - , ET AL. 1989. Palaeomagnetic constraints on the early histroy of the MCre-TrCndelag Fault Zone, Central Norway. In: KISSEL,C. & -

-

,

,

E X H U M A T I O N OF U H P ROCKS BY TRANSTENSION LAJ, C. (eds) Paleomagnetic Rotations and Continental Deformation. Kluwer, Dordrecht, 431-457. , , , KISCH, H. J. & BERING, D. 1986. The tectonic implications of Solundian (Upper Devonian) magnetization of the Devonian rocks of Kvamshesten, western Norway. Earth and Planetary Science Letters, 80, 337-347. , SWENSSON,E., ANDERSEN,T. B. & DEWEY, J. F. 1992. Paleomagnetic dating of fault rocks: evidence for Permian and Mesozoic movements and brittle deformation along the extensional Dalsfjord Fault, western Norway. Geophysical Journal International, 109, 565-580. VAN ROERMOND, H. L. M. 1985. Eclogites of the Seve Nappe, central Scandinavian Caledonides. In: GEE, D. G. & STURT,B. A. (eds) The Caledonide Orogen - Scandinavia and Related Areas. John Wiley, New York, 293-316. VON BLANCKENBURG, F. & DAVIES, J. H. 1995. Slab

181

breakoff: a model for syncollisional magmatism and tectonics in the Alps. Tectonics, 14, 120-131. WAIN, A. 1997. New coesite-eclogite occurrences in western Norway: the nature of an ultrahigh pressure province in the Western Gneiss Region. Geology, 25, 927-930. WHEELER, J. 1991. Structural evolution of a subducted continental sliver: the northern Dora Maira massif, Italian Alps. Journal of the Geological Society, London, 148, 1101-1113. & BUTLER, R. W. H. 1994. Criteria for identifying structures related to true crustal extension in orogens. Journal of Structural Geology, 16, 1023-1027. WILKS,W. J. & CUTHBERT,S. J. 1994. The evolution of the Hornelen Basin detachment system, western Norway: implications for the style of late orogenic extension in the southern Scandinavian Caledonides. Tectonophysics, 238, 1-30.

The Trans Mojave-Sierran shear zone and its role in Early Miocene collapse of southwestern North America R O Y K. D O K K A 1, T I M O T H Y

M. R O S S 2 & G A N G L U 3

1Department of Geology and Geophysics, Louisiana State University, Baton Rouge, L A 70803, USA 2Department of Geological Sciences, California State University, San Bernardino, CA 92407, USA 3Marathon Oil Company, P.O. Box 3128, Houston, T X 77253-3128, USA Abstract: The relative motion between the Pacific and North American plates in early Miocene time was not parallel to the overall NW strike of the transform, but was instead oblique and transtensional. It has recently been proposed that in response to this divergence, the western edge of the North American plate east of the transform gravitationally collapsed and moved 100-150 km to the southwest ($50~176 The region of collapse covered an area of nearly 106 km 2 and included what is now southern California, southwestern Arizona, and northwestern Mexico. A major structure facilitating collapse between 21 and 18 Ma was the Trans Mojave-Sierran shear zone (TMSSZ). This east-west shear zone linked the classic detachment fault terranes and metamorphic core complexes of the Mojave desert, southeastern California, southern Arizona, and Sonora, Mexico, to the transtensional plate boundary. To more fully understand the nature and kinematic significance of the TMSSZ and its role in facilitating early Miocene fragmentation of the North American plate, a palinspastic reconstruction of the Mojave desert was performed to remove the disruptive effects of the TMSSZ and younger tectonic events. Features formed just before movements along the TMSSZ were used as markers to assess the TMSSZ deformation. Our analysis indicates that TMSSZ deformation was distributed across a c. 90 km wide band; restoration of markers to their original positions implies that >80 km of dextral shear occurred along the TMSSZ. First-order dextral shear deformation within the TMSSZ is expressed internally by clockwise vertical axis rotations of large areas that were facilitated by second-order zones of sinistral shear that separated the blocks. These second-order sinistral zones apparently exploited older transfer zones of the 24-21 Ma Mojave Extensional Belt.

Plate tectonic principles have been widely and successfully applied to explain Neogene deformation of coastal regions of the s o u t h e r n Cordillera (e.g. Atwater 1970; Ingersoll 1982; Stock & Molnar 1988; Severinghaus & Atwater 1990; Nicholson et al. 1994; B o h a n n o n & Parsons 1995; Fig. 1). Early attempts to apply these principles to explain the distribution and origin of the classic extensional terranes within the interior of the N o r t h A m e r i c a n plate were thwarted because of unresolved uncertainties of older plate models and by the lack of clear-cut geometric and kinematic links to the global plate circuit. Recent plate tectonic reconstructions (Stock & Molnar 1988; Atwater 1989; Severinghaus & Atwater 1990) have improved our comprehension of the uncertainties and behaviour of the P a c i f i c - N o r t h A m e r i c a n portion of the global plate circuit, and have resulted in bringing us closer than ever to fulfilling the promise of fuller tectonic u n d e r s t a n d i n g set forth in Atwater's (1970) plate tectonic tour de force.

The nagging mechanical question of direct, physical linkage of the global plate circuit with c o n t e m p o r a n e o u s early Miocene tectonic systems of the southern Cordillera (e.g. extension in southern California-Arizona, the San Andreas fault system, the Sierran orocline) has been apparently solved by the recognition of the Trans Mojave-Sierran shear zone (TMSSZ; Fig. 1), a broad (c. 90 km), approximately E - W dextral shear zone that passes t h r o u g h the s o u t h e r n Sierra N e v a d a region and Mojave desert (Dokka & Ross 1995; this paper). This link was detected and identified from regional analysis of field structural and palaeomagnetism data. D o k k a & Ross proposed that the TMSSZ formed along the northern edge of a large fragment of the North American plate that detached in early Miocene time in response to transtension developed along the Pacific-North American plate boundary. In addition to providing the kinematic linkage between the global plate circuit and c o n t e m p o r a n e o u s early Miocene

DOKKA, R. K., ROSS,T. M. & LU, G. 1998. The Trans Mojave-Sierran shear zone and its role in Early 183 Miocene collapse of southwestern North America. In: HOLDSWORTH,R. E., STRACHAN,R. A. & DF~WEY,J. E (eds) 1998. Continental Transpressionaland Transtensional Tectonics. Geological Society, London, Special Publications, 135, 183-202.

184

R.K. DOKKA E T A L .

Fig. 1. Index map of the present-day positions of tectonic features and localities of the southwestern USA and northern Mexico that are mentioned in the text. Plate tectonic elements from Severinghaus & Atwater (1990). Dark grey areas, early Miocene detachment terranes and metamorphic core complexes of southeastern California and Arizona. Blackened areas, early Miocene Mojave Extensional Belt (including offset portion in the Salinian Block of central California (Dokka 1989)). Ruled area, Trans Mojave-Sierran shear zone (TMSSZ); sense of shear shown by arrows. Lightly shaded area, late Neogene Eastern California shear zone (ECSZ). Fp, Farallon plate; Mp, Monterey plate; S, Salinian Block; SCT, Santa Catalina and Tortolita Mountains; SM, South Mountain; WM, Whipple Mountains; F.Z., fracture zone; T.J., triple junction; state abbreviations standard. tectonic systems of the southern Cordillera (e.g. extension in southern California-Arizona, the San Andreas fault system, the Sierran orocline), the TMSSZ has been important in rearranging the position of older palaeogeographical and tectonic elements. This paper presents the results of a structural analysis of the Mojave desert that was performed to gain insights into the original geometry and kinematic history of the TMSSZ, and its role in the early Miocene transtensional

collapse of southwestern North America. Our study consisted of two steps. Before we could analyse the structure of the TMSSZ, we first needed to restore the Mojave desert back to its early Miocene (c. 18 Ma) configuration, just after the end of movement along the TMSSZ. This required that subsequent translations, rotations, and strains associated with the 0-13 Ma Eastern California shear zone be accounted for and restored (Fig. 2). U p o n reaching this point, we used well-constrained 24-21 Ma markers to

THE TRANS MOJAVE-SIERRAN SHEAR ZONE

185

Fig. 2. The Pacific-North American transform boundary in the western USA highlighting the location of the Eastern California shear zone (lightly shaded; modified from Dokka (1993)). observe and measure the deformational effects of the TMSSZ.

Early Miocene palinspastic resonstruction of the Trans Mojave-Sierran shear zone The need f o r reconstruction It is fundamental to the concept of structural analysis that the effects of younger events in an area be sequentially removed before one can accurately assess the nature of older events. In intricately deformed areas such as the Mojave desert, this requires that all translations, vertical-axis rotations, and strains associated with each event be known and explained. As will be discussed below, several models for Mesozoic and Cenozoic tectonics of the Mojave desert region are incorrect because they fail to account for the disruptive effects of all Miocene and younger deformations. During late Cenozoic time, the Mojave experienced three structurally different and temporally separated intervals of deformation: (1) approximately N-S directed opening of the c. 24-21 Ma Mojave Extensional Belt (Dokka 1989; Ross 1995); (2) approximately E - W striking, dextral shearing (c. 21-18 Ma) along the TMSSZ (Dokka & Ross 1995, 1996; this paper); (3) the c. 13-0 Ma Eastern California shear zone (Dokka & Travis 1990a, b; Dokka 1993). Subsequent to its time of activity at 21-18 Ma, the TMSSZ was truncated and disrupted by the 0-13 Ma Eastern California shear zone (ECSZ (Fig.

2); D o k k a & Travis 1990a; D o k k a 1993). Approximately 65 km of slip (resolved along an approximately N40~ line) has occurred within this c. 80 km wide belt of distributed righ t shear since its inception (Dokka & Travis 1990a); Pezzopane & Weldon (1993) proposed that the ECSZ continues through northern California and Nevada to eastern Oregon, where it may connect with the Cascade Range. Faults of the southern portion of the ECSZ join with those of the San Andreas system near the Pinto Mountain fault and final merger is completed in western Sonora, Mexico. The physical connection of faults of the ECSZ with the San Andreas fault system demonstrates that the ECSZ is a key element of the Pacific-North American plate boundary; the ECSZ has accommodated, and continues to accommodate, between 9% and 23% of the total relative plate motion (Sauber et al. 1986, 1994; Dokka & Travis 1990b; Savage et al. 1990). Given these disruptive effects, palinspastic reconstruction of the region is thus essential before meaningful structural analysis of the TMSSZ can be carried out.

Methodology Reconstruction of the Mojave desert region to remove the effects of the 0-13 Ma ECSZ followed the protocol established by Dokka & Travis (1990a) and Dokka (1993). The improved reconstruction presented here, as well as the original Dokka & Travis (1990a) model, is constructed primarily to explain regional, two-dimensional, surface (x,y) relations; vertical strain implications of the ECSZ such as local crustal extension and

186

R.K.

DOKKA

ETAL.

.., /

..~L~)

/.

D,! i t

~":

">"

/

,'

!7

.

...

/ .-"

I

.;

;~.

I/~ I I

.,: '~'" , ,

' ~.,, 9 "

e,,,.'" " ~

~-~'Z :-,.

I~<~

I:::

..:

D

f

-~<

~

,9

."

"

;,

..:', ..... :'7

.W,.'?*-.<:::. '5':. ,:t.j~.,.D.,) ,;t

..

,

'~-

-:.'"

, ..,.-.~ .~.

., ,""

-

.)

"'~',' .;, ,;,"

-,

<

.... #

"

9 ~'~

,>,',>.-.

I ""' >' ~ ..:: " .' ..i~ , ( 7 / ..'.~ "., " ' ""~. ". ~" ' , ~:/~".,~-I :'.. ."4~ ,., " ~ " ' ~, ', s. ." , . - . x"s "r . . . .

i

. ::

,':"

~ , , ; ~ +-

.', ~<~: .,~..," .4,~ 9' ." .%.~ /...-:. .,,,~..4, , . ,.ii~', v . - . .'

.

~_."

~

4,~>~.,,

#,"~*,

.."

~: ..--,~'

: .:.;5

' ..

"%.;

;

<."

/

:

,~--'-,~..

~ l " /,,.I": f ! : 9...5,'.=
,/..-

~UI :,. 't" r

: ~.,I -$-

,"

, ..~'

,'" , " ..

/-.

:::~

," ".

. .!,/~.:.,.t <.

:,

/,

~.-->

~'-' >"..I " ,

-:-~ ~.

~ 7

' ~\I"

:

~":~

..

9

"

"....,

,"

. :~" '

:

"

~..--'T

.

,., ~x'~/

.-.'/.~::~: '

:

" t ,.' '

"-::;~d

".

>" , ~ ~

~

'.~

"":-:-..~,

::

~;,..,

I

I

',: ..

.r

.

....i ....'<

~:-:.

,

I

.,~

.i

"-:'L ,~:' ~-s..'.

~

..-'"""

"2"~"':: ,j"'

,?'-#d.

" 4" " ":'~

:; '~,,

.::; . .! ., ~I" I~l i .:.4. , o

-;<

] ,

///

" ".'7;" " ~ t

'

:'/

: ."

n,"

""

,,,..,%-.7

~../ _. . ~ ,,.. l~J/,, ,

.

r~,;.7-..: ~ - , ~] ,;;-'::.'; ~..,,,?-)~:,'<, ....~ <..'" ..,,:.:.x".,,'7~: ", " i -, ' : " : , - ~ ..... ".,.:';.j,.~'" I" ,,r. ,';' .~ " I- <'. ~-~ / ........ --:--'k ~~' . . . . .~". . '..'..I-,9 ~-~ , i ," . . ~ 9 ."" .- ."" "% ,.,. ."". . :" .~ ...... ). , ,,.,.,..:,,/ .... .. . - . '"". I " ,, ,. , . ~ , . i,',. .,,.:_~>-. --S ..,~.." .>.>, .... , . . . . . . / ,, 9 . <~,, 9 | ...... ~ ...~......~-,, (-...... . ,.- . . . . . < :t-~. ~--.>\ .. - , :i .~I . ~ - , ~ "re' .~ ,,;;",.,,, , :"~,..:7 r....." ' ~, - , - ,'<" ,',"/ l ' ,.-<~,./_"~", ~ .... ~ ' " * ~ < ,.., , ~" '."~' "_"j'~" ~ t .;-'-~ii ~"+ " "...... ' " " :-,t" '< "" " " 'iY I,.,_..#~ :Y~".,, ..... ." .:~,t:~,~, ,... , # " ~ " > > : ~ f ~ . , ~ ..~ .'<;:~ A ' . --..'~,,/

-:

~,~,

,,~,,.,~.'.,"" ..'f" . . . .

.<.

:,

/,.-+5 .....

.,

...

,

'

~.... {. ,

<.~

~o~..o . N

~,~4:~o~

,~ ~ ~ ~ = <> " ~u ~ ,~,

.; t~.

. "'JU.'."

..~ ,/:, 9

\

~7-..

,

~ ,,'..

,,r .:'

'.,,.,, " "%" : ~ @ 9: ~. " . :".~',i~"~, ~

.~

2./

"

v,,..

, ~::. ,. !

!! .. / ,,s i ~

. ~.

,:

,, . .

... .... 7< ...... ', :../....: "" +7....-:

,S.,~"~:~, .,

H'-'

"

. i i,,.':'.,> ~' "

o, ~

.

,, :..- '

' ".~, " .... '~ ' - ' ...

'

/

I

;

#/

.,,,

I ...~/

o

~t ; ~ , 0 I ' ~h~..~; .'% ./;(.4~

~.: :: ~~:

,I~--. o

k"

><-.\:

i
'

"~ ~ : ~ ] ~ < ~ z

l

~, ~ ~ = ~ :~.u ~ o

~IJ=

, '(,J# I

t

=.X..=

THE TRANS MOJAVE-SIERRAN SHEAR ZONE contraction have been are discussed by Dokka (1993). The resolution of the reconstruction presented here is c. _+1km. Analysis of the ECSZ has shown that strain in the Mojave desert region is inhomogeneous and is partitioned into several domains of similar deformation that are separated by strike-slip faults and extensional zones (Garfunke11974; Dokka & Travis 1990a; Dokka 1993). Strike-slip translation along dominantly NW striking faults, along with tectonic rotations of blocks bounded by approximately E-W, left-slip faults are considered to have occurred within this broad, regional zone of right shear. The earlier models, as well as this reconstruction, are founded on several explicit and implicit assumptions that need to be discussed. First, we assume that all deformation is brittle and involves translations and rotations of rigid, fault-bounded bodies. Internal strain within individual fault blocks is assumed to be negligible. To simplify the reconstruction, fault systems composed of multiple strands are considered as single faults with a composite net slip. For example, the Camp Rock fault system, which includes several, locally named strands (Camp Rock, Emerson, Old Woman Springs, etc.), is treated as one continuous fault. The model considers that the amount of lateral translation of one fault block relative to an adjacent block is equal to its strike-slip. In addition, the amount of vertical axis rotation of a fault block is reckoned equivalent to the declination anomaly implied from palaeomagnetic studies carried out within that block or within a kindred block in the same structural domain; for example, in a group of tilted dominoes deformed in simple shear, the rotation of all can be inferred from the measurement of any one domino. Finally, the recognition of the style of deformation in the region (overall inhomogeneous deformation composed of domains of homogeneous simple shear) permits us to use simple geometry and trigonometry to facilitate and describe the restoration of translations and rigid body rotations. The following steps were conducted during the reconstruction. Fault blocks were first defined on the basis of mapped, late Cenozoic faults (Fig. 3); where such data were lacking, the positions of the boundaries were inferred on the basis of seismicity or topography. Fault blocks were then restored to prefaulting and prerotation positions using fault slip (Table 1) and palaeomagnetic declination vector constraints described by Dokka & Travis (1990a); these constraints, as well as more recent contributions, include data presented by Golombek & Brown (1988), Ross (1988), Ross et al. (1989), Wells & Hillhouse (1989), MacFadden et al. (1990a,b), Valentine et al. (1993), MacConnell et al. (1994), Lu & Dokka (1995) and Ross (1995). The geometry of and motion along unconstrained boundaries were then adjusted to eliminate the few geologically unexplained overlaps or gaps. The model was then run forward to test how well it could predict topography, present-day deformation patterns, and geological relationships. Successive iterations of this procedure were performed to obtain a geometric best fit to observed relationships. To assess the effects of the TMSSZ, we assembled a

187

group of older markers deformed by the TMSSZ and restored them to their pre-ECSZ configuration according to the reconstruction. An index map showing the locations of these markers is presented in Fig. 4. The most important of these markers include structures and kinematic indicators of the 24-21 Ma Mojave Extensional Belt (Dokka 1989) as well as palaeomagnetic vectors recorded in lower Miocene strata (see Ross (1995) and Geological Society of America Data Repository item 9530 for more complete discussion of palaeomagnetic data for the Mojave desert). Results

T h e results of o u r r e c o n s t r u c t i o n are p r e s e n t e d in Fig. 5. Figure 5a shows the p r e s e n t - d a y configuration of the early M i o c e n e m a r k e r s that we used to d e t e c t the T M S S Z in the M o j a v e desert. Figure 5b is a r e c o n s t r u c t e d view that r e m o v e s the effects of the E C S Z at c. 1 Ma, a n d Fig. 5c displays the configuration of the T M S S Z just after its time of activity (c. 18 Ma). Figure 5d depicts the M o j a v e d e s e r t block at c. 21 Ma, showing the position of o l d e r features b e f o r e d e f o r m a t i o n by the T M S S Z . T h e r e c o n s t r u c t i o n p r o d u c e s several interesting effects. First, the r e s t o r a t i o n of 65 k m of dist r i b u t e d right shear along the E C S Z returns the now fragmented Mojave Extensional Belt ( M E B ) into a single c o h e r e n t unit (Fig. 5c). Structural domains and intervening transfer zones of the M E B such as the K a n e Springs, B a x t e r Wash, and L a n e M o u n t a i n faults b e c o m e w h o l e a n d c o n t i n u o u s . All a r e a s t h a t are i n f e r r e d to be y o u n g extensional basins c r e a t e d by the E C S Z (in black, Fig. 5) during its history b e c o m e c l o s e d in t h e r e s t o r a t i o n . S e c o n d , a l t h o u g h intact, the M E B r e m a i n e d severely dist o r t e d in the reconstruction, as i n d i c a t e d by the w a r p e d traces of transfer faults of the M E B and associated 24-21 M a k i n e m a t i c indicators (Fig. 5c). P a l a e o m a g n e t i c declination vectors in 24-21 M a rocks as well as k i n e m a t i c indicators f r o m the M E B show similar disorientation (Fig. 5c). All of t h e s e d i s t o r t e d l i n e a r e l e m e n t s o c c u r within an a p p r o x i m a t e l y E - W , c. 90 k m wide zone that defines the region of effect of t h e T M S S Z (Fig. 5c). We c o n c l u d e that this 18 M a r e c o n s t r u c t i o n reflects the configuration of the T M S S Z just after its time of activity.

Structure of the Trans Mojave-Sierran shear zone Geometry

T h e 18 M a r e c o n s t r u c t i o n (Fig. 5c) indicates that the T M S S Z was o r i e n t e d a p p r o x i m a t e l y E - W

188

R.K. D O K K A E T A L .

Table 1. N e t s l i p s a f o r faults used to reconstruct the Eastern California shear z o n e in the M o ] a v e desert Fault Gravel Hills-Harper Lake Lockhart Helendale Lenwood Camp Rock Calico Rodman-Pisgah Sleeping B e a u t y Cady Manix Coyote L a k e

Ludlow Broadwell Lake Bristol Mountains Granite Mountains Ford Valley Calico-Blackwater Goldstone Lake (north)

Observed

Reference

This paper

<3.2 >0 3 1.5-3.0 1.6-4.0 9.6 6.4-14.4

b c d e e e,f e

2 2 3 2 3 10 10

n/a <1 n/a n/a

c -

12 1 12 11

small? n/a >6.0 n/a n/a 8.5 n/a

e g h -

3 <1 14 22 19 11 2

Bicycle L a k e

3.2

i

2

Paradise Coyote Canyon Tiefort Mountain Garlic Spring Desert King Spring (north) Red Pass Lake

<1 rda n/a 1 n/a n/a

j k -

1 1 1 1 24 10

Fort I r w i n Drinkwater Lake

n/a n/a

-

6 8

<1 8-20

c l,m

0 8, n

16 8 11 3 1.5 2.4

o o o o o o

27 11 10 p p p

East Paradise Southern Death Valley Pinto M o u n t a i n B l u e Cut Chiriaco Substation S m o k e Tree Porcupine W a s h

aAll values indicate the amount of lateral slip in kilometres; faults with left-slip shown in italics. bEstimate of right-lateral displacement by Dibblee (1968). cR. K. Dokka (unpublished mapping 1988-1992). aMiller & Morton (1980). eDokka (1983). fIncludes 1.4 km of right shear expressed as strain. gMiller et al. (1982) documented >6 km of right separation. hDokka (1989) documented 8-9 km of right separation in the Calico Mountains-Mud Hills. iLeft separation of distinctive marble unit. JR. K. Dokka, D. MacConnell & J. Ford (1992, unpublished map and report). kyount et al. (1994). lWright and Troxel (1967, 1970). mBrady (1990). nUse of higher model value for Southern Death Valley fault zone requires that the cumulative net slip for faults of Domain V be equal to 49.5 km. This also suggests that the total net slip for the entire ECSZ is 77 km. ~ (1981). PNet slip included with other named left faults of the domain.

(N82~ at its final stage of m o v e m e n t a n d t h a t it s p a n n e d t h e M o j a v e d e s e r t a n d t h e s o u t h e r n Sierra N e v a d a region. C o m p a r i s o n of t h e 18 M a a n d 21 M a r e s t o r a t i o n s suggests t h a t t h e T M S S Z

may have rotated 30-40 ~ clockwise during d e f o r m a t i o n f r o m an original t r e n d t h a t was W S W - E N E (Fig. 5d). F i g u r e 5d also suggests t h a t t h e o r i g i n a l t r e n d a n d p o s i t i o n of t h e

THE TRANS MOJAVE-SIERRAN SHEAR ZONE

189

Fig. 4. Location map of the early Miocene markers used to assess effects of the Trans Mojave-Sierran shear zone. Kinematic indicators associated with 24-21 Ma Mojave Extensional Belt and data source in parentheses: 1, Kramer Hills (R. K. Dokka, unpublished data); 2, Hinkley Hills (Dokka 1989); 3, Waterman Hills (Dokka 1989); 4, Mitchel Range (Dokka 1989); 5, Newberry Mountains (Dokka 1989); 6, Rodman Mountains (Dokka 1989); 7, southwestern Cady Mountains (Ross 1994 1995); 8, central Cady Mountains (Dokka 1989); 9, Baxter Wash fault (Dokka 1989; Temple 1997). Palaeomagnetic declination vectors in lower Miocene volcanic rocks and data source in parentheses: a and b, Goldstone district (MacConnell et al. 1994); c, north of Barstow (Burke et al. 1982); d, Kramer Hills (Lu & Dokka 1995; this paper; compare Golombek & Brown 1988); e, Newberry Mountains (Ross 1988, 1994; Ross et al. 1989); f, Rodman Mountains (Ross 1988; Ross et al. 1989); g, southwestern Cady Mountains (Ross 1995); h, Lava Bed Mountains (Ross 1988,1994; Ross et al. 1989); i, central Cady Mountains (Ross et al. 1989; Ross 1995); j, Bristol Mountains (Ross 1988, 1994; Ross et al. 1989); k, Alvord Mountain (Ross et al. 1989); localities 1, m, and n are unrotated rocks of the Jurassic Independence Dike Swarm. Asterisk denotes the location of the town of Barstow, California (B).

TMSSZ were inherited from the slightly older Mojave Extensional Belt. Early Miocene regional extension had w e a k e n e d the lithosphere through tectonic thinning (10-17 km vertical) and magmatism (Dokka 1989; Henry & Dokka 1992). Additional evidence for reactivation (usage from Holdsworth et al. (1997)) is presented below. Figure 5c shows that early Miocene markers such as palaeomagnetic declination vectors and the MEB kinematic indicators and transfer faults are consistently and regularly deflected across the central Mojave desert. The sense of this deflection implies dextral shear and is thus consistent with the conclusion of D o k k a & Ross (1995 1996) regarding the nature of first-order

shear along the TMSSZ. A l t h o u g h the now approximately E - W belt was dominantly rightslip in character, dextral shears parallel to the TMSSZ are few. This, of course, is not a requirement of a right-slip zone (see Glazner et al. 1996). Closer examination of the TMSSZ shows that internal strain is complex and facilitated subregionally by several mechanisms (Dokka & Ross 1995, 1996). Figure 6 illustrates our concept of how deformation occurred w i t h i n the TMSSZ. Throughout much of the Mojave, dextral shear deformation is expressed by clockwise rotations of large areas about vertical axes (Golombek & Brown 1988; Ross et al. 1989; Ross 1995). Across the central Mojave, deformation is generally

190

R.K. DOKKA ETAL.

T H E TRANS MOJAVE-SIERRAN S H E A R Z O N E

191

Fig. 5. Palinspastic reconstruction of the Mojave desert region at various times from present day to the beginning of the Miocene. The method of reconstruction and associated assumptions are discussed in the text. The reconstruction steps include: (a) 0 Ma; (b) c. 1 Ma; (c) c. 18 Ma; (d) c. 21 Ma; (e) c. 24 Ma. Single-headed arrows are palaeomagnetic declination vectors in lower Miocene volcanic rock. Double-headed arrows are kinematic vectors from structures of the 24-21 Ma Mojave Extensional Belt. Ruled line pattern and blackened areas represent regions of contraction and extension associated with the Eastern California shear zone, respectively. Black lines are active faults and grey lines are inactive or future faults. SDVF, Southern Death Valley fault; DKSF, Desert King Spring fault; GMF, Granite Mountain fault; BMF, Bristol Mountains fault; PWF, Packard Well fault; PvF, Panamint Valley fault; RLF, Rodgers Lake fault; LMF, Lane Mountain fault; BWF, Blackwater fault; KSF, Kane Springs fault.

192

R.K. DOKKA E T A L .

Fig. 6. Conceptual diagrams illustrating the relationship between clockwise and counter-clockwise vertical axis rotations and sinistral faults within (a) a hypothetical shear zone and (b) the TMSSZ. In each case, the sense of first-order shear is dextral. It should be noted that, in most areas, internal deformation occurs by clockwise rotations about vertical axes. Second-order sinistral faults play a critical role by facilitating the rotation of the intervening blocks. Initial motions along these left shear zones produce local counter-clockwise vertical axis rotations, folding, and left-slip faulting. Later, as soon as a through-going fault is propagated, all areas between the sinistral faults rotate clockwise. BWF, Baxter Wash fault; KH, Kramer Hills; LMF, Lane Mountain fault; TMSSZ, Trans Mojave-Sierran shear zone.

continuous and the distribution of shear strains across the TMSSZ inferred from vertical axis rotations is asymmetrical with respect to the centre of the shear zone; the maximum apparent shear strain occurs south of the centre of the TMSSZ (Fig. 7). Because of the lack of E - W faults and the observation that shear strains (rotations) gradually and continuously decrease towards the outer limits of the shear zone, we conclude that much of the dextral shear in the TMSSZ was accomplished by continuous oroclinal folding about vertical axes. This, however, was not the entire story. A growing body of palaeomagnetic and field structural data indicates that sinistral shear zones played a critical role in the development of the TMSSZ. We propose that these cryptic left shear zones facilitated the clockwise rotation of large intact crustal blocks within the TMSSZ

(Fig. 6). Motions along these now N E trending zones also produced local effects such as counter-clockwise rotations about vertical axes, megascopic folding, and faulting (Valentine et al. 1993; D o k k a & Ross 1995, 1996; Lu & D o k k a 1995; Temple 1997). These shear zones were first detected by Valentine et al. (1993), who showed that rocks along the now N E striking Lane Mountain fault of the west-central Mojave were rotated 23-57 ~ counter-clockwise between 22 and 18 Ma (Fig. 8). Re-evaluation of the Kramer Hills, a well-exposed area that lies adjacent to the Lane Mountain fault and a key area cited by Golombek & Brown (1988) and VaLentine et al. (1993), reveals i m p o r t a n t details on the sequence of 21-18 Ma rotational events (Fig. 8; Lu & D o k k a 1995; this paper). Our studies show that rocks of the Kramer Hills were rotated c. 80 ~ counter-clockwise about a vertical axis just

THE TRANS MOJAVE-SIERRAN SHEAR ZONE

193

Fig. 7. Shear strain profile across the Trans Mojave-Sierran shear zone (TMSSZ) in the central Mojave desert. Shaded area is the TMSSZ at c. 18 Ma. The shear strain is inferred from rotated lines, i.e. palaeomagnetic declination vectors from lower Miocene rocks (see Figs 4 and 5 for descriptions of lines used). The skewed profile suggests that the displacement field across TMSSZ is not regular and continuous as required by ideal simple shear. Instead, shearing appears to have been more concentrated in the southern portion of the shear zone. BWF, Baxter Wash fault; LMF, Lane Mountain fault; RLF, Rodgers Lake fault; KSF, Kane Springs fault.

before 21 Ma (R2, Fig. 8). This occurred during the same time interval that the area to the east (central Mojave) was rotating clockwise (Ross 1995). At c. 21 Ma and continuing until c. 18 Ma, the Kramer Hills area rotated c. 45 ~ clockwise; again, this occurred synchronously with clockwise vertical axis rotations in the central Mojave desert (Ross 1995). These relations support our view that the early counter-clockwise rotation was due to local strain during initial sinistral shearing. As soon as a through-going fault was formed, local strain ceased and the Kramer Hills rotated clockwise along with the rest of the region. Similar relations have recently been observed along the Baxter Wash transfer zone in the northern Cady Mountains (Temple 1997). Here, lower Miocene rocks yield palaeomagnetic declination vectors suggesting >80 ~ of local counterclockwise rotation that occurred in association with development of large folds along the now NE striking Baxter Wash fault (Temple 1997). Higher in this section, upper lower Miocene rocks (>18 Ma) yield evidence of a large (20-67 ~) regional clockwise rotation (MacFadden et al. 1990b; Ross 1995; Temple 1997). Finally, re-evaluation of folds along the Kane Springs fault in the Newberry Mountains sug-

gests that yet another MEB transfer zone was reactivated during 21-18 Ma and facilitated TMSSZ deformation. These megascopic folds were originally explained as reverse drag folds that formed along the Kane Springs fault during extension ( D o k k a 1980, 1986). The problem with this earlier interpretation, however, is that the limbs of some folds are locally overturned, a situation unlikely to have occurred exclusively as a result of reverse drag. Even though the argument for initial formation because of reverse drag remains compelling, a more likely scenario is one where these earlier formed folds were modified and tightened by sinistral shear during TMSSZ time. Comparison of these sinistral shear zones reveals several common characteristics worth noting. First, each of these sinistral shear zones coincides spatially with major transfer faults of the slightly older MEB; these older faults originated as right-slip faults (Dokka 1986,1989). We propose that these older discontinuities may have influenced the deformation as weak zones reactivated by the younger deformation. Previously, Valentine et al. (1993), as well as Bartley & Glazner (1991), speculated that early Miocene clockwise and counter-clockwise rotations were coincident and that both occurred

R.K. DOKKA ETAL.

194 I

I

I

"~'~ a 7 ~ 1

R

I~

--3

3

basalt limest. . . . .

.......

iltstone ~O26 ,+

O

N R3 = 45 ~ C W (regional) 2

N

7

:,- ::= ...... :. . . .

E siltstone

E

9 28 + 'O14 +

6 -- --"---~----"---

E

4

)

)

//

/

/

/

/,

/

/

/

/

[itll]llll

Red Buttes Andesite

I 9 17 -49 29 -4-

limestone, silicified dolomite

O18 l- 2 0 19 __

L

, /

Upper Tropico Group observed directions

R

f

N

] ~

R2=80occW(Iocal)~

~020

L

032 unexposed (about 300rn)

--~_

_

Upper Tropico Group after removing regional rotation

- .

E IO21

3

sandstone, siltstone I

i

I

I

I

2 L

~ i I I i I Li ....... :'~'7Q ~ ' ~ 5S~ ~

}E

limestone

O 33 IO34

~ c"L' ~-~

tuffa. . . . . . .

dstone

Lower Tropico Group observed directions

(b) Lower Tropico Group after removing regional rotation

N

N

0 2 3 --

R~ 1

5~-m Cretaceous

, ~ ~ e ~ ~ ~ v ~ ~~ + + ~ § + ++ - +++++ +§ + . . + +

silicified tuff, tuffaceous siltston~ 9 24 + 025 + ,: Quartz monzonite

E

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

Bottom of Lower Tropico Group observed directions

~ (b) Bottom of Lower Tropico Group after removing younger rotations

Fig. 8. Evidence for multiple vertical axis rotations in the Kramer Hills, west-central Mojave desert. Composite stratigraphic column of lower Miocene Tropico Group and observed group mean declination anomalies with the interpreted multiple episodes of vertical axis rotation. +, Clockwise vertical axis rotation; - , counter-clockwise rotation. R1 is related to development of the c. 24-21 Ma Mojave Extensional Belt, whereas Re and R3 are related to the 21-18 Ma Trans Mojave-Sierran shear zone (see text for discussion). Site means (O) are projected onto the lower-hemisphere projection of an equal-area net with 95% confidence limit (a95) of group means (11) for the three groups. (a) Observed directions in bedding coordinates; (b) directions after removing the effect of younger rotations. Triangle in the equal-area projection is the Miocene reference direction calculated from the Miocene pole of Diehl et al. (1988). CW, Clockwise; CCW, counter-clockwise.

concurrently with regional extension. This interpretation was shown to be incorrect by Ross (1995), who found that early Miocene extension and regional vertical axis rotations were not coincident. This was accomplished through examination of palaeomagnetic declination anomalies in a well-exposed stratigraphic section containing extended and post-extensional rocks of the Cady Mountains. Second, the slip on each fault is similar (sinistral) and some (all?) have substantial displacement. For example, the width (>16 km) of the shear zone and degree of shearing inferred from palaeomagnetism and structural studies along of the Lane Mountain fault in the Kramer Hills (c. 80 ~ counter-clockwise) suggest that >22 km of distributed left-slip occurred. This is based on the simple relationship, NS>w+~r/180, where NS is the net slip, w is the width of the shear zone, and

+ is the rotation produced by shear. Preliminary studies in the northern Cady Mountains suggest that c. 13-25 km of left-slip occurred along the Baxter Wash fault (Temple 1997). Finally, these sinistral shear zones die out to the north and south at the inferred outer limits of the TMSSZ.

A m o u n t o f displacement Determination of the amount of dextral displacement along the TMSSZ between 21 and 18 Ma is not easy or straightforward because of the lack of appropriate piercement lines with which to gauge the integrated net slip. Previously, D o k k a & Ross (1995) proposed that 41-66 km of right-slip occurred along the T M M S Z , based on the assumption of regional simple shear; they calculated this value using an average vertical axis rotation as a measure of the shear strain. This

THE TRANS MOJAVE-SIERRAN SHEAR ZONE

195

Fig. 9. Approach used to determine the net slip along the TMSSZ using palinspastic reconstructions. The method compares the positions of selected points at times before and after movement along the TMSSZ; straight-line movement is assumed. The outline of the Mojave Extensional Belt is shown in each reconstruction for reference. Single-headed arrows are palaeomagnetic declination vectors in lower Miocene volcanic rock. Double-headed arrows are kinematic vectors from structures of the 24-21 Ma Mojave Extensional Belt (see Fig. 4 for source of data). Lightly shaded area, 18 Ma position of TMSSZ; heavily shaded area, 21 Ma configuration. Mesozoic tectonic elements (Last Chance and eastern California thrust systems) from Snow (1992). (a) Reconstructed view of TMSSZ and reference features at 18 Ma. (b) Reconstructed view of TMSSZ and reference features at 21 Ma. (e) Comparative view; large arrows are net slip vectors based on changes of position of two selected points in the TMSSZ. (See text for discussion.) Abbreviations as in Fig. 5.

method probably underestimates the total net slip by a factor of perhaps two because it neglects the contribution made by sinistral faults to the overall deformation. To at least partially address this shortcoming, we present an improved method to estimate net slip that relies on the palinspastic reconstruction presented above (Fig. 9). This m e t h o d compares the positions of selected elements of the MEB at 18 Ma and 21 Ma. The net slip is taken as the average displacement of selected points in the shear zone. For this approach to be valid, the restoration must deal effectively with several geological constraints. First, variably oriented kinematic indicators of the M E B should become parallel and oriented north. Because detailed palaeomagnetism studies have demonstrated that orig-

inal N - S extension of the M E B was not accompanied by rotations about vertical axes ( D o k k a 1989; Ross 1995), we conclude that the observed disorientation of these features at 18 M a is an artefact of T M S S Z deformation. Second, transfer zones of the M E B should also become aligned and parallel to the extension direction implied by M E B kinematic indicators. Third, p a l a e o m a g n e t i c declination vectors recorded in 24-21 Ma rocks of the region should become reoriented to north. Fourth, the severely warped traces of Mesozoic thrust systems (e.g. Snow 1992; Fig. 9) and older P r e c a m b r i a n - P a l a e o z o i c palaeogeographical trends that cross the region (e.g. Burchfiel & Davis 1981) should straighten; although the degree of straightening n e e d e d is uncertain

196

R.K. DOKKA E T A L .

because we do not know the original configuration of the Mesozoic and older features, the complex form of the features suggests that some smoothing is required. Fifth, restoration should involve and undo sinistral motions along the 'reactivated' transfer zones of the MEB. As shown in Fig. 6a, these sinistral shear zones can significantly contribute to the overall dextral shear along the TMSSZ; because each left-slip zone adds its own contribution to the total rightslip, we expect the net slip along the TMSSZ to increase to the west. At present, we lack net slip data on all of these faults to make a complete restoration. Our reconstruction at 21 Ma (Figs 5d and 9) uses the values calculated for the Lane Mountain and Baxter Wash faults, and suggests that the TMSSZ has a minimum dextral net slip of 80 km; because we assume that no slip occurred on the Rodgers Lake and Kane Springs faults, the resultant total net slip value for the TMSSZ is understated.

Discussion Recognition of the TMSSZ is significant for two main reasons. First, the TMSSZ is a major, crustal-scale break whose net slip is among the largest in the southwestern Cordillera. Surprisingly, this major structure lay undetected until recently in one of the world's most intensively studied areas. This was probably the result of the masking by younger deformations, and the shear zone's broad and subtle effects. It is clear that discovery would probably not have occurred were it not for the application of palaeomagnetic declination analysis pioneered in the region by Luyendyk et al. (1980), Golombek & Brown (1988), Ross et al. (1989), MacFadden et al. (1990a,b), Wells & Hillhouse (1989), Valentine et al. (1993), and Ross (1995). Perhaps the true significance of the TMSSZ is more philosophical in nature than structural. As discussed below, modern plate tectonic reconstructions tell us that the Pacific-North American plate boundary was transtensional in early Miocene time. These models specify the total amount of integrated displacement that should be apparent in the rocks across the boundary. Such models, however, offer no insights on how, where, and when the continent responded. In our view, the TMSSZ represents the last remaining major piece of a long-standing tectonic puzzle that includes the metamorphic core complexes and detachment faults of the California-Arizona and Sonora region, the San Andreas fault system, and the Sierra Nevada

orocline. When viewed collectively, the geometry and kinematics of this co-ordinated system of moving parts is consistent with motions along the transtensional plate boundary. We expand on this theme below. Plate tectonic controls on continental deformations

The early Miocene plate tectonic history of the northeastern Pacific region offers several important insights into the origin of concomitant deformation of the North American plate (e.g. Atwater 1970; Stock & Molnar 1988; Severinghaus & Atwater 1990; Fig. 10). It is now clear that the relative motion between the Pacific and North American plates was not purely strike-slip, but was instead oblique to the plate boundary (e.g. Atwater 1970, 1989; Stock and Molnar 1988; Severinghaus & Atwater 1990). These movements resulted in transtension across the transform and the W N W migration of the Pacific plate away from the NW striking transform (Bohannon & Parsons 1995; Dokka & Ross 1995). As illustrated in Fig. 10, 340 __ 200 km of transform-normal migration should have occurred since c. 30 Ma (Atwater 1989), with as much as c. 175 km developing between 20 and 16 Ma. Evidence for such divergence would be expected to be observed within the Pacific plate, along the boundary between the two plates, or within the North American plate (Atwater 1989). Magnetic anomaly mapping of the Pacific plate offshore of central California shows no indication of extension, i.e. sea-floor spreading, within the Pacific plate or between the two plates (Lonsdale 1991), whereas the North American plate is highly extended (e.g. Davis & Coney 1979; Davis et al. 1980; Wernicke 1981; Davis & Lister 1988; Dokka 1989). On this basis, Dokka & Ross (1995) concluded that, as the Pacific plate migrated W N W away from the transform during the Miocene (24-5 Ma), the North American plate extended laterally to maintain contact. Dokka & Ross viewed this divergence as the triggering mechanism that set up gravitydriven movement of a large fragment of the North American plate to the southwest toward the transform. Through what they termed the 'Collapse', they sought to integrate all early Miocene deformational events in the continent and relate them to the divergence occurring along the transform. As will be discussed below, the TMSSZ was an integral structure of this system of co-ordinated early Miocene deformations.

THE TRANS MOJAVE-SIERRAN SHEAR ZONE

Fig. 10. Plate tectonic evidence for transtension along the Pacific-North American (P-NA) plate boundary during Miocene time (data from Stock & Molnar (1988), their table 16). Transtension along the Pacific-North American plate boundary created the instability that triggered the southwesterly directed collapse of the edge of North America. Lightly shaded area is the envelope for the displacement path of a reference point (A) on the Pacific plate located south of the Mendocino triple junction that moves relative to the North American plate. Heavily shaded boxes define position of point A at magnetic anomaly times A7 (25.98 Ma), A6 (19.90 Ma), A5 (10.59 Ma), and A3 (5.5 Ma). The implausible early Miocene position of point A within the North American plate is explained by the westward migration of the Pacific plate and resultant extensional collapse of the western margin of North America. (See text for discussion.)

A system o f co-ordinated continental deformations Dokka & Ross (1995) considered that all major early Miocene tectonic systems active in the southwestern part of the North American plate were physically interconnected and kinematically linked to the global plate circuit (Figs 11 and 12). Collapse began near 24 Ma and continued until c. 18 Ma, resulting in southwesterly translation of the lithospheric fragment. The region of collapse covered an area of nearly 106 km 2 and included what is now southern California, southwestern Arizona, and northwestern Mexico. Strain, both laterally and vertically, was not uniform. Strain at the surface was concentrated in belts around the perimeter of the large collapsed fragment (Arizona metamorphic core complexes, Colorado River Extensional Corridor, MEB, TMSSZ), whereas areas within the fragment away from its highly extended margins remained largely intact (i.e. no major lateral

197

lengthening). The collapsed fragment decoupled in the middle and lower crust and moved southwest along subhorizontal detachment faults and ductile shear zones; recognition of this mid-crustal zone of shearing is based on extensive industry and university seismic reflection profiles of southeastern California and southern Arizona that show that the middle crust and deeper is characterized by low-angle reflectors (e.g. Cheadle et aL 1986; Frost et al. 1987; Hamilton 1987; Dokka 1989; McCarthy et al. 1991; Morris 1993). In the northern Mojave desert, middle-crustal reflectors correlated with collapse-related detachment faults do not continue at depth past the region of surface extension (Serpa & Dokka 1992). Major brittle-ductile detachment faults that breach the surface in the highly extended Colorado River Extensional Corridor and southern Arizona can be traced downward to the northeast where they merge with the zone of decoupling (e.g. Frost et al. 1987; Hamilton 1987). Figure 11 shows the sequential tectonic evolution of the southwestern USA and Mexico during early Miocene time and the role played by the TMSSZ. Both northern and eastern edges of the collapsed region underwent extension between c. 24 and c. 21 Ma. Although extension in southeastern California and Arizona was coaxial with the overall southwest collapse direction, the local kinematics along the northern boundary (MEB) was complicated by the combined effects of the SW collapse and the independent approximately NW translation of the region to the north. These two motions resulted in approximately N-S extension across the MEB. Collapse continued during the interval 21-16 Ma, with major extension of southeastern California and Arizona occurring between 20 and 18 Ma. Cessation of NW motion of the region to the north at c. 21 Ma caused the kinematics of the northern boundary to change to dextral shear and minor extension (Fig. 11b). This aspect of the model is confirmed by the results of this study, which demonstrate that motions along the TMSSZ were dextral and substantial (>80 kin). Distributed right shear along the TMSSZ resulted in oroclinal folding of rocks about vertical axes along an c. 90 km wide, approximately E - W belt that included the southern Sierra Nevada and the central Mojave desert.

A m o u n t s o f Early Miocene deformation Evidence presented in this paper for >80 km of 21-18 Ma dextral slip along the TMSSZ

198

R.K. D O K K A E T A L . i

\

Farallon

plate

Utah iColo.

/

2-~..........

MTJ

',Mex.

MEB -'WM

ac'--iurezone Pioneer_..... fracture zone

"7~'+ i

|

I

Pacific plate

x

A

9

}he

RTJ

N

A

300 km Utah Colo. Ariz. New IMex.

Farallon 1 ~ plate MTJ ~,~ondocino fracture zone

. . . . . .

. . . . . .

i i

;N

. . . . . . . .

WM

MD

fracturezone

~-

~

t-

1

',

f Pacific plate

-"~,

SMB

zone

N

+Okm@

IB /

. Xboun Sd:~ue~ertain ~

Fig. 11. Model for early Miocene tectonics associated with collapse of the southern Cordillera (modified from Dokka & Ross 1995). Divergence at the plate boundary created a gravitational instability that triggered southwest directed movement of a large mass of the North American plate (noted as 'Region of Collapse'). Areas of extension (blackened) formed along the northern and eastern edges of the region of collapse (see Fig. 12 for 3D character). As the western edge of the collapsed region crept west past the line connecting the two triple junctions, i.e. the Pacific-North American transform, the leading edge came under the influence of the NW motion of the Pacific plate. Eventually, pieces of the leading edge of North America were cut off and began to move with the Pacific plate. This process permanently transferred crustal fragments such as the Salinian Block to the Pacific plate. (a) Situation at 24-21 Ma. (b) Situation at 21-18 Ma. Extended terranes (blackened) shown diagramatically. DV, Death Valley; IB, Inner borderland; MTJ, Mendocino triple junction; MD, Mojave desert; MEB, Mojave Extensional Belt; Mp, Monterey plate; RTJ, Rivera triple junction; NIF, Newport-Inglewood fault; S, Salinian Block; SAF, San Andreas fault; SM, South Mountain; SMB, Santa Maria basin; SN, Sierra Nevada; SCT, Santa Catalina and Tortolita Mountains; TMSSZ, Trans Mojave-Sierran shear zone; WM, Whipple Mountains; WT, Western Transverse Ranges block; state names abbreviated. Single-barbed arrows, local kinematics; heavy black arrows, motions relative to the North American plate (Stock & Molnar 1988).

THE TRANS MOJAVE-SIERRAN SHEAR ZONE

199

Fig. 12. A 3D rendering of the collapse of southwestern North America at c. 18 Ma (see Fig. 11 for discussion and guide to abbreviations).

provides an important test of the notion that events in the continent were kinematically linked to the divergence along the transform. The Dokka & Ross (1995) theory requires that the magnitudes and rates of displacement in each of the interconnected tectonic elements should be equivalent. These elements include: (1) SW directed extension in southeastern California and southern Arizona; and (2) the amount of the North American plate truncated and transferred to the Pacific plate (e.g. Salinian Block, etc.). The >80 km of slip on the TMSSZ reported here falls within the limits of the amount of divergence predicted between Pacific and North American plates (175 _+ 100 km; Stock & Molnar 1988); our comparison does not include the amount contributed by the 24-21 Ma MEB. The magnitude and timing of TMSSZ slip are also similar to those associated with extension occurring along the eastern edge of the collapsed region. The amount of extensional lengthening recorded in rocks of southern Arizona-southeastern California from the edge of the stable Colorado Plateau in early Miocene time is 86___13 km (Spencer & Reynolds 1991). Similar structures lying to the southwest along the Salton trough that are linked in the middle and lower crust by detachment faults imply that the total lengthening is even greater. The Dokka & Ross (1995) model predicts that the amount of TMSSZ net slip and the amount of the North American plate truncated

and transferred to the Pacific plate should be equivalent; the rationale for this is illustrated in Figs 11 and 12. According to Dokka & Ross, the transtensional nature of the Pacific-North American plate boundary led to the profound fragmentation of the west coast of California and Mexico by strike-slip faults of the San Andreas fault system. They reasoned that, as the margin of North America collapsed to the west, the continental edge moved west past the locus of transform shear between the Mendocino and Rivera triple junctions. Thus, the portion of the plate lying west of the locus of shear became increasingly subject to the motion of the Pacific plate. Eventually, a new fault was propagated between the triple junctions. Over time, slivers of the North American continent such as the Salinian Block (Page & Engebretson 1984) were progressively truncated in meat-slicer-like fashion and transferred to the Pacific plate (Figs 11 and 12). If the Dokka & Ross (1995) model is correct, then the TMSSZ net slip should be equivalent to the total map width of the fault blocks that were cut off and transferred to the Pacific plate by the San Andreas system. In central California, the combined width of these fault blocks lying between the edge of the continent and the San Andreas fault is c. 150 kin. We believe this discrepancy reflects the incomplete status of our analysis of the TMSSZ (i.e. >80 kin) and suggest that the true value of the net slip along the TMSSZ may be closer to 150 km.

200

R.K. DOKKA E T A L .

Conclusions

The following conclusions were reached in this analysis of the TMSSZ: (1) The overall motion along the 21-18 Ma TMSSZ was dextral with at least 80 k m of net slip. Although the first-order character of the TMSSZ was right-slip, internal d e f o r m a t i o n occurred by large clockwise vertical rotations of regions that were facilitated by m o v e m e n t s along syntectonic sinistral shear zones (second order); motions along individual left-slip zones may be large (>20 kin) and are thus important contributors to the total deformation. The spatial coincidence of these sinistral shear zones with transfer zones of the 24-21 Ma MEB suggests that the older structures may have controlled and guided TMSSZ deformation. It is also likely that the original trend and position of the entire TMSSZ may have b e e n inherited from the MEB. (2) The original trend of the TMSSZ was probably E N E but rotated 30-40 ~ clockwise during its evolution. (3) The g e o m e t r y and kinematics of the TMSSZ determined here are consistent with the unified theory of D o k k a & Ross (1995), who proposed that the TMSSZ is a regional dextral shear zone that served as the kinematic link b e t w e e n inland extensional belts ( A r i z o n a metamorphic core complexes, Colorado River Extensional Corridor, Mojave Extensional Belt) and the transtensional Pacific-North American plate boundary (Atwater 1970, 1989; Ingersoll 1982; Stock & Molnar 1988; Severinghaus & Atwater 1990). Careful and constructive reviews were provided by T. Pavlis, B. Tikoff, and M. Woodburne. Insightful discussions were provided by Z. Garfunkel, C. Travis, D. Henry, R. Holdsworth, D. MacConnell, B. Temple, G. Sella, and C. Christensen. Special thanks are extended to R. Holdsworth, J. Dewey, and R. Strachan, for their hard work and good cheer in the creation of this volume. This study was supported by grants from the National Science Foundation (EAR-9004339, EAR9219191, and EAR-9418834). References

ATWATER,T. 1970. Implications of plate tectonics for the Cenozoic tectonic evolution of western North America. Geological Society of America Bulletin, 81, 3513-3536. 1989. Plate tectonic history of the northeast Pacific and western North America. In: Winterer, E.L. et al. (eds) Geology of North America, Volume N. Geological Society of America, Boulder, Co, 21-71. BARTLEY,J. & GLAZNER,A. 1991. En echelon Miocene

rifting in the southwestern United States and model for vertical axis rotation in continental extension. Geology 19, 1165-1168. BOHANNON, R. & PARSONS,T. 1995. Tectonic implications of post-30 Ma Pacific-North American relative motions. Geological Society of America Bulletin, 107, 937-959. BRADY, R. a., III. 1990. Neogene stratigraphy of the Avawatz Mountains between the Garlock and Death Valley fault zones, southern Death Valley, California. Implications as to the late Cenozoic tectonics. Sedimentary Geology, 38, 127-157. BURCHFIEL,B. C. & DAVIS,G.A. 1981. Mojave Desert and surrounding environs. In: ERNST,W. G. (ed.) The Geotectonic Development of California. Prentice-Hall, Englewood Cliffs, NJ, 217-252. BURKE,D. B., HILLHOUSE,J.W., MCKEE,E. H., MILLER, S. T. & MORTON,J. L. 1982. Cenozoic rocks in the Barstow Basin area of southern California Stratigraphic relations, radiometric ages and paleomagnetism. US Geological Survey Bulletin, 1529-E, 1-16. CHEADLE, M. J., CZUCHRA,B. L., BYRNE,T., ANDO, C. A., OLIVER,J. E., BROWN,L. D., KAUFMAN,S., et al. 1986. The deep crustal structure of the Mojave desert, California, from COCORP seismic reflection data. Tectonics, 5, 293-320. DAVIS, G. A. & LISTER, G. 1988. Detachment Faulting in Continental Extension. Perspectives from the Southwestern U.S. Cordillera. Geological Society of America, Special Paper, 218, 133-159. , ANDERSON,J. L., FROST,E. G. & SHACKELFORD,T. J. 1980. Mylonitization and detachment faulting in the Whipple-Buckskin-Rawhide Mountains terrane, southeastern California and western Arizona. In: CRITI~ENDEN,M. D., CONEY, P. & DAVIS,G. H. (eds) Cordilleran Metamorphic Core Complexes. Geological Society of America, Memoir, 153, 79-130. DAVIS, G. H. & CONEY,P. J. 1979. Geologic development of Cordilleran metamorphic core complexes. Geology, 7, 120-124. DIBBLEE,T. W., JR 1968. Geology of the Fremont Peak and Opal Mountain quadrangles, California. Bulletin of the California Division of Mines and Geology, 188, 1-64. 1975. Late Quaternary uplift of the San Bernardino Mountains on the San Andreas and related faults. California Division of Mines and Geology Special Report, 118, 127-135. DIEHL,J. E, MCCLANNAHAN,K. M. & BORNHORST,T. J. 1988. Paleomagnetic results from the Mogollon-Datil volcanic field, southwestern New Mexico and a refined mid-Tertiary reference pole for North America. Journal of Geophysical Research, 93, 4869--4879. DOKKA, R. K. 1980. Late Cenozoic tectonics of the central Mojave Desert, California. Ph.D. dissertation, University of Southern California. -1983. Displacements on late Cenozoic strike-slip faults of the central Mojave Desert, California. Geology, 11, 305-308. 1986. Patterns and modes of early Miocene

THE TRANS MOJAVE-SIERRAN SHEAR ZONE extension of the central Mojave desert, California. In: MAYER,L. (ed.) Continental Extension Processes. Geological Society of America, Boulder, Special Paper, 208, 75-95. 1989. The Mojave Extensional Belt of southern California. Tectonics, 8, 363-390. 1993. The Eastern California shear zone and its role in the creation of young extensional zones in the Mojave desert region. In: CRAIG, S. D. (ed.) Structure, Tectonics and Mineralization of the Walker Lane. Geological Society of Nevada, Reno, Walker Lane Symposium Volume, 161-187. & Ross, T. M. 1995. Collapse of southwestern North America and the evolution of early Miocene detachment faults, metamorphic core complexes, the Sierra Nevada orocline and the San Andreas fault system. Geology, 23, 1075-1078. --- & -1996. Reply-Collapse of southwestern North America and the evolution of early Miocene detachment faults, metamorphic core complexes, the Sierra Nevada orocline and the San Andreas fault system. Geology, 24, 859-860. & Travis, C. J. 1990a. Late Cenozoic strike-slip faulting in the Mojave desert, California. Tectonics, 9, 311-340. & -1990b. Role of the Eastern California shear zone in accommodating Pacific-North American plate motion. Geophysical Research Letters, 17, 1323-1326. FROST, E., et al. 1987. Crustal transect. Colorado Plateau-detachment terrane-Salton Trough. In: DAVIS, G. H. & VANDENDOLDER,E. (eds) Geologic Diversity of Arizona and its Margins: Excursions to Choice Areas. Arizona Bureau of Geology and Mineral Technology, Tucson, Special Paper, 5, 398-422. GARFUNKEL,Z. 1974. Model for late Cenozoic tectonic history of the Mojave Desert, California, and for its relation to adjacent areas. Geological Society of America Bulletin, 85 1931-1944. GLAZNER, A., BARTLEY, J. & INGERSOLL, R. 1996. Comment-Collapse of southwestern North America and the evolution of early Miocene detachment faults, metamorphic core complexes, the Sierra Nevada orocline and the San Andreas fault system. Geology, 24, 858. GOLOMBEK,M. & BROWN,L. 1988. Clockwise rotation of the western Mojave desert. Geology, 16, 126-130. HAMILTON,W. 1987. Crustal extension in the Basin and Range province, southwestern United States. In: Coward, M., et al. (eds) Continental Extensional Tectonics. Geological Society, London, Special Publications, 28, 155-176. HENRY, D. J. & DOKKA, R.K. 1992. Metamorphism of exhumed middle to lower crustal rocks in the Mojave Extensional Belt, southern California, USA. Journal of Metamorphic Geology, 10, 347-364. HOLDSWORTH,R. E., BUTLER, C. A. & ROBERTS,A. M. 1997. The recognition of reactivation during continental deformation. Journal of the Geological Society, London, 154, 73-78. -

-

-

-

-

-

201

INGERSOLL, R. 1982. Triple-junction instability as a cause for late Cenozoic extension and fragmentation of the western United States. Geology, 10, 621-624. LONSDALE, P. 1991. Structural patterns of the Pacific floor offshore of Peninsular California. In: DAUPHIN,J. & SIMONEIT,B. (eds) Gulf and Peninsular Province of the Californias. Memoirs, American Association of Petroleum Geologists, 47, 87-125. Lu, G. & DOKKA, R. K. 1995. Early Miocene tectonic history of the Kramer Hills of the western Mojave desert. Paleomagnetic evidence for multiple episodes of vertical axis rotations. Los Transactions, American Geophysical Union, 76, F177. LUYENDYK, B., KAMERLING M. & TERRES, R. 1980. Geometric model for Neogene tectonic rotations in southern California. Geological Society of America Bulletin, 91, 211-217. MACCONNELL, D. E, MCCABE, C., DOKKA, R. K. & CHU, M. 1994. Paleomagnetic and structural evidence for localized tectonic rotation associated with fault drag in the northeastern Mojave Desert. Implications for the late Cenozoic tectonic evolution of the Eastern California shear zone. Earth and Planetary Science Letters, 126, 207-216. MACFADDEN, B., SWISHER, C., III, OPDYKE, N. & WOODBURNE, M. 1990a. Paleomagnetism, geochronology, and possible tectonic rotation of the middle Miocene Barstow Formation, Mojave Desert, southern California. Geological Society of America Bulletin, 102, 478-493. - - , WOODBURNE, M., & OPDYKE,N. 1990b. Paleomagnetism and Neogene clockwise rotation of the northern Cady Mountains, Mojave Desert of southern California. Journal of Geophysical Research, 95, 4597-4608. MCCARTHY, J., LARKIN, S., FUIS, G., SIMPSON, R. & HOWARD, K. 1991. Anatomy of a metamorphic core complex: seismic reflection/wide-angle reflection profiling in southeastern California and western Arizona. Journal of Geophysical Research, 96, 12259-12292. MILLER, D., HOWARD,K. & JOHN, B. 1982. Preliminary geology of the Bristol Lake region, Mojave Desert, California. In: COOPER, J. D. (ed.) Geologic Excursions in the California Desert. Geological Society of America, Anaheim, CA, 91-100. MILLER, E K. & MORTON, D. 1980. Potassium-Argon Geochronology of the Eastern Transverse Ranges and Southern Mojave Desert, Southern California. US Geological Survey Professional Paper, 1152. MORRIS, R. 1993. Tertiary basin structure revealed in seismic reflection profiles from Milpitas Wash, southeastern California. US Geological Survey Bulletin, 2053, 217-234. NICHOLSON, C., SORELIEN,C., ATWATER,Z., CROWELL, J. & LUYENDYK,B. 1994. Microplate capture, rotation, of the western Transverse Ranges and initiation of the San Andreas transform as a low-angle fault system. Geology, 22, 491-495.

202

R.K. D O K K A E T A L .

PAGE, B. & ENGEBRETSON, D. 1984. Correlation between the geological record and computed plate motions for central California. Tectonics, 3, 133-155. PEZZOPANE, S. • WELDON, R., II 1993. Tectonic role of active faulting in central Oregon. Tectonics, 12, 1140-1169. POWELL, R. 1981. Geology of the crystalline basement

complex, eastern Transverse Ranges, southern California. PhD dissertation, California Institute of Technology, Pasadena. Ross, T. M. 1988. Neogene tectonic rotations in the central Mojave desert, California. MS thesis, University of California, Santa Barbara. -1994. Neogene extension and regional rotation of the central Mojave desert, California. PhD dissertation, Louisiana State University, Baton Rouge. 1995. North-south directed extension and timing of extension and vertical-axis rotation in the southwest Cady Mountains, Mojave desert, California. Geological Society of America Bulletin, 107, 793-811. , LUYENDYK, B. E & HASTON, R. B. 1989. Paleomagnetic evidence for Neogene clockwise tectonic rotations in the central Mojave desert, California. Geology, 17, 470-473. SAUBER, J., THATCHER,W. & SOLOMON,S. 1986. Geodetic measurement of deformation in the central Moj ave desert, California. Journal of Geophysical Research, 91, 12683-12694. , , & LISOWSKI, M. 1994. Geodetic slip rate for the Eastern California shear zone and the recurrence time of Mojave Desert earthquakes. Nature, 367, 264-266. SAVAGE, J. C., LISOWSKI,M. & PRESCOTT,J. 1990. An apparent shear zone trending north-northwest across the Mojave desert into Owens Valley, eastern California. Geophysical Research Letters, 17, 2113-2116. SERPA,L. & DOKKA, R. 1992. Geometry of the Garlock fault zone based on seismic reflection data. Journal of Geophysical Research, 97, 1529715306. SEVERINGHAUS, J. & ATWATER, Y. 1990. Cenozoic geometry and thermal state of the subducting slabs beneath western North America. In: -

-

WERNICKE, B. (ed.) Basin and Range Extensional

Tectonics near the Latitude of Las Vegas, Nevada. Geological Society of America, Memoir, 176, 1-22. SNOW, J. 1992. Large-magnitude Permian shortening and continental-margin tectonics in the southern Cordillera. Geological Society of America Bulletin, 104, 80-105. SPENCER, J. & REYNOLDS, S. J. 1991. Tectonics of midTertiary extension along a transect through west-central Arizona. Tectonics, 10, 1204-1221. STOCK,J. & MOLNAR,P. 1988. Uncertainties and implications of the Late Cretaceous and Tertiary position of North America relative to the Farallon, Kula and Pacific plates. Tectonics, 7, 1339-1348. TEMPLE, B. 1997. Structural geology of a portion of the

northern Cady Mountains, Mojave desert, California. MS thesis, Louisiana State University, Baton Rouge. VALENTINE, M., BROWN, L. & GOLOMBEK, M. 1993. Cenozoic crustal rotations in the Mojave desert from paleomagnetic studies around Barstow, California. Tectonics, 12, 666-677. WELLS, R. & HILLHOUSE,J. 1989. Paleomagnetism and tectonic rotation of the lower Miocene Peach Springs Tuff: Colorado Plateau, Arizona, to Barstow, California. Geological Society of America Bulletin, 101, 846-863. WERNICKE, B. 1981. Low-angle normal faults in the Basin and Range province-nappe tectonics in an extending orogen. Nature, 291, 645-648. WRIGHT, L. A. & TROXEL, B. 1967. Limitations on right-lateral, strike-slip displacement, Death Valley and Furnace Creek fault zones, California. Geological Society of America Bulletin, 78, 933-949. & 1970. Discussion on Summary of regional evidence for right-lateral displacement in the western Great Basin. Geological Society of America Bulletin, 81, 2167-2174. YOUNT, J., SCHERMER,E., FELGER, T., MILLER, D. & STEPHENS, K. 1994. Preliminary geologic map of

Fort Irwin basin, north-central Mojave Desert, California. US Geological Survey Open-file Report 94-173, 1-20.

Transtension in southeastern Africa associated with Gondwana break-up M. K. W A T K E Y S 1 & D. S O K O U T I S 2

1Department of Geology and Applied Geology, University of Natal, Durban 4041, South Africa 2Hans Ramberg Tectonic Laboratory, Institute of Earth Sciences, Uppsala University, Villaviiggen 16, S-752 36 Uppsala, Sweden Abstract: The southeastern margin of Africa is marked by a zone of complex faulting associated with crustal extension related to the Mesozoic break-up of Gondwana. The oldest age of faulting (180-175 Ma) displays differences in style in regions with Archaean as opposed to Proterozoic basement. Faulting in the Archaean region shows extension normal to the proto-plate boundary whereas transtension prevailed in the Proterozoic region. The north-south trending normal faulting in the Archaean craton, which cuts across basement trends, displays an asymmetry indicating simple shear extension of the lithosphere. In contrast, transtension occurred in the region of Proterozoic basement by ENE strike-slip movement of blocks along terrane boundaries. As these blocks moved, normal faults developed behind them as well as within them at high angles to the terrane boundaries, to accommodate extension. Analogue modelling reproduced the observed different fault patterns. It is proposed that sinistral movement along the interconnected Gastre Fault System and Falkland-Agulhas Fracture Zone, together with the presence of a releasing bend, are the reasons for extension in the Lebombo region. This movement may have been caused by plate edge events between the Patagonia block and the proto-Pacific Ocean.

During continental break-up, the development of continental margins was, until relatively recently, generally considered to be the product of either stretching normal to the margin or strike-slip movement parallel to the eventual coastline. These two e n d - m e m b e r situations yield a predictable tectonic history which can be tested against observations. W i t h the everincreasing knowledge about passive continental margins, it has become clear that not every margin represents one of the end-members. Instead, the e n d - m e m b e r mechanisms may effectively operate in tandem, resulting in transtension where stretching is not normal to the eventual margin. Again, this mechanism can be modelled to reveal the general tectonic evolution that might be expected for different situations. One of the parameters that needs to be considered in such modelling is the lithospheric architecture of the region being affected and how reactivation of pre-existing structures influences the tectonic development. The southeastern margin of Africa is an area in which this has been investigated (Fig. 1). It is also important in terms of Gondwana break-up because it represents the region where Africa, South America and Antarctica were once connected (Martin & H a r t n a d y 1986; Groenewald et al. 1991). The zone of faulting in southeast Africa consists of two parts. The n o r t h e r n

portion is known as the Lebombo 'monocline' and occurs along the eastern margin of the Archaean Kaapvaal craton. It has long been considered to represent early continental rifting during 180 Ma Karoo volcanism (du Toit 1929), with the magmatism ascribed to a plume (Burke & D e w e y 1972; White & McKenzie 1989; Campbell & Griffiths 1990; Storey 1995), subduction-related processes (Cox 1978, 1988) or lithospheric stretching (Gallagher & Hawkesworth 1992). In this northern portion, crustal stretching has always been considered normal to the continental margin whereas the southern portion, sometimes termed the Natal 'monocline' (King 1972), is represented by the coastal faulting in KwaZulu-Natal (Beater & Maud 1960; Maud 1961; von Veh & Andersen 1990) and is generally considered to be associated dextral strike-slip movement related to extraction of the Falkland Plateau at about 132 Ma (Martin 1983). Thus, up to now, the same structural feature has been ascribed to two different events and there has been no satisfactory model which unifies the development of the northern and southern portions. It is clear now that evolution of the southeastern African margin is not due to either the Karoo event or the extraction of the Falklands, but rather is the cumulative product of a prolonged tectonic history that may be subdivided

WATKEYS,M. K. & SOKOUTIS,D. 1998. Transtension in southeastern Africa associated with Gondwana

In: HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. E (eds) 1998. Continental Transpressionaland TranstensionalTectonics.Geological Society, London, Special Publications, 135,

break-up. 203-214.

203

204

M . K . WATKEYS & D. SOKOUTIS

~!i! !!!

.CRATON."

, ..............,..,......,.,,......

9. . . , . . . . . . , . . . . . . . . . . . . . , . . , . . , . . -

I..::..:>.::.-::.,::..::--::..::.."..'

!?!?i!!!?~?!i~?!i~!!i?!!?i!!ii~!ii!!i?iiii!i?i!i??!ii~iiiii?iii!i!i!iiil~.~ !!;il;il;iil;iilliiiiil;iliiiiiiiiii)iii~:" ,:,,..: 7"""'"'""'"'""""'"'""""""'""'"'~ ~:"::"::" ,.....,.

9.......,...........,..,................, ~

9z :i!!!i!!ii!ii!iii!iii!!ii!iii!i!iii!!ii!!ii!:::: ..............................

Ld "1-

-.,,.___

iiiiiiiiiiiiiiiiiiiiiiiiiiiiii 0 !ii!i!i!!ii!!!!i!!!i!!!!i!!iiii!!i!!!iiiiiiiii !!? TT

;ii !ii!!i!ii~~ r

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

-~C RETACEOU S &YOUNGER ,AN

~KAROO VOLCANICS 30~ SEDIMENTS ~ & NATAL GROUP KAROO

---~ NATAL META. PROVINCE ~

PONGOLA

- ~ GRAN ITE GREENSTONE

30

TRANSTENSION IN SE AFRICA into five stages (Watkeys 1997). The first stage (180-175 Ma) is marked by faulting with some rifting along the Lebombo 'monocline' but no actual continental separation at the time of Karoo volcanism (Cox 1992). This was followed (175-155 Ma) by dextral strike-slip movement along the Gastre Fault System-Falkland Agulhas Fracture Zone (Rapela & Pankhurst 1992), which resulted in microplate rotations, most notably of the Falkland Islands (Adie 1952; Mitchell et al. 1986; Ben-Avraham et al. 1993; Marshall 1994). The third stage (155-135 Ma) involved further dextral movement, this time between East and West Gondwana along a fracture system linking Tethys and the Somali basin to the Weddell Sea and proto-Pacific Ocean (Martin & Hartnady 1986; Leitchenkov et al. 1996). In the fourth stage (135-115 Ma), the arrival of the Tristan da Cunha plume beneath a proto-South Atlantic rift caused the onset of sea-floor spreading (Hawkesworth et al. 1992; Turner et al. 1994; Vandecar et al. 1995) and the extraction of the Falkland Plateau from the Natal Valley (Goodlad et al. 1982; Martin & Hartnady 1986). The fifth stage (115-90 Ma) saw the split and final severing of any continental connection between South America and Africa as the whole Atlantic Ocean became linked (Fairhead 1988). The recognition of this protracted history has allowed the faulting of the different stages to be analysed. This paper describes initiation of faulting during the first stage (180-175 Ma) in the southern part of the Lebombo 'monocline' and the coastal faulting of KwaZulu-Natal. It demonstrates that the fault style changes across a major crustal province boundary, with transtension occurring within one of the crustal provinces. The explanation for the faulting appeals to plate edge processes rather than to the plume event which is generally considered to be the cause of the voluminous Karoo continental flood basalts and associated events of this time.

The Lebombo 'monocline' This feature defines the eastern margin of the Archaean Kaapvaal craton, which consists of a granite-greenstone basement overlain by the late Archaean Pongola Supergroup (Fig. 1). In the south the Palaeozoic sediments of the Natal

205

Group partially cover the Precambrian rocks (Thomas et al. 1992), whereas the late Palaeozoic-Mesozoic Karoo Supergroup sediments are widespread, extensively intruded by subvolcanic Karoo dolerite sills (Tankard et al. 1982). Karoo palaeocurrents for the Ecca and Beaufort Groups are derived from the northeast, and then reverse direction towards the northeast for the succeeding Triassic-Jurassic Stormberg Group (Dingle et al. 1983). During the interval of change, the c. 200 Ma Dokolwayo kimberlite was intruded (Allsopp & Roddick 1984), setting a minimum age of deposition for the Stormberg in this region. Along the eastern Kaapvaal craton margin, Karoo basalts and rhyolites dip eastward beneath Cretaceous basalts and sediments of the Mozambique coastal plain. The Rooi Rand sheeted dyke swarm intrudes the Karoo sediments and basalts immediately west of the rhyolites, extending from northern KwaZulu-Natal into central Swaziland, subparallel to the Lebombo 'monocline' (du Toit 1929; Armstrong et al. 1984). The swarm is 10-22 km wide and c. 200 km long, consisting of sheeted dolerite dykes, generally dipping steeply to the west, with variable amounts of intervening countryrock. The dolerites are tholeiites with an 'Ferich' evolution trend, with one generation having a mid-ocean ridge basalt (MORB)-like affinity, which has led to the oft-quoted supposedly 'failed MOR' origin for the dyke swarm. The dykes cut layered intrusions which occur along the Lebombo 'monocline', and the analogy with an early stage in the evolution of the east coast of Greenland has had a strong influence on the models for the Lebombo as well as the persistence of the term 'monocline', which is a misnomer because it is actually a zone of complex faulting and block tilting. The structural grain of the Kaapvaal craton changes in the study area (Fig. 1). The Pongola Supergroup contains north-south trending folds of Archaean age which swing around to east-west close to the contact with the overthrust Natal Metamorphic Province (Tankard et al. 1982). The 'monocline' also strikes north-south but it can be seen further north outside the study area, where the basement trends are ENE-WSW, that its development is not strongly influenced by older structures of the Kaapvaal craton. The 'monocline' causes the

Fig. 1. Simplified geological map of the southern Lebombo and coastal area of KwaZulu-Natal. Faults are shown as continuous lines, with ticks on the downthrown side of normal faults. MgT, Margate terrane; MzT, Mzumbe terrane; TT, Tugela terrane. Inset shows the locality of the map in southern Africa.

206

M. K. WATKEYS & D. SOKOUTIS

9

.

9

.

9

.

9

. ' '

9

.

.

. " .--7.

Fig. 2. Block diagram showing, in the background, the style of faulting along the southern Lebombo on the Archaean Kaapvaal craton and the position of the Rooi Rand dyke swarm 9The foreground illustrates the transtensional coastal faulting in the region of Proterozoic basement and schematically shows preferential dyke intrusion in a zone of normal faulting 9Stippled zone represents the mantle. Karoo Supergroup to dip and be downthrown eastward and the faulting is concentrated in a zone about 25 km wide, with the north-south fault traces being generally straight (Fig. 1). The region is segmented into a series of horsts and graben by planar conjugate faults. Along the eastern side, blocks are tilted mainly towards the east by domino-style rotation, implying a simple shear extensional model (Wernicke 1985) rather than pure shear (Fig. 2, background). This overall fault pattern changes c. 50 km north of the contact with the Natal M e t a m o r p h i c Complex. The fault traces become orientated SW-NE, connecting the north-south faults with the east-west faults that separate the Archaean Kaapvaal craton from the Proterozoic Natal Metamorphic Province. Coastal

faulting

of

KwaZulu-Natal

The Natal Metamorphic Province forms the basement south of the Kaapvaal craton. It

consists of Meso-Proterozoic lithologies that were deformed and metamorphosed at about 1000 Ma; it has been subdivided into three main terranes (Thomas 1989) (Fig. 1). In the north, the Tugela terrane is about 30 km wide, formed from a series of nappes that were overthrust northwards onto the Archaean rocks. Extending for about 200 km south of this is the Mzumbe terrane, which consists of E N E - W S W trending belts of supracrustal rocks intruded and separated by a variety of granitic gneisses. The Margate terrane in the extreme south contains a distinctive suite of metamorphic rocks. These basement rocks are exposed as windows in the Natal Group which covered large parts of the region (Thomas et al. 1992). As with the Kaapvaal craton, this Palaeozoic sequence is overlain by sediments of the Karoo Supergroup which are intruded by Karoo dolerite sills and dykes. The coastal strip north of D u r b a n is intruded by a distinctive suite of younger quartzbearing dolerites (Effingham-type). To the west,

TRANSTENSION IN SE AFRICA the Karoo sediments are topped by the >1500 m of 180 Ma Karoo basalts exposed in the Drakensberg and Lesotho. The zone of coastal faulting in this region is over 70 km wide with fault traces being arcuate, changing from an E N E - W S W strike to north-south (Fig. 1). The interpretation of this fault pattern has been hampered by a number of factors. They are generally not well exposed, as the region is affected by subtropical weathering, which preferentially affects the fault zones except where they contain a silicified breccia (Thomas 1988). Establishing the exact age of the faults is difficult inland because the youngest units present are the 180 Ma Karoo dolerites. However, along the coast and offshore, Cretaceous units provide constraints on the timing. The faults have undergone a number of reactivations and, in places, are still active today. Consequently, the kinematic indicators preserved do not necessarily reflect anything other than the last period of movement. The superficial overall appearance of the fault pattern is of listric normal faults convex towards the coast. However, it is clear that the upthrown block is the apparent hanging-wall block, and the fault planes are steep and often dip inland rather than towards the coast. Superimposed on this pattern is a zone of younger normal faulting along the coast which fits a dip-slip listric model (von Veh & Andersen 1990). This zone, within which the Effingham-type dolerites largely occur, is about 20 km wide in the north tapering southwards to a point about 80 km south of Durban. The older faulting reveals a different pattern. Maud (1961) recognized that all the faults have a normal sense of m o v e m e n t , and T h o m a s (1988) pointed out that the downthrows are variable in side and amount, There is, however, a more systematic variation. In the areas underlain by basement of Mzumbe terrane, the E N E segments of the faults, although having a normal sense of throw, tend to have horizontal to subhorizontal slickenlines indicating an oblique strike-slip movement, with offset markers revealing a sinistral movement. These faults also contain dip-slip slickenlines, so the overall normal movement may be the culmination of more than one event. In contrast, the north-south segments do not have horizontal slickenlines, displaying only dip-slip slickenlines, with offset stratigraphy indicating that they are normal faults. The larger arcuate faults have a downthrow towards the west; in the case of the R e n k e n Fault (Fig. 3), about 600-700 m of movement is indicated (du Toit 1946; Thomas 1988).

207

i

N

o , o krn

:9~ .

N

Fig. 3. The fault pattern at the Mzumbe-Margate terrane boundary, with a schematic block diagram viewed from the southeast. This overall pattern is interpreted as indicating that discrete blocks are pulling out towards the ENE, parallel to the trend of the supracrustal rocks in the basement, utilizing such zones of weakness (Fig. 2, foreground). As the blocks move, extension is accommodated behind the block through normal faulting, resulting in the development of approximately n o r t h - s o u t h grabens. Furthermore, as the sides of the moving blocks are not parallel but tend to diverge eastward, extension occurred approximately at right

208

M.K. WATKEYS & D. SOKOUTIS

Fig. 4. (a) A plan view of an analogue model simulating the fault pattern of Fig. 3 in the vicinity of the Mzumbe-Margate terrane boundary (note that north is down in this photograph). (b) A cross-section through the half of the model containing the plastic sheet separating the ductile-brittle components. This is considered to represent the faulting in the Mzumbe terrane. (e) A cross-section through the half of the model lacking the plastic sheet separating the ductile-brittle components. This is considered to represent the faulting in the Margate terrane.

TRANSTENSION IN SE AFRICA

angles to block movement, resulting in the development of normal faults along the block margins (Fig. 2). It is this interaction of events that has given rise to the complex pattern observed today.

Analogue modelling Analogue modelling was undertaken to simulate the development of the fault patterns mentioned above by scaling the brittle-ductile regimes of the Archaean and Proterozoic provinces. A series of experiments were undertaken in a typical 3D 'sandbox' used for extensional fault experiments (Mandl 1988). In this case a 19 cm • 7 cm Plexiglas container was used which had one end open. A vertical Plexiglas wall was inserted into this opening and held in position while the model was constructed. The brittle upper regime was modelled using sand and the lower ductile regime was represented by silicone putty. The support for the backstop was then removed and the model was allowed to extend by gravitational collapse. The rate of strain and the development of faults were recorded during movement. Once extension had ceased, sections were cut through the model to establish the final structures. The fault pattern in the southern part of the study area is of particular interest because there

209

is a change across the contact between the Mzumbe and Margate terranes (Fig. 3). The Nhlazuka Fault is the main bounding fault on the west side of the moving block within the Margate. It displays a 'lazy-S' trace as it changes strike from SW-NE to north-south where it becomes the R e n k e n Fault. Together with another normal fault to the east that dips eastward, the moving block tapers to a SSW-NNE trending horst in the Mzumbe terrane. This fault pattern was simulated by constructing an analogue model in which the ductile basement was represented by two halves of polydimethylsiloxane (PDMS) (4 cm thick) placed together. A plastic sheet was left on top of one half when the sand (2.1 cm thick) was placed on top of this ductile layer. The two halves were then allowed to extend together, and are considered to represent a situation where 15 km of brittle crust is underlain by 25 km of ductile crust. During extension, the sand overlying the plastic sheet remained a coherent mass, oblivious to extension occurring in the ductile lower portion of the model. This could be represented by a number of geological situations: an extremely competent layer at the brittle-ductile transition in the crust; magma injected horizontally at this level; a region of high fluid pressure; a major structural discontinuity. In all cases, the

210

M.K. WATKEYS & D. SOKOUTIS

requirement is that the extensional strain beneath the layer is not transmitted upwards to the brittle layer directly above the layer. This decoupling resulted in two bounding faults on either side of the coherent block with extension in this half of the model entirely confined beyond them. The other half of the model developed a more evenly distributed extension pattern with a small central horst. The development of this feature was controlled by the bounding faults on the decoupled half. The resulting overall fault pattern in the model is similar to the regional map (compare Fig. 3 with Fig. 4a) although the eastern part is missing, presumably because of later extraction of the Falkland Plateau. The part of the model with the plastic sheet represents the block movement of the Mzumbe terrane (Fig. 4b) whereas the more evenly extended half of the model simulates the structure of the Margate terrane (Fig. 4c). This model demonstrates that the interface between the brittle and ductile portions of the crust plays a controlling role in the development of fault patterns. In the case of the Mzumbe terrane, the decoupling is probably along Proterozoic thrust planes. These are absent in the Margate terrane, allowing a coupling between the brittle and ductile crust in this southernmost region.

Palaeogeography at the time of faulting As mentioned earlier in the paper, the development of the southeastern African margin has been ascribed to various causes. Without wishing to debate the strengths and weaknesses of each model, we present here another possibility, which integrates the geological observations of southeastern Africa, the analogue modelling undertaken in this study, and the more regional studies on this part of Gondwana which have revealed important links that were previously unrecognized. It is well established that, at the time of faulting (180-175 Ma), Gondwana was united. However, whereas the South America-Africa fit is fairly well constrained (Unternehr et al. 1988), the situation with regard to the Africa-Antarctica fit is far from settled. Martin & Hartnady (1986) provided a longitudinally constrained but latitudinally flexible fit, with the continental margin of Dronning Maud Land, Antarctica, juxtaposed with that of southern Mozambique and northern KwaZulu-Natal. De Wit et al. (1988) had a looser fit on their map of Gondwana which still involved some overlap of the continental shelves, whereas Lawver& Scotese (1987) had a tighter fit. Groenewald et al. (1991) also

suggested a tighter fit based on regional geology, whereas Lawver et al. (1991) and Grunow et aL (1991) used computer-fitting based on microplate analysis for their fits. Although geographically the difference between the loose and tight fits may seem rather small, the problem with the tight fits is that they require elimination of the Mozambique Ridge (Fig. 5), which is largely covered by post-Cenomanian sediments with pockets of Neocomian sediments (Simpson et al. 1979). Neither the geophysical evidence (Chetty & Green 1972; Darracott 1974; Scrutton 1976) nor the geochemistry of basalts (Erlank & Reid 1974; Thompson et al. 1982) proved a continental or oceanic origin for the ridge but dredging has confirmed the presence of continental crust (Mougenot et al. 1991). Continental rocks with ages of 1100 Ma and 500 Ma have also been recovered from the southern part of the Agulhas Plateau (Allen & Tucholke 1981; Tucholke et al. 1981). When the northern oceanic part of the Agulhas Plateau is removed, the remainder can be restored against the southern edge of the Mozambique Ridge, lending further support to a continental origin. In light of this, the looser fit is preferred here. It is envisaged that the Patagonia block and Falkland Plateau formed a coherent wedge between Africa and Antarctica (Fig. 5). One side of this triangular piece of crust was defined by a large fault represented by the Gastre Fault System and the Agulhas-Falklands Fracture Zone (hence termed the Gastre-Agulhas Fault Zone when considered to be operating as a single entity). The second side was bordered by fairly wide and loose zone of microplates in the Weddell Sea region (Storey et al. 1996), and along the third side oceanic crust of the protoPacific Ocean was being subducted beneath the Patagonia block (Uliana & Biddle 1987). The Gastre-Agulhas Fault Zone was active during the late Triassic-early Jurassic (Rapela & Pankhurst 1992) and provides the link between geographically widely spaced events. The Gastre Fault System records movement in the Triassic, and, before Karoo volcanism, palaeocurrents in the Stormberg Group sediments indicate subsidence to the east of the Lebombo after about 200 Ma. The 178 +_ 1 Ma silicic volcanism of the Marifil Group, Patagonia, is of exactly the same age as the Lebombo rhyolites (Allsopp et al. 1984). Furthermore, there is evidence of this event in an intervening area through the 177 + 2 Ma partial overprint of Pan-African greywackes in the southwest Western Cape Province, South Africa (Gresse et al. 1992).

TRANSTENSION IN SE AFRICA

211

{!

0 1:13

! ,,,~ %, o;,

0 m uJ _J

;!/J

!

AFRICA

.~"

< w

...._.~~:

I--

u <

/ ,

/

X":.

b-

.',

/

/

EW

f

I

I;

,%

I

G' 0

500 krn-

~

/ /

Fig. 5. Sketch map illustrating the position of the Patagonia block. The inset shows that sinistral movement along the Gastre-Agulhas Fault Zone, combined with a releasing bend, would cause extension along the Lebombo transtension further south in the region with Proterozoic basement. AP, Agulhas Plateau; EW, Ellsworth Mountains; FAFZ, Falklands-Agulhas Fracture Zone; FI, Falkland Islands; GFS, Gastre Fault System; MR, Mozambique Ridge.

This implies relative movement of the Patagonia block along the Gastre-Agulhas Fault Zone with respect to Africa and the rest of South America during this time. Rapela & Pankhurst (1992) suggested major dextral movement during this time on the basis of evidence from Patagonia. However, an analysis of the intrusion mechanism of Karoo dolerite sills between Durban and the Lebombo indicates that the maximum compressional stress was horizontally directed approximately NW-SE at about 180 Ma (Kattenhorn 1994). Such an orientation is more in keeping with the Gastre-Agulhas Fault Zone (or at least the northeastern end of it) undergoing sinistral strike-slip movement at this time. The fault zone defines a small circle, except at the northeastern end near the southern Lebombo where the strike is more northerly

than the projected small circle. The extension direction in the study region indicates a sinistral sense of movement so that this change in strike results in a releasing bend which opened up the Lebombo. The asymmetry faulting of the Lebombo indicates extension by simple shear, domino-style deformation along the eastern part of the fault zone (Fig. 3). The Rooi Rand dyke swarm intruded at the position of thinnest crust where, because of extension, mantle upwelling was the greatest. This explanation accounts for the relatively restricted transtensional faulting further south in the Proterozoic region. It was, essentially, a marginal part of the releasing bend; further southwest, where the Gastre-Agulhas Fault Zone lay on a small circle, only strike-slip movement took place. During extension, pre-existing weak zones were utilized, which, because of

212

M.K. WATKEYS & D. SOKOUTIS

their orientation, assisted with producing a transtensional regime. The amount of extension was less than in the Archaean region, so there was no intrusion of a s h e e t e d dyke swarm. H o w e v e r , the thickest k n o w n K a r o o sills (Insizwa-Mt Currie sills) lie just inland of the normal fault zones that developed as the blocks pulled out. It appears that these zones were preferentially utilized by Karoo magmas, again because of crustal thinning and m a n t l e upwelling. The reason for the m o v e m e n t of the Patagonia block remains obscure. One possibility that requires investigating is the group of events along the b o u n d a r y with the proto-Pacific Ocean. The arrival of a suspect terrane might be responsible for the w e d g e - s h a p e microplate being pushed towards the northeast. A slight r e a r r a n g e m e n t of the loose alliance of microplates in the Weddell Sea could have accommodated such movement. The rotation of the Ellsworth Mountains before 180 Ma (Curtis & Storey 1996) may be an indication of such events taking place. Whatever the reason, the implication is that the initial faulting and transtension along the southeast margin of Africa is ultimately related to plate edge processes affecting the Patagonia block rather than to a mantle plume rising b e n e a t h s o u t h e r n Africa. Conclusions

The reactivation of structures is a fundamental feature of deformation in the continental lithosphere ( H o l d s w o r t h et al. 1997). A l o n g the southeast margin of Africa, some of the older structures of the Proterozoic basement were not only reactivated during the early stages of Mesozoic faulting (180-175 Ma) but also controlled the d e v e l o p m e n t of younger fractures. The t r a n s t e n s i o n a l r e g i m e o b s e r v e d in this region is due to the d e v e l o p m e n t of a releasing b e n d on the sinistral G a s t r e - A g u l h a s Fault Z o n e in the early Jurassic caused by the northeast m o v e m e n t of the Patagonia block with respect to Africa. The p r e s e r v a t i o n of this initial fault pattern is due to its position close to the t e r m i n a t i o n of the G a s t r e - A g u l h a s Fault Z o n e so that it was not overprinted by further m o v e m e n t except along a n a r r o w coastal strip. Discussions with J. McCarthy and R. Uken about the geology of KwaZulu-Natal helped stimulate this study. The comments of two anonymous reviewers on the first draft are greatly appreciated, as was the mediating role of R. Holdsworth.

References

ADIE, R. J. 1952. The position of the Falkland Islands in a reconstruction of Gondwanaland. Geological Magazine, 89, 401-410. ALLEN,R. B. & TUCHOLKE,B. E. 1981. Petrography and implications of continental rocks from the Agulhas Plateau, southwest Indian Ocean. Geology, 9, 463-468. ALLSOPP, H. L. & RODDICK, J. C. 1984. Rb-Sr and 4~ age determinations on phlogopite micas from the pre-Lebombo Group Dokolwayo kimberlite pipe. In: ERLANK,A. J. (ed.) The Karoo Igneous Province. Geological Society of South Africa, Special Publication, 13, 267-271. - - , MANTON,W. I., BRISTOW,J. W. & ERLANK,A. J. 1984. Rb-Sr geochronology of Karoo felsic volcanics. In: ERLANK,A. J. (ed.) The Karoo Igneous Province. Geological Society of South Africa, Special Publication, 13, 273-280. ARMSTRONG,R. A., BRISTOW,J. W. & Cox, K. G. 1984. The Rooi Rand dyke swarm, southern Lebombo. In: ERLANK, A. J. (ed.) The Karoo Igneous Province. Geological Society of South Africa, Special Publication, 13, 77-86. BEATER,B. E. & MAUD,R. R. 1960. The occurrence of an extensive fault system in SE Zululand and its possible relationship to the evolution of part of the coastline of southern Africa. Transactions of the Geological Society of South Africa, 63, 51-64. BEN-AVRAHAM,Z., HARTNADY,C. J. H. & MALAN,J. A. 1993. Early extension between the Agulhas Bank and the Falkland Plateau due to the rotation of the Lafonia microplate. Earth and Planetary Science Letters, 117, 43-58. BURKE,K. & DEWEY,J. E 1972. Plume-generated triple junctions: key indicators in applying plate tectonics to old rocks. Journal of Geology, 81, 406-433. CAMPBELL,I. H. & GRIFFITHS,R. W. 1990. Implications of mantle plume structure for the origin of flood basalts. Earth and Planetary Science Letters, 99, 79-93. CHEqq'Y,19.& GREEN,R. W. E. 1977. Seismic refraction observations in the Transkei Basin and adjacent areas. Marine Geophysical Research, 3, 197-208. Cox, K. G. 1978. Flood basalts, subduction and the break-up of Gondwanaland. Nature, 274, 47-49. - - 1988. The Karoo Province. In: MACDOUGALL,J. D. (ed,) Continental Flood Basalts. Kluwer, Dordrecht, 239-271. - 1992. The Karoo igneous activity, and the early stages of the break-up of Gondwanaland. In: STOREY,B. C., ALABASTER,T. & PANKHURST,R. J. (eds) Magmatism and the Causes of Continental Break-up. Geological Society, London, Special Publications, 68, 137-148. CURTIS, M. L. & STOREY,B. C. 1996. A review of the geological constraints on the pre-break-up position of the Ellsworth Mountains within Gondwana: implications for Weddell Sea evolution. In: STOREY, B. C., KING, E. C. & LIVEMORE, R. A. (eds) Weddell Sea Tectonics and Gondwana Break-up. Geological Society, London, Special Publications, 108, 11-30.

TRANSTENSION IN SE AFRICA DARRACOTT, B. W. 1974. On the crustal structure of southeastern Africa and the adjacent Indian Ocean. Earth and Planetary Science Letters, 24, 282-290. DE WIT, M. J., JEFFERY,M., BERGH, H. & NICOLAYSEN, L. 1988. Geological map of sectors of Gondwana reconstructed to their dispositions c. 150 Ma, 1 : 10 000 000. American Association of Petroleum Geologists, Tulsa, OK. DINGLE, R. W., SIESSER,W. G. & NEWTON,A. R. 1983. Mesozoic and Tertiary Geology of Southern Africa. Balkema, Rotterdam. Du TOIT, A.L. 1929. The volcanic belt of the Lebombo - a region of tension. Transactions of the Geological Society of South Africa, 18, 189-217. -1946. The geology of parts of Pondoland, East Griqualand and Natal Explanation to Sheet 119. Geological Survey of South Africa. ERLANK,A. J. & REID, D. L. 1974. Geochemistry, mineralogy and petrology of basalts, Leg 25 Deep Sea Drilling Project. In: Initial Reports, Deep Sea Drilling Project, 25, US Government Printing Office, Washington, DC, 543-551. FAIRHEAD, J. D. 1988. Mesozoic plate tectonic reconstructions of the central South Atlantic Ocean: the role of the West and Central African rift systems. Tectonophysics, 155, 181-191. GALLAGHER, K. & HAWKESWORTH,C. J. 1992. Dehydration melting and the generation of continental flood basalts. Nature, 358, 57-59. GOODLAD, S., MARTIN,A.K. & HARTNADY,C.J.H. 1982. Mesozoic magnetic anomalies in the southern Natal Valley. Nature, 295, 686-688. GRESSE, P. G., THERON, J. N., FITCH, E J., & MILLER, J. A. 1992. Tectonic inversion and radiometric resetting of the basement in the Cape Fold Belt. In: DE WIT, M. J. & RANSOME,I. G. D. (eds) Inversion Tectonics of the Cape Fold Belt, Karoo and Cretaceous Basins of Southern Africa. Balkema, Rotterdam, 217-228. GROENEWALD,P. B., GRANTHAM,G. H. & WATKEYS,M. K. 1991. Geological evidence for a Proterozoic to Mesozoic link between southeastern Africa and Dronning Maud Land, Antarctica. Journal of the Geological Society, London, 148, 1115-1123. GRUNOW,A. M., KENT,D. V. & DALZIEL,I. W. D. 1991. New palaeomagnetic data from Thurston Island: implications for the tectonics of West Antarctica and Weddell Sea opening. Journal of Geophysical Research, 96, 17935-17954. HAWKESWORTH, C. J., GALLACHER, K., KELLEY, S., MANTOVANI, M., PEATE, D. W., REGELOUS, M. & ROGERS, N. W. 1992. Paranfi magmatism and the opening of the South Atlantic. In: STOREY,B. C., ALABASTER,T. & PANKHURST,R. J. (eds) Magmatism and the Causes of Continental Break-up. Geological Society, London, Special Publications, 68, 221-240. HOLDSWORTH,R. E., BUTLER, C. A. & ROBERTS,A. M. 1997. The recognition of reactivation during continental deformation. Journal of the Geological Society, London, 154, 73-78. KATTENHORN,S. K. 1994. Mechanisms of sill and dyke intrusion. MSc thesis, University of Natal, Durban.

213

KING, L. C. 1972. The Natal Monocline: Explaining the Origin and Scenery of Natal. University of Natal, Durban. LAWYER,L. A. & SCOTZSE,C. R. 1987. A revised reconstruction of Gondwanaland. In: MCKENZIE, G.D. (ed.) Gondwana Six: Structure, Tectonics and Geophysics. Geophysics Monograph, American Geophysical Union, 40, 17-23. - - , ROYER, J.-Y., SANDWELL,D. T. & SCOTESE,C. R. 1991. Evolution of the Antarctic continent. In: THOMSON, M. R. A., fRAME, J. A. & THOMSON,J. W. (eds) Geological Evolution of Antarctica. Cambridge University Press, 533-539. LEITCHENKOV, G. L., MILLER, H. & ZATZEPIN, E. N. 1996. Structure and Mesozoic evolution of the eastern Weddell Sea, Antarctica: history of early Gondwana break-up. In: STOREY,B. C., KING, E. C. & LIVEMORE,R. A. (eds) Weddell Sea Tectonics and Gondwana Break-up. Geological Society, London, Special Publications, 108, 175-190. MANDL, G. 1988. Mechanics of Tectonic Faulting. Models and Basic Concepts. Elsevier, Amsterdam. MARSHALL,J. E. A. 1994. The Falkland Islands: a key element in Gondwana paleogeography. Tectonics, 13, 499-514. MARTIN,A. K. 1983. Fault pattern of coastal Natal - a result of strike-slip motion during Gondwanaland break-up? Joint Geological Survey Technical Report, 14. Marine Geoscience Group, University of Cape Town, 194-214. & HARTNADY, C. J. H. 1986. Plate tectonic development of the south west Indian Ocean: a revised reconstruction of East Antarctica and Africa. Journal of Geophysical Research, 91, 47674786. MAUD, R. R. 1961. A preliminary review of the structure of coastal Natal. Transactions of the Geological Society of South Africa, 64, 247-256. MITCHELL, C., TAYLOR, G. K. & Cox, K. G. 1986. Are the Falkland Islands a rotated microplate? Nature, 319, 131-139. MOUGENOT, D., GENNESSEAUX, M., HERNANDEZ, J., et al. 1991. La ride du Mozambique (Oc6an Indien): un fragment continental individualis6 lors du coulissement de l'Am6rique et de l'Antarctique le long de l'Afrique de l'Est? Comptes Rendus de l'AcadOmie des Sciences, SOrie 2, 312, 655-662. RAPELA, C. W. &; PANKHURST,R. J. 1992. The granites of northern Patagonia and the Gastre Fault System in relation to the break-up of Gondwana. In: STOREY,B. C., ALABASTER,T.& PANKHURST,R. J. (eds) Magmatism and the Causes of Continental Break-up. Geological Society, London, Special Publications, 68, 209-220. SCRUTTON, R. A. 1976. Crustal structure at the continental margin south of South Africa. Geophysical Journal of the Royal Astronomical Society, 44, 601-623. SIMPSON,E. S. W., SCLATER,G. J., PARSONS,B., NORTON, I. & MEINKE, L. 1979. Mesozoic magnetic lineations in the Mozambique Basin. Earth and Planetary Science Letters, 43, 260-264. STOREY, B. C. 1995. The role of mantle plumes in

214

M.K. WATKEYS & D. SOKOUTIS

continental break-up: case histories from Gondwanaland. Nature, 377, 301-308. --, KING, E. C. & LIVEMORE, R. A. (eds) 1996. Weddell Sea Tectonics and Gondwana Break-up. Geological Society, London, Special Publications, 108.

TANKARD, A. J., JACKSON,M. R A., ERIKSSON, K. A., HOBDAY,D. K., HUNTER,D. R. & MINTER,W. E. L. 1982. Crustal Evolution of Southern Africa. 3.8 Billion Years of Earth History. Springer-Verlag, New York. THOMAS, R. J. 1988. The geology of the Port Shepstone area. Explanation of Sheet 3030. Geological Survey of South Africa. - 1989. A tale of two tectonic terranes. South African Journal of Geology, 92, 306-321. , MARSHALL,C. G. A., WATKEYS,M. K., FITCH, E J. & MILLER,J. A. 1992. K - A r and 4~ dating of the Natal Group, southeast Africa: a post PanAfrican molasse? Journal of African Earth Sciences, 15, 453-471. THOMPSON, G., BRYAN,W. B., FREY, E A., DICKEY,J. S. & DAVIES,H. 1982. Petrology, geochemistry and original tectonic setting of basalts from the Mozambique Basin and Ridge (DSDP sites 248, 249 and 250) and from the southwest Indian Ridge (DSDP site 251). Marine Geology, 48, 175-195. TUCHOLKE,B. E., HOUTZ, R. E. & BARRETr,D. M. 1981. Continental crust beneath the Agulhas Plateau, southwest Indian Ocean. Journal of Geophysical Research, 86, 3791-3806. TURNER, S., KELLEY, S., HAWKESWORTH, C.J. & MANTOVANI,M. 1994. Magmatism and continental

break-up in the South Atlantic: high precision 4~ geochronology. Earth and Planetary Science Letters, 121, 333-348. ULIANA, M. A. & BIDDLE, K.Y. 1987. Permian to late Cenozoic evolution of northern Patagonia, main tectonic events, magmatic activity, and depositional trends. In: MCKENZIE, G. D. (ed.) Gondwana Six: Structure, Tectonics and Geophysics. American Geophysical Union, Geophysical Monograph, 40, 271-286. UNTERNEHR, P., CURIE, D., OLIVET,J. L. GOSLIN, J., & BEUZART, R 1988. South Atlantic fits and intraplate boundaries in Africa and South America. Tectonophysics, 155, 169-179. VANDECAR,J. C., JAMES,D. E. & ASSUMP~AO,M. 1995. Seismic evidence for a fossil mantle plume beneath South America and implications for plate driving forces. Nature, 378, 25-31. VON VEH, M. W. & ANDERSEN,N. J. B. 1990. Normalslip faulting in the coastal areas of northern Natal and Zululand, South Africa. South African Journal of Geology, 93, 574-582. WATKEYS,M. K. 1997. A South African perspective of Gondwana break-up. In: HATTON,C. J. (compiler) Plumes, Plates and Mineralisation Abstracts. University of Pretoria, South Africa, 120-121. WERNICKE,B. 1985. Uniform sense of normal simple shear of the continental lithosphere. Canadian Journal of Earth Science, 22, 108-125. WHITE, R. & MCKENZIE,D. R 1989. Magmatism at rift zones: the generation of volcanic margins and flood basalts. Journal of Geophysical Research, 94, 7685-7729.

Transtensional deformation in the evolution of the Bohai Basin, northern China MARK

B. A L L E N 1, D A V I D I. M. M A C D O N A L D VINCENT 1 & CHRISTINE

1, Z H A O

X U N 2, S T E P H E N

BROUET-MENZIES

J.

1

1Cambridge Arctic Shelf Programme, Department of Earth Sciences, Cambridge University, Downing Street, Cambridge CB2 3EQ, UK 2Chinese Academy of Geological Sciences, Beijing, People's Republic of China Abstract: Extensional basins with an element of strike-slip deformation can form because

of a perturbation in a strike-slip fault zone (pull-apart and fault wedge basins), or where extension is oblique to the margins of the deforming zone (transtensional basins). Transtensional basins are characterized by en echelon arrays of normal faults which are individually oblique to the basin margins. The Bohai Basin, northern China, has previously been modelled as either (1) a giant pull-apart between NNE-SSW trending dextral strike-slip faults, or (2) a rift basin caused by WNW-ESE extension, without significant strike-slip deformation. We present a model for the Bohai Basin's rift history in which the basin formed as a result of dextral transtension. The Bohai Basin is one of a family of early Tertiary extensional basins present within eastern Asia from northeastern Russia to southeast China. The structural grain in this basin was inherited from a phase of late Mesozoic sinistral transpression. Tertiary extension began in the Paleocene. Most half-grabens in the eastern and western regions of the Bohai Basin have master faults with a NE-SW or NNE -SSW orientation. Secondary normal faults strike oblique to the main structures, in en echelon arrays which indicate a component of dextral transtension. The central part of the basin, the Bozhong Depression, became a significant depocentre for the first time in the middle Eocene. It formed when activity on transtensional zones to its east and west created an extensional overlap between them. Thus the basin as a whole resembles a giant pullapart basin, with the Bozhong Depression as its central depocentre, but dextral transtension rather than simple strike-slip controlled the deformation. The component of dextral deformation in the Bohai Basin is shared by other early Tertiary east Asian extensional basins, and is consistent with the sense of shear implied by the oblique convergence of the Pacific and Asian plates: an east-west convergence vector applied to a NE-SW trending plate boundary. The consistency of this dextral shear along the Asian margin, and the fact that several of these basins pre-date the India-Asia collision, supports an origin by subduction roll-back of the oceanic Pacific plate from Asia. The extrusion model for east Asian basin formation, whereby extension was caused by lateral transport of lithospheric blocks out of India's northward path following the India-Asia collision, is not applicable to major basins east and northeast of the Red River Fault. T h e Bohai Basin, n o r t h e r n China (Fig. 1), is China's s e c o n d m o s t productive h y d r o c a r b o n basin, with annual oil production in the region of 60 x 106 tonnes. Its structure (Fig. 2) and Paleog e n e stratigraphy are k n o w n f r o m several decades of h y d r o c a r b o n exploration, despite a post-rift fill of N e o g e n e and Q u a t e r n a r y clastic deposits, which ranges in thickness from a few h u n d r e d metres to c. 4 km. T h e N e o g e n e and Q u a t e r n a r y strata obscure early Tertiary n o r m a l and strike-slip faults, which controlled the deposition of a syn-rift succession which reaches a m a x i m u m thickness of c. 6 k m in the B o z h o n g D e p r e s s i o n (Li G u o y u & Lu Minggang 1988). G r e a t e r thicknesses have b e e n claimed (11 k m by H u Jianyi et al. (1989)) but published interpretations of seismic sections do not support this (e.g. Lu B a n g g a n 1991). The dimensions of the

basin are c. 1100 k m N N E - S S W by c. 400 k m W N W - E S E . In plan view, the basin is a giant 'dog-leg', with a central, e q u a n t region and elongate rift zones to the northeast and southwest. This simple shape hides a m o r e complex internal structure, with six regions of significant extensional d e f o r m a t i o n separated from o n e a n o t h e r by five regions relatively unaffected by Tertiary d e f o r m a t i o n (Fig. 2). Chinese p e t r o l e u m industry practice is to refer to these regions as depressions and uplifts, respectively. H y d r o c a r b o n reserves are c o n c e n t r a t e d in the depressions. Smaller regions of subsidence or relative uplift within the depressions are k n o w n as sags and rises (Chang C h e n g y o n g 1991). E a c h sag is typically a graben or half-graben structure. Existing m o d e l s for the basin's structure t e n d either to simplify it as a giant dextral pull-apart

ALLEN, M. B., MACDONALD,D. I. M., ZHAO XUN.,VINCENT,S. J. • BROUET-MENZIES,C. 1998. Transtensional deformation in the evolution of the Bohai Basin, northern China. In: HOLDSWORTH, R. E., STRACHAN,R. A. & DEWEY,J. E (eds) 1998. ContinentalTranspressionaland Transtensional Tectonics. Geological Society, London, Special Publications, 135, 215-229.

215

216

M.B. ALLEN E T A L . 110~

...-

.

9

40 ~

9 '

"

"

.

'

-

"

'

"

"

'

-

"

'

"

"

. - '. ' ,- ". "

" . ' . " " " ' ' ~-

" "

115 ~

..-..:..

.:

'

'

'

120 ~ .

"

'

"

"

"

.

"

.

'

"

"

'

"

i

.

- -. . . - ~ . . ' ' . . . -

'

'

"

125 ~

-..

.

"

....., .--..:...-:.../

' /

"

. . ~"

:

"

"

/- '-. - .-

' -~=~-Bi~iji

i

. -=

' ".

' . -.

, ,

::. :: " i

..... -.

.,.

~

.--

'

9 ..

. ,

.

9. 9 .. - ,

,

.

::?-:,:"..::?

- -T :::.y:::,:

..

.

-

'

-:-~-

9 .

. . .

. . . .

~:~=~

"30 ~

-

9

"

". . . . . . . . .

'."

_:

----

-d-.

~

.

. .

.

"

' -,

~X~-~-~;~'~':W=

......

."

..

. . . .

" '~'

S e a

,

-

..

o~-'--~-

"

'

"

'

/

""

~

~ i x..

-\,

- .

.........

" I

9 ": )\ , - -. h

<-~' . " 9 -... \

. . . .

-

"- "'

~ = : ~

..

C/~ " -

~ : ~ - .

9

....

~~/

/~i~ =

-

~

-

.

J

f

\.--

i'

!'~;, i"- C

. . . . . . .

: .:~ .

":

'+~.:",~

.~r~~:~../ (~~/ d

4"~

. . . . . . . . .

~<-/~:~

I .-. '._.-~:::~ I ~/?::~ ,<~,',u~

~-

Hangjiahu

>~--~~

~

I

"" ~ ~

- . . . . .

"'"'

l

_-~

. . . . . . . :~.~

~ao,

-

"

" >

"~:. ,:ii,

-.-#/;-~"-?'--S,ang,ai

'~ '

~ .9

"

~

~':~"I""

i" ". ~ ? ~

:

9

..

" -- " " ~ ' - " "

"." . " " .

' '

~ ~ ~V~:<

~ ~ ' ~

'"""""-

" .- " -

l'

.... .I " / ~->-----~

' ~. ~"~-,

~ "

.

'

~ - - - ' - - ~ - - ~

:".~.~ i . : ~ ~ : : """""

'

(/ .....- . . 9 .)/ -. . . . .

.... . .

~

:.\--. "

"....

. -....

L - ~ - - ~

'

" . , ..

Yellow

:i .:?'~/'~:::::i,i:;'),-,'~

,. ,.. " . 9

~--:

"

' "

f.' North

.... . . . ...... . .... . .... :.::-~!.: " .... ~

9

'

':~).. ..i : .: -.~

" '. . ,

9 '.' i ' '' : ~ - :. - " : i - . - " ' i . " . 9 " . 9 " . 9 " " 9

L .

35 ~

" . ,.

"

. -- ".' - ~;-~"

~ ; ~ - - '

. : . .. .:: . . -,i : .:I :.i : : . ..i : .:. 9-.

"

:i

130 ~

'"

-

.-

.... =

25 ~

""" ~ : ~ : ~

:

:.

'" "" " "''" n

7 i ." .j> ''~ :.~:~..u~..,oogKoo~ ~i~.>W 'n

"'~"'~!~

0

:',,

.

20 ~

~r "

: Luzoln k,

500

i!-::"

".'

~

Early

.).....

~

Land

Tertiary

extensional

basins

15~

Fig. 1. Location of early Tertiary extensional basins in eastern China. basin (Fig. 3a), between the Tan-Lu Fault along its eastern margin and the Taihang Shan Fault System to its west (Klimetz 1983), or regard it as a rift basin created by roughly W N W - E S E extension (Fig. 3b), with little or no role for strike-slip deformation (e.g. Hong Ye et al. 1985; Liu Guodong 1987). A variant of the pull-apart model was proposed by Nabelek et al. (1987), who postulated that the basic right-stepping geometry of the basin is repeated on several scales within the basin, with a series of nested pull-aparts. Each of these models has obvious difficulties. The pull-apart model and its variations neglect the presence of rift zones in the Liaohe and Linqing depressions at the northeast and southwest of the basin (Fig. 2). Extension has occurred in these regions across c. 100 km, which is comparable with the width of the Central Graben of the North Sea. However,

simple W N W - E S E extension cannot explain the east-west normal faults in the Bozhong Depression, nor the major strike-slip faults parallel to normal faults in the Liaohe Depression (Fig. 2). In this paper we review the structure of the Bohai Basin, by integrating Chinese industry data published for the six petroleum producing areas. These data are structural maps (typically on scales of 1 : 50 000 to 1 : 1 000 000, i.e. from oilfield to depression scale) compiled from 2D and 3D seismic and well data, individual seismic sections, isopach maps and summary logs. Our aim is to understand the mode of basin formation, and reconcile the conflicting models previously proposed. Our preferred model, which combines elements of transtensional and pull-apart deformation, has wider implications for strikeslip related basin formation and emphasizes transtensional basins as a separate category of

EVOLUTION OF BOHAI BASIN, CHINA

217

Fig. 2.

Early Tertiary structures of the Bohai Basin. Uplifts are in large italics, depressions are in largest type, sags and rises are in small type. Names of sags follow Chang Chengyong (1991), names of rises are taken from many sources. Structure compiled from the 15 sources listed by Allen et aL (1997).

basin, in contrast to some schemes which do not distinguish between transtensional basins and pull-apart basins s e n s u s t r i c t o (Ingersoll & Busby 1996), or do not mention transtensional structures at all (Nilsen & Sylvester 1996). For a more detailed account of the syn-rift stratigraphy of the Bohai Basin and its relation to structure, the reader is referred to Allen et al. (1997).

Regional setting The Bohai Basin is one of a family of early Tertiary extensional basins, which are present within the eastern part of Asia from northeastern Russia to offshore southeast China. The locations of the Chinese basins in this group are shown in Fig. 1, and their stratigraphy is summarized in Fig. 4. The regional setting of these basins is in dispute, and schemes for their origin typically fall into one

of two main categories. One group of models invokes far-field effects of the India-Asia collision, and in particular linked strike-slip and extensional faulting as a result of the postulated extrusion of lithospheric blocks to the south and southeast, away from the path of the Indian plate as it indents Asia from the south (Fig. 5a; Tapponnier et al. 1986; Jolivet et al. 1990; Worrall et al. 1996). Support for the extrusion model is claimed from analogue experiments, designed to recreate the tectonic indentation on the laboratory scale (e.g. Cobbold & Davy 1988; Davy & Cobbold 1988). Variants of the experiments deemed to be successful in recreating the pattern of Cenozoic deformation across central and eastern Asia always utilize an unconfined boundary to the east of the indenter, to represent the Pacific-Asia convergent boundary as a 'free face' which does not oppose extrusion. What these

218

M.B. ALLEN E T A L .

(a) Pull-apart model

Fig. 3. Contrasting models for the evolution of the Bohai Basin: (a) dextral pull-apart; (b) rift basin with negligible strike-slip component. (See text for discussion.)

rate in the early Tertiary led to the east Asian extension. Subduction roll-back might also have been achieved by increasing the density of the descending slab (subducting older lithosphere). However, the age of the oceanic lithosphere entering the subduction zone along the Asian margin in the Mesozoic-Cenozoic is not well constrained, so this particular mechanism must remain speculation. We do not examine early Tertiary regional tectonics in detail in this paper, but note that the inception of extension in several east Asian basins, including the Bohai Basin, pre-dated the India-Asia collision at c. 55 Ma (Fig. 4). Biostratigraphic indications of Paleocene or early Eocene ages are supported by radiometric ages from basic volcanic rocks intercalated with the syn-rift sediments (e.g. Xu Lunxun et al. 1995; Yao Yimin et al. 1995). In most basins along the east Asian margin as far south as the Red River Fault (Fig. 1), extension had definitely begun by Eocene times (c. 57-35 Ma), if not before (Gnibidenko & Khvedchuk 1982; Zhang Yuchang et al. 1989; Anon. 1990; Yao Yimin et al. 1994). This makes the extensional deformation in these east Asian basins distinctly older than much of the recorded compressional deformation in the Himalayas, Tibet and Tien Shan, all of which are closer to the India-Asia suture zone than the east Asian basins. For this reason alone, we discount the India-Asia collision as the cause of the early Tertiary extension within eastern Asia, north of the Red River Fault. To the south and southwest of the Red River Fault there are numerous basins sited within or adjacent to strike-slip fault systems, many of which began extension as late as the Oligocene. These basins are more likely to be related to major strike-slip deformation caused by the India-Asia collision, although even in this region the extrusion model has been challenged (Daly et aL 1991).

experiments do not include is an extensional stress across this boundary, which could assist or even cause the rifting process. The second group of models regards the extension as a result of subduction roll-back of the oceanic Pacific Plate from the convergent boundary with the Asian continent (Fig. 5b; Watson et al. 1987). It is difficult to determine the vector of the subducting oceanic slab in the early Tertiary, although the relative motions of the Pacific and Asian plates for the Mesozoic-Cenozoic have been determined (Engebretson et al. 1985). The vector azimuth changed from roughly north-south to east-west between the Late Cretaceous and the Eocene. Northrup et al. (1995) suggested that a drop in the convergence

The Bohai Basin's basement consists of Archaean and Lower Proterozoic metamorphic and igneous rocks, part of a cratonic nucleus known as the North China Block. These rocks are overlain by Middle and Upper Proterozoic marine clastic rocks and carbonates, in turn paraconformably overlain by a Cambrian to Middle Ordovician carbonate succession. A regional unconformity separates the Middle Ordovician strata from mid-Carboniferous to Permian deltaic and fluvial clastic deposits. The Jurassic-Cretaceous Yanshanian orogeny produced widespread magmatism and deformation throughout eastern China. The causes

Pre-Tertiary structure and stratigraphy

EVOLUTION OF BOHAI BASIN, CHINA

219

3 &

&

Age (Ma 9

('3 :...

Plio. :,,:.

10 I1) e-

._o 20

9. . . 9. .

oV

:. "..:.

: ,,.:.

i. '..:!

: ...:.

::

:..:. .:.. ....

~

:. 9 ....

E

,

".'i : . :. '..:.

....

....

'./ .

:.,..

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

. . ...., ... . . .... ] . ....

./:. 9.'.~.

,4

.

::

9

-.

~ i ..

::: J

.-!.

::

: ......

. - . . . . . . . ..

].

......

rv~

30 --

,...

9. . . . . . . . . .

...-. :..:~

9

...

:..' ..':.

... .....

9:-

..:

.i I ',~

-..~.

.;...

,

... :..:.

0 ._

I

I

I

I1

l

0

I

I

I

I

!

,

,4

I

I1

!

.

r

9

(1)

....

I

I

I

!

.

' I t

I

I

l

.

"

' |

' I

"

1 I

!

l |

.

.

.

.

9

!

. 9

' '

=1[

I

l

5o

:

i ......

9

4O

I

I

I

|

.

i

I !

.

,

il =

i

i

.i

,

"

il

i

i

,

,

i

9

I

!

I

i[

|

60 ~ --I1~

,

'4

I

I1

1 [

I

E

!

~J

I

I

I

,

I

I

l

I .

,

i ,

i

i ,

.

"

Ii

9 .

.

It

.

i

.

l

.

I I

1

9 .

.

,

_

J i

,

,1

,

.

Fig. 4. Tectonostratigraphic chart for Cenozoic basins in eastern Asia. Compiled from numerous sources, including Zhang Yuchang et aL (1989), Anon. (1990), Yao Yimin et al. (1994), Wang Guangming et al. (1995) and Xu Lunxun et al. (1995).

and nature of this tectonism are not fully understood, but alternative phases of sinistral transtension and transpression led to the formation of rift basins such as Songliao, and large folds and thrusts, respectively. Examples of the latter are at present exposed west of Beijing in the Taihang Shan, and are also present in the sub-surface beneath the western part of the Bohai Basin (Liu Hefu 1986). The timing of this deformation is not well constrained, probably because of a polyphase deformation history, with both Jurassic and late Cretaceous deformation recognized (Chern & Hsuing 1935; Watson et al. 1987). At least locally, the region of the Bohai Basin underwent Mesozoic extension: over 2 km of Upper Jurassic-Lower Cretaceous non-marine clastic rocks were deposited in halfgrabens in the Jiyang Depression (Li Guoyu & Lu Minggang 1988), which were reactivated during the early Tertiary. The Tan-Lu Fault (Fig. 2) and sub-parallel strike-slip faults northeast of Bohai were active in the Cretaceous, as witnessed by numerous small Cretaceous basins adjacent to these

structures (Bureau of Geology and Mineral Resources of Liaoning Province 1989). Wang Tonghe (1995) described Mesozoic NE-SWtrending anticlines from the eastern side of Bohai; like the Taihang Shan structures, these are arranged in an en echelon array which indicates sinistral transpression. Between these regions with a N E - S W or NNE-SSW structural grain there are examples of Mesozoic folds and thrusts with east-west or E N E - W S W strikes (Fig. 6). These are exposed in the Yanshan, north of the Bohai Basin, and have been mapped in the sub-surface of the Bozhong Depression (Bureau of Geology and Mineral Resources of Hebei Province 1989; Anon. 1990).

Initial Tertiary deformation Tertiary rifting began in the Bohai Basin with the deposition of the Kongdian Formation in half-grabens (Figs 2 and 4). The exact age of rift initiation is not well constrained, but Yao Yimin et al. (1995) quoted a K - A r age of 55 Ma for volcanic rocks near the base of the Kongdian

M.B. ALLEN E T A L .

220

(a) 70=E I

90=E !

110"E \

130"E .... ~,

40~ I

60"N

150*E

Minor Tertiary oceanic basin

Pacific Plate

60*N J

J

50"N

Baikal

150"E 30*N

130"E 20~

"/X'

40~

Philippine Sea Plate 30*N

20*N

y

|

1

70~

~

80"E

90~

',

\

100*E

110~

(b) X

Bohai Basin

South Yellow Sea

,.c,,.,,~. East China Sea-'~,~,:.

Pacific Plate

X~

Asian lithosphere 9 / 9

.:::

:: :::~:: :~::::~:~::~:::~

Subduction ' ' : : : r roll-back ~

Fig. 5. Contrasting models for the evolution of Tertiary extensional basins in east Asia. (a) Extension is a product of lithospheric extrusion out of the path of the Indian indenter. In the model of Tapponnier et al. (1986), Indochina was extruded in the mid-Tertiary (arrow labelled 1), followed by eastern China (arrow 2) and Baikal in the late Cenozoic. More recent models include the Sea of Japan and Sea of Okhotsk as extrusion-related structures. Diagram modified from Tapponnier et al. (1986). (b) Extension results from subduction roll-back of the Pacific plate from Asia (Watson et aL 1987). The location of section X-X' is shown in (a).

EVOLUTION OF BOHAI BASIN, CHINA

~/r

s/

//

Yanshan

S

~/7,g/

~/

I/~'~

.~"~

5"/

.....

, aOOkrn

,

'

,

Fig. 6. Sketch map of Mesozoic compressional structures within and around the Bohai Basin. From Anon. (1990) and Wang Tonghe (1995). Formation (no error quoted). This suggests that the onset of rifting was near the PaleoceneEocene boundary. Some workers place the initial rifting at the base of the Paleocene (Li Desheng et al. 1988; Dai Xianzhong et al. 1993), based on biostratigraphic analysis, but we know of no published radiometric ages which support this. The Kongdian Formation consists predominantly of clastic rocks deposited in alluvial and shallow lacustrine environments. Deposition did not occur across the entire basin, but was mainly restricted to the Jizhong, Huanghua and Linqing depressions (Fig. 2). Less than 300 m of strata were deposited in isolated parts of the Jiyang Depression, and there was almost no sedimentation in the Bozhong Depression (Yao Yimin et al. 1994). The age and assignment of the oldest syn-rift strata in the Liaohe Depression are not agreed; although some workers show local deposits of the Kongdian Formation in this region (Yao Yimin et al. 1994), seismic interpretations typically place the basal Tertiary sediments in the Shahejie Formation (Lu Banggan 1991). The major depocentre during the Paleocene and early Eocene was in the Jizhong Depression, although the maximum thickness of sediments is interpreted differently by different researchers. Li Desheng (1981) showed a thickness of over 6 km of Kongdian Formation in the west of the basin, whereas Xu Jie et aL (1985) depicted part of this same area as having between 2 and 3 km. The structural style in the Jizhong and Linqing depressions is of NE-SW or NNE-SSW trending normal faults, which define the margins of individual half-grabens or grabens. Secondary

221

faults within these sub-basins are typically clockwise oblique to the boundary faults, and either splay off them or are individually isolated within the sedimentary fill of the main half-graben (Fig. 7; Wu Jilong & Lu Xueqi 1986), without necessarily penetrating its basement (Lu Banggan 1991). Variations in thickness of the Kongdian Formation across these secondary faults indicate that they were active during deposition. Direct measurement of kinematic indicators on these faults is not possible because they are nowhere exposed at the surface. Their oblique configuration relative to the major normal faults at the sub-basin boundaries is consistent with a component of dextral transtension during rifting, in accord with current structural models of transtensional basins (Roberts et al. 1990; Seng6r 1996). East-west trending half-grabens in the Jiyang Depression are anomalous to the structural trend elsewhere in the basin at this time, but, perhaps uniquely for the basin, the Tertiary normal faults in this area are reactivations of late Jurassic-early Cretaceous normal faults (Li Guoyu & Lu Minggang 1988). Northwest-southeast trending faults in the basin appear to have acted as transfer zones between different rift segments. These are interpreted to have sinistral displacements (Wang Tonghe 1995), although detailed evidence for this displacement is not provided.

Middle Eocene to Oligocene deformation The Shahejie Formation overlies the Kongdian Formation (Fig. 4). Its age is not precisely constrained, with estimates for the base of the lowermost (fourth) member varying from 50.5 to 42 Ma (Li Desheng et al. 1988; Yao Yimin et al. 1994), i.e. from early to middle Eocene in the time scale of Harland et al. (1990). The nature of sedimentation within the basin changed with the deposition of the third member of the Shahejie Formation, which most workers put as beginning at c. 43-45 Ma (Allen et al. 1997), i.e. middle Eocene. This member consists predominantly of deep lacustrine sediments, including lake-floor turbidites. Extensive fan deltas fed into the lakes within the basin, supplied from both the hanging walls and footwalls of half-grabens. For the first time in the Tertiary there was extension and subsidence across all areas of the Bohai Basin, with all the major faults shown in Fig. 2 being active by the middle Eocene. Peripheral sub-basins show a pattern of secondary normal faults oblique to the main boundary fault of each sub-basin; for example, in the Liaoxi and Xibu sags of the Liaohe Depression, in the northeast of the basin

222

M.B. ALLEN E T A L .

" ~ Early Tertiarynormal fault

~

Basementhigh

Fig. 7. Structural map of the Liaohe Depression. From Qi Jiafu &Chen Fajing (1992) and Anon. (1993). (Fig. 8). We interpret this as representing dextral transtension in these regions, continuing from the early Eocene in the Jizhong and Linqing depressions, but for the first time across most of the eastern part of the basin. Discrete N E - S W or NNE-SSW trending strike-slip faults have also been identified, especially in the Liaohe

/ 120OE

. _ -B- ~ og ~

Depression and the eastern part of the Bozhong Depression (Qi Jiafu & C h e n Fajing 1995), i.e. along the eastern side of the basin (Figs 8 and 9). These have a consistent dextral sense of motion, deduced from the orientation of secondary fault splays to the main fault zone and the offset of other Tertiary fault blocks (Qi Jiafu & C h e n Fajing 1992; Wang Xinyun & Zhang Yamin 1993). Sub-vertical strike-slip faults cut the preTertiary strata of the Dongpu Sag and Liaohe Depression, but splay upwards into negative flower structures which deformed the Tertiary succession (Wang Tonghe 1995). During the deposition of the third member of the Shahejie Formation the Bozhong Depression became the major depocentre in the basin, with the accumulation of a maximum of over 3 km of strata. Deformation in the Bozhong Depression took place predominantly by normal faulting on east-west trending faults (Figs 2 and 10), although there is a considerable variation in strike (Wang Shangwen 1983; Anon. 1990; Li Yancheng 1993). Isolated basement highs within the depression are faulted on most of their margins; the Haizhong Rise is the largest example (Fig. 2). An en echelon system of N E - S W striking normal faults is present in the eastern part of the depression, and is the continuation of the dextral fault system in the Liaohe Depression to its north (Yan Junjun & Ma Qiangrui 1992). This distributed dextral shear zone is approximately along strike from the Mesozoic Tan-Lu Fault Zone to the south (Xu Jiawei 1993), but there is no clear evidence for a single, major strike-slip fault along the eastern side of the Bohai Basin. Extension

11~I=E /~, 41ON

~ ~

122=E /

Liaodong,,

;

~, 42ON

123~ /

! e ....... ........

/

/" /

121~

i I

Early Tertiary strike-slip fault ~ , Early Tertiary normal fault

",.r

r 122"E f

'~

40=N

#,~

~ . Tertiary pinchout at basin margin ~

123~E

~

0

50km

i

r

41"N

k,.. Modern coastline

Structural high within basin

Fig. 8. Structural map of the northern part of the Jizhong Depression. After Wu Jilong & Lu Xueqi (1986) and Anon. (1988).

EVOLUTION OF BOHAI BASIN, CHINA

223

NW

SE Liaoxi Sag

Liaozhong Sag

LiaodongSag

(kin)

8 ~

|174

~

l ,~b

Neogene & Quaternary

~

ShahejieFm.

DongyingFm.

~

Pre-Tertiary

Mud diapir

Fig. 9. Cross-section through the Liaohe Depression. Location shown in Fig. 2. The boundary within the Shahejie Formation represents the top of the third member. From Qi Jiafu & Chen Fajing (1992). S

N

Neogene & Quaternary Lower TertiarySyn-riftfill Pre-Tertiary

L

50km

J

Fig. 10. North-south cross-section through the Bozhong Depression. Location shown in Fig. 2. From Wang Shangwen (1983). The original figure does not show a vertical scale. continued on east-west faults in the Jiyang Depression from the early Eocene into the late Eocene. Final stages of syn-rift sedimentation in the Bohai Basin resulted in the deposition of the Dongying Formation (Fig. 4). This unit contains major deltaic systems, and suggests that sedimentation was able to keep up with and exceed tectonic subsidence rates, in contrast to the sediment-starved half-grabens developed during deposition of the lower part of the Shahejie Formation. The pattern of faulting did not change markedly through the Oligocene. Major rifting ceased at or near the Oligocene-Miocene boundary, when there was mild inversion and exhumation of the syn-rift strata (Guo Suiping et al. 1996; Allen et aL 1997), before the deposition of a postrift, Miocene-Quaternary alluvial succession. These Neogene and Quaternary rocks represent deposition during overall thermal subsidence of the basin (Hong Ye et al. 1985), although there was minor, localized rifting, e.g. on the south side of the Bozhong Depression (Fig. 10). Historical seismicity, and the accumulation of up to 1000 m of Quaternary sediment in well-defined regions

in the northwest of the basin, suggest there may have been a resumption of strike-slip and extensional deformation (Chen Wangping & Nabelek 1988). Discussion Characteristics o f transtensional basins

Sedimentary basins can be related to strike-slip deformation in several ways, as emphasized by reviews (Biddle & Christie-Blick 1985; Seng6r 1996). Some are created where the geometry of a strike-slip fault zone is more complex than a single, linear feature. Pull-apart (stepover) and fault wedge basins both belong to this category. There is zero extension across the deformation zone in a true pull-apart basin. Transtensional basins form where the deformation zone is oblique to the extension direction, producing a component of slip parallel to the zone, and a greater-than-zero component of extension across it. In this latter respect transtensional basins are crucially different from pull-aparts sensu stricto. Many transrotational basins

224

M.B. ALLEN ETAL.

Fig. 11. Idealized structure of a dextral transtensional basin, with sedimentary fill removed. Individual normal faults within the basin are oblique to a major normal fault at the basin boundary, and collectively form an en echelon array. Both the en echelon faults and the fault blocks they define rotate clockwise as deformation proceeds. Localized strike-slip faulting occurs within the basin and/or at the basin margins. Large arrows indicate overall motion sense between left and right margins of deformation zone.

(Ingersoll 1988), defined by the rotation of faultbounded crustal blocks, are also transtensional basins. Basins formed by transtension are commonly characterized by en echelon arrays of normal faults oblique to the boundaries of the deformation zone (Fig. 11). One or both margins of the transtensional zone may be a normal fault system (Fig. 11), such as in the Triassic-Jurassic basins in the High Atlas of Morocco (Beauchamp 1988), or a strike-slip fault, such as the Loreto region of Baja California in the Pliocene-Quaternary (Zanchi 1994). However, the boundaries of the transtensional deformation zone may not be m a r k e d by any major structures, with the zone ending at the lateral tips of the en echelon arrays. The neotectonics of Central Greece is an example of this structural style (McKenzie & Jackson 1986). Hybrid styles are possible, and Fig. 11 illustrates the case where the d e f o r m a t i o n zone resembles a half-graben, with one faulted and one unfaulted margin along its length. In all cases outlined above, a component of slip parallel to the deformation zone can be accommodated by the rotation of the oblique fault blocks within it; an interesting consequence is that slip on the oblique faults includes a strike-slip element, which is antithetical to the main sense of shear (McKenzie & Jackson 1986).

Kinematics of Tertiary transtension in the Bohai Basin Individual regions of the Bohai Basin commonly have a transtensional motif, in that secondary normal faults are orientated in en echelon arrays within half-grabens, with the array as a whole orientated parallel to the bounding fault of the half-graben, but with individual faults oblique to it. The resemblance of the schematic fault array in Fig. 11 to parts of the Langfang-Gu'an and Baxian sags (Fig. 7) and the Liaohe Depression (Fig. 8) should be noted. The sense of strike-slip implied by the en echelon fault arrays in the Bohai Basin is consistently dextral, and suggests that dextral transtension operated throughout the rift phase of the basin. Because we cannot derive direct kinematic data from exposed fault planes, or fault plane solutions from earthquakes for an unexhumed basin which finished major rifting c. 25 Ma ago, this geometrical comparison is perhaps the only tool available for the identification of ancient transtensional deformation. The overall shape of the Bohai Basin resembles a dextral pull-apart, with east-west normal faults in the central part of basin, the Bozhong Depression, flanked by elongate rift systems where the dextral transtension is most pronounced. The early phase of deformation was focused on the western part of the basin (Fig.

EVOLUTION OF BOHAI BASIN, CHINA b Middle Eocene: 3rd Member

aPaleocene-early Eocene: Kongdian Formationj/'/

Shahejie

225 Formation,

-/.C.,o0 Liaohe//Iranstension ,. / ~".'5,~176 f

no extension in this region

Iocal~sed~/1 H{a~nc.ua / g

.:

9

1 u

Z Y'~'. / ~ ~-" --,,-/r /rapid southward i/'_ f f ~ ~ / .~'n propagation or " " t r162176 / z"ffffnon - / ~ /"' L " faultsIrom ho~ng"..,,_ kiaohe Depression

.''"

.'"

transtension,

N

/ ..;~ Linqi

.

9

Z'_/~,ff' j~/. ..-/.

, lOOkm,

I Z~, Huangnua -"~--"--J- / ~

'Linqing/ L / ~ M a o r j . -S~extens,o~n ' In ~ J in Bozhong and ,~, Jiyang depressions ~,.~ beginsat this time

Fig. 12. Schematic diagram for the Tertiary evolution of the Bohai Basin. In (a) (Paleocene-early Eocene) sedimentation occurs as result of dextral transtension in narrow, isolated rifts in the Jizhong, Liaohe, Linqing and southern Huanghua depressions. In (b) (middle Eocene) rifting has propagated southwards from the Liaohe Depression, and possibly northwards from the eastern Jiyang Depression, to create an extensional overlap that is the Bozhong Depression. Extension on east-west normal faults in this region led to rapid subsidence and deposition of the thickest successions of the Shahejie Formation known from the Bohai Basin. From Allen et al. (1997). 12a). The rarity of the Kongdian Formation in the Bozhong Depression and along the entire eastern side of the basin suggests that deformation in these areas did not begin until later, when the Shahejie Formation was deposited. This may be c. 10 Ma after initial rifting in western regions. When dextral transtension was established in separate regions in the east and west of basin, it created an overlap zone between them. We suggest that the east-west faults in this intervening central region developed to accommodate the extensional strain created by the overlap, in a manner analogous to, but more complex than, classic pull-apart basins (Fig. 12b), and that this occurred over a much larger area than typical pull-aparts. The first Tertiary transtensional structures formed where there were preferentially orientated N E - S W or NNE-SSW Mesozoic crustal fabrics. East-west trending normal faults in the Bozhong Depression only formed during the later pull-apart phase. Thus the combined late Mesozoic-Cenozoic tectonics of the Bohai region is one of ?late Cretaceous sinistral transpression, the kinematics of which was reversed by early Tertiary dextral transtension.

Cause o f Tertiary transtension We prefer subduction roll-back of the Pacific plate as the cause of extension in the Tertiary

basins of east Asia north of the Red River Fault, for the reasons outlined in the description of its regional setting, and emphasize that the timing of initial extension pre-dates the India-Asia collision in several major basins (Fig. 4). The component of dextral deformation in Bohai can also be explained as a Pacific effect. The Pacific-Asia convergence vector switched azimuth from N W - S E to east-west between the late Mesozoic and the early Tertiary (Engebretson et al. 1985). The east Asian continental margin has a rough N E - S W alignment, so whether the stresses across the boundary were tensional or compressional a dextral component of shear would be imparted (Fig. 13). A major test of this model is the presence or absence of similar dextral shear in other east Asian basins. From a brief survey of available data on the distribution and motion sense of faults in other basins, we conclude that dextral shear has been involved in the formation of the Sea of Okhotsk, Sea of Japan, East China Sea, Pearl River Mouth, Beibu Gulf and Qiongdongnan basins, and relate this consistent d e f o r m a t i o n style to shear imparted at the Pacific-Asia plate boundary (Fig. 13). There has been much research on strain partitioning at oblique compressional convergent plate boundaries (e.g. McCaffrey 1996). A typical structural style in an overriding continental plate is for the strain to be partitioned by a thrust system and a discrete strike-slip fault

226

M.B. ALLEN

/~.fj

j.j\

.,..)

/i

~

l

,-

/

:

"

i

9

ASIA Bohai ,~

"%:,, )

I-C./// PACIFIC

,~

""

,/.?

9 .

..1//

Chin~

proto South

, /i

7. y 9 /

\

40Ma

Fig. 13. Reconstruction of Pacific-Asia plate boundary in the early Tertiary (c. 40 Ma). East-west convergence took place across a plate boundary aligned roughly NE-SW, imparting a component of dextral shear to the extensional deformation in the Asian continental crust. Dextral shear interpreted from structural maps of Jin Qinghuan (1989), Jolivet et aL (1990), Du Deli (1991), Liu Guangding (1992) and Worrall et at. (1996). Names of individual Chinese basins shown in Fig. 1. Barbed line indicates inferred position of Pacific subduction zone; arrows indicate Pacific-Asia convergence vector (Engebretson et aL 1985). zone, both striking parallel to the boundary, but accommodating the b o u n d a r y - n o r m a l and boundary-parallel components of the strain respectively. The Sumatra active margin is an excellent example (Malod & Kemal 1996). There has been less work on convergent plate boundary systems where the overriding continental plate is in extension. The example of the east Asian continent in the early Tertiary suggests that the boundary-parallel component of strain may occur by diffuse zones of transtension, rather than major, sharply defined strikeslip faults.

Conclusions Previous studies of the structure of the Bohai Basin have proposed origins as either a giant

ETAL.

pull-apart, or an orthogonal rift basin with no contribution from strike-slip faults. Our analysis of the location, orientation, style and timing of faulting indicates that dextral transtension controlled the early Tertiary deformation, with the central part of the basin, the Bozhong Depression, developed as a result of extensional overlap b e t w e e n t r a n s t e n s i o n a l fault systems to its east and west. As a whole, the basin resembles a classic pull-apart structure, but it is crucially different in that its size is much greater than typical pull-aparts, and the structures controlling d e f o r m a t i o n are transtensional, not simple strike-slip faults. A possible factor in the distribution of the early Tertiary normal faults is the presence of late Mesozoic compressional structures in the basement to the Tertiary basin. These have a distribution which suggests they originated by sinistral transpression, with the b a s e m e n t to the Bozhong D e p r e s s i o n being a large push-up zone on east-west thrusts between elongate zones of en echelon N E - S W and N N E - S S W trending thrusts and folds (Fig. 6). The scarcity of N E - S W or N N E - S S W faults in the basement of the Bozhong Depression may explain why this area did not respond to initial Tertiary transtension in the manner of areas to its east and west (Fig. 12a). Transtensional basins are distinct from pullapart basins because of their c o m p o n e n t of extension orthogonal to the strike-slip direction. There are features which are typical of transtensional basins in general (Fig. 11). In particular, there are en echelon arrays of normal faults within the deformation zone, which are oblique to its boundaries, and rotate as deformation proceeds. Displacement on each of these en echelon faults contains a strike-slip element, which is antithetical to the sense of slip parallel to the deformation zone. The boundaries to the deformation zone may be normal faults or strike-slip faults, or the en echelon faults may die out along strike, so that no individual structure is parallel to the margin of the resultant basin. Hybrids of these conditions are possible, and many of the individual sags in the Bohai Basin resemble halfgrabens, with a master normal fault along one side, and an array of oblique normal faults within the sag which die out away from the master fault (Fig. 11). Similarities between the timing of extension in the Bohai Basin and other early Tertiary basins within east Asia suggest that they share a common origin. Because several basins, including the Bohai Basin, pre-date the onset of the India-Asia collision we reject the hypothesis that they arise from the collision. Subduction

EVOLUTION OF BOHAI BASIN, CHINA roll-back of the Pacific plate f r o m the A s i a n m a r g i n is a b e t t e r e x p l a n a t i o n for r e g i o n a l extension. D e x t r a l shear is a p p a r e n t in the structure of several m a j o r basins b e t w e e n the Sea of O k h o t s k a n d the S o u t h C h i n a Sea. We i n t e r p r e t this as the result of e a s t - w e s t extension b e t w e e n the Pacific and A s i a n plates acting on a contin e n t a l m a r g i n aligned r o u g h l y N E - S W . This work originated as part of the CASP China Basins Project, which has been supported by the following companies: AGIP, Amoco, Anderman-Smith, Apache, Arco, Chevron, Conoco, Deminex, Exxon, JNOC, Louisiana Land and Exploration, Mobil, Phillips, Texaco and Unocal. We thank C. Northrup and J. Reijs for their constructive reviews. This is Cambridge University Department of Earth Sciences Contribution 4961.

References

ALLEN, M. B., MACDONALD, D. 1. M., ZHAO XUN, VINCENT,S. J. & BROUET-MENZIES,C. 1997. Early Cenozoic two-phase extension and late Cenozoic thermal subsidence and inversion of the Bohai Basin, northern China. Marine and Petroleum Geology, in press. ANON. 1988. Petroleum Geology of China, Vol. 5 (Huabei). Petroleum Industry Press, Beijing. ANON. 1990. Petroleum Geology of China, Vol. 16 (Bozhong). Petroleum Industry Press, Beijing. ANON. 1993. Petroleum Geology of China, Vol. 3 (Liaohe). Petroleum Industry Press, Beijing. BEAUCHAMP,J. 1988. Triassic sedimentation and rifting in the High Atlas (Morocco). In: MANSPEIZER,W. (ed.) Triassic-Jurassic Rifting. Elsevier, Amsterdam, 477-497. BIDDLE,K. Y. & CHRISTIE-BLICK,N. 1985. Deformation and basin formation along strike-slip faults. In: BIDDLE, K. T. & CHRISTIE-BLICK,N. (eds) StrikeSlip Deformation, Basin Formation and Sedimentation. Society of Economic Paleontologists and Mineralogists Special Publications, 37, 1-34. BUREAU OF GEOLOGY AND MINERAL RESOURCES OF HESEI PROVINCE 1989. Regional Geology of Hebei Province. Geological Publishing House, Beijing. BUREAU OF GEOLOGY AND MINERAL RESOURCES OF LIAONING PROVINCE 1989. Regional Geology of Liaoning Province. Geological Publishing House, Beijing. CHANG CHENGYONG 1991. Geological characteristics and distribution patterns of hydrocarbon deposits in the Bohai Bay Basin, east China. Marine and Petroleum Geology, 8, 98-106. CHEN WANGPING & NABELEK, J. 1988. Seismogenic strike-slip faulting and the development of the North China Basin. Tectonics, 7, 975-989. CHERN, M. K. & HSUING, Y. H. 1935. Note on some thrusts in the western hills of Peiping. Bulletin of the Geological Society of China, 14, 535-567. COBBOLD,P. R. & DAVY,P. 1988. Indentation tectonics

227

in nature and experiments, 2, Central Asia. Bulletin of the Geological Institute of Uppsala University, 14, 143-162. DAI XIANZHONG, DU YUNHUA, WANG QINGLONG & WAN JINGLIN 1993. Fission track dating of authigenic apatite from Tertiary sediments in the Jiyang Depression, Shandong. Scientia Geologica Sinica, 28, 283-287. DALY,M. C., COOPER, M. A., WILSON,I., SMITH,D. G. & HOOPER, B. G. D. 1991. Cenozoic plate tectonics and basin evolution in Indonesia. Marine and Petroleum Geology, 8, 2-21. DAVY,P. & COBBOLD,P. R. 1988. Indentation tectonics in nature and experiments, 1, Experiments scaled for gravity. Bulletin of the Geological Institute of Uppsala University, 14, 130-141. D u DELI 1991. Characteristics of the geological structure and hydrocarbon potential of the southwest Taiwan Basin. Marine Geology and Quaternary Geology, 11, 21-33. ENGEBRETSON,D. C., Cox, A. & GORDON, R. C. 1985. Relative Motions between Continental and Oceanic Plates in the Pacific Basin. Geological Society of America, Special Papers 206. GNIBIDENKO,H. S. & KHVEDCHUK,I. I. 1982. The tectonics of the Okhotsk Sea. Marine Geology, 50, 155-198. Guo SUIPING, SHI XIAOBIN,WANG LIANGSHU,WANG JIE & ZHOU GUANGJIA 1996. Thermal history analysis of Dongying Depression, Shengli Oil Region: evidence from apatite fission track. Oil and Gas Geology, 17, 32-36. HARLAND,W. B., ARMSTRONG,R. L., COX,A. V., CRAIG, L. E., SMITH,A. G. & SMITH, D. G. 1990. A Geologic Time Scale. Cambridge University Press, Cambridge. HONG YE, SHEDLOCK, K. M., HELLINGER, S. J. & SCLATER, J. G. 1985. The North China Basin: an example of a Cenozoic rifted intraplate basin. Tectonics, 4, 153-169. H u JIANYI, XU SHUBAO, TONG XIAOGUANG & Wu HUAYUAN 1989. The Bohai Bay Basin. In: ZHU XIA (ed.) Chinese Sedimentary Basins. Elsevier, Amsterdam, 89-105. INGERSOLL, R. W. 1988. Tectonics of sedimentary basins. Geological Society of America Bulletin, 100, 1704-1709. --, & BUSBY, C. J. 1996. Tectonics of sedimentary basins. In: BusBY, C. J. & INGERSOLL,R. V. (eds) Tectonics of Sedimentary Basins. Blackwell, Cambridge, 1-51. JIN QINGHUAN 1989. Geology and Petroleum Resources of the South China Sea. Petroleum Industry Press, Beijing. JOLIVET,L., DAVY,P. & COBBOLD,P. 1990. Right-lateral shear along the northwest Pacific margin and the India-Eurasia collision. Tectonics, 9, 1409-1419. KLIMETZ, M. P. 1983. Speculations on the Mesozoic plate tectonic evolution of Eastern China. Tectonics, 2, 139-166. LI DESHENG 1981. Geological structure and hydrocarbon occurrence of the Bohai Gulf oil and gas basin (China). In: MASON, J. E (ed.) Petroleum Geology in China. Pennwell, New York, 180-192.

228

M.B. ALLEN ETAL.

, Du YONGLIN & HU GUONONG 1988. Petroleum geology of the Langfang-Gu'an Basin, Beijing-Tianjin region, China. In: WAGNER,H. C., WAGNER,L. C.,WANG, F. F. H. &WONG, E L. (eds) Petroleum Geology of China and Related Subjects. Circum-Pacific Council for Energy and Mineral Resources, Earth Science Series, 10, 219-234. LI GuoYu & Lu MINGGANG 1988. Atlas of Oil and Gas Basins in China. Petroleum Industry Press, Beijing. Ll YANCHENG1993. Geological evolution and faulting of the Bahai Sea. Marine and Quaternary Geology, 13, 25-34. LIU GUANGDING 1992. 1:5 000 000 maps series of geology and geophysics of China seas and adjacent regions. Geological Publishing House, Beijing. LIU GUODONG 1987. The Cenozoic rift system of the North China Plain and the deep internal process. Tectonophysics, 133, 277-285. LIU HEFU 1986. Geodynamic scenario and structural styles of Mesozoic and Cenozoic basins in China. Bulletin, American Association of Petroleum Geologists, 70, 377-395. Lu BANGGAN1991. Atlas of Typical Seismic Sections in China. Petroleum Industry Press, Beijing. MALOD, J. A. & KEMAL, B. M. 1996. The Sumatra margin: oblique subduction and lateral displacement of the accretionary prism. In: HALL, R. & BLUNDELL, D. J. (eds) Tectonic Evolution of SE Asia. Geological Society, London, Special Publications, 106, 19-28. McCAVFREY, R. 1996. Slip partitioning at convergent plate boundaries of SE Asia. In: HALL, R. & BLUNDELL, D. J. (eds) Tectonic Evolution of SE Asia. Geological Society, London, Special Publications, 106, 3-18. MCKENZIE, D. & JACKSON,J. 1986. A block model of distributed deformation by faulting. Journal of the Geological Society, London, 143, 349-353. NABELEK,J., CHEN WANGPING & YE HaNG 1987. The Tangshan earthquake sequence and its implications for the evolution of the North China Basin. Journal of Geophysical Research, 92, 12615-12628. NmSEN, T. H. & SYLVESTER,A. G. 1996. Strike-slip basins. In: BusBY, C. J. & INGERSOLL,R. V. (eds) Tectonics of Sedimentary Basins. Blackwell, Cambridge, MA, 425-457. NORTHRUP, C. J., ROYDEN, L. H. & BURCHFIEL, B. C. 1995. Motion of the Pacific plate relative to Eurasia and its potential relation to Cenozoic extrusion along the eastern margin of Eurasia. Geology, 23, 719-722. QI JIAFU • CHEN FAJINO 1992. Structural style in Liaodongwan-Xialiaohe Basin. Oil and Gas Geology, 13, 272-283. -& -1995. Analysis of the Structure of Cenozoic Rift Basins from Xialiaohe to Liaodongwan. Geological Publishing House, Beijing. ROBERTS, A. M., BADLEY,M. E., PRICE, J. D. & HUCK, I. W. 1990. The structural history of a transtensional basin: Inner Moray Firth, NE Scotland. Journal of the Geological Society, London, 147, 87-103.

SengGr, A. M. C. 1996. Sedimentation and tectonics of fossil rifts In: BUSBY,C. J. & INGERSOLL,R.V. (eds) Tectonics of Sedimentary Basins. Blackwell, Cambridge, MA, 53-117. TAPPONNIER, P., PELTZER, G. & ARMIJO, R. 1986. On the mechanics of the collision between India and Asia. In: COWARD,M. R & RIEs, A. C. (eds) Collision Tectonics. Geological Society, London, Special Publications, 19, 115-157. WANG GUANGMINO,COWARD,M. P.,YUANWENGUANG, L~U SHENSHU & WANG WENQIANG 1995. Fold growth during basin inversion - example from the East China Sea. In: BUCHANAN, J. G. & BUCHANAN, P. G. (eds) Basin Inversion. Geological Society, London, Special Publications, 88, 493-522. WANG SHANGWEN1983. Petroleum Geology of China. Petroleum Industry Press, Beijing. WANGTONGHE 1995. Oil-gas distribution regularity of the sedimentary basin to the east of Taihang Shan. Journal of Geology and Mineral Resources of North China, 10, 399-414. WANG XINYUN& ZHANGYAMIN1993. Petroleum geological characteristics and hydrocarbon trap types of the Lanliao Fault Belt. Oil and Gas Geology, 14, 229-237. WATSON, M. R, HAYWARD,A. B., PARKINSON,D. N. & ZHANG, Z. 1987. Plate tectonic evolution, basin development and petroleum source rock deposition onshore China. Marine and Petroleum Geology, 4, 205-225. WORRALL, D. M., KRUGLYAK, V., KUNST, E & KUZNETSOV,V. 1996. Tertiary tectonics of the Sea of Okhotsk, Russia: far-field effects of the India-Eurasia collision. Tectonics, 15, 813-826. Wu JmONG & Lu XUEQI 1986. The structural framework of the eastern Jizhong Depression and its evolution. Oil and Gas Geology, 7, 20-31. Xu JIAWEI 1993. Basic characteristics and evolution of the Tancheng-Lujiang Fault Zone. In: Xu JIAWEI (ed.) The Tancheng-Lujiang Wrench Fault System. John Wiley, Chichester, 17-50. Xu JIE, HaNG HANJING ~; ZHAO GUOZHE 1985. The evolution and kinematics of Cenozoic rift basins in the Huabei plain. Research on Crustal Movement, 1, 26-40. Xu LUNXUN, NAN CHUNDE, YU HUILONG, WANG BAOQING,Yu FANGQUAN& WANG DIANFU 1995. Chronology of Paleogene volcanic rocks in the Jianghan Basin. Oil and Gas Geology, 16,132-137. "fAN JUNJUN & MA QIANGRUI 1992. The Tancheng-Lujiang wrench fault system in the Bahai Bay area. Scientica Geologica Sinica, 17, 31-38. YAO YIMIN, LIANa HONGDE, CAI ZHIGUO, et al. 1994. Tertiary in Petroliferous Regions of China (IV): The Bahai Gulf Basin. Petroleum Industry Press, Beijing. , DUAN ZHIBIN,ZHENG HERONG, Ltu SHOUYI,XIA YURONG & XIANGWEIDA 1995. Guide to the Core Workshop of the International Symposium on Oil-Gas Exploration in Lacustrine Basins. Geological Research Institute of Shengli Petroleum Administration, Shandong.

EVOLUTION OF BOHAI BASIN, CHINA ZANCHI,A. 1994. The opening of the Gulf of California near Loreto, Baja California, Mexico: from basin and range extension to transtensional tectonics. Journal of Structural Geology, 16, 1619-1639. ZHANG YUCHANG, WEI ZILI, Xu WELLING, TAO

229

RUIMING & CHEN RUIGENG 1989. The North Jiangsu-South Yellow Sea Basin. In: ZHU XIA (ed.) Chinese Sedimentary Basins. Elsevier, Amsterdam, 107-123.

Transpressional tectonics and strain partitioning during oblique collision between three plates in the Precambrian of southeast Brazil HANS DIRK EBERT & YOCITERU HASUI Departamento de Petrologia e Metalogenia, Instituto de GeociOncias e Ci~ncias Exatas, U N E S P - U n i v e r s i d a d e Estadual Paulista, P. O. B o x 178, 13.506-900 Rio Claro-SP, Brazil Abstract: Two orogenic belts have been recognized in south-east Brazil, which are interpreted to have been formed as a product of diachronous collisions between three continental plates. Wide crustal-scale shear belts have developed both between and inboard of the collided and amalgamated plate borders. These shear belts record frontal, oblique or lateral displacements during oblique plate convergence and A-type subduction. The overall structural style of each belt depends on the angle subtended between the plate boundary and the convergence vector. The E-W branch between the S~o Paulo and Brasflia plates, the Campo do Meio strike-slip shear belt, has undergone dominantly sinistral wrench dominated transpression along a set of folds and shear zones dipping southwards. The NE-SW branch between the S~o Paulo and Vit6ria plates, the Para~a do Sul strike-slip shear belt, has undergone a partitioned dextral transpression, whereas the north-south branch between the Brasflia and Vit6ria plates is essentially a frontal thrust system with only a weak component of dextral strike-slip. These complex structural patterns, formed at deep to mid-crustal levels, reflect temporal and spatial partitioning at all scales between flattening and non-coaxial deformation, and down-dip and strike-slip shearing, in tangential as well as in transcurrent structural domains. Additionally, this area demonstrates that regional flower structures, lateral extrusion and other secondary deformations across the yz sections of transpressional belts are important in accommodating shortening in obliquely convergent orogens.

The Precambrian of south-eastern Brazil is characterized by two orogenic belts, one bordering the coastal zone and the other branching towards the inner country (Fig. 1). They comprise high-grade belts and medium-grade gneissic-migmatitic basement rocks, lowto medium-grade m e t a v o l c a n o - s e d i m e n t a r y sequences, and Proterozoic granitoid rocks (Fig. 2). The supracrustal assemblages have been respectively assigned to the Ribeira and Brasflia belts bordering the Silo Francisco and Rio de La Plata cratons. Both orogenic belts were widely overprinted by two ductile thrust systems along the collisional borders of three Archaean to Palaeoproterozoic continental palaeoplates of Gondwanaland, named the Brasflia, Silo Paulo and Vit6ria plates (Fig. 1). They constitute only the continental portions of former larger lithospheric plates, whose oceanic portions may have been consumed by B-type subduction, and whose continental borders may have been partially subducted and melted. In the light of global tectonics, we prefer the term plate for the collided continental masses. The term cratons, as generally used in the literature referring to Western Gondwana, corresponds only to the central portions of the plates, preserved from the collisional tectonics.

The Alterosa suture zone (ASZ, Figs 1 and 4, below) separates the northern Brasflia plate (BP), also described as the Silo Francisco Plate by Alkmim et al. (1993), from the Silo Paulo plate (SPP) in the south, whose northern border is the upthrust Alfenas granulite belt (Haralyi & Hasui 1982). U n d e r n e a t h this lithospheric suture zone and squeezed between the juxtaposed crustal blocks are highly dismembered rocks which have an ophiolitic signature (Roig 1992; Zanardo et al. 1996) and other rocks of an accretionary derivation. The Vit6ria plate (VP) in the east is separated from the previously collided SPP and BP by the A b r e - C a m p o suture zone (ACSZ). The arcuate Atlantic granulite belt (Fyfe & Leonardos 1974) constitutes the western border of the VP, where portions of the lower crust have been thrust to the NW and to the W, respectively onto the margins of the already welded SPP and BP (Fig. 2). The sinistral Campo do Meio strike-slip shear belt (CMSB, Hasui & Oliveira 1984) and the dextral Rio P a r a ~ a do Sul strike-slip shear belt (RPSSB, Braun 1975; Ebert et al. 1991a) (Figs 2 and 3) overprint the thrust systems, respectively. They represent the late tectonic expression of the oblique collision of continental plates, which evolved to form, after strong crustal shortening

H. D. EBERT& Y. HASUI. 1998. Transpressional tectonics and strain partitioning during oblique collision 231 between three plates in the Precambrian of south-east Brazil. In: HOLDSWORTH,R. E., STRACHAN,R. A. ~fr DEWEY,J. E (eds) 1998. Continental Transpressionaland Transtensional Tectonics. Geological Society, London, Special Publications, 135, 231-252.

232

H.D. EBERT & Y. HASUI 50 ~

44 ~

:36 '~

I

I

..........

"~-J

BASIN

+ /I

+' {

i "

/

.

)]II .

.

!'/'

~' ti'~"'/) '

"~--='--..

i

., .

/I

/I

)I[

_

,

"~-'

)- :_

,,,t/,1)

+

~-

', +

-)-

+(

+" ~]:

...*_

~

s

#/

[

] PHANEROZOIC COVER

,4"-'

1 / /Y ( t

"-'~4

/

(?

-.

~' ,,:'

'/

+

_ , p., > / < s

24"-- /'

I

/ 71". /,. ,i.J' / l ] # )/ ~

/' 4'

i'ASIN~:'~__lJ_--~/~,,s/ ,

." V

U + +'4-+

~,,,D,,,.r~ r }17~ r~---.-a,,-,,- 1

§ ~

.... "- "--

....+ /'~\//~,

+7 )$. i~/ J / ,' / , v .....

+ +

~.

L--'L--- I '

", ~ ' > - ~

+ +

'-"

"

. .. ,.o

,I"

~ I

!l [ / ~ ~ \

~ ~

REGIONS .AFFECTED BY THE BRASILIANO EVENT

I n'--}-l- t BRASlLIANO CRATONIC AREAS

/ /

~)

Fig. 2

]

BP SPP VP s

v4---. ;@5

500 I

SA0

~FC

,r

s

7

km I

s

FRANCISCO CRATON

S~.O PAULO PLATE VITORIA PLATE SUTURE ZONES TECTONIC VERGENCE STRIKE-SLIP SHEAR ZONES

Fig. 1. Simplified regional tectonic map of East Brazil with location of studied area. The NE-SW trending orogenic belt along the coast includes the lineaments of the transpressional Rio Para~a do Sul shear belt. Bold dashed lines indicate the limits between Vit6ria (VP), Sio Paulo (SPP) and Brasflia (BP) palaeoplates (Fig. 4). Suture zones: ACSZ, Abre-Campo; ASZ, Alterosa. Cratons: CC, Congo; KC, Kalahari; RPC, Rio de La Plata; SFC, Sio Francisco. and thickening, transpressional ramps of welded and widely dissected continental borders. Regional deformation is diffusely accommodated by ductile to ductile-brittle strike-slip shear zones, lateral thrusts and isoclinal folds along a zone tens of kilometres wide around the palaeoplate boundaries. The RPSSB comprises a 200 km wide anastomosing network of N N E - S S W to E N E - W S W trending, ductile, dextral, strike-slip shear zones extending over 1000 km along the coast (Figs 3 and 4). It truncates and deflects the previous tangential structures and the sinistral structures of the narrower E - W trending CMSB, thereby testifying to the younger age of the RPSSB. The suture zones that link the plates together (Fig. t) are marked

by prominent gravimetric signatures (Figs 4 and 5) and also magnetic anomaly patterns associated with upper-plate granulite belts; these are interpreted as lower crust exhumed during continental collision (Haralyi & Hasui 1985; Haralyi et aI. 1985; Hasui et al. 1993). The low-angle lateral ramps and the sub-vertical mylonitic foliation in both strike-slip belts exhibit sub-horizontal stretching lineations which record plate-border-parallel motions. Nevertheless, there are also oblique and downdip stretching lineations revealing the predominance of local shortening t h r o u g h reverse ductile shearing. Large and gentle inflections of previous regional structures along diffuse zones tens of kilometres wide, operating together with

233

TRANSPRESSIONAL OROGENS OF SE BRAZIL 43

47

46 ~7"~ ~-~

0

50

SUPRACRUSTALSEQUENCES

100 kr~

~

GRANITIC COMPLEXES

UII~sU'rlUN OF uONVI~RGENCE

~

PIIANEROZOICCOVER

GNEISSIC- MIGMATITIc COMPLEXES

Fig. 2. Regional geological structural map of SE Brazil. Continuous lines indicate the major shear zones of the E-W trending Campo de Meio and the NE-SW trending Rio Para~a do Sul Shear belts. The low-angle metamorphic foliation (tangential domains) occurs intercalated between these strike-slip shear zones. These relations are represented along the transects A-B and C-D in Fig. 6. The triangular structure corresponds to the Guaxup6 Syntaxis, which is bounded by the bifurcation of both belts. transcurrent zones of high shear strain, reveal lateral strike-slip displacements of at least 200 km between the plate margins (Hasui et al. 1975; Ebert et al. 1991b, 1993). The bifurcation of both shear belts along an inverted 'triple junction' (Hasui et al. 1993) and the northern domain of the $5o Paulo plate delineates the Guaxup6 Syntaxis (sensu Davison (1994)), which is a remarkable regional, synformal, thrust wedge, with a northern border which gently dips to the SSW and a southern border dipping to NNW (Figs 2 and 3). The axes dip gently to the W, and the apex of this triangular structure points E. The kinematic interpretation of the associated structures has been controversial, in particular with regard to the main direction of plate convergence and the relative importance of thrusting, folding and strike-slip shearing. Geochronological constraints have been summarized by Cordani et al. (1988), Tassinari

(1988), Sato & Campos Neto (1996), Valladares (1996) and S611ner & Trouw (1997), who attributed mainly Neoproterozoic ages to the associated plutonism and to the metamorphic overprinting of Archaean to Palaeoproterozoic basement rocks that constitute the framework of the above-mentioned continental plates. This research aimed to clarify the complex regional structural architecture, and to provide a model for the tectonic evolution of the area in terms of the strain and kinematics responsible for the present framework. Additionally, we hoped to gain insights into the complex relationships between dip-slip and strike-slip deformations, and between coaxial and non-coaxial strains caused by strain partitioning in deeply eroded transpressional orogens, as has been described in this belt (e.g. Campanha 1981; Lammerer 1987; Ebert et al. 1988, 1991a, 1993; Chrispim & Tupinamb~i 1989; Ebert & Hasui 1989; Dayan & Keller 1990; Correa Neto et aL

234

H.D. EBERT & Y. HASUI

Fig. 3. Simplified tectonic map of SE Brazil. The kinematics of the main shear zones of the E-W trending sinistral CMSB and the NE-SW trending dextral RPSSB records approximate E-W convergence during two collisions. The conjunction between NW-SE contraction structures and NE-SW strike-slip along the RPSSB reflects the partitioning of this main vector along the NE-SW oblique plate boundaries (represented in Fig. 4).

1993; Machado & Endo 1993; Sadowski 1991; Fassbinder & Machado 1996) and in many other regions (e.g. Harland 1971; Sanderson & Marchini 1984; Holdsworth & Strachan 1991;Tikoff & Teyssier 1994; Krantz 1995; Teyssier et al. 1995). This paper summarizes the structural framework and the tectonic evolution of a transpressional triple plate collision. It is based on regional and semi-detailed structural field investigations conducted during the last 15 years. The results of structural analysis, strain analysis, the study of microfabrics and metamorphic assemblages associated with deformation, processing of gravity and magnetic surveys, and analogue experiments have been integrated with the information available from regional geological project reports, other literature and several MSc and PhD theses (Braun 1975; Hasui et al. 1975, 1988, 1989, 1990, 1993; Cavalcante et al. 197.9; D R M 1981-1982; Wernick & Fiori 1981; Trouw et al. 1982, 1994; Hasui 1983; Machado Filho et al. 1983; Campos Neto et al. 1984; Ebert 1984; Artur 1988; Gon~alves 1988; Ribeiro et al. 1990; Schrank et

al. 1990; Soares et al. 1990; Campos Neto 1991; Ebert et al. 1991b; Sadowski 1994; Artur & Wernick 1993; Heilbron 1993; Heilbron et al. 1993; Morales & Hasui 1993; Vauchez et aL 1994; Campos Neto & Figueiredo 1995; Mesquita et al. 1995; Almeida 1997; Braga & Ebert 1997).

Structural framework The regional structures can broadly be described in terms of domains with (1) dominantly lowangle foliation associated with tangential deformation and (2) dominantly steeply dipping foliation associated with strike-slip deformation. This subdivision is generally adequate, but is fairly artificial in some limited areas where one may observe a gradation between them or the existence of an inverse kinematic relationship, such as gently dipping lateral ramps and steeply dipping reverse shear zones. Figure 6 shows two representative regional sections across the CMSB and the RPSSB where the relation between tangential and sub-vertical structures is shown.

TRANSPRESSIONAL OROGENS OF SE BRAZIL

235

Fig. 4. Regional gravity map showing the approximate outline of the Brasflia (BP), Silo Paulo (SPP) and Vitdria (VP) plates and their boundaries (ASZ, Alterosa Suture Zone; ACSZ, Abre-Campo suture zone). NE-SW shear zones of the RPSSB (fine continuous lines) running parallel to the ACSZ dextrally dissect the ASZ, the SPP and the BP. Isogal spacing 5 mgal. Values higher than -35 mgals which increase to 40 regal over the continental shelf, are omitted. Interpolation based on 907 of our own data points, 505 data points from Fernandes (1993) and 985 from Hasui et al. (1989) and Observatdrio Nacional, Rio de Janeiro (obtained from the database of the Instituto Astron6mico e Geoffsico, Silo Paulo).

Tangential d o m a i n s The main regional syn-metamorphic planar and linear fabrics are pervasive in most rocks (apart from late-orogenic plutonic intrusions) and consist of gneissic foliation, compositional layering, mylonitic foliation, schistosity, chocolatetablet boudinage, stretching and mineral lineations, intrafolial, sheath and extension parallel isoclinal folds, and kinematic indicators such as S-C foliation, mica-fish, asymmetrical boudins and shear bands. These structures record thick-skinned, thrust tectonics at deep to mid-crustal levels, driven by two different east-west collisions: initially, overthrusting of the SPP eastward onto the BP (the Brasflia collision), and later, oblique eastward subduction of these already collided blocks under the VP (the Ribeira collision). Tangential domains are widely preserved from wrenching in the central portions and to the south and south-east of the Guaxup6 Syntaxis, within regional strike-slip duplex structures of the RPSSB, where they are

truncated or gradually replaced by strike-slip shear zones (Fig. 6b). Shallow-dipping structures are widespread and exhibit an alternation between slivers where either coaxial or non-coaxial deformations predominate (Fig. 7). Horizons of coaxial deformation are dominant and they comprise granoblastic, blastomylonitic and symmetrical, flattening type fabrics (S and SL tectonites) with low strain ratios (Rs average typically between 1.5 and 3), and often preserve high-grade (>600~ metamorphic assemblages in the deeper portions of the orogens. Quartz is always more strongly orientated, stretched or flattened (as ribbons) and recrystallized than the feldspar porphyroclasts and the mafic minerals (hornblende and biotite). Strain analysis using Rf/phi and Fry techniques on quartz crystals has yielded higher strain ratios than on feldspars in the same samples. This indicates the importance of quartz in controlling crustal rheology, with crystal-plastic strains substantially accommodating the regional stress field.

236

H.D. EBERT & Y. HASUI

Fig. 5. Gravity model of a north-south crustal profile (indicated as section A-B in Fig. 4) showing the collisional framework of the SPP onto the BP (Mallagutti Filho et al. 1996). (ASZ, Alterosa suture zone).

These horizons are regarded as lenses of deep crustal material, which were protected from hydration during cooling and exhumation to higher levels. They are separated by other shallow-dipping and narrow zones of predominantly non-coaxial deformation, with qualitatively higher shear strains and quantitatively higher strain ratios (R,x-z), typically between three and ten. The analysis of hundreds of thin sections has revealed asymmetrical (S-C foliation), mylonitic fabrics and an intense strainsoftening of physical (dynamic recrystallization of quartz and feldspars) and chemical (break-up of feldspars in mica and quartz) characteristics (schematized in Fig. 7). Medium- to low-temperature deformation fabrics, including retrograde metamorphic assemblages triggered by fluid percolation, record decreasing temperatures during exhumation of these rocks. These represent gliding surfaces of dominantly ductile simple shear which have concentrated differential motion during subduction and thrusting. Concordant quartz veins are also widely developed, indicating intense fluid percolation.

Stretching lineations Figure 2 shows the average direction of the stretching lineations (alignment of minerals, of stretched grains and of mineral aggregates)

within, and between the juxtaposed continental masses. In the northern and eastern borders of the SPP, E - W sub-horizontal stretching lineations, associated with sub-horizontal foliations, prevail. Apart from local sheath and oblique folds, most isoclinal folds in tangential domains are recumbent and overturned towards the N, or S, with their hinges parallel to the E - W transport direction (related to the SPP against BP collision). Along the moderately to steeply dipping foliations of the RPSSB, shallowly plunging NE-SW trending stretching lineations are dominant. The isoclinal fold hinges and intersection lineations (Figs 8 and 10, below) run parallel to these, and to strike, reflecting border-parallel displacement during the late VP collision westwards, against already collided SPP and BP. Inside the SPP and portions of the RPSSB where fiat-lying foliation is preserved, lineations and kinematic indicators show overthrusting to the NW. The NW-SE and E - W trending lineations are often gradually inflected along the CMSB (sinistrally) and the RPSSB (dextrally). The significance of the lineations and kinematic indicators showing top-to-the-NW thrusting in the widespread tangential domains of the SPP is not yet clearly understood. They seem to be older than the E - W structures, because a NW-SE convergence is not compatible with the

TRANSPRESSIONAL OROGENS OF SE BRAZIL late sinistral sense of displacement recorded between the SPP and BP plate borders on the CMSB. This could only arise as a result of a topto-the-NE thrusting. In a different context, along the RPSSB where it affects collisional borders, they reflect an inboard NW-thrust component (orogen perpendicular extrusion) of the transpression during the oblique collision of the VP to the west, onto the S~o Paulo and Brasflia plates. Mapping of the lineation trends also shows that many stretching lineations, orientated approximately N-S, represent older NW-SE or E - W lineations from the first collision (SPP against BP). During the second collision (VP against SPP and BP), these lineations, together with the tangential foliation, suffered a dextral borderland inflection, evidence of which can be seen up to 20 km away from the main RPSSB lineaments (Hasui et al. 1990; Braga & Ebert 1997). Strike-slip d o m a i n s

The N E - S W trending RPSSB, which is the primary focus of this paper, comprises a 100-200 km wide belt, of parallel to anastomosing dextral transcurrent shear zones, extending along the Atlantic Ocean coast (Figs 2 and 3). They are recognized in the field as steeply dipping mylonitic rocks that affect all the Precambrian units and the previously described foliation (Figs 6 and 9-13). The metamorphic assemblages associated with the strike-slip foliation range from lower granulite facies along the A16m Para~a lineament (Campanha 1981), through to zones of partial melting, and down to low-grade facies. In general, they reflect a lower-temperature deformation than the adjacent flat-lying foliation. The interaction between the main discontinuities and synthetic P shear zones forms contractional strike-slip duplexes and typically shows an anastomosing pattern. At the northern edge, the RPSSB bends to a N-S direction and the shear zones dissipate along a regional contractional 'horse-tail' fault geometry. Here, the already oblique-slip kinematic character of the shear zones is gradually replaced by WNWvergent thrust zones and large isoclinal folds (Lammerer 1987), also described further north by Cunningham et aL (1996). As an exception, three NW-SE trending shear zones, at the SW edge of the belt close to the Paran~i Basin, form an extensional 'horse-tail' which controls the emplacement of the Itu granite pluton (23.5~ 47~ in Fig. 2). Several features, such as S-type fabrics, oblate strain ellipsoids, symmetrical fabrics, upright isoclinal folds, hinges of en

237

echelon folds and other extension-parallel features, are orientated at an angle to the strike-slip shear zones that is lower than expected for a simple shear model. This reflects a general compression field associated with the strike-slip regime. The upright isoclinal folds overprinting regional foliation are always related to the proximity of a strike-slip shear zone (Figs 10 and 11) and suggest a close relationship between shortening and wrenching. The adjacent regional flat-lying metamorphic foliation converges downwards into these strikeslip shear zones and duplexes, forming largescale, transpressional, ductile, flower structures. Such flowers can be inferred in structural maps and transects as kilometric-scale structures, or they are seen at metre-scale in outcrop (Fig. 9). They differ from those at higher crustal levels by displaying a greater complexity in response to a more ductile behaviour and penetrative grainscale deformation. The regional strain is not only accommodated by simple shear along discrete fault planes separating different 'petals' of each flower, but also by mechanisms of ductile flow, such as crystal-plastic strain, dynamic recrystallization and diffusive mass transfer. By analysing the microstructures in outcrop and in hundreds of thin sections, a systematic parallelism between metamorphic fabrics, small-scale kinematic and strain data, and meso- and largescale structures was identified. The main rock components, such as quartz, feldspars, hornblende and biotite, exhibit orientations and strain patterns (ductile flattening or stretching, recrystallization on pressure shadows) that are geometrically coincident with the shear zones, fold hinges and axial planes. Figure 9 shows schematically the patterns of a typical midcrustal transpressional flower structure, commonly observed at various scales throughout these belts. Secondary shortening and vertical extrusion within each segment ('petal') of the flowers is recorded by folding of the internal foliation, flattened grains and oblate strain ellipsoids. Nevertheless, weakly plunging stretching lineations indicate that the direction of principal strain and displacement was still predominantly sub-horizontal. The mutual interaction between adjacent flowers, during progressive deformation, has often produced such complex geometric patterns that it is difficult to delimit each flower from the neighbouring ones. The transcurrent shear zones separate intermediate zones with shallow-dipping foliation, low-angle lateral ramps and contemporaneous, en echelon and sub-parallel, open to isoclinal upright folds (Figs 6, 10 and 11). A new axialplane crenulation can be developed where

238

H . D . EBERT & Y. H A S U I

~.o

~Qa oo

0

r~

r.~

r ~:::I .IU 0

X~ 0

9

9

N ~.~

.~ "~

*...~ ,.....~

.~

0

m

rm

. ~

~ ~~= 0"0

(__) ii

~

TRANSPRESSIONAL OROGENS OF SE BRAZIL

SIMPLE SHEAR ASYMMETRIC FABRICS ( S-C ) STRETCHING FABRICS ( LS, SL ) MYLONITES RETROGRADE METAMORPHISM (MEDIUM GRADE) HIGH FLUID PRESSURE HIGH STRAIN SOFTENING TRANSPORT SURFACES

239

PURE SHEAR SYMMETRIC FABRICS FLATTENING FABRICS ( S, SL BLASTOMYLONITES OR GRANOBLASTIC FABRICS PRESERVED HIGH GRADE METAMORPHISM LOW FLUID PRESSURE LOW STRAIN SOFTENING PASSIVELY EXHUMATED LENSES

Fig. 7. Schematic sketch synthesizing the typical features of the tangential structures of the investigated area. They are partitioned into zones of dominantly coaxial and non-coaxial deformation. Location of this 50 m wide outcrop in Fig. 2 is about 22~ 45~ metapelites have been strongly affected by tight upright folding. These structures led to the identification of a 'Ribeira fold belt', which was traditionally interpreted to be the product of a N W - S E frontal contraction. However, these features are all compatible and belong to the same regional strain field, where x is horizontal, NE-SW, parallel to strike, z is horizontal and perpendicular to it, and y is vertical and of extensional character. Penetrative synmetamorphic fabrics are also developed on shallow-dipping (<30 ~) lateral ramps that link to sub-horizontal surfaces, typical of tangential deformation. Where stretching lineations are parallel to the highangle transcurrent shear zones, they can be ascribed to plate border wrenching. However, where down-dip lineations and reverse to oblique thrusts occur, it is often difficult to distinguish wrenching concurrent with thrusting from older tangential structures (Fig. 12). There are many other places where it is difficult to define the kinematic character of the penetrative foliation, because the strong flattening fabrics prevent recognition of any stretching lineation. Younger overprinting strike-slip shear zones (D2) are well recognized where they are steeply dipping and show cooler deformation fabrics of greenschist-facies conditions, dissecting penetrative higher-temperature low-angle foliation. This occurs primarily within the cooler subducted

plates underneath the suture zones (Fig. 4), such as where the W - E trending CMSB dissects the BP, and within the SPP, south of the Guaxupd Syntaxis, where the N E - S W trending RPSSB obliquely dissects the older fabrics (Fig. 13a and b). Along the previously described low-angle lateral ramps, which are characterized by a strike-parallel stretching lineation, the deformation fabrics are preserved as assemblages of the main amphibolite- to granulite-facies metamorphism, which includes recrystallization of feldspar. These features are most prevalent on the hanging wall of the suture zones where upthrust plates have brought hotter rocks from deeper levels, and in metasediments, which are squeezed between the overthrust plates. In these zones it is often difficult to distinguish between the earlier, more ductile, plate border parallel, tangential D 1 phase and a later D2 strike-slip deformation, especially where a gradation from one to the other exists. Most structures along the N E - S W to E N E - W S W trend of the RPSSB correspond to dextral, strike-slip oblique, contractional, shear zones. However, along oblique ( N E - S W to NNW-SSW direction) contractional bends, they assume a preferentially more northwestwards dextral reverse character. To determine the local kinematic significance of the regional foliation, it was necessary to classify the stretching

240

H.D. EBERT & Y. HASUI

7

J

J/

/

ss s s -

~X '~:~',

+

+

"~:~

".......

+

+--.~+ +

+

+

+~

+

+ + + L ,-.,..

+ +

+

"....

+ +

+ +

+

+ + +

+ +

+

+ +

+

+ +

+

+

Fdsp I

2m

i

Fig. 8. Sub-horizontal foliation with well-developed stretching lineations parallel to strike. Arrows indicate subsidiary shear strain along the kinematic yz-plane, whereas vorticity is not identified in the xz-plane. This figure represents a small-scale kinematic example of recumbent folds and thrusts which are perpendicular to the main transport direction, along low-angle lateral ramps of the E-W trending CMSB and the NE-SW trending RPSSB. lineations in accordance to their rake on the foliation plane. Thus, the shear zones and faults could be classified kinematically, either in extensional or contractional tectonic setting, as frontal, frontal- oblique, oblique, strike-slip oblique and strike-slip, for angles of 75-90 ~, 55-75 ~ 35-55 ~ 15-35 ~ and 0-15 ~ respectively. Nevertheless, most shear zones display a concentration of N E - S W shallow-plunging lineations on stereoplots. Some transcurrent shear zones of the RPSSB display sinistral displacements. This opposite shear sense can be explained with reference to an overprinting N-S contraction (Machado & Endo 1993) or by differential horizontal stretch-

ing of strongly shortened partitioned segments. Aligned syntectonic granitic intrusions and small pull-apart basins indicate local transtensional structures concurring within the bulk transpressional regime. These occur along releasing bends, or are related to a reversal in the shear sense, during a transtensional intervening period (S. Molyneux pers. comm.). The ductile strike-slip shear zones formed at depth in both belts were active along narrow steeply dipping brittle-ductile shear zones during exhumation processes concentrating the latest orogenic movements. T h e y were also reactivated as intracratonic faults during the Phanerozoic, in particular during the opening of

TRANSPRESSIONAL OROGENS OF SE BRAZIL

A"

' ..

J

./

' ,

" -

/ / ~

" /' " "

./" , " "'

' ,'

/ i

.

.

.

/ ." Y"

" .

.

"

./ //

/ ..

'

./"

/-.

-/"

i

";

" ,, /

/ I. "

241

~

..-.,'~'/"

/ /

~

/

. .,//

i" LIQUE T H R U S T S

Y>I

Y Fig. 9. Schematic sketch of a typical ductile transpressional flower structure (10 m to 10 km scale) commonly exposed in NW-SE transects across the RPSSB. Strike-slip shearing is accompanied by simultaneous ductile cross-shortening caused by folding and reverse shearing within each petal.

_

Fig. 10. Sketch of NE-SW trending small strike-slip shear zones bounding upright folds with axial planes and hinges parallel to the horizontal stretching lineation. They affect adjacent horizontal foliation related to eastward thrusting of SPP onto BP. Location is the uppermost north lineament of the RPSSB, on the eastern edge of the Guaxupd Syntaxis, south of the city of Lavras (about 21 ~ 45~ Fig. 2). These structures represent in miniature the regional intercalation between folds and shear zones, regionally also observed in Fig. 6b, which together accommodate the partitioning between strike-slip and shortening across the RPSSB.

242

H.D. EBERT & Y. HASUI

S2 / / S 1

D2

Fig. 11. Sketch of the same NE-SW trending, ductile, dextral, strike-slip shear zone of the transpressional RPSSB as shown in Fig. 10, dissecting and folding the older regional flat-lying foliation of the Brasflia collision. Outcrop view is 5 m wide. the Atlantic Ocean, associated with NNW-SSE tear faults.

Regional strain regime Although the term transpression has been applied for simultaneous shortening and strikeslip deformations, the field evidence and strain analysis conducted in these large mid-crustal orogenic belts shows that, in terms of strain regime, for both strike-slip and tangential structures (Northrup 1996), pure shear and simple shear coexist. The analysis of microstructures and of the orientation of strain ellipses in the x z plane in the investigated area attests to a broad range of kinematic vorticity numbers Wn (Simpson & DePaor 1993). These are generally lower than 0.25 in zones of predominantly coaxial strain, where fabrics are almost symmetrical and the 0' angle between the x-axes of the strain ellipse is around zero. Other intercalated zones of higher simple shear record typical asymmetrical fabrics (used as shear sense indicators) and x-axes of the strain ellipses oblique to the foliation which correspond to higher vorticity numbers (Wn between 0.25 and 0.75). In the transpressional RPSSB, crustal thickening is generally achieved by vertical extension (regional y>l). In areas where cross-shortening prevails over strike-slip simple shear, dip-slip stretching lineations indicate (Fig. 12) that vertical extension (now bulk kinematic x-axes) became larger than the horizontal strain.

Deformation fabrics and strain analysis that characterize the tangential domains of that region point to a heterogeneous general noncoaxial flow, similar to the transpressional regime, but with simultaneous flattening (pure shear) and simple shear on sub-horizontal planes. Local horizontal extension could be invoked, but none the less the regional structures indicate a contractional event. The vertical flattening of rock slices in outcrop scale, which would theoretically promote crustal thinning, is regionally compensated by ductile accretion of thrusted crustal segments. Although the regional penetrative structures are related to regional ductile non-coaxial deformation there are surprisingly few asymmetrical fabrics on x z sections either in tangential or in strike-slip domains. The small angle 0' between the longest x-axis of the finite strain ellipses and the main foliation is usually negligible. This feature cannot be ascribed to only simple shear, which would require a much higher strain ratios (Ramsay & Huber 1983) than those measured in our studies, where Rsx-z ranges from 1.5 to 5 in either strike-slip or tangential domains. On tangential horizons, this is lower than for predicted simple shear Rs • 0' angular relations (Ramsay & Huber 1983), and is better explained by the action of simultaneous sub-vertical pure shear shortening on individual overstacking thrust slices, triggered during the increasing weight of the thickening crust. Along the strike-slip domains with steeply inclined foliation this

243

TRANSPRESSIONAL OROGENS OF SE BRAZIL yX

Z

-.

x

x x

x

-.

x x

x x

"-.

x x

x

x x

•

NW~

1 - 10 m

, SE

Fig. 12. Temporal and spatial partitioning between dip-slip, oblique and strike-slip stretching lineations in one outcrop of the transect of the RPSSB shown in Fig. 6b. Here, it is difficult to distinguish rotated older lineations from those related to strike-slip and coeval reverse shearing.

feature is explained by the strong cross-shortening during the transpressional regime. It is often easier to recognize asymmetrical fabrics on the kinematic y z sections than on the principal xz-planes (Fig. 8) either in tangential or strike-slip domains of the investigated area. This indicates that subsidiary deformation and displacement (lateral extrusion) perpendicular to the x direction of main stretching and motion is important to accommodate local heterogeneities in the strain field, and may also account for the complex structural scenario. Along the E - W trending CMSB and the NE-SW trending RPSSB, thrusts, and upright or reclined isoclinal folds have hinges parallel to the regional stretching lineation (Fig. 9), and are considered here as by-products of the oblique plate border convergence.

Tectonic setting of plutonism Many deformed synkinematic granites of calc-alkaline to alkaline signature expose

cross-cutting relationships between tectonic, magmatic, and metamorphic features that indicate an emplacement sequence related progressively to thrusting, to transpression and then to transtension along shear zones. In the later plutons, many examples show low tectonic deformation, and primary magmatic fabrics are preserved (Vlach et al. 1990; Artur et al. 1991; Janasi & Ulbrich 1991; Hackspacher 1994), whereas others have been rotated and obliquely compressed between the strike-slip shear zones during regional transpression. Emplacement related to transpressional zones has also been recognized in some elongated granites (i.e. Serra do Lopo Granite) in the SW portion of the RPSSB in Fig. 2 (23~ 45~ Its recognition is more difficult, because confinement and higher strains have led to much stronger solidstate fabrics overprinting the magmatic fabrics (Ebert et al. 1996). Migmatites, charnockites and granites, which were emplaced along active strike-slip shear zones, are well exposed along the Atlantic coast.

244

H.D. EBERT & Y. HASUI

..... 22.......................... 7<: ...... ;;7"---. .... 2"--,, "

X

ZQ

~

~

......... ~ _,5....... _ J_'2.

"'-,<,

L2

~ 7-"

$2 W,

5m

,

-

,E

S1 500\/ 0

S2

700

NW J

100m

I SE

Fig. 13. (a) NE-SW dextral shear zone (steep mylonitic foliation and sub-horizontal stretching lineation) of the RPSSB dissecting the penetrative tangential foliation with associated E-W stretching lineation associated with overthrust quartzites on the BP. The location of this shear zone is just south of Lavras (about 21~ 45 ~ Fig. 2). This relation corresponds on a small scale to the regional interaction between the RPSSB and the CMSB delineating the Guaxup6 Syntaxis. (b) Sigmoidal pattern formed by dextral strike-slip RPSSB shear zones overprinting the older tangential foliation in migmatitic gneisses inside the Guaxup6 Syntaxis. Location is about 23~ 47~

Physical e x p e r i m e n t s The reconstruction of the crustal framework and tectonic evolution of the region has been tested using analogue experiments (Ebert et al. 1995a). These have been carried out in a sandbox (62 cm • 35 cm • 6 cm) at the laboratory of C E N P E S P E T R O B R A S in Rio de Janeiro. Three solid (Plexiglas) plates on its base represent the shape of the BP, SPP and VP continental plates (scale 1:2 500 000). The continental crustal thickness of 40 km was represented by eight layers of quartz sand (grain size 0.42 mm), each 2 mm thick. In other experiments the more ductile lower crust was r e p r e s e n t e d by silicone putty. Initially

115-200 km wide basins were considered between the plates. The thinner crust of the basins were set up by only six layers, each 2 mm thick. To test the tectonic model based on field investigations, the experiments simulated E - W compression which led to the closure of the basins. First, the Silo Paulo plate moved against the Brasflia plate from west-to-east and then the Vitdria plate collided against both plates in an east-to-west fashion (Fig. 14). The E - W far-field convergence vector was reconstructed based on the integration of structural and kinematic features such as: (a) the direction of the stretching lineations along strike-slip and thrust shear

TRANSPRESSIONAL OROGENS OF SE BRAZIL

{o]

(b}

/ 7

,Y

7 /-~

~-.... "~', '-~ [cl

245

f

I

i

I

I

i

Y /

f

I (c~)

Fig. 14. Initial (a), intermediate (b) and final (e) stages of a sandbox simulation of an E-W oblique collision of the Silo Paulo plate (SPP) against the Brasflia plate (BP) and subsequently of the Vit6ria plate (VP) against both plates. ARGB, RB and DSB are respectively the Alto Rio Grande, Ribeira and Don Silv6rio basins before deformation. The initial and final positions of three point pairs along each branch (E-W CMSB, NE-SW and NNE-SSW RPSSB) show the relative shortening and the strike-slip displacement between the plate boundaries (quantified in Fig. 16). (d) shows the incremental displacement of each block. The differences between transects A-B and C-D (e) across the RPSSB are shown in Fig. 15.

zones; (b) the associated shear sense related to tectonic contraction or transpression; (c) the orientation of the plate borders (Figs 2-4). An ESE-directed convergence vector would not have produced such strong contractional structures, in particular the NW-directed thrusts along the N E - S W boundary between SPP and VP. On the other hand, a NW convergence vector, as assumed by many workers, would produce an overall sinistral displacement along the N-S branch of the RPSSB, which is not confirmed by field evidence. The aim of the physical modelling was to analyse the relations between strike-slip shearing and simultaneous contraction during oblique convergence of three plate margins of different orientation to the convergence vector. The deformation structures produced by the horizontal, non-coaxial strain component were recognized on the surface of the sandbox as dextral (NE-SW branch) and sinistral ( E - W branch) strike-slip shear zones displacing a perpendicular grid (Fig. 14). The simultaneous shortening along the three branches between the

plate margins is recognized in cross-section (Fig. 15) and reflects the transpressional regime deduced from field investigations. The amount of shortening, shear strain and displacement were measured along each of the three branches for the different stages of the experiment. They show a simultaneous increase in shortening and strike-slip, but with distinct relative intensities in accordance with their orientation in relation to the convergence vector (Fig. 16). Along the east-west and the northern branches, where only one of the plates of the model was moving against the fixed Brasflia plate, the contractional structures (folds and overthrusts) are asymmetrical (Fig. 15a). In the field this feature corresponds to the N-S branch (thrust system) and to the E - W striking CMSB, where a strong cratonward vergence of the regional structures against the S~o Franscisco Craton is characteristic. Along the southern branch, where the S~o Paulo and Vit6ria plates obliquely converge against each other, two parallel systems of shear zones with opposite dipping directions are produced (Fig. 15b). They have

246

H.D. EBERT & Y. HASUI

Fig. 15. Three-dimensional geometric models of two transects in the analogue experiment across the transpressional RPSSB (Fig. 14c): (a) at the northern branch, separating the more frontal collided plates, shortening, accommodated by asymmetrical folds and reverse faults, is more important then strike-slip (compare with Figs 14c and 16c); (b) at the southern branch, more oblique to the collision, flower-like structures correspond to a well-developed partitioning between contraction and strike-slip (compare with Figs 14c and 16b).

oblique strike-slip reverse character, being separated by a large synformal fold. In the field, they correspond kinematically to large, ductile, positive flower structures observed in the transpressional Rio P a r a ~ a do Sul shear belt. A lower degree of vergence also characterizes this area. Synformal folds are delimited in the experiment from antiformal structures by fault planes with oblique displacement (Fig. 15). This partitioning because of transpression corresponds to

the framework of the RPSSB, where folds are intercalated with contemporaneous subparallel strike-slip shear zones (Figs 6b, 9 and 10). Until recently, these distinct structural styles have been ascribed to different tectonic events, first to a frontal thrust-fold belt as a result of N W - S E plate convergence, then overprinting by N E - S W strike-slip shear zones. The simulation of the geometry and the movements of the three E - W obliquely converging

TRANSPRESSIONAL OROGENS OF SE BRAZIL 1 0 0 - -

8o

/

(a)

~

(b)

Y

~ 40

~ . . . .

i ~

20 0

~ ~'='~ 0

5o

(-W)

~RPSSB (N-S)

I 100

1,50

20e

strike-slip displacement

25o

3oo

(km)

Fig. 16. Shortening v. strike-slip displacement along each of the three colliding plate borders of the sandbox experiment. Curve (a), E-W trending CMSB; curve (b), oblique NE-SW branch of the RPSSB; curve (c), the frontal N-S branch of the RPSSB. plates produced simultaneous strike-slip and shortening structures which are consistent with the present regional tectonic arrangement. The N-S belt exposes contractional, asymmetrical structures typical of frontal collision, such as thrusts and reclined folds with west-vergence to the foreland. Along the N E - S W and N W - S E oblique plate margins, shortening is recorded by large en echelon folds, divergent foliation and thrusting onto both margins of the orogens.

Discussion on the evolution of regional strain regime At early collisional stages, the bulk horizontal incremental shortening z-axis parallel to a frontal convergence was accommodated by stretching along the vertical x-axis as a result of crustal thickening. The increasing weight of the overstacking crustal slices causes a reaction against this now more slowly developing process. At a certain stage, when it becomes easier for the continental crust to accommodate the horizontal tectonic contractional forces by a perpendicular but horizontal extension, a switch between the x- and y-axes takes place. The previous bulk intermediate horizontal y-axis becomes larger than the former vertical x-axis, giving rise to a regional strike-slip regime where the x-axis is now horizontal, and sub-parallel to the plate borders, and the y-axis is still extensional but vertical (transpression). At the latest stages, when decreasing horizontal tectonic contractional forces can no longer cope with the weight of the crustal pile (vertical bulk strain), tectonic relaxation and collapse begin to act as

247

transtension, or as extension. This kinematic evolution is similar to that of other orogenic belts which have begun with thrust tectonics, evolved to transpression and finished with transtensional or extensional collapse. Local heterogeneities in the orientation of the continent boundaries in relation to plate convergence, as observed in the studied area, led to the distinct partitionings and consequently to different, more complex, structural patterns. Along the N E - S W border between SPP and VP, the highly oblique convergence led to a strong partitioning between shortening and strike-slip, when the bulk incremental strain had already larger horizontal (x) then vertical (y) stretching from the early stages of collision.

Interpretations and conclusions The two deeply eroded Precambrian orogenic belts in southeastern Brazil expose regional- and local-scale structures that reveal the mechanisms of accommodating three-dimensional deformation in deep to mid-crustal levels along collisional margins of different orientations in relation to the angle of plate convergence. The geometric and kinematic parameters examined in the field and laboratory investigations were applied in analogue experiments to test the adopted W - E far-field convergence vector between the converging plates. The structural framework achieved is compatible with the observations made on a regional scale for each of the differently orientated margins, and confirms the model. The penetrative tangential foliation of the northern and eastern border (SPP) initially formed during its extremely oblique collision to the ENE onto the BP. Later-stage deformations of the transpressional CMSB were concentrated along narrow strike-slip E - W trending shear zones that preferentially dissected the cooler subducted plate (BP). The open sea that existed to the south-west of the now-assembled continental margins of SPP plus BP was progressively closed by westward convergence of the third plate (VP) that evolved into the second collision, the Ribeira Orogen. Deep to mid-crustal tangential penetrative fabrics formed within these plates and along their mutual boundaries during oblique thickskinned thrust tectonics and border-parallel, low-angle, lateral ramps. Subsidiary thrusting was mainly towards the NW and secondarily to the SE, along the N E - S W trending margin of the former SPP, and to the W, along the N N E - W S W margin of the former BP. Large ductile positive flower structures

248

H.D. EBERT & Y. HASUI

accommodated vertical and orogen transverse extrusion during transpression. After substantial crustal thickening and exhumation, lateral thrusts and penetrative flattening of sub-horizontal rock fabrics were gradually replaced during continuing transpression by a clear partitioning between upright to NW-vergent folds, oblique reverse shear zones, and narrower strike-slip shear zones, along the RPSSB. The structural bifurcation of two arcuate transpressional shear belts around the S~o Paulo plate (SPP) delineate the Guaxup6 Syntaxis, a regional-scale, gently west plunging synformal thrust wedge. Its eastern edge, nevertheless, does not have the features of a crustal indentation wedge (Wernick et aL 1981), as this would have produced much stronger crustal thickening in that area. These belts were produced as a result of two differently orientated and distinct collisional events, involving three continental plates. In studying this area, we have observed that the phrase 'pure-shear-dominated transpression' is inadequate to describe the features of this three-dimensional deformation. As the crustal shortening that accompanied wrenching was accommodated both by pure shear (upright folds or vertical flattening) and by simple shear (transverse to oblique reverse thrusts) along the cross-section, we prefer the term shortening or contraction-dominated transpression. It is difficult to establish the boundary conditions at such crustal levels, as the transpressional thickskinned deformation was widely and heterogeneously distributed inboard of the plate margins. The limits between deformation zones and their walls are diffuse. Consequently, it has been difficult to determine precisely the outer boundaries of the S~o Francisco Craton. The primary controls on the structural style along and across the investigated transpressional belts are: (1) their different orientations in relation to plate convergence, and (2) the variation in rheological conditions because of their inclined boundaries. The footwall was the cooler subducting plate and the hanging wall was the hotter overriding plate, which behaved in a more ductile fashion. Across the RPSSB, this feature could have promoted a progressive propagation from NW to SE of the regional shear zones and may explain a relative decrease in the isotopic ages in that direction (Tassinari 1988). Many granites to the south of the regional Guaxup6 structure (Fig. 2) have been interpreted as related to subduction of an oceanic crust along the Ribeira orogen (Porada 1979) to the NW (Wernick 1984). This model was always

difficult to explain, because the 'magmatic arc' is in the lower plate (SPP) (Fig. 4). We suggest that the origin of some of those granites intruded along the intersection of the strike-slip shear zones of the RPSSB with the tangential structures of the SPP may be ascribed to partial melting of lower crust or upper mantle during subduction of the BP under the SPP (first collision). Afterwards, during the subsequent collision of the VP to the west, the N E - S W trending strike-slip shear zones of the RPSSB deeply dissected the continental crust of the SPP. The confined magmas were collected from deep crustal sources, then channelled to higher crustal levels, along the deep crustal discontinuities, and finally frozen as a result of reducing temperatures during their ascent, as described in other continents (Hutton & Reavy 1992). As a consequence of this model, the genesis and emplacement of these granitoids should be considered as resulting from two distinct collisional events. We regard the granites emplaced along transpressional shear zones of the RPSSB as melts which were expelled from crustal depths to higher levels. This was triggered by the vertical pressure gradient during tectonic contraction, in a manner similar to the neighbouring positive flower structures in the country rocks, but it occurred more rapidly because of their lower viscosity (Ebert et aL 1995b). Space was created by slices of country rocks which were constantly being extruded laterally and to higher crustal levels during transpression. Instead of a material that only fills free spaces, we see granites as crustal material which was moving upwards faster than the more viscous country rocks, substantially aiding the decrease in volume and vertical extrusion in crustal depths during crustal shortening and thickening. The original basins separating the three plates (BP, SPP and VP) must have been, before the collision, connected along a 'triple junction' (Hasui et al. 1993) (see Fig. 14a) and they had common sedimentary features, particularly along the passive margin (Faixa Alto Rio Grande) surrounding the south border of the BP (Paciullo et al. 1993). However, the distinct tectonic regime in each branch, during orogenic inversion, produced particular structural patterns in accordance to their orientation relative to the E - W convergence. Each segment of these shear belts reveals a distinct structural style which was controlled by the subtended angle between E - W convergence and the orientation of each plate boundary. This also determined the degree of partitioning between regional coaxial and non-coaxial, as well as between dip-slip and strike-slip deformations. Most transcurrent

TRANSPRESSIONAL OROGENS OF SE BRAZIL shear systems are d e v e l o p e d in E - W and in N E - S W striking branches. C o m p r e s s i o n a l structures increase in significance in N E - S W striking branches and are d o m i n a n t in the N - S striking thrust belt (see also A r t u r & Wernick 1993). T h e described structural f r a m e w o r k d e m o n strates the i m p o r t a n c e of t e m p o r a l and spatial partitioning, which occurred on a regional to a small scale, along tangential and sub-vertical ductile tectonic structures, in a c c o m m o d a t i n g the c o m p l e x t h r e e - d i m e n s i o n a l strains within the m i d d l e crust. Partitioning b e t w e e n coaxial a n d n o n - c o a x i a l d e f o r m a t i o n s , and b e t w e e n r e g i o n a l strike-slip and c o n t r a c t i o n a l d e f o r mation, is widespread. Orogen-transverse shortening was a c c o m m o d a t e d by pure shear (upright folding) and by simple shear (oblique to dip-slip shearing). Structures with non-coaxial deform a t i o n patterns in the k i n e m a t i c yz-planes indicate the i m p o r t a n c e of secondary d e f o r m a t i o n and d i s p l a c e m e n t (vertical or lateral extrusion). These features t o g e t h e r form complex, flowerlike structures, w h i c h a c c o m m o d a t e d large orogen-parallel movements and shortening during oblique continental collision. The results presented in this paper were achieved by the integration of data obtained in various projects of the authors supported by Conselho Nacional de Desenvolvimento Cientffico e Tecnol6gico (CNPq), Fundaqfio de Amparo ~ Pesquisa do Estado de Silo Paulo (FAPESP) and Program de Apoio ao Desenvolvimento Cientffico e Tecnol6gico (PADCT)/Finaciadora de Estudos e Projetos (FINEP) (43-89.0115.00 and 65 92 0040 00). We also express our thanks to S. J. Molyneux for reviewing the manuscript, and to I. A. M. Isler and our research students for drawing figures and maps.

References ALKMIM,F. F., BRITO NEVES,B. B. & CASTROALVES,J. A. 1993. Arcabou~o tect6nico do Crfiton do S~o Francisco - uma revis~o. Anais, 2 ~Simp6sio sobre o Crdton do Sao Francisco e suas Faixas Marginais, Salvador, 72-77. ALMEIDA, S. H. S. 1997. Estruturaf8o tect~nica da margem norte da Cunha de Guaxupd e suas relafOes corn o Cintur6o Transcorrente Campo do Meio-MG. Dissertaq~o de Mestrado, IGCEUNESP, Rio Claro. ARTUR, A. C. 1988. Evoluf~to policiclica da infraestrutura da p o r ~ o sul do Estado de Minas Gerais e regiOes adjacentes do Estado de S~o Paulo. Tese de Doutoramento, IG-USP, S~o Paulo. & WERNICK, E. 1993. Modelos geotect6nicos aplicados ao Pr6-Cambriano superior do NE do Estado de S~o Paulo e fireas adjacentes do Estado de Minas Gerais: uma discussao. Geoci~ncias, 12, 155-185. - - , EBERT, H. D. & WERNICK, E. 1991. Magmatismo -

-

249

e tect6nica no Complexo Socorro (SP/MG). Anais, 2 ~ Simp6sio Geologia do Sudeste, S~o

Paulo, 105-110. BRAGA, I. E & EBERT, H. D. 1997. Relaq~o entre a tect6nica tangencial e transcorrente na porq~o central da Cunha de Guaxup6-MG. Anais, 6 ~ Simp6sio Nacional de Estudos Tect6nicos, Piren6polis, 259-261. BRAUN, O. P. G. 1975. The coastal orogenic-belt structure of south-eastern Brazil. International Symposium on Continental Margins of Atlantic Type, S~o Paulo. CAMPANHA, G. m. C. 1981. O Lineamento de Al6mParafba na firea de Tr4s-Rios (R J). Revista Brasileira de Geoci~ncias, 11, 159-171. CAMPOS NETO, M. C. 1991. A porf~o ocidental da Faixa Alto Rio Grande - ensaio de evolufao tectOnica. Tese de Doutoramento, IG-USP, S~o Paulo. CAMPOS NETO, M. DA C. • FIGUEIREDO,M. C. H. 1995. The Rio Doce Orogeny, Southeastern Brazil. Journal o f South American Earth Sciences, 8, 143-162. - - , BASEl, M. A. S., ALVES,E R. & DE VASCONCELOS, A. C. B. 1984. A Nappe de Cavalgamento Socorro (SP-MG). Anais, 33 ~ Congresso Brasileiro de Geologia, Rio de Janeiro, 4, 1809-1822. CAVALCANTE,J. C., CUNHA, H. C. S., CHIEREGATI,L. A., et al. 1979. Projeto Sapucai, Estados de S~o Paulo, Rio de Janeiro e Minas Gerais. Brasflia, Departamento Nacional da Produ~fio Mineral, Serie Geologia n. 4; Seato Geol6gica n. 2. CHRISPIM, S. ~ TUPINAMBA,M. A. 1989. Observa~6es sobre o carfiter transpressivo da deforma~o entre as zonas de cisalhamento do Para~a do Sul (RJ) e Gua~uf (ES). Anais, 1 ~Simp6sio de Geologia do Sudeste, 120-121. CORDANI, U. G., TEIXEIRA, W., TASSINARI, C. C. G., KAWASHITA, K. ~ SATO, K. 1988. The growth of the Brazilian Shield. Episodes, 11, 163-167. CORRt~A NETO, A. V., DAYAN, H., VALEN(~A, J. G. & RAFAELCABRAL,A. 1993. Geologia e estrntura da zona de cisalhamento do Rio Para~a do Sul entre Tr~s Rios (RJ) e Sapucaia (R J). Atas, 3 ~Simp6sio de Geologia do Sudeste, 194-200. CUNNINGHAM, W. D., MARSHAK, S. ~r ALKMIM, E E 1996. Structural style of basin inversion at midcrustal levels: two transects in the internal zone of the Brasiliano Araquai Belt, Minas Gerais, Brazil. Precambrian Research, 77, 1-15. DAVISON, I. 1994. Linked fault systems; extensional, strike-slip and contractional. In: Hancock, P. L. (ed.) Continental Deformation. Pergamon Press, Oxford, 121-142. DAYAN, H. 8r KELLER, J. V. A. 1990. A zona de cisalhamento do Rio Para~a do Sul nas vizinhanqas de Tr~s Rios (RJ): uma anfilise da deformaq~o dada pot algumas feiq6es estruturais. Revista Brasileira de Geoci~ncias, 19, 494-506. DRM 1981-1982. Projeto Carta Geol6gica do Estado do Rio de Janeiro, Escala 1:50 000. Departamento de Recursos Minerais, Rio de Janeiro. EBERT, H. 1984. Os Paraibides entre S~o Jo~o del Rei, Minas Gerais, e Itapira, S~o Paulo, e a bifurca~o

250

H.D. EBERT & Y. HASUI

entre Paraibides e Araxaides. Publicacfto Sociedade Brasileira de Geologia-SP-Documento, 12, 72-103. - & HASUI, Y. 1989. Tect6nica transpressiva no Sub-Bloco Tunas, Paranfi. Atas, 2 ~ Simp6sio Nacional de Estudos Tectdnicos, Fortaleza, 255-257. - - , CHEMALE,JR, E, BABINSKY,M., ARTUR, A. C. & VAN SCHMUS, W. R. 1996. Tectonic setting and U/Pb zircon dating of the plutonic Socorro Complex in the transpressive Rio Paral'ba do Sul Shear Belt, SE Brazil. Tectonics, 15, 688-699. --, HASUL Y. & COSTA, J. B. S. 1991a. O car~iter transpressivo do Cinturgo de Cisalhamento Rio Parafba do Sul. Boletim, 3 ~ Simp6sio Nacional de Estudos Tect6nicos, Rio Claro, 139-141. , & QUADE, H. 1988. Aspectos da evoluq~o estrutural do Cintur~o M6vel Costeiro na regi~o da Mina do Perau, Vale do Ribeira. Anais, 35 ~ Congresso Brasileiro de Geologia, Bel6m, 5, 2318-2331. - - , SARTORATO,G., ALMEIDA,S. H. & COSTA,J. t3. S. 1993. Arcabouqo estrutural e tect6nica transpressiva das faixas m6veis das bordas sul e sudeste do Cr~iton do S~o Francisco e da Sintaxe de Guaxup6. Anais, 4 ~ Simp6sio Nacional de Estudos TectOnicos, Belo Horizonte, 166-171. - - , MUTO, H. K. & ALMEIDA,S. H. 1991b. Quantificaqao da deformaq~o cisalhante e dos deslocamentos em zonas de cisalhamento de alto ~ngulo do sul de Minas Gerais. Boletim, 3 ~ Simp6sio Nacional de Estudos TectOnicos, Rio Claro, 184-186. , NEVES, M. A., HASUI,Y., LOPES,J. A. & GUERRA, M. 1995a. Compartimentac~o crustal e evolucao cinem~itica da Provincia Mantiqueira atrav6s de modelagem fisica. Boletim de Resumos Expandidos, 5 ~ Simp6sio Nacional de Estudos TectOnicos, Gramado, 26-28. , , HuvroN, D. H. W. & GUERRA, M. 1995b. Alojamento de rochas plut6nicas em regimes transpressivos: uma modelagem experimental. Boletim de Resumos Expandidos, 5 ~ Simp6sio Nacional de Estudos TectOnicos, Gramado, 186-188. FASSBINDER, E. & MACHADO R. 1996. Reavaliaq~o sobre os modelos tect6nicos existentes sobre o Grupo Aqungui: uma proposta de modelo alternativo. Anais, 39 ~ Congresso Brasileiro de Geologia, Salvador, 6, 54-56. FERNANDES, L. E 1993. Arcabou(o estrutural e evolu(6o da Bacia de Taubat~ - SP. Dissertaqfio de mestrado, Escola de Minas, Universidade Federal de Ouro Preto. FYFE, W. S. & LEONARDOS, O. H., JR 1974. Ancient metamorphic migmatite belts of the Brazilian Atlantic Coast. The African connection. Revista Brasileira de Geoci@ncias, 4, 247-251. GONr M. L. 1988. O estudo da falha de empurrao na regi~o de Santana do Garamb6u Sul de Minas Gerais. Anais, 35 ~ Congresso Brasileiro de Geologia, Bel6m, 5, 2332-2342. HACKSPACHER,P. C. 1994. TectOnica transtensiva/transpressiva e alojamento de rochas plut6nicas, a

exemplo da Folha Cabre~va. Tese de Livre Doc~ncia, Universidade Estadual Paulista, Rio Claro. HARALYI, N. L. E. & HASUI, Y. 1982. Compartimentaq~o geotect6nica do Brasil Oriental corn base na informaqao geofisica. Anais, 32 ~ Congresso Brasileiro de Geologia, Salvador, 1, 374-385. - & -1985. Interpretation of gravity and magnetic data of old cratonic areas in Central-East Brazil. In: Hinze, W. J. (ed.) Utility of Regional Gravity and Magnetic Anomaly Maps. Society of Exploration Geophysicists, Tulsa, OK, 124-131. , MIOTO,J. A., HAMZA,g. M. & RODRIGUES, C. V. 1985. Ensaio sobre a estrutura~o crustal do Estado de Minas Gerais com base na informaq~o geofisica e geol6gica. ContribuiqOes d Mineralogia e Petrologia, Volume Djalma Guimargtes. Sociedade Brasileira de Geologia: Belo Horizonte, Boletim Especial, 71-93. HARLAND,W. D. 1971. Tectonic transpression in Caledonian Spitsbergen. Geological Magazine, 108, 27--42. HASUI, Y. 1983. Aspectos geol6gicos essenciais da seq~o Caconde-Caraguatatuba e suas implicaq6es na reconstituiq~o da organizaq~o e evoluq~o do Pr6-Cambriano do leste paulista. Atas, la. Jornada sobre a Carta Geol6gica do Estado de Sao Paulo em 1:50 000, S~o Paulo, 227-252. - • OLIVEIRA, M. A. F. 1984. Prov/ncia Mantiqueira. Setor Central. In: ALMEIDA, E E M. & HASUI, Y. (eds) O Pr~-Carnbriano do Brasil. Edgard Bliicher, S~o Paulo, 265-281. - - . , CARNEIRO, C. D. & COIMBRA,A. M. 1975. The Ribeira folded belt. Revista Brasileira de Geoci@ncias, 5, 257-266. - - , EBERT,H. D. & COSTA,J. B. S. 1990. Estruturaq~o tect6nica da extremidade oriental da chamada Cunha de Guaxup6: dados preliminares. Anais, 36 ~ Congresso Brasileiro de Geologia, Natal, 5, 2296-2308. - - , HARALYI,N. L. E., CAMPANHA,g. A., HAMZA,g. M., GALLI g. A., FRANGIPANI, A. & PULEGUIN~ FILHO, P. 1989. Compartimentafao estrutural e evolufao tectOnica do Estado de Sgto Paulo. Instituto de Pesquisas Tecnol6gicas do Estado de Silo Paulo, Relat6rio, n. 27394. ,-& COSTA, J. B. S. 1993. Megaestruturaqfio pr6-cambriana do territ6rio brasileiro baseada em dados geofisicos e geol6gicos. Geoci@ncias, 12(1), 7-31. - - , RODRIGUES,E. R., QUADE,H., NIK, R., MORALES, N., ZANARDO,A. & KULLER,M. L. 1988. Processo de cisalhamento dt~ctik modelo de transformaqOes em rochas prO-cambrianas antigas do leste de Sgto Paulo. S~o Paulo, Instituto de Pesquisas Tecnol6gicas do Estado de Silo Paulo, Relat6rio, n. 25908. HEILBRON, M. 1993. Evoluqgto tectono-metam6rfica da Seqfto Bom Jardim de Minas (MG)-Barra do Pirai (R J). Setor Central da Faixa Ribeira. Tese de Doutorado, IG-USP, Silo Paulo. --, VALERIANO, C. M., TUPINAMB,4_, M., et al. 1993. Compartimentaqfio tect6nica e evoluq~o

T R A N S P R E S S I O N A L O R O G E N S OF SE B R A Z I L geol6gica do segmento central da faixa Ribeira, a sul do CrOton do S~o Francisco. Anais, 2 ~ Simp6sio do Crdton do S~o Francisco, Salvador, 263-265. HOLDSWORTH, R. E. & STRACHAN,R. A. 1991. Interlinked system of ductile strike-slip and thrusting formed by Caledonian sinistral transpression in northeastern Greenland. Geology, 19, 510-513. HUTTON, D. H. W. & REAVY,R. J. 1992. Strike-slip tectonics and granite petrogenesis. Tectonics, 11, 960-967. JANASI, V. A. & ULBRICH, H. G. J. 1991. Late Proterozoic granitoid magmatism in the state of Silo Paulo, southeastern Brazil. Precambrian Research, 51, 351-374. KRANTZ,R. W. 1995. The transpressional strain model applied to strike-slip, oblique-convergent and oblique-divergent deformation. Journal o f Structural Geology, 17, 1125-1137. LAMMERER, B. 1987. Short notes on a structural section through the Ribeira Mobile Belt - Minas Gerais and Esp/rito Santo, Brazil. Zentralblattfiir Geologie und Paliiontologie, 7-8, 719-728. MACHADO FILHO, L., RIBEIRO, M. W., GONZALEZ, S. R., et al. 1983. Geologia. Folhas SF23 Rio de Janeiro e SF24 Vit6ria. Projeto R A D A M BRASIL, Levantamento de Recursos Naturais, v. 32. MACHADO, R. & ENDO, 1. 1993. O Cinturfio de Cisalhamento Atlfintico: um exemplo de tect6nica transpressiva neoproteroz6ica. Anais, 4 ~ Simp6sio Nacional de Estudos TectOnicos, Belo Horizonte, 188-191. MALAGUTrI FILHO, W., EBERT, H. D., HASUI Y., et aL 1996. Gravimetria e compartimentaq~o crustal do sul de Minas Gerais. Geoci~ncias, UNESP, 15 (n. esp.), 199-217. MESQUITA, M. P. S., EBERT, H. D. & HASUI, Y. 1995. Anfilise estrutural de um segmento das zonas de cisalhamento transcorrente de Monte Silo e Ouro Fino, sul de Minas Gerais. Boletim de Resumos Expandidos, 5 ~ Simp6sio Nacional de Estudos TectOnicos, Gramado, 57-58. MORALES,N. & HASUI, Y. 1993. Evoluqfio tect6nica da porq~o oriental do Cintur~o de Cisalhamento Campo do Meio. Anais, 4 ~ Simp6sio Nacional de Estudos TectOnicos, Belo Horizonte, 292-295. NORTHRUP, C. J. 1996. Structural expressions and tectonic implications of general noncoaxial flow in the midcrust of a collisional orogen: the northern Scandinavian Caledonides. Tectonics, 15, 490-505. PACIULLO,F. V. P., RIBEIRO,A. & ANDREIS, R.A. 1993. A reconstruqfio de uma bacia fragmentada - o caso do Ciclo Deposicional Andrelfindia. Anais, 2 ~ Simp6sio Crdton do Sao Francisco, 224-227. PORADA, H. 1979. The Damara-Ribeira orogen of the Pan-African-Brazilian cycle in Namibia (South West Africa) and Brazil as interpreted in terms of continental collision. Tectonophysics, 57, 237-265. RAMSAY,J. C. & HUBER, M. I. 1983. The Techniques o f Modern Structural Geology, Vol.1, Strain Analysis. Academic Press, London. RIBEIRO, A., PASCIULLO,F. g. P., ANDREIS, R. A., TROW,

251

R. A. J. & HEILBRON,M. 1990. Evoluq~o policfclica proteroz6ica no sul do Cr~iton de Sgo Francisco: anglise da regigo de Sao Jogo del Rei e Andrel~ndia, MG. Anais, 36 ~ Congresso Brasileiro de Geologia, Natal, 6, 2605-2614. ROlG, H. L. 1992. Caracterizaqgto da 'zona de sutura' Jacui-Conceir da Aparecida, MG - limite norte do Cinturao Alto Rio Grande: implicaqOes geotectOnicas e metalogen~ticas. Dissertaq~o de Mestrado, IG-Universidade de Campinas. SADOWSKI, G. R. 1991. A Megafalha de Cubaffto no Sudeste Brasileiro. Boletim do Instituto de Geoci~ncias da USP, SOrie Cientifica, 22, 15-28. SANDERSON, D. J. & MARCHINI,W. R. D. 1984. Transpression. Journal o f Structural Geology, 6, 447-466. SATO, K. & CAMPOS NETO, M. 1996. Dados dos principais eventos de acreqao manto-crosta continental no sudeste brasileiro, baseado em idades modelo Sm-Nd. Anais, 39 ~ Congresso Brasileiro de Geologia, Salvador, 5, 528-530. SCHRANK, A., ABREU, E R., ROIO, H. L., CHOUDHURI, A., SZABO, G. J. A. & CARVALHO,E. D. R. 1990. Determina~fio dos vetores de transporte tect6nico na borda sudoeste do Cr~ton do Silo Francisco. Anais, 36 ~ Congresso Brasileiro de Geologia, Natal, 5, 2276-2283. SIMPSON, C. & DEPAOR, 1993. Strain and kinematic analysis in general shear zones. Journal o f Structural Geology, 15, 1-20. SOARES,P. C.,FIORI, A. P. & DE CARVALHO,S. G. 1990. Tect6nica colisional obliqua entre o bloco Paran~ e a margem sul do Cr~iton do S. Francisco, no Maciqo de Guaxup6. Anais, 36 ~ Congresso Brasileiro de Geologia, Natal, 6, 2332-2342. SOLLNER, E & TROUW, R. A. J. 1997. The Andrelhndia depositional cycle (Minas Gerais/Brazil), a posttransamazonic sequence south of the S~o Francisco Craton: evidence from U - P b dating on zircons of a metasediment. Journal o f South American Earth Sciences, 10, 21-28. TASSINARI, C. C. G. 1988. As idades das rochas e dos eventos metam6rficos da porr sudeste do Estado de S6o Paulo e sua evolur crustal. Tese de Doutorado, IG-USP, S~o Paulo. TEYSSIER, C., TIKOFF, B. & MARKLEY, M. 1995. Oblique plate motion and continental tectonics. Geology, 23, 447-450. TI~OFF, B. & TEYSSIER,C. 1994. Strain modeling of displacement-field partitioning in transpressional orogens. Journal o f Structural Geology, 16, 1575-1588. TROUW, R. A. J., PASCIULLO,E g. P., CHRISPIM, S. J. & DAYAN,H. 1982. Amilise de deformaq~o em urea ~irea a SE da Lavras, Minas Gerais. Anais, 32 ~ Congresso Brasileiro de Geologia, Salvador, 1, 187-198. ~, RIBEIRO, A. E. & PASCIULLO,F. V. P. 1994. A Faixa Alto Rio Grande interpretada como a zona de interfer6ncia entre a Faixa Brasflia e Faixa Ribeira. Boletim de Resumos Expandidos, 38 ~ Congresso Brasileiro de Geologia, Camborifi, 1, 234-235. VALLADARES, C. V. 1996. Evolur geoldgica do

252

H . D . EBERT & Y. HASUI

Complexo Paraiba do Sul, no segmento central da Faixa Ribeira, corn base em estudos de geoquimica e geocronologia U-Pb. Tese de Doutoramento. IG-USP, S~o Paulo. VAUCHEZ, L. A. D., TOMMASI,A. & EGYDIO-SILVA,M. 1994. Self-indentation of a heterogeneous continental lithosphere. Geology, 22, 967-970. VLACH, S. R. F., JANASI,g. A. & VASCONCELOS,C. A. C. B. 1990. The Itu-belt: associated calc-alkaline and aluminous A-type late Brasiliano granitoids in the states of Silo Paulo and Paranfi, SE Brazil. Anais, 36 ~ Congresso Brasileiro de Geologia, Natal, 4, 1700-1711. WERNICK, E. 1984. Caracteriza~fio gen6tica de alguns granit6ides brasilianos dos Estados de S~o Paulo e Minas Gerais e implica~6es geotectSnicas preliminares. Anais, 33 ~ Congresso Brasileiro de Geologia, Rio de Janeiro, 6, 2902-2918.

-

& FIORI, A. E 1981. Contribuiqfio ~ geologia da borda sul do Crfiton do Silo Francisco. Anais, Simp6sio sobre o Crdton do Sgto Francisco e suas Faixas Marginais, Salvador, 169-179. - - , BITFENCOURT,J. DA S. • CHOUDHURI,A. 1981. A tect6nica rfgida no tim do Ciclo Brasiliano e sua implicaq~o na estruturaq~o da borda sul e sudoeste do Craton do S~o Francisco: tentativa de um modelo preliminar. Anais, Simp6sio sobre o Crdton do Sflo Francisco e suas Faixas Marginais, Salvador, 164-168 ZANARDO,A., DE OLIVEIRA,M. A. F., DEE LAMA, E. A. & CARVALHO,S. G. 1996. Rochas mNicas e ultram~ificas da faixa Jacui-Bom Jesus da Penha-Conceiqao da Aparecida (MG). Geoci~ncias, 15 (n.esp.), 143-168. -

-

-

Transpressionally driven rotation in the external orogenic zones of the Western Carpathians and the SW British Variscides R O D G A Y E R 1, T A N Y A H A T H A W A Y 1 & M I C H A L N E M C O K 1'2

1Laboratory for Strain Analysis, Department of Earth Sciences, University of Wales, Cardiff CF1 3YE, UK (e-mail: gayer@cardiffac, uk) 2present address: Institute for Geology, University of Wurzburg, Pleicherwall D-97070 Wurzburg, Germany Abstract: Analysis of two examples Of obliquely convergent external orogenic zones, the western part of the Western Carpathians and the northern Variscan margin in southwest Britain, indicates the operation of two dominant stress rotation mechanisms in the transpressionally deformed thrust wedge: (1) the rotation of an inferred stress field; (2) the rotation of a deforming body within a constant stress field. In the thinnest, external parts of the thrust wedge, % stress trajectory rotations of up to 90~ occur with deformation having a relatively small component of pure shear. Towards the hinterland, ~r1 stress trajectories in thicker parts of the wedge are progressively less rotated but develop a larger component of pure shear. Resultant cr1 trajectories are curvilinear, lying parallel to the orogenic convergence vector in the hinterland but diverging progressively from this direction towards the foreland, where they lie at high angles to the external margin in frontal parts of the thrust wedge. It is argued that balanced cross-sections should be constructed parallel to the curved trace of the ~1 stress trajectories.

The direction of tectonic transport is a parameter required for many aspects of structural analysis, e.g. in the construction of balanced cross-sections (Elliott 1983). However, in the external zones of obliquely convergent orogens, where transpressional d e f o r m a t i o n is developed, the tectonic transport vector is commonly either unclear or of uncertain significance. In many studies it has been assumed to lie perpendicular to the regional trend of fold axes or of thrust surface trajectories (the 'bow and arrow rule' of Boyer & Elliott (1982)). In these cases, strain and/or stress determinations can help to clarify the regional deformation, allowing strain (displacement) partitioning to be characterized within thin-skinned fold and thrust belts. As pointed out for the Alps (Lacassin 1989) and in orogenic belts in general (Oldow et al. 1990), partitioning of displacement during oblique convergence is a common occurrence (Fitch 1972; Walcott 1978; Beck 1983; Jarrard 1986; Cashman et al. 1992). Regions undergoing transpressional deformation typically result in simultaneous displacements along contractional and transcurrent fault systems. In such cases, stress or strain determinations may provide a better indication of the overall mechanism than scattered determinations of tectonic transport at localities where commonly either contractional or transcurrent faults are more dominant.

Major orogenic belts are frequently curvilinear zones with their length one or two orders of magnitude greater than their width, reflecting the geometry of the generating convergent plate boundary. Large-scale deflections of the orogen are frequently directly related to constructive plate margin geometry before convergence, e.g. the p r o m o n t o r i e s and e m b a y m e n t s of the Palaeozoic Appalachians (Thomas 1977). In general, plate convergence will be oblique; only rarely will ocean closure directly reverse the vectors of original opening, and even here convergence will be oblique locally around original margin irregularities. Curved orogenic belts have been shown by Doglioni (1992) to show a complex 3D geometry which changes and evolves through time, where the shape of the foreland fold and thrust belt, and the arcuation of the orogen are balanced by zones of extension and strike-slip deformation in the orogenic hinterland. Transpression is developed in the deflected margins of the arcuate orogenic front, producing significant rotations that emphasize the original curvilinear geometry, or generating deviations from the regional trend of the orogen. Palaeostress trajectories within these external zones obtain their present o r i e n t a t i o n by a combination of (1) the rotation of an inferred stress field (e.g. Tapponier 1977; Gamond 1987) (hereinafter termed 'stress deflection') and (2) the r o t a t i o n of a deforming body within a

R. GAYER,T. HATHAWAY& M. NEMCOK.1998. Transpressionally driven rotation in the external orogenic 253 zones of the Western Carpathians and the SW British Variscides. In: HOLDSWORTH,R. E., STRACHAN,R. A & DEWEY,J. E (eds) 1998. ContinentalTranspressionaland TranstensionalTectonics.Geological Society, London, Special Publications, 135, 253-266.

254

R. GAYER ETAL.

constant stress field (hereinafter t e r m e d 'passive rotation'). T h e a m o u n t of passive rotation can be d e t e r m i n e d from the reconstructed successive c h a n g e s in o r i e n t a t i o n of p r e - o r o g e n i c structure, and p a l a e o m a g n e t i c vectors, etc. By subtracting the a m o u n t s of these rotations from the deviation b e t w e e n the inferred stress field and the a s s u m e d c o n v e r g e n c e direction an estim a t i o n of the c o m p o n e n t of stress deflection can be made. In a progressively d e f o r m i n g zone, it is i m p o r t a n t to m e a s u r e the a m o u n t of passive rotation and the inferred stress field from coeval structures because structures from which the latter is m e a s u r e d will b e c o m e passive m a r k e r s and rotate during later increments of passive rotation. In this c o n t r i b u t i o n we analyse two cases of orogenic transpressional rotation. T h e first is f r o m the w e s t e r n part of the Tertiary W e s t e r n Carpathians, which d e m o n s t r a t e s h o w stress deflection and passive rotation can be determ i n e d in a relatively y o u n g orogenic belt w h e r e plate vectors are fairly well c o n s t r a i n e d f r o m the r e g i o n a l p a l a e o s t r e s s t r a j e c t o r y p a t t e r n s and from the poles of rotation of related plates (e.g. N e m c o k 1993; N e m c o k et al. 1993). Most of the literature relating to this case study is in Slovak (see references in N e m c o k (1993)) but it has b e e n i n c l u d e d as an e x a m p l e of t h e a p p r o a c h w h e r e c o n v e r g e n c e d i r e c t i o n s are relatively well known. T h e o t h e r case study is f r o m the Palaeozoic Variscan orogenic belt in s o u t h w e s t e r n Britain, w h e r e t h e Variscan conv e r g e n c e vectors are less well c o n s t r a i n e d f r o m regional d e f o r m a t i o n studies. T h e f o r m e r provides an e x a m p l e of sinistral transpression in an accretionary wedge, w h e r e a s t h e latter r e p r e sents dextral transpression in a fold and thrust sheet wedge.

placed within the movement plane, such that the e3 axis lies at 30 ~ to the fault plane, oriented so as to produce the sense of slip recorded for the fault. The E1 axis is perpendicular to E3, and the E2 axis is perpendicular to the movement plane.

Palaeostress determination The states of palaeostress associated with faulting were analysed from the following measurements: fault size, fault plane-slip vector orientation, sense of the fault displacement, and polyphase slip and its chronology. Fault sizes were measured to identify first-order faults and to allow a comparison of faults measured at outcrop with those interpreted from maps. The recognition of variously oriented striae, showing crosscutting relationships, allows, in favourable cases, the separation of superimposed palaeostress configurations. Slip sense determination was based on the criteria of Hancock (1985), Means (1987) and Petit (1987). The stress inversion methods of Carey & Brunier (1974) and Hardcastle & Hills (1991) were used to compute palaeostress configurations, defined by the orientations of the three principal stress axes together with the ratio of their magnitudes [O0 = (0.2-0.3)/ (oq-0-3)]. The stress configurations were calculated by the grid search method (Hardcastle & Hills 1991). For each stress configuration a check of the associated faults was made to determine whether slip, obeying the Coulomb criteria, would occur. Tensors were considered acceptable if slip would occur (1) on a minimum chosen percentage of the faults in the population, and (2) so that the deviation between the measured and calculated slip vector is less than a chosen limit. For palaeostress analysis in the South Wales region, data in the form of thrust lineations were separated into families using a statistical program CLUSTER (Nemcok & Lisle 1995). Stress ratios and orientations were calculated using the TENSOR program of Angelier (1988) which utilizes a method based on the direct inversion method of Carey & Brunier (1974).

Methods

Western Carpathians

Field data were collected, mainly from mesoscale brittle faults, from a range of localities in each of the areas described. These data were used for palaeostress and strain analysis, to determine stress trajectories and deviations of tectonic transport from the regional directions.

T h e external zones of the W e s t e r n Carpathians (Fig. 1) f o r m a f o r e l a n d tapering Tertiary accretionary w e d g e in front, and to the n o r t h e a s t of t h e ancestral p r e - T e r t i a r y C a r p a t h i a n s . T h e latter lie above a s o u t h w e s t w a r d s subducting oceanic slab a t t a c h e d to the E u r o p e a n platform. T h e arcuate w e d g e f o r m e d a b o v e failed M e s o z o i c rifts in the t h i n n e d E u r o p e a n platf o r m that w e r e i n v e r t e d during t h e Tertiary o r o g e n i c e v e n t ( N e m c o k 1993). F r o m foreland to h i n t e r l a n d , t h e w e d g e consists of: (1) a N e o g e n e molassic f o r e l a n d basin; (2) a flysch belt f o r m e d in an accretionary prism; and (3) the Pieniny k l i p p e n belt. T h e I n n e r Carpathians r e p r e s e n t the o l d e r a n d t h i c k e r part of t h e

Tectonic transport and strain Tectonic transport was determined as the mean value of the slip vectors at each locality. The slip vector is one of the vectors in Arthaud's (1969) movement plane, which is defined by the pole to the fault plane and the slickenline. Strain axes were determined utilizing the movement plane, following Hardcastle (1989). In this method the strain axes ~1, and e3 are

T R A N S P R E S S I O N A L L Y D R I V E N ROTATION

140

16~

18~

20~

220

24~

255

~

2

~

q~

East EuropeanPlatform 50~ ~_ 9Praha

., .

~oi@i:)))))io))i,)i~ Carpathians ~ ) i @ ) \

I ~50 o

100 km

Brne Inner West Carpathian~

Eastern Alps ~ l ,~+~+~+~+~§ ~ z ~ ~

('4 Debrecen - ~Carpathians~ +-* § ,~2; ~Budapest e Inner i ~,+~, East "Szolnok __~j

t ~

46~

, ,','+-+*+'+',~,_-,-,_-,-~-,~§

16~

18~

Molassicforedeep

~

Flysch belt

~

9Szeged ~+

+~Apuseni

46~ untains "Zag[eb +~ + ~ + ~ ~ Southern Alps ~ ~ ~ _ i+ +~i::::::: ,+ ~'~ South Carpa t ~ ~ : ~ . . . . . . . . . . . :::o ++~::~'.f~ Beograd : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : Adratc '+'-'~ . . . . . . . . . . -~' ' + ' + ~+ + ~ _, +++ :o.,o,::. ..................... ~ Bucuresti 1 44~ / .~o= . . .~'%'+'-' . . . Dinanc . . . .Alps . . .. ' _. ' .+ +. +. +. +.~ . . . . . #:'++++?~ . ~ ~Moesian Platform 440 14~

~:~ Pecs

20~

+

+~+~+~+o:)~ _ . x _ - - - - - - - -

22~

Pieninyklippen belt Inner Alpine, Carpathian & Dinaric units

24o

26o

280

~

volcanicNe~

~

InnerCarpathian Front

Fig. 1. Regional geological map of the Carpathians, showing regional convergence vector and area of Fig. 2.

mountain range, which accommodated the shortening in the external Carpathians by progressive, foreland-ward, propagating extensional collapse. The external accretionary wedge consists dominantly of siliciclastic sedimentary rocks with low cohesive strengths, which were shortened by folding and thrusting during the E o c e n e - E a r l y Miocene (e.g. Swidzinski 1948; Ksiazkiewicz 1956, 1960; Ksiazkiewicz & Lacko 1959; Kov~ic et al. 1989; Str~in~ et al. 1993). Regional 0.1 stress trajectories determined in the external Carpathian zone, following the method of Huchon et al. (1986), appear to be a sensitive indicator of convergence trajectories during the Eocene-Lower Miocene (Nemcok 1993). The movement vector for the Western Carpathians during this period was towards 026-038 ~. Palaeostress analysis

Reconstruction of the driving palaeostress fields (e.g. Nemcok 1993; Nemcok et al. 1993) shows a

3D and time-related development. The Eocene-Lower Miocene regional o1 stress trajectories established in this zone (Fig. 2) show that the Western Carpathians deformation front advanced towards the northeast from Paleogene through Early Miocene time (Nemcok 1993). In the eastern parts of the Western Carpathians, the Inner Carpathian front and the depocentre of the Neogene molassic sediments are roughly perpendicular to this SW-NE trajectory. Tectonic transport directions here are subparallel to the o1 stress trajectories. To the west, however, the Carpathians underwent a significant curvature as the Tertiary Carpathian arc was formed. The o-1 stress trajectories in the external zone along this western side of the Carpathians form angles between 0 ~ and 90 ~ with the movement vector of the Carpathians. Figure 2 shows an enlargement of the area where this deviation is first observed westwards along the arc. In the easternmost parts of this area, movement vectors are subparallel to the regional Carpathian vector. To the west, the vectors vary

256

R. GAYER E T A L .

Fig. 2. Geological map of part of Western Carpathians (location shown in Fig. 1), showing stress trajectories in transition zone between frontal compression and sinistral transpression.

progressively from 032 ~ to 027 ~ through 017 ~ to 005 ~, and finally to 345 ~. This area shows the transition from frontal compression to sinistral transpression, although only in the extreme west of the study area does the structural grain show a complementary swing. It is likely that a part of this deviation is due to stress deflection produced by transpression as shown, for example, by Tapponier (1977) and G a m o n d (1987). The remainder will have resulted from a passive rotation in the zone affected by transpression as discussed below. Numeric modelling of stress trajectories at overstepping strike-slip faults in transpression (e.g. Gamond 1987) has shown that significant stress deflection occurs in the contact zone between adjacent blocks, such that cr1 stress trajectories tend to rotate to make a higher angle with the zone boundary, and this appears to have occurred in the western external zones of the Carpathians. Palaeomagnetic data from the regions of the external zone affected by transpression indicate values of the sinistral (anticlockwise) passive rotation c o m p o n e n t for three stratigraphic

levels (Krs et al. 1991; Tfinyi & Kovfic 1991). Computed palaeostress configurations indicate the combined passive rotation and stress deflection for the same stratigraphic levels (Nemcok 1993). From these data the components of rotation of the ~r1 stress trajectory related to transpression in the westernmost parts of the Carpathians have been determined (Table 1).

Variscan northern margin in southwest Britain The Variscan orogen in western and central Europe is formed from a number of terranes that originated as rifted fragments of North Africa, and subsequently accreted to the European margin during Late Palaeozoic orogenic convergence (Franke 1989). A n approximately northwestwards convergence is recorded along the northern external Variscan margin to the north of these terranes. The northern margin (Fig. 3) consists of a northwards tapering fold and thrust belt comprising a southerly Rhenohercynian zone of Late Palaeozoic extensional basins above stretched European continental

257

TRANSPRESSIONALLY DRIVEN ROTATION Table 1. Components of passive rotation and stress deflection for the westernmost parts of the Carpathians between 50 and 18 Ma, determined from palaeomagnetic data and palaeostress configurations Stage (age) Karpatian (18 Ma) Eggenburgian (22-20 Ma) Middle Eocene (50-45 Ma)

Passive rock mass rotation during interval

Stress deflection acting during stage

28~ (18-0 Ma) 15~ (20-18 Ma) 17~ (45-22 Ma)

11~ (18 Ma) 47~ (22-20 Ma) 30~ (50-45 Ma)

crust, inverted during Variscan convergence, and a northerly belt of coal-bearing foreland basins in which Variscan deformation diminishes northwards (Raoult & Meilliez 1987; Gayer et al. 1993). In the west, the Variscan external zone is buttressed against the E - W trending WalesLondon-Brabant basement massif, generating, in southwest Britain, a major zone of dextral transpression (Gayer & Nemcok 1994). From north to south the zone consists of four units. The first, in South Wales, is a shortened foreland basin (the South Wales coalfield) with Namurian-Westphalian fill (Gayer & Jones 1989), which in the east unconformably overlies Dinantian platform carbonates that pass downwards into a thick fluvial and alluvial Devonian Old Red Sandstone sequence. The second unit, in North Devon and North Cornwall, consists of a strongly shortened Namurian-Westphalian turbidite sequence (De Raft et al. 1965) resting above a condensed siliciclastic Lower Carboniferous sequence that in turn lies above a thick interdigititating fluvial and shallow marine Devonian succession (Tunbridge 1986). The third unit, in southern Devon and central Cornwall, represents a Devonian continental shelf; and the fourth unit comprises Devonian flyschoid deposits in southwestern Cornwall. The zone is affected by very low grade metamorphism, increasing southwards. The Lizard complex, in southern Cornwall, represents the incorporation of oceanic basement into the orogen (Le Gall 1990; Gibbons & Thompson 1991). The four units of the zone are incorporated into a northwards tapering thrust wedge (Coward & Smallwood 1984; Le Gall 1990, 1991). Variscan thrusts in the zone verge NNW, and an oblique convergence is indicated by such features as rotation of fold hinges from SW-NE towards W N W - E S E trends, oblique folds and thrusts, and the predominance of major dextral strike-slip faults in the hinterland subparallel to the W - E to NW-SE trending orogenic front (e.g. Coward & Smallwood 1984; Leveridge el al. 1984).

Palaeostress a n d strain analysis

The northern margin of the external zone is represented by two contrasting styles of deformation; the first in the coal-bearing basin and its immediate surrounds (localities 1-14, Fig. 3b), and the second in the area to the south of the Bristol Channel (localities 15-18, Fig. 3b). In the first area, palaeomagnetic data from the Old Red Sandstone at locality 5 and from areas north of the Carreg Cennen Disturbance (Fig. 3b) show no evidence for rotation (Setiabudidaya 1991). However, the foreland basin fill and immediately underlying Dinantian limestones have been strongly deformed by folds and thrusts that have accommodated variable amounts of shortening (Fig. 3b).The results of the analysis of strain and tectonic transport for two localities in the foreland basin are presented in Fig. 4 and the results of palaeostress analysis in Fig. 5. The deformation in the foreland basin fill and in the underlying Dinantian limestones has resulted in significant dextral (clockwise) rotation. In the eastern parts of the foreland basin (north and east of locality 12, Fig. 3b) NE-SW trending fold axes and NW-SE striking crossfaults (dextral strike-slip faults), interpreted as lateral thrust ramps by Jones (1991), are oriented perpendicular and parallel respectively to the regional Variscan convergence vector (Fig. 3a) and suggest no rotation in this area. In the western parts of the foreland basin (localities 1-6, Fig. 3b) fold axes trend E - W and cross-faults strike N-S, implying up to 38 ~ of clockwise rotation. Tectonic transport in this area, however, is variable, with that at localities 1 and 2 showing major clockwise rotation but at localities 3-5 apparently no rotation. At the latter localities the tectonic transport was determined from major N E - S W trending, steeply inclined folds that deform both strata down to Old Red Sandstone and also coal seam parallel detachments. It was suggested by Frodsham & Gayer (1997) that these folds were formed by late Variscan reactivation of N E - S W trending basement structures and therefore the tectonic transport registers a

258

R. GAYER E T A L .

a

I~"

"-

~

LO~o, ' ~,

,,

~u~

/

I~i~djlffj ~andBerkshire ~ ~b'~l<,~= ~. ~ ~ ~ I s ~ ~ v : . ~ 9 \~"~ "-~&~, I ~ J""~_~_7~ Londonk. "N~"I " ~ ~~ ~~ V,~-~

/

"~

"

~

(

_ ~ . ~ ' ~ , ~

. ~.'~defeld 4~"

~

Lidge

~

~---- ~

~ -

-

250

1

bearing

~

f~

~

it of MAJORVariscan thrusts ~

b

"~ 5o:

'asin~oic land massifs

Variscanconvergence

300

,,;

-~

~

.~.~

_

20( ~" ! ! ! ! ! i ~ ~ [ ~ w a n s e a ~ ~

UpperPennantMeasures LowerPennantMeasures Lower&MiddleCoalMeasures Pre-andPost-CoalMeasures

~ y~v

0

~

~ ~ Cardiff U ,._..~J 7v -

I'~

0 km 10

15(

TRANSPRESSIONALLY DRIVEN ROTATION late, post-rotation direction sub-parallel to the regional Variscan convergence. The area of foreland basin between these two extremes shows intermediate rotation of cross-faults, tectonic transport and o'1 (Figs 3-5) The o'1 data for locality 13 are from Lisle & Vandycke (1996), and those for locality 1 are from Srivastava et al. (1995), who used shear zone fabrics to determine a N E - S W oriented 0-1 stress orientation. Palaeomagnetic data from deformed Old Red Sandstone in southwest Wales, along tectonic strike to the west (Fig. 3) (McClelland Brown 1983; McClelland 1987; Setiabudidaya 1991), indicate a passive dextral rotation of 10-45 ~ These variable orientations suggest a thin-skinned deformation of the foreland basin as a result of oblique convergence. A second set of dip-slip thrusts in parts of the external zone (localities 6-12, Fig. 3) have atypical directions. These thrusts are a set of broadly N-S striking and W - E transporting structures which are believed to have formed during a later (Variscan) deformation phase (Hathaway 1996), and are the subject of a further publication. The N-S striking thrusts were first described by Roberts (1972) and a full account and palaeostress analysis have been given by Hathaway (1996). Corfield et al. (1996) described N-S trending inversion structures in the Variscan foreland, and concluded that their orientation was greatly influenced by basement grain. Similarly oriented thin-skinned thrusts have also been described by Peace & Besly (1997) in the buried Variscan foreland of Oxfordshire, and these structures were also ascribed to reactivated basement faults. This does not appear to be the case in South Wales, where there is no obvious relationship to basement structures. At locality 6 (Nant Helen) the anomalous stress orientations may have been the result of the close proximity of the sinistral strike-slip Swansea Valley Disturbance (Owen & Weaver 1983) facilitating passive rotation. In the second area, along the southern margin of the Bristol Channel, a regionally consistent

259

tectonic pattern is developed, with oblique thrusts and associated structures (Figs 3, 6 and 7) indicating up to 11 ~ of dextral (clockwise) rotation. This rotation is indicated at locality 16 by the geometry of mineral fibres within two sets of extensional veins (Fig. 6); the first parallel and the second perpendicular to the fold axes. Initially, fibres grew perpendicular to the vein walls. After dextral rotation of the whole fold-thrust structure relative to the stress field, the perpendicular veins developed a sinistral component of displacement and the parallel veins a dextral component. The angular deviation between the fibres before and after rotation is about 11 ~. The tectonic transport in this area (localities 15-18, Fig. 3b) is more nearly parallel to the regional Variscan convergence shown in Fig. 3a. The analyses suggest that the brittle fault populations sampled do not show any significant strain partitioning. The azimuths of e3, o.a and tectonic transport (Figs 4, 5 and 7) differ by only a few degrees, within the limits of error. In none of the localities sampled was a first-order fault present, and the results suggest that at such localities strain partitioning is not a factor. Partitioning between wrenching and thrusting, so typical in zones of oblique convergence, seems to be restricted to distinct first-order fault structures (e.g. Tikoff & Teyssier 1992). Blocks between these structures simulate more isotropic conditions, with azimuths of e3, o1 and tectonic transport more or less coincident. S a n d e r s o n i a n strain c o m p u t a t i o n To understand how finite strain has been partitioned between pure and simple shear components within the foreland thrust wedge, an analysis of the transpressional strain has been undertaken. Two regions of the wedge have been chosen to characterize the transpressional deformation. The first region, 1 in Fig. 8, is situated in the centre of the foreland basin and contains localities 6, 7 and 10. The second region, 2 in Fig. 8, is in the thicker part of the wedge

Fig. 3. (a) Map of the northern Variscan margin in western Europe showing the location of coal-bearing foreland basins and the regional direction of tectonic convergence. (b) Geological map of part of Variscan external margin in southwestern Britain showing Variscan transport directions and values of shortening at numbered localities. Shortening estimates based on data from Jones (1991), Miliorizos (1992), and our observations. High values of shortening represent deformation in major detachments. Cross-faults represent strike-slip, dextral lateral ramps. CCD, Carreg Cennen Disturbance; SVD, Swansea Valley Disturbance; LLD, Llannon Disturbance; ND, Neath Disturbance; TD, Trimsaran Disturbance. Numbered locations are: 1, Caswell Bay; 2, Ffos Las opencast coal site; 3, Gilfach Iago opencast coal site; 4, East Pit opencast coal site; 5, Cribarth; 6, Nant Helen opencast coal site; 7, Ffyndaff opencast coal site; 8, Morlais limestone quarry; 9, Black Rock limestone quarry; 10, Park Slip opencast coal site; 11, Llanilid opencast coal site; 12, Taffs Well limestone quarry; 13, Ogmore coast; 14, Barry Island; 15, Croyde Bay; 16, Wild Pear Beach, Coombe Martin; 17, Woody Bay; 18, Portishead Point.

260

R. GAYER E T A L .

i) Tectonic Transport Direction

ii) Epsilon 3 Principal Strain

_ _ J

iii) Epsilon 1 Principal Strain

accommodated by the development of cleavage, tight chevron folds and steeper oblique thrusts. Following Sanderson & Marchini (1984), it is assumed that there has been no volume change and that the wedge is laterally confined, i.e. no lateral extrusion has occurred. Under these conditions the shortening across the wedge results in an area change which must be compensated by vertical thickening to conserve volume. If lateral extrusion has occurred then the amount of observed vertical thickening would be underestimated. D a t a have been collected from these two regions to allow factorization of the deformation in the form of a matrix following Sanderson & Marchini (1984) into pure shear and simple shear components:

~ (i 0 !) n=7 i) Tectonic Transport Direction

ii) Epsilon 3 Principal Strain

N

'

X

iii) Epsilon 1 Principal Strain

\ n=6 Fig. 4. Lower-hemisphere equal-area stereoplots of

strain axes and tectonic transport at (a) Nant Helen opencast coal site (locality 6) and (b) Taffs Well (locality 12). (i) Tectonic transport; (ii) e3; (iii) ~. towards the hinterland and contains locality 16. In this second region a much larger tectonic shortening is apparent than in region 1, which is

(i ~ !) : (i !)

This factorization and the parameters eLand are defined in Fig. 9, where o~-1 specifies the shortening across the zone, a the vertical stretch and ~ the shear strain parallel to the zone. In the foreland basin (region 1, Fig. 8) the vertical stretch, oL, has been estimated at 1.24 from differential erosion of the Westphalian strata laterally across the zone (Fig. 3) and the shear strain, % has been determined as 0.78 from the local 38 ~ rotation of the regionally consistent cross-faults (Fig. 3). In the thicker, hinterlandward parts of the thrust wedge (region 2, Fig. 8) the shortening, cx-1, has been estimated at 0.45 from geological cross-sections in the region and the shear strain, % determined as 0.19 from the attitudes of mineral fibres in extensional veins indicating 11 ~ rotation (Fig. 6b). To compute the shape of the finite strain ellipsoid (Fig. 10), the matrix D obtained from factorization must be multiplied by its transpose, resulting in a D D T matrix. The eigenvalues of the tensor D D T are equal to the principal quadratic elongations. The first eigenvalue from the characteristic equation is solved using Cardano's approach. The remaining two eigenvalues are computed after reduction of the third- to second-degree polynomial and after the first eigenvalue has been calculated. A comparison between the transpressional strains experienced in the two regions is made in the triangular diagram of Fig. 10. The finite strain of both regions is constrictional, with the strain in the external thinner parts of the wedge being higher than in the thicker more internal region. However, the pure shear component of

TRANSPRESSIONALLY Nant

DRIVEN ROTATION

F ~

Helen

i

i

I

'

261

i

. . . . .

M~

........................ i

~'

A+

............................. . ............................................ I O~ 10-242 03 78-098 0 = 0.579

o~ 12-073 o, 77-230

0 = 0.563

Black Rock

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

i

o~ 17-181 O~ 71-334 0 = 0603

1

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

O, 14-358 o, 75-158

Taffs Wells

= 0.848, ...................................

!

o, 12-355 a~ 78-185 0 = 0.628

o~ 6-100 O~ 80-228 = 0.465

o~ 5-002 03 10-27!

O, 16-068 o~ 70-286

r = 0.675

0 = 0.440

Fig. 5. C o m p i l a t i o n o f stress tensors calculated f r o m fault groups identified by cluster analysis at localities in S o u t h Wales (location s h o w n in Fig. 3). O n e set of data shows a N - S o r i e n t e d or1 and a s e c o n d set shows a N E - S W to W N W - E S E o r i e n t e d or1, i n t e r p r e t e d to be a y o u n g e r event.

J

262

R. GAYER E T A L . N

i) Tectonic Transport Direction

ii)

Epsilon 3 Principal Strain

a)

iii) Epsilon 1 Principal Strain

n=41

N

b) Fig. 7. Lower-hemisphere equal-area stereoplots of strain axes and tectonic transport at Coombe Martin (locality 16; location shown in Fig. 3). (i)-(iii) as in Fig. 4.

432

1

Fig. 6. Lower-hemisphere stereoplots of structural data from Old Red Sandstone at Coombe Martin (locality 16; location shown in Fig. 3). (a) Great circles of oblique thrusts showing slip vectors. (b) Great circles of numbered extensional veins with plots of associated quartz fibres. A sinistral component of displacement along the veins perpendicular to the hinge of dextrally rotated folds is indicated by a change from original (o) to younger (y) orientations.

)nic Front

,, trajectories

Hi the finite strain is m a r k e d l y less in the external region than in the internal region, implying that there is a significant increase in simple shear in the outer zone.

Fig. 8. Simplified model of curvilinear era stress trajectories in Variscan transpressional wedge, from foreland, through coal-bearing foreland basin (1), and thicker part of wedge (2), to hinterland (H).

Discussion and conclusions The two case studies indicate the o p e r a t i o n of a c o m b i n a t i o n of 0-1 stress deflection and passive rotation related to transpression in both external

orogenic zones. This process results in atypical thrust orientations in the t h i n n e r parts of the wedge, which allow significant lateral extrusion.

TRANSPRESSIONALLY DRIVEN ROTATION

Greatest rotation

263

Strike-Slip Faults ", //

Fore x

'

m

Y Van~ =~/'~

1

D= 0 T 0

~-'y

0

0 [ ' "1

0

0

Fig. 9. Diagrammatic representation of dextral transpression, showing an originally cubic block deformed by a vertical stretch of oLand a dextral angular shear of ~p.Horizontal shortening normal to the transpressional zone margin, assuming no lateral extrusion, is the reciprocal of vertical stretch (c~-1). Shear strain is tan ~p= 7. The transpression matrix is after Sanderson & Marchini (1984). 1 0.63 0 0 0.81

0

D T1

-

0 0

1.24

D

1 0.087 0 0 0.45 0 0 0 2.22

_ T2

D

1 0

0

=

0 0.81

0

P1

0 [1 0

D

= P2

1.24 0

0 I 00 0.45 0 2.22

sphere

(y-z)/

/x

cigar / 0

v

v

v

w V V V V V \ pancake

.1 .2 .3 .4 .5 .6 .7 .8 .9 1 R = Strain Ratio (0) q) = y-z/x-z

Fig. 10. Triangular diagram representing the shapes of strain ellipsoids, plotting (x-y) and ( y - z ) , both normalized with respect to x, along the right- and lefthand axes, respectively. The ratio of the principal strains qb= (y-z)/(x-z) is plotted along the horizontal axis. All ellipsoids lying along a line radiating from the upper apex have the same strain ratio, indicated by the intersection of the line with the horizontal axis. (See Nemcok (1995) for a detailed explanation). The diagram shows the strain ellipsoids from regions i and 2 (located in the text) of the Variscan transpressional wedge. II, Finite strain ellipsoids; 0, pure shear component. Ellipsoids have been calculated from deformational matrices for transpression (DT) and its pure shear component (De) for the two regions.

Hinterland

i [

Oblique Convergence

Fig. 11. Schematic representation of obliquely convergent accretionary thrust wedge. Bold arrows indicate convergence vectors; curved arrows represent component of passive rotation experienced by each thrust slice during latest increment of transpression. The deformation in the thin-skinned wedge of the Variscan foreland basin resulted in rotations of up to 38 ~ with a relatively small pure shear component of strain. In this outer region of the thrust wedge the palaeomagnetic data suggest that only the foreland basin fill has rotated, presumably on detachments near the base of the basin sequence (e.g. Frodsham & Gayer 1997). The lateral variation in rotation was accommodated by the numerous dextral lateral ramps that produced a relatively smooth (soft) change in orientation of fold axes. The situation progressively changes towards the hinterland. Modelling by Mulugeta & Koyi (1987) and Colletta et al. (1991) suggests that shortening will systematically increase, and our data indicate that the component of pure shear increases, whereas passive rotation decreases. This implies that individual thrust slices are progressively more constrained within the thickening wedge and lose their ability to rotate about the vertical axis (Fig. 11). As the orogenic wedge propagates foreland-ward, individual thrust slices progress towards the hinterland and acquire (1) a steeper attitude and (2) a rotated position, with progressively smaller increments of rotation. Wilkerson & M a r s h a k (1991) have d e m o n s t r a t e d that thrust-sheet loading, syntectonic burial and/or penetrative pure shear can affect greatly the slip vector on the thrust planes within the wedge. The slip vectors on thrusts in the more internal parts of an obliquely shortened wedge systematically change attitude from dip-slip towards strike-slip as the pure shear component increases. The extreme case of strike-parallel slip vectors is shown for some localities in the western margin of the Western Carpathians (Nemcok et al. 1989). Oblique convergence deflects o-1 stress trajectories from a regional orientation in the internal parts of the wedge

264

R. GAYER E T A L .

towards a higher angle to the external margin in the thinner, frontal parts of the wedge. Thus the present orientation of o-1 stress trajectories along the external zones of an obliquely shortened orogen should be treated very carefully when considering regional tectonics. This study shows that various divergent trajectories can be the result of different stages of the same process: transpression along the external zone. A n understanding of the mechanisms involved in transpression has implications for the construction of balanced cross-sections. In an obliquely convergent orogenic wedge, deformation paths will be complex, and in general curvilinear. If a cross-section is constructed parallel to an assumed rectilinear tectonic transport, material will move in and out of the section where the section line is oblique to the deformation path. A t t e m p t s at overcoming this p r o b l e m have resulted in so-called pseudo-3D balancing, whereby segments of the section constructed normal to fold and thrust packets are offset along identified strike-slip zones. This practice will result in reasonable approximations to the deformation path in the very thin frontal portions of the wedge, and in thicker parts towards the hinterland (e.g. sites 1 and H in Fig. 8). However, in regions of intermediate wedge thickness (e.g. site 2, Fig. 8) the stress trajectories are curvilinear, and give a better approximation to deformation paths than that of pseudo-3D balancing. Thus in regions between major faults, which should be treated specially, a more realistic balance could be achieved by constructing the cross-section parallel to the curved stress trajectories. M. N. was supported by a research fellowship from the Royal Society. We are indebted to L. Pick for mathematical assistance. J. Jones, R. Lisle, M. L. Curtis and an anonymous referee improved an early version of the manuscript. References

ANGELIER, J. 1988. Program MESURE. Quantitative Tectonics. Universit6 Pierre et Marie Curie, Paris. ARTHAUD, F. 1969. M6thode de ddtermination graphique des directions de raccourcissement, d'allongement et intermGdiaire d'une population de failles. Bulletin de la SociOtO Gdologique de France, 7, 729-737. BECK, M. E., JR 1983. On the mechanism of tectonic transport in zones of oblique subduction. Tectonophysics, 93, 1-11. BOYZR,S. E. & ELLIOrr, D. 1982. Thrust systems. Bulletin, American Association of Petroleum Geologists, 66, 1196-1230. CAREY,E. & BRUNIER,B. 1974. Analyse thGorique et numGrique d'un moddle mGcanique 616mentaire appliqu6 ?a l'Gtude d'une population de failles.

Comptes Rendus de l'Acad~mie des Sciences, S~rie D, 279, 891-894. CASHMAN, S. M., KELSEY, H. M., ERDMAN, C. E, 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, 242-257. COLLETrA, B., LETOUZEY,J., PINEDO,R., BALLARD,J.E & BALI~,P. 1991. Computerized X-ray tomography analysis of sandbox models: examples of thin-skinned thrust systems. Geology, 19, 1063-1067. CORFIELD, S. M., GAWTHORPE, R. L., GAGE, M., FRASER,A. J. & BESLY,B. M. 1996. Inversion tectonics of the Variscan foreland of the British Isles. Journal of the Geological Society, London, 153, 17-32. COWARD,M. R & SMALLWOOD,S. 1984. An interpretation of the Variscan tectonics of SW Britain. In: HUTrON, D. H. W. & SANDERSON, D.J. (eds) Variscan Tectonics of the North Atlantic Region. Geological Society, London, Special Publications, 14, 89-102. DE RAFF, J. E M., READING,H. G. & WALKER,R. G. 1965. Cyclic sedimentation in the Lower Westphalian of North Devon, England. Sedimentology, 4, 1-52. DOGLIONI, C. 1992. Main differences between thrust belts. Terra Nova, 4, 152-164. ELLIOTT, D. 1983. The construction of balanced crosssections. Journal of Structural Geology, 5, 99. FITCH, T. J. 1972. Plate convergence, transcurrent faults, and internal deformation adjacent to southeast Asia and the western Pacific. Journal of Geophysical Research, 77, 4432-4461. FRANKE, W. 1989. Tectonostratigraphic Units in the Variscan Belt of Central Europe. Geological Society of America, Special Papers, 230, 67-90. FRODSHAM, K. & GAYER,R. A. 1997. Variscan compressional structures within the main productive coal-bearing strata of South Wales. Journal of the Geological Society, London, 154, 195-208. GAMOND,J. F. 1987. Bridge structures as sense of displacement criteria in brittle fault zones. Journal of Structural Geology, 9, 609-620. GAYER, R.A. & JONES,J. A. 1989. The Variscan foreland in South Wales. Proceedings of the Ussher Society, 7, 177-179. - & NEMCOK, M. 1994. Transpressionally driven rotation in the external Variscides of south-west Britain. Proceedings of the Ussher Society, 8, 317-320. - - - , GREILING,R. 0., HECHT, C. & JONES,J. A. 1993. Comparative evolution of Variscan coal-bearing foreland basins. In: GAYER,R. A., GREILING,R. O. & VOGEL,A. K. (eds) Rhenohercynian and Subvariscan Fold Belts. Vieweg, Braunschweig, 47-82. GIBBONS, W. & THOMPSON, L. 1991. Ophiolitic mylonites in the Lizard complex: ductile extension in the lower oceanic crust. Geology, 19, 1009-1012. HANCOCK,EL. 1985. Brittle microtectonics: principles and practice. Journal of Structural Geology, 7, 437-457.

TRANSPRESSIONALLY DRIVEN ROTATION HARDCASTLE, K. C. 1989. Possible paleostress configurations derived from fault-slip data in Eastern Vermont and Western New Hampshire. Tectonics, 8, 265-284. -& HILLS, L. S. 1991. BRUTE3 and SELECT: Quickbasic 4 programs for determination of stress tensor configurations and separation of heterogeneous populations of fault-slip data. Computers and Geoscience, 17, 23-43. HATHAWAY,T. M. 1996.3D geometry and displacement variation of thrust faults. PhD thesis, University of Wales, Cardiff. HUCHON,P., BARRIER,E., BREMAECKER,J.-C. & ANGELIER, J. 1986. Collision and stress trajectories in Taiwan: a finite element model. Tectonophysics, 125, 179-191. JARRARD, R. D. 1986. Terrane motion by strike-slip faulting of forearc slivers. Geology, 14, 780-783. JONES, J.A. 1991. A mountain front model for the Variscan deformation of the South Wales coalfield. Journal of the Geological Society, London, 1 4 8 , 881-891. KOVAC, M., CiCHA, I., KRYSTEK, I., SLACZKA, A., STRANIK, Z., OSZCZYPKO, N. & VASS, D. 1989.

Palinspastic maps of the Western Carpathian Neogene, scale 1:1000000. Geological Survey, Prague. KRS, M., KRSOVA,M., CHVOJKA,R. & POTFAJ,M. 1991. Paleomagnetic investigations of the flysch belt in the Orava region, Magura unit, Czechoslovak Western Carpathians. Geologicke Prace Zpravy, 92, 135-151. KSIAZKIEWICZ, M. 1956. Geology of the northern Carpathians. Geologische Rundschau, 4 5 , 3 6 9 - 4 1 1 . 1960. Pre-orogenic sedimentation in the Carpathian geosyncline. Geologische Rundschau, 5 0 , 8-31. - & LACKO,B. 1959. On the relation between the Krosno- and Magura-Flysch. Bulletin de l'Acaddmie Polonaise des Sciences, 7, 773-780. LACASSIN,R. 1989. Plate-scale kinematics and compatibility of crustal shear zones in the Alps. In: COWARD,M. E, DIETRICH,D. & PARK,R. G. (eds) Alpine Tectonics. Geological Society, London, Special Publications, 45, 339-352. LE GALL, B. 1990. Evidence of an imbricate crustal thrust belt in the southern British Variscides: contributions of South-Western Approaches traverse (SWAT) deep seismic reflection profiling recorded through the English Channel and the Celtic Sea. Tectonics, 9, 283-302. 1991. Crustal evolutionary model for the Variscides of Ireland and Wales from SWAT seismic data. Journal of the Geological Society, London, 148, 759-774. LEVERIDGE, B. E., HOLDER, M. T. & DAY, G. A. 1984. Thrust nappe tectonics in the Devonian of south Cornwall and the western English Channel. In: HUTrON, D. H. W. & SANDERSON, D. J. (eds)

Variscan Tectonics of the North Atlantic Region. Geological Society, London, Special Publications, 14, 103-112. LISLE, R. J. & VANDYCKE,S. 1996. Separation of multiple stress events by fault striation analysis: an

265

example from Variscan and younger structures at Ogmore, South Wales. Journal of the Geological Society, London, 153, 945-953. MCCLELLAND, E. 1987. Palaeomagnetic results from the Lower Devonian Llandstadwell Formation, Dyfed, Wales - Discussion. Tectonophysics, 143, 335-336. MCCLELLAND BROWN, E. 1983. Palaeomagnetic studies of fold development and propagation in the Pembrokeshire Old Red Sandstone. Tectonophysics, 98, 131-149. MEANS,W. D. 1987. A newly recognised type of slickenside striation. Journal of Structural Geology, 9, 585-59O. MILTORIZOS,M. 1992. Tectonic evolution of the Bristol Channel borderlands. PhD thesis, University of Wales, Cardiff. MULUGETA, G. & KOYI, H. 1987. Three-dimensional geometry and kinematics of experimental piggyback thrusting. Geology, 15, 1052-1056. NEMCOK, M. 1993. Transition from convergence to escape: field evidence from the West Carpathians. Tectonophysics, 217, 117-142. - 1995. Representation of tectonic stress. Geologica Carpathica, 46, 13-17. - & LISLE,R. J. 1995. A stress inversion procedure for polyphase fault/slip data sets. Journal of Structural Geology, 17, 1445-1453. - - , HOK, J., KovAI~, P., MARKO, E, MADARAS,J. & BEZAK, V. 1993. Tertiary West Carpathian tectonics (in Slovak). In: RAKOS, M. & VOZAR, J. (eds) Geodynamicky model a hlbinnd stavba Zdpadnych Karpdt. GUD, Bratislava, 263-267. --, MARKO, E, KOVAE, M. & FODOR, L. 1989. Neogene tectonics and paleostress changes in the Czechoslovakian part of the Vienna Basin: Jahrbuch Geologische Bundesanstalt, 132, 443-458. OLDOW, J. S., BALLY,A. W. & Avt~ LALLEMANT,H. G. 1990. Transpression, orogenic float, and lithospheric balance. Geology, 18, 991-994. OWEN,T.R. & WEAVER,J.D. 1983. The structure of the main South Wales coalfield and its margins. In: HANCOCK,P.L. (ed.) The Variscan Fold Belt in the British Isles. Adam Hilger, Bristol, 47-73. PEACE, G. R. & BESLY,B. M. 1997. End-Carboniferous fold-thrust structures, Oxfordshire, UK: implications for the structural evolution of the late Variscan foreland of south-central England. Journal of the Geological Society, London, 154, 225-238. PETIT, J. P. 1987. Criteria for the sense of movement on fault surfaces in brittle rocks. Journal of Structural Geology, 9, 597-608. RAOULT,J. F. & MEILLIEZ,F. 1987. The Variscan Front and the Midi Fault between the Channel and the Meuse River. Journal of Structural Geology, 9, 473-479. ROBERTS,J. C. 1972. Minor thrust faults in the Avonian of the North Crop of the South Wales Coalfield. Geological Journal, 8, 23-28. SANDERSON, D. J. & MARCHINI,W. R. D. 1984. Transpression. Journal of Structural Geology, 6, 449-458.

266

R. GAYER E T A L .

SETIABUDIDAYA, D. 1991. Magnetostratigraphy and Tectonics in South Wales. PhD thesis, University of Liverpool. SRIVASTAVA,D. C., LISLE, R. J. & VANDYCKE, S. 1995. Shear zones as a new type of palaeostress indicator. Journal of Structural Geology, 17, 663-676. STRANiK, Z., DvoRAK, J., KREJt~I, O., MOLLER, P., PI~ICHYSTAL,A., SUK, M. & TOMEK, C. 1993. The contact of the North European Epivariscan Platform with the West Carpathians. Journal of the Czech Geological Society, 38, 21-29. SWIDZINSKI, H. 1948. Stratigraphical index of the Northern Flysch Carpathians. Bulletin du Service GOologime de Pologne, 37, 1-128. TAPPONIER, P. 1977. Evolution tectonique du systbme alpin en M6diterran6e: poinconnement et 6crasement rigide-plastique. Bulletin de la SociOtO GOologique de France, 19, 437-460. THOMAS, W. A. 1977. Evolution of AppalachianOuachita salients and recesses from re-entrants and promontories in the continental margin. American Journal of Science, 277, 1233-1278.

TIt(OFF, B. & TEYSSIER, C. 1992. Crustal-scale, en echelon 'P-shear' tensional bridges: a possible solution to the batholic room problem. Geology, 20, 927-930. TUNBRIDGE, I. P. 1986. Mid-Devonian tectonics and sedimentation in the Bristol Channel area. Journal of the Geological Society, London, 143, 107-116. T~3NYI, I. & KovAc, M. 1991. Palaeomagnetic investigation of the Neogene sediments from the Little Carpathians (Lower Miocene of the SW part of the Western Carpathians). Contributions of the

Geophysical Institute, Slovak Academy of Sciences, 21, 125-146. WALCOTT,R. I. 1978. Geodetic strains and large earthquakes in the axial tectonic belt of North Island, New Zealand. Journal of Geophysical Research, 83, 4419-4429. WILKERSON, M. S. & MARSHAK, S. 1991. Factors controlling slip-lineation orientation on thrust-fault planes. Tectonophysics, 196, 203-208.

The main phase of the Hercynian orogeny in the Pyrenees is a dextral transpression G. G L E I Z E S , D. L E B L A N C & J. L. B O U C H E Z POtrophysique et Tectonique, U M R 5563 C N R S , UniversitO Paul-Sabatier, 38 rue des Trente-Six-Ponts, 31400 Toulouse, France Abstract: Synthesis of several recent structural studies in the Pyrenees points to a syntectonic emplacement of granitic plutons, more or less early during the D2 main phase of the Hercynian orogenesis. A dextral E-W component of simple shear is demonstrated for D2, which was previously considered as a pure N-S compression. This component is recorded in the granites by large shear bands and sigmoidal features, and in the country rocks by asymmetrical strain shadows of foliation trajectories in map view and by systematically dextral sense of shear in xz sections. The compressive component of D2, well known in the country rocks where it induced a strong flattening related to ubiquitous upright isoclinal tight folds, was also observed in the granites. There, we documented a reverse movement within magmatic or solid-state shear bands. The addition of all these data constrains a new interpretation for the D2 main Hercynian phase in the whole Pyrenees: this phase was a dextral transpression.

Although the Pyrenees have been much studied, no consensus exists on the number and the nature of the tectonic phases in this chain during the Hercynian orogeny. However, many workers now accept that the following main phases occurred: an early phase of southward thrusting followed by a compressional north-south phase, the whole being followed by a late phase of extension. The north-south compression event, which we will name D2 in accordance with the classical terminology, is u n d o u b t e d l y responsible for the map-scale structures. Depending on the phase considered the most important, very different geodynamical contexts were proposed to account for the geological data: continuous convergence (Guitard 1970; Zwart 1979; Matte & Mattauer 1987), rifting or extension (Wickham & Oxburgh 1985, 1986; Soula et al. 1986), or convergence followed by extension (Van den Eeckhout & Zwart 1988; Vissers 1992). These models, however, were exclusively based on observations in the Palaeozoic metasediments and did not take into account the numerous granitoid plutons which crop out in the Hercynian Pyrenees (Fig. 1). These plutons were somewhat disregarded, as they did not seem to relate to H e r c y n i a n tectonics, on account of numerous R b - S r ages at about 275 Ma (compilation by Vissers (1992)), implying for the plutonism a post-tectonic, Permian age. Moreover, the internal structures of the plutons were difficult to unravel by field measurements. The broad outlines of the planar fabrics were described in some plutons (Marre

1973; Lagasqui6 1984) but the linear fabrics remained unknown. The recent development of the anisotropy of magnetic susceptibility (AMS) technique now allows rapid accurate measurement of the linear and planar fabrics of the granitoids (Hrouda 1982; Borradaile 1988; Bouchez 1997). In the Pyrenees, the granites are devoid of magnetite and their paramagnetic susceptibility and anisotropy is essentially carried by the biotites (Gleizes et al. 1993). The long axis of the AMS ellipsoid therefore matches with the mineral lincation, the zone axis of biotites, whereas the short axis matches with the pole of the foliation, the mean planar disposition of biotite (001) cleavages. Application of the magnetic fabric technique to several granitoid plutons of the Pyrenees has resulted in very detailed cartographic pictures of the internal structures of plutons (Santana et al. 1992; Gleizes et al. 1991; Leblanc et al. 1994; Bouchez & Gleizes 1995). These results were complemented by structural studies in the country rocks of some of the plutons (Leblanc et aL 1996a; Evans et aL 1997, 1998). The present paper is a synthesis at the scale of the whole Pyrenean chain of the structural data obtained so far by this new approach. We demonstrate the syntectonic character of the granitic plutonism, which was mostly coeval with the D2 main phase. We also present evidence constraining a new interpretation of this phase, which was not only a north-south compression but also had a strong dextral strike-slip component. D2 was therefore transpressional. This

GLEIZES,G., LEBLANC,D. & BOUCHEZ,J. L. 1998. The main phase of the Hercynian orogeny in the 267 Pyrenees is a dextral transpression. In: HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. E (eds) 1998. Continental Transpressional and Transtensional Tectonics. Geological Society, London, Special Publications, 135, 267-273.

268

G. GLEIZES E T A L .

Fig. 1. The main Hercynian plutons of the Pyrenees.

result better constrains the geodynamic evolution of the chain during the Hercynian orogeny.

Evidence for the syntectonic character of the granitoids The structural patterns of the plutons strongly indicate syntectonic emplacement during the Hercynian orogeny. Some of them, such as Bassi6s, were emplaced early, because, during D2, they behaved like rigid bodies which were more or less deformed in the solid state (Evans et al. 1997). The fact that most plutons of the Pyrenees are coeval with D2 is demonstrated, however, by the regional foliation-lineation patterns and associated kinematics, which are alike in the granitoid bodies and their neighbouring country rocks (Leblanc et al. 1996a, b; Gleizes et al. 1997). Additional evidence of syn-D2 emplacement of plutons is provided by contact metamorphic minerals. For example, the scapolites from the aureole of the Maladeta pluton include internal sigmoids trapping the D2 foliation, but these crystals were also boudinaged with development of pressure shadows during the same D2 event (Evans et al. 1998). That the plutons are syntectonic is now consistent with recent U - P b zircon studies giving emplacement ages at about 310 Ma (Romer & Soler 1995; Paquette et al. 1997). These ages, which fit with the period of late orogenic movement of the Hercynian belt, provide a base for the dating of D2 in the Pyrenees, and invalidate the post-tectonic Rb-Sr Permian ages previously obtained from several plutons. As a conclusion of these recent studies, field arguments such as obliquity of pluton borders

across D2 structures (e.g. Garcia Sansegundo 1992), should no longer be put forward as evidence for post-tectonic emplacement. We think that the emplacement of a magma lasts a very short time during a tectonic event and that a pluton may be locally discordant with respect to the structures of its wall-rocks which had previously begun to form during the same event.

Evidence for the dextral shear component of D2 A D2 dextral shear component is supported by various data from both the granite bodies and the country rocks. I n the g r a n i t e s

As simulated experimentally by Arbaret et al. (1997), the magnetic fabrics in a granite pluton record magma deformation, in the magmatic state principally. Overprints in the solid state may also affect the fabric to some extent (Benn 1994). The structural pattern of a given pluton therefore records its strain history during emplacement. The latter is no more than a snapshot during a tectonic phase, only a few increments of that phase being recorded by each pluton. The granite plutons of the Pyrenees were probably emplaced at different ages during the orogenesis, but their structures invariably attest to a D2 dextral shearing which induced asymmetries of different types in early emplaced plutons and in plutons fully coeval with D2. This consistency of the tectonic regime allows us to consider them as tectonic markers of the orogen and more precisely as markers of

TRANSPRESSION IN THE HERCYNIAN PYRENEES

269

Fig. 2. Structural map of the Mont Louis-Andorra pluton and its country rocks (after Bouchez & Gleizes 1995; Gleizes et al. 1997). The within-pluton structures derive from AMS measurements. The shaded corridors in the granite are D2 reverse dextral shear zones with solid-state microstructures.

the permanent dextral shear component during D2. The early emplacement of the Mont Louis-Andorra pluton (Bouchez & Gleizes 1995) is recorded by a first set of structures with a NE-SW trend (Fig. 2). Its D2 overprint is recorded by a second set of structures in NW-SE trending shear corridors. Within these corridors (shaded in Fig. 2), the rock microstructures display a continuum of solid-state deformation from high to low temperatures. These corridors are characterized by foliation planes steeply dipping to the north and clearly deflecting the former NE-SW trending structures into dextral sigmoids. The Cauterets-Panticosa and Trois Seigneurs plutons are coeval with D2 and illustrate differently the dextral shear component of this phase. On the one hand, the three juxtaposed Cauterets-Panticosa granite bodies (Fig. 3) display a remarkable lineation pattern (Leblanc et al. 1996b) characterized, in map view, by NE-SW trajectories deflected into dextral sigmoids. In the West Cauterets body, the sigmoidal pattern includes N l l 0 ~ trending magmatic shear corridors (shaded in Fig. 3) which contribute to the bulk dextral movement. In the Panticosa body, the foliation trajectories are also organized into dextral sigmoids, parallel to those of the lineations (Santana et al. 1992). On the other hand, the Trois Seigneurs granodiorite displays subvertical foliations parallel to the pluton border, and the associated subhorizontal lineations are arranged in an eastward divergent fan (Fig. 4). Such structures indicate an E - W stretching of the magma and, together with the arcuate

shape of the pluton, they are attributed to an overall dextral movement during D2 (Leblanc et al. 1996a), in accordance with the kinematic data collected in the country rocks. In the metasediments

The $2 foliation trajectories are mostly N110~ trending and subvertical. In the areas where stretching lineations have been observed, they are slightly plunging to the west on average. Such a geometry implies a strike-slip component for D2, and its dextral sense is indicated by the consistent criteria observed in numerous x z thin sections, particularly around the plutons of Trois Seigneurs (Fig. 4) (Leblanc et al. 1996a), Bassi~s (Evans et al. 1997) and Lisse-Droite (Gibert & Bouchez 1991). The dextral component of D2 can also be demonstrated on a regional scale. On the structural maps, the foliation trajectories wrap around the plutons, forming dextrally asymmetrical neutral points. Such patterns are particularly obvious around the Cauterets (Fig. 3) and Nrouvielle (Gleizes et al. 1997) plutons. A neutral point also appears at the southwestern tip of Bassibs, the northeastern one being truncated by an Alpine fault. In addition, sigmoidal foliation trajectories, along with C-S patterns characteristic of dextral shear, can be traced regionally around the pluton of Bassirs (Evans et al. 1997).

Evidence for compression during D2 As for the dextral shear component, evidence for a compressive component is strengthened by

270

G. GLEIZES E T A L .

Fig. 3. Structural map of the Cauterets-Panticosa plutonic complex (AMS measurements) and its country rocks (after Santana et al. 1992; Leblanc et al. 1996b). The two shaded corridors in the West Cauterets body are D2 magmatic dextral shear zones.

data from both the granite bodies and their country rocks. I n the g r a n i t e s

Compression coeval with dextral shear expresses itself by shear bands with a thrust component. These shear bands occurred in the solid state within the early plutons and mostly in the magmatic state within the fully syn-D2 plutons. In the early Mont Louis-Andorra pluton, the thrust component of the dextral shear bands is indicated by the overall NW plunges of the D2 lineations within D2 foliations with overall moderate to steep dips to the north. This structural pattern therefore characterizes a tectonic regime which is compressive and dextral simultaneously. The reverse movement resulted in a stronger uplift of the eastern part of the pluton, compared with its Andorra western part. This agrees with the observation of the gneissic pluton floor metamorphosed at sillimanite grade at the eastern extremity of the pluton, contrasting with the anchizonal sediments surrounding the flattened dome-shaped pluton, which corresponds to the roof at the western extremity.

Contrasting with the latter case, emplacement of the Maladeta plutonic complex occurred during D2. The complex is made of four individual bodies separated by magmatic contacts (Leblanc et al. 1994). The southern border zone of three of them is characterized by high anisotropy ratios and by magmatic lineations whose mean trend is nearly N-S and sub-perpendicular to the neighbouring border. This pattern is interpreted as the result of magma upwelling during a southward directed thrust event. I n the m e t a s e d i m e n t s

Almost all workers agree that the D2 main Hercynian phase of the Pyrenees was strongly compressive. This is based on the ubiquity in the metasediments of upright, isoclinal and very tight F2 folds with a well-developed axial planar schistosity. In some areas of the Pyrenees, the D2 flattening is nearly coaxial, marked by either totally lacking or poorly imprinted and disorganized stretching lineations. By contrast, as already mentioned, well-imprinted horizontal to west-plunging lineations point to a dextral strike-slip component of D2 strain in other

TRANSPRESSION IN THE HERCYNIAN PYRENEES

271

Fig. 4. Structural map of the Trois Seigneurs granodiorite and its country rocks (after Leblanc et aL 1996a). The data result from AMS measurements in the pluton, and from field measurements and microstructural observations for shear sense in the country rocks. areas. Such variations in the strength and plunges of the linear structures indicate that strain was heterogeneously distributed, and possibly partitioned into dominant compressive and dominant strike-slip domains.

Condusion We have argued that the main D2 Hercynian tectonic phase of the Pyrenees, classically considered as a compression responsible for a large N-S shortening, is also characterized by a dextral parallel-to-chain strike-slip component of strain. Regional dextral shearing is well documented by the magmatic or solid-state structures recorded in the granite bodies and by the asymmetrical features of the metasediments from the map scale down to the microscopic scale of observation. This overall coexistence during the same geological event, and all along the chain, of coaxial and non-coaxial signatures of strain allows us to conclude that the D2 phase of the

Pyrenees was transpressive. This conclusion agrees with the numerous theoretical and field studies which have shown that transpressional strain in the crust can be partitioned into separate domains of dominant pure shear and simple shear (e.g. Cobbold et al. 1991; Girard 1993; Jones & Tanner 1995). These results have fundamental implications for the understanding of the Hercynian plutonism in the Pyrenees, which cannot be linked to a late extensive geodynamical context as previously proposed (Vissers 1992) but to the D2 transpression. In this regard, this emplacement mode is completely different from that of the late granites of the nearby French Massif Central, which relate to massif-scale extension during post-thickening gravity collapse (Faure 1995). In the Pyrenees, the collisional stage (D1 event) probably induced a moderate crustal thickening and generation of calc-alkaline granodioritic magmas. During the late collisional stage (D2 event) large amounts of magmas were

272

G. GLEIZES E T A L .

t r a n s f e r r e d t h r o u g h the crust, r e l a t e d to strain h e t e r o g e n e i t i e s during transpression. In agreem e n t with D ' L e m o s et al. (1992), R o s e n b e r g et al. (1995) a n d H u t t o n (1997), a m o n g others, it is therefore concluded that a transpressional e n v i r o n m e n t does not rule out the e m p l a c e m e n t of granitic magmas. Local tensional sites r e l a t e d to shear zones in the ductile crust and to jog openings in the brittle u p p e r crust are likely sites of m a g m a transfer and ponding. We thank B. Holdsworth for his numerous suggestions, and A. McCaig and R. D'Lemos for their constructive reviews. The Centre National de la Recherche Scientifique and the Universit6 PaulSabatier are thanked for their financial support.

References ARBARET, L., DIOT, H., BOUCHEZ,J. L., LESPINASSE,P. & DE SAINT-BLANQUAT,M. 1997. Analogue 3D simple shear experiments of magmatic biotite subfabrics. In: BOUCHEZ,J. L. HUTI"ON,D. H. W. & STEPHENS. W. E. (eds) Granite: from Segregation of Melt to Emplacement Fabrics. Kluwer Academic, Dordrecht, 129-143. BENN, K. 1994. Overprinting of magnetic fabrics in granites by small strains: numerical modelling. Tectonophysics, 233, 153-162. BORRADAILE, G. J. 1988. Magnetic susceptibility, petrofabrics and strain. Tectonophysics, 156, 1-20. BOUCHEZ, J. L. 1997. Granite is never isotropic: an introduction to AMS studies of granitic rocks. In: BOUCHEZ, J. L. HUTTON, D. H. W. & STEPHENS,W. E. (eds) Granite: from Segregation of Melt to Emplacement Fabrics. Kluwer Academic, Dordrecht, 95-112. -& GLE~ZES, G. 1995. Two-stage deformation of the Mont Louis-Andorra granite pluton (Variscan Pyrenees) inferred from magnetic susceptibility anisotropy. Journal of the Geological Society, London, 152, 66%679. COBBOLD, P. R., GAPAIS, D. & ROSSELLO,E. A. 1991. Partitioning of transpressive motions within a sigmoidal foldbelt: the Variscan Sierras Australes, Argentina. Journal of Structural Geology, 13, 743-758. D'LEMOS, R. S., BROWN,M. & STRACHAN,R. A. 1992. Granite magma generation, ascent and emplacement in a transpressional orogen. Journal of the Geological Society, London, 149, 487-490. EVANS,N. G., GLEIZES,G., LEBLANC,D. & BOUCHEZ,J. L. 1997. A new interpretation of the Hercynian tectonics in the Pyrenees based on the detailed examination of structures around the Bassi6s granite. Jounml of Structural Geology, 19, 195-208. , ,-& -1998. Syntectonic emplacement of the Maladeta granite (Pyrenees) deduced from relationships between Hercynian deformations and contact metamorphism. Journal of the Geological Society, London, 155, 209-216.

FAURE, M. 1995. Late orogenic Carboniferous extensions in the Variscan French Massif Central. Tectonics, 14, 132-153. GARC1A-SANSEGUNDO,J. 1992. Estratigrafia y estructura de la zona Axial pirenaica en la transversal del Valle de Arhn y de la Alta Ribagorqa (Parte III). Boletin Geologico y Minero, 103, 253-290. GLEIZES, G., LEBLANC, D. & BOUCHEZ, J. L. 1997. Variscan granites of the Pyrenees revisited: their role as syntectonic markers of the orogen. Terra Nova, 9, 37-40. --, Nt~DI;LEC, A., BOUCHEZ, J. L., AUTRAN, A. & ROCHETFE, P. 1993. Magnetic susceptibility of the ilmenite-type granite Mont-Louis-Andorra (Pyrenees): a new tool for the petrographic characterization and the regional mapping of zoned granite plutons. Journal of Geophysical Research, 98, 4317-4331. GIBERT, P. & BOUCHEZ, J. L. 1991. Le granite syntectonique de Lisse-Droite (Luchonais): nouveau jalon d'une phase cisaillante tangentielle pr6coce dans les Pyr6n6es hercyniennes. Comptes Rendus de l'Acad~mie des Sciences, Paris, 312, 407-414. GIRARD, R. 1993. Orogen-scale strain partitioning and an analogy to shear-bands in the Torngat orogen, northeastern Canadian Shield. Tectonophysics, 224, 363-370. GU1TARD, G. 1970. Le m~tamorphisme hercynien m~sozonal et les gneiss oeill~s du massif du Canigou (PyrOn&s orientales). M6moire du Bureau de Recherches G6ologiques et Mini6res, Orldans, 63. HROUDA,E 1982. Magnetic anisotropy of rocks and its application in geology and geophysics. Geophysical Survey, 5, 37-82. HuTroN, D. H. W. 1997. Syntectonic granites and the principle of effective stress: a general solution to the space problem?. In: BOUCHEZ,J. L., HUTrON, D. H. W. & STEPHENS,W. E. (eds) Granite: from Segregation of Melt to Emplacement Fabrics. Kluwer Academic, Dordrecht, 189-197. JONES, R. R. & TANNER,P. W. G. 1995. Strain partitioning in transpression zones. Journal of Structural Geology, 17, 793-802. LAGASQUIt~,J. J. 1984. G Oomorphologie des granites: les massifs granitiques de la moitid orientale des PyrdnOes franfaises. Editions CNRS, Paris. LEBLANC,D., GLEIZES, G., LESPINASSE,P., OLIVIER,PH. & BOUCHEZ, J. L. 1994. The Maladeta granite polydiapir, Spanish Pyrenees: a detailed magneto-structural study. Journal of Structural Geology, 16, 223-235. , Rogx, L. & BOUCHEZ,J. L. 1996a. Variscan dextral transpression in the French Pyrenees: new data from the Pic des Trois-Seigneurs granodiorite and its country rocks. Tectonophysics, 261, 331-345. - - , SANTANA,V., OLIVIER,PH., BOUCHEZ,J. L. & ARANGUREN,A. 1996b. Estructuras sigmoides en el macizo granitico hercinico de Cauterets-Panticosa (Pirineos). Geogaceta, 20, 778-781. MARRE, J. 1973. Le complexe Oruptif de Qu&igut: pOtrologie, structurologie, cindmatique de raise en place. PhD Thesis, Universit6 de Toulouse.

TRANSPRESSION IN THE HERCYNIAN PYRENEES MATI'E, PH. & MATrAUER, M. 1987. Hercynian orogeny in the Pyrenees was not a rifting event. Nature, 325, 739-740. PAQUETFE,J. L., GLEIZES,G., LEBLANC,D. & BOUCHEZ, J. L. 1997. Le granite de Bassi~s (Pyr6n6es): un pluton syntectonique d'gge westphalien. G6ochronologie U-Pb sur zircons. Comptes Rendus de l'AcadOmie des Sciences, 324, 387-392. ROMER, R. L. & SOLER, A. 1995. U-Pb age and lead isotopic characterisation of Au-bearing skarn related to the Andorra granite (central Pyrenees, Spain). Mineralium Deposita, 30, 374-383. ROSENBEr~G,C. L., BZRGzR, A. & SCHMmT,S. M. 1995. Observations from the floor of a granitoid pluton: inferences on the driving force of final emplacement. Geology, 23, 443-446. SANTANA,V., BOUCHEZ, J. L., GLHZES, G. & TUBIA, J. M. 1992. Estructura del pluton granitico de Panticosa (Pirineos). III Congreso Geologico de Espaha, Salamanca, 2, 179-185.

-

273

SOULA, J. C., DEBAT,P., DI~RAMOND,J., GUCHEREAU,J. Y., LAMOUROUX,C., POUGET, R & ROUX, L. 1986. l~volution structurale des ensembles m6tamorphiques, des gneiss et des granitoYdes dans les Pyrdn6es centrales. Bulletin de la Socidtd GOologique de France, 8, 79-93. VAN DEN EECKOUT,B. & ZWART,H. J. 1988. Hercynian crustal-scale extensional shear zone in the Pyrenees. Geology, 16, 135-138. VISSERS, R. L. M. 1992. Variscan extension in the Pyrenees. Tectonics, 11, 136%1384. WICKHAM,S. M. & OXBURGH,E. R. 1985. Continental rifts as a setting for regional metamorphism. Nature, 318, 330-333. & -1986. A rifted tectonic setting for Hercynian high-thermal gradient metamorphism in the Pyrenees. Tectonophysics, 129, 53-69. ZWART, H. J. 1979. The geology of the Central Pyrenees. Leidse Geologische Mededelingen, 50, 1-74. -

Three-dimensional retro-modelling of transpression on a linked fault system: the Upper Cretaceous deformation on the western border of the Bohemian Massif, Germany D A V I D C. T A N N E R 1, J A N H . B E H R M A N N

1, O N N O O N C K E N 2 & K L A U S

WEBER 3

1Geologisches Institut, Universitiit Freiburg, Albertstrafle 23b, D-79104 Freiburg, Germany 2GeoForschungsZentrum, Telegrafenberg A17, D-14473 Potsdam, Germany 3IDGL, Universitiit GOttingen, Goldschmidtstrafie 3, D-37077 GOttingen, Germany

Abstract: The Zone of Erbendorf-Vohenstraug (ZEV) on the western margin of the Bohemian Massif was deformed by an Upper Cretaceous intra-plate deformation event. Dextral transpression was caused by the reactivation of pre-existing structures. Using the extensive geological database available, we have constructed a three-dimensional virtual model of the ZEV. The model was deformed in reverse, to remove the effects of the Upper Cretaceous event. This involved moving the hanging wall (the ZEV) in a sinistral transtensive sense northwards above a composite active fault surface composed of two steep faults, perpendicular to each other in strike, and a detachment intersecting both faults at 9.5 km depth. Hanging-wall deformation was accommodated by antithetic inclined shear. Seven kilometres heave of the hanging wall fulfilled the geological constraints. Calculated uplifts range from 2 to 6 km. Deformation is mostly only contained within the ZEV. The hangingwall deformation above a linked fault system was highly complex, causing rollover above one fault and drag-folding above the other. The most important control on the vertical movement and deformation of the hanging wall was a 30~change in the strike of one of the coupled faults.

The Zone of Erbendorf-Vohenstraug ( Z E V ) is a small Variscan tectonometamorphic unit on the western edge of the Bohemian Massif (Fig. 1). On its western boundary, the Z E V is thrust over Upper Cretaceous and older sediments. However, it was not until the drilling of the German deep continental borehole (KTB) in the n o r t h e r n ZEV, and the assimilation of associated geophysical data, that the full extent of the Upper Cretaceous deformation was realized. As a result of the intensive work done in the last 10 years, the Z E V represents one of the most highly studied crystalline areas in the world. There exists now an extensive geophysical and geological database that allowed us to model the Z E V in virtual space. The U p p e r Cretaceous event r e w o r k e d a linked system of Variscan structures, causing dextral transpressional deformation of the Z E V as the hanging wall. It was therefore necessary to analyse this event in three dimensions. The method employed here was three-dimensional (3D) retro-deformation, in which the current geological situation was deformed with the reverse kinematics to those of the U p p e r Cretaceous. Finite displacements were constrained by several pieces of geological evidence. Until now, retro-deformation has only

been achieved using 2D cross-sections, matching an undeformed template. This paper shows that the retro-modelling in 3D space, without a lithostratigraphic template, generates insights into both the retro-deformed event itself and the pre-deformation geometry.

Geological background The B o h e m i a n Massif forms part of the M o l d a n u b i a n Zone, a region of para- and orthogneisses with large volumes of posttectonic granites, formed during the Variscan Orogeny (Franke 1989). In western Bohemia, along the border between G e r m a n y and the Czech Republic, the Moldanubian Zone is subdivided into a number of lithostratigraphically and tectonically distinct units (Fig. 1); the Teplgt-Barrandian Zone, the Z E V and the Moldanubian senso stricto (s.s.). A complete description of the Variscan tectonic development of this region has been given by Franke (1989, 1990) and B e h r m a n n & Tanner (1997). The m e t a m o r p h i c d e v e l o p m e n t has been reviewed by B10mel (1984, 1990) and with special reference to the Z E V by Bltimel et al. (1988) and Reinhardt et al. (1989).

D. C. TANNER,J. H. BEHRMANN,O. ONCKENd(ZK. WEBER. 1998. Three-dimensional retro-modelling of 275 transpression on a linked fault system: the Upper Cretaceous deformation on the western border of the Bohemian Massif, Germany. In: HOLDSWORTH,R. E., STRACHAN,R. A. c~zDEWEY,J. t (eds) 1998. Continental Transpressionaland Transtenxional Tectonics.Geological Society, London, Special Publications, 135, 275-287.

276

D.C. TANNER E T A L .

Fig. 1. The geological framework of the Variscan units along the western margin of the Bohemian Massif. Granite bodies are cross-patterned, amphibolites are black. The white area is the South German Basin. Dashed lines are the traces of seismic profiles mentioned in text. The bold black box marks the area shown in Fig. 2.

S u r f a c e s t r u c t u r e o f the m o d e l l e d area

The Z E V is approximately wedge shaped, 35 km in length and 25-10 km in width (from north to south). It is bounded on all sides by steep faults or shear zones (Fig. 2). On the western border, the Z E V is thrust over the Carboniferous- Mesozoic sediments of the South German Basin along a fault known as the Frankonian Lineament. This is a major N W - S E striking, NE-dipping fault that can be traced northwestwards for c. 160 km. To the south, the E - W trending steep Luhe Lineament juxtaposes the Z E V against the Moldanubian s.s. The eastern boundary with the Moldanubian s.s. is largely obscured by late Variscan plutons, but in the few outcrops available it is seen as a steeply east-dipping shear zone subparallel to the foliation on both sides (Heinicke et al. 1986; Stein & Kleemann 1992). The Erbendorf Lineament forms the northern border to the ZEV, although it is also mostly

obscured by late Variscan granitoids. It separates the Z E V (and on a larger scale the Moldanubian Zone) from the Saxothuringian Zone and is postulated as a major plate suture in most Variscan tectonic models. The most obvious feature of the Z E V is the large amount of granitoid plutons. These granitoids are part of the composite Northwestern Oberpfalz Pluton (Wendt 1993). They intruded and cooled to less than 400~ between c. 340 and 295 Ma. Weakly metamorphic, probably Lower Palaeozoic metasediments (the 'Wetzldorf Series'; R6hr & Zulauf 1992) crop out along the southern and northern boundaries of the Z E V (Fig. 2). They may represent Saxothuringian sediments (Voll 1960; Weber 1992; Lapp 1996). Structural (Variscan) fabrics were mapped by Stettner (1990) and described in detail by Weber (1992) and Lapp (1996). To summarize the work of the latter two workers, the first deformation,

THREE-DIMENSIONAL RETRO-MODELLING OF TRANSPRESSION

277

Fig. 2. Geological map of the area around the ZEV. Major tectonic boundaries are shown by bold lines, u/c, erosional unconformity. Modified after Stettner (1990), ROhr & Zulauf (1992) and Lapp (1996). Bold dashed line represents the extent of the ISO'89 seismic array. Line A is the line of section in Figs 3a and 4; lines B and C are the lines of section in Figs 7 and 8, respectively. German Gauss-Krtiger coordinate system based on 12~ longitude, in kilometres. D1, is found only in gneisses west of the granites, and is characterized by a vertical to steep, N W - S E striking foliation and sub-horizontal lineation. D1 is correlated with Devonian mediumpressure and high-temperature (6 to >10 kbar, 600--800~ Bliimel et al. 1988; Reinhardt et al. 1989) metamorphism (M1). A second deformation, D2, is correlated with Upper Carboniferous low-pressure and high-temperature (2-5 kbar, 600-800~ Kleemann 1992) metamorphism (M2). D2 foliation is sub-vertical, striking N E - S W with a shallowly SW-plunging stretching lineation in the south, east and north, and steeply plunging in the middle of the ZEV.

Geological information from the K T B borehole The German Continental Deep Drilling Project (KTB, see E m m e r m a n n & Wohlenberg 1989) drilled two boreholes 200 m apart (the pilot and main boreholes), down to depths of 4 km and 9.1 kin, respectively, in the northern Z E V (see Fig. 2 for location). The project has produced a wealth of geological, geophysical and geochronological data. The KTB borehole encountered a series of garnet-sillimanite-(kyanite)-biotite gneisses interspersed with, in order of decreasing

278

D.C. TANNER E T A L .

Fig. 3. (a) Geological section through the ZEV (see Fig. 2 for location) based on the KTB borehole data, modified after Hirschmann et al. (1994). (b) Dip direction and dip angle of the metamorphic foliation in the KTB borehole, recognized by the Formation Micro Scanner. Grey area represents total range of data, bold line the harmonic mean over 100 m intervals; n = 1683. (e) The apparent fission-track ages of sphene, plotted against depth, from Coyle & Wagner (1994). Error bars indicate standard deviation. Line drawn at 7200 m represents the large offset in ages across the Frankonian Lineament.

amount, amphibolites, ophitic metagabbros, calc-silicates and metatuffites (Duyster et al. 1995), all of which also crop out on the surface. Amphibolites occurred mainly between 3000 and 7000 m (Fig. 3a). Between the surface and a depth of c. 7000 m, there are a number of thin, late Variscan lamprophyre dykes (see later for relevance). Metamorphic foliation in the KTB borehole is consistently steeper than 40 ~ with a variable strike (Fig. 3b; Duyster et al. 1995). A number of cataclastic shear zones were discovered by the KTB borehole, most of which were parallel to the foliation. The Frankonian Lineament is probably related to a set of cataclastic shear zones at depths of 6850-6950 m, 7000-7020 m, 7060-7100 m and 7190-7260 m (Duyster et al. 1995). Chlorite geothermometry suggests that the fault fabrics were formed at temperatures between 250 and 350~ (Schi3ps & Friedrich 1994). Geophysical information Seismic information. A large number of seismic profiles and experiments was completed around the KTB site (see summaries by D E K O R P Research Group (1988), Dtirbaum et al. (1990, 1992) and Meissner & Bortfeld (1990)). The first

reflection seismic profiles (Fig. 1) were the D E K O R P 4 profile, three short lines parallel to the D E K O R P 4 (lines KTB8504-6) and three short lines perpendicular to the D E K O R P 4 (lines KTB8501-3). The short lines only traversed the Z E V ( D E K O R P Research Group 1988; Meissner & Borffeld 1990). In general, the D E K O R P 4N and associated lines demonstrated only shallowly dipping structures, which were not verified by the KTB borehole. In 1989, a detailed 3D seismic survey, ISO89, was centred on the KTB site and oriented orthogonally with respect to the F r a n k o n i a n Lineament (Fig. 2; see Dt~rbaum et al. 1992). It covered an area of 17.85 km • 19.1 km (357 inlines • 382 crosslines) with a spacing of 50 m and an average 15-fold cover. The Frankonian Lineament or SE1 (as it is known in seismic interpretations) is clearly seen as a 55 ~ NEdipping reflector down to a depth of c. 8 kin. Other reflectors, parallel to SE1 (SE2-4A) can be traced down to 9.5 km. SE4 probably corresponds to the Fichtelnaab Fault (Hirschmann 1992; Wiederholt 1992). From the many different projections of the D E K O R P data, the most interesting are those of Simon & Gebrande (1994a, b). Figure 4 shows the pre-stack migrated seismic section KTB8502

THREE-DIMENSIONAL RETRO-MODELLING OF TRANSPRESSION

279

Fig. 4. Pre-stack migration, amplitude-density (higher amplitudes are darker) plot of the seismic section KTB8502 (see Figs 1 and 2 for location) after Simon & Gebrande (1994b). Dashed white line shows the interpretation of the Frankonian Lineament and of the continuation of the fault at depth used in this work. EB, Erbindort body.

(Simon & Gebrande 1994b), in which the Frankonian Lineament can be traced down to a depth of 9-10 km. We postulate (as did Duyster et al. 1994) that the Frankonian Lineament is listric in form, as it cannot be projected through the Erbendorf body. Instead, the Frankonian Lineament detaches along the top of the body, a surface that dips shallowly east (Fig. 4).

Gravity and magnetic anomaly modelling. Bosum et al. (1994; see also Bosum et al. 1993) produced a 3D integrated model of the gravity and magnetic anomaly data of an area 20 km x 20 km around the KTB location. Both the gravity and magnetic surveys demonstrate the sub-surface extent of the amphibolite bodies and the positions of the Frankonian and Fichtelnaab Faults. The Bouguer anomalies were matched by postulating the depth of the Falkenberg Granite to be 7-8 km (see also Fig. 4). Other geological information Sedimentary record of the Mesozoic basin adjacent to the ZEV. The sediments of the South German Basin west of the Bohemian Massif are well documented (see Schr6der 1975, 1978, 1988). The deepest sediments recorded in the Weiden borehole (see Fig. 2 for location) are dated as Westphalian-Stephanian (Dill 1990). Sedimentation was nearly continuous throughout the Permian, Triassic and Jurassic; these sediments amount to over 2-3 km in thickness (Schr0der 1975, 1978). During the Lower Cretaceous, there was no deposition. Schr6der (1987) suggested that c. 1.5 km of the Jurassic and

Triassic cover was eroded during this period. In the area adjacent to the ZEV, no Jurassic sediments remain (Fig. 2). Above an unconformity, coarse clastic sedimentation began again in the Cenomanian (Fig. 5), with the deposition of c. 500 m of sandstones and conglomerates (Klare & Schr6der 1990). The Upper Cretaceous sediments thicken and coarsen rapidly towards the Frankonian Lineament, suggesting the ZEV as a source for these clastic deposits. Another unconformity in the Maastrichian marks the end of Cretaceous sedimentation (Fig. 5).

The record of Mesozoic and Cenozoic crustal movements. Wemmer & Ahrendt (1992, 1994) showed that muscovite fine-fractions in many of the cataclastic zones within the KTB borehole have Lower Cretaceous K - A r cooling ages (Fig. 5). Newly grown adular grains in extensional fractures have K - A r ages which range from 113 to 84 Ma (Wemmer & Ahrendt 1994). Preliminary apatite fission-track analyses for various locations within the Z E V have 60~ cooling ages of around 40-60 Ma near the KTB and the major granites, and ages between 120 and 180 Ma in the southern ZEV (Wagner et al. 1991). The rock suite drilled by the KTB was analysed for apatite, sphene and zircon fissiontrack ages (Fig. 5; Jacobs et al. 1993; Coyle & Wagner 1994). All the data (exemplified by the sphene data, Fig. 3c) point to 3-4 km of vertical uplift above the Frankonian Lineament during the Upper Cretaceous. Some of the most interesting aspects of the KTB borehole were the observations that K - A t metamorphic mica and amphibole cooling ages

280

D.C. TANNER E T A L .

stratigraphic age

Ma

Era

Period

60 -

Paleocene

70 -

Maastrichtian

~ Campanian

80-

~

90- m O

sedimentology of the South German Basin Schr6der (1992; 1987; 1988); Klare & Schr6der (1990)

uplift history from fission track analysis

K-Ar radiometric ages of retrograde events

130-

within the ZEV

Mesozoic sediments

(Zulauf 1993)

(Petere k et al. 1994)

,,•/c conglomerate]

Sant./Coniac-

O4

1.13 03

~:___ -~>~

Turon./Cenom

4--

0 < 110- t-

120- 0

palaeostress vectors. Arrrows show direction of the maximum stress in the horizontal plane

Jacobs et al. & Ahrendt' Duyster et al. (1993); Coyle & Wemmer (1994) (1992, 1994) Wagner (1994)

100- w

I.U nr"

tectonic uplift of the KTB

Albian -~ Aptian ~o Barremian "J Hauterivian

{3. "o

Valanginian 140 -

(/) -I--t>

thrusting +

T

._.9. --O

Berriasian Rthonian

Fig. 5. Compilation of geological data relevant to the Upper Cretaceous event, u/c, unconformity. (See text for discussion.) (e.g. Kreuzer et al. 1993; Henjes-Kunst et al. 1994), variscan metamorphic P - T conditions (Reinhardt 1992) and the mineral facies of retrograde metamorphism do not vary greatly with depth (Zulauf 1993; Duyster et al. 1994; Zulauf et al. 1994). This led Duyster et al. (1994) to conclude that the ZEV crust was thickened at least three-fold by post-Variscan events, of which the most important was during the Upper Cretaceous. They postulated that the Frankonian Lineament formed a frontal ramp to a detachment at 9.5 km. In a later publication (Zulauf & de Wall 1994), these workers placed the detachment at 12 km. Zulauf (1993), Peterek (1994) and Peterek et al. (1994) analysed the directions of palaeostress axes from slickensides on structures of different ages in the ZEV and South German Basin, respectively (Fig. 5). Both groups of workers concluded that the maximum stress direction during the Lower and Upper Cretaceous event was horizontally north- south, although Zulauf (1993) divided the Upper Cretaceous into two deformation stages, both causing strike-slip structures, but one with north-south and one with N W - S E maximum stresses. P a l a e o m a g n e t i c vectors. Upper Carboniferous

or Lower Permian palaeomagnetic data are extremely useful in constraining the tectonic rotation caused by later deformations. Palaeomagnetic studies on the KTB pilot borehole (Worm & Rolf 1994; de Wall et al. 1995) reveal that lamprophyre dykes ( K - A r hornblende ages

of 305-295 Ma; Kreuzer et al. 1993; Harms & HUlzl 1994) below the depth of 2600 m have palaeomagnetic vectors consistent with the Permian palaeomagnetic field. This implies there has been no fabric rotation in the northern Z E V since this period.

The Upper Cretaceous deformation event Using the information above, we envisage the following scenario for the Upper Cretaceous development of the ZEV. At 95 Ma, subhorizontal, north-south oriented maximum compressive stress was present along the western border of the Bohemian Massif, part of a plate stable since the end of the Variscan orogeny. Preexisting Variscan faults, the Frankonian and Luhe lineaments, were reactivated, detaching along an east-dipping mid-crustal d6collement at a depth of 10-11 km (the current depth plus 1 km of Tertiary uplift). We suggest that the major faults were simultaneously reactivated (at least during the Upper Cretaceous) because in the Mesozoic sediments of the South German Basin, faults parallel to the Luhe Lineament curve into faults with the orientation of the Frankonian Lineament (see maps by Schr6der (1978) and Bayerisches Geologisches Landesamt (1981)). As a consequence, the ZEV was dextrally transpressively deformed and thrust over both active faults. Uplift of the ZEV near the KTB was about 3-3.5 km, together with internal thickening of the hanging wall (see above). During this intra-plate deformation, large

THREE-DIMENSIONAL RETRO-MODELLING OF TRANSPRESSION volumes of coarse clastic deposits were locally shed onto the footwall fault block. The end of the deformation is marked by the cessation of sedimentation (c. 75 Ma). Although c. 1 km of topography was eroded during the Tertiary (evidence from apatite fission-track, structural and sedimentological data; see Schr6der 1992; Peterek & Schr0der 1997) this was equal on both the footwall and hanging wall. Thus the present topographical surface is sub-parallel to the equivalent surface at the end of the Upper Cretaceous.

Retro-modelling in three dimensions Aim of the modelling. The aim of the modelling exercise was to deform the ZEV to remove the Upper Cretaceous deformation. This would not only reproduce the pre-Upper Cretaceous crustal geometry, but also yield more consistent information on the deformation event itself. This modelling procedure is termed retro-deformation. Three-dimensional retro-deformation was necessary because of the oblique movement vector on the two faults and because the two faults, which are perpendicular to each other, were active simultaneously. The point along the retro-deformation path when the pre-deformation geometry was reached was not constrained by the lithostratigraphy as in balanced cross-sections (e.g. Dahlstrom 1969; Hossack 1979), but indirectly by fission-track ages, palaeomagnetic vectors and petrological evidence. This modelling was carried out using the software program 3D-Move 3.1 ( 9 Midland Valley Exploration Ltd) running on a Silicon Graphics Indigo ( 9 Silicon Graphics Inc.) workstation. The program allows visualization of a model in 3D virtual space. A model can be composed of surfaces and volumes; the former can represent geological features, e.g. faults, stratigraphic or other marker horizons, and the latter can define geological bodies, such as igneous intrusions. A variety of deformation algorithms, such as flexural slip and inclined shear, can be used to deform these objects above a fault, either forward or backwards. Various parameters can be varied in 3D space to suit the geological problem.

Known and projected elements required for modelling. The faults, the Frankonian and the Luhe lineaments, were traced from geological maps. The Frankonian Lineament was in part picked from the 3D seismic sections (ISO '89 seismic survey) and in part projected from the surface outcrop along a vector with azimuth 055 ~

281

and dip 55~ which was found to match the interpretation of the 3D seismic data set. As a result of the genetic relationship of the Luhe Lineament to the Frankonian Lineament, the former was also projected from the surface outcrop (strike approximately east-west; Fig. 2) along the same vector. This produced a surface dipping NNE by c. 70 ~ The listric shape of the Frankonian Lineament was modelled by attaching the fault at a depth of 9.5 km to a 15~ detachment surface, as interpreted in Fig. 4. Because one of the aims of this modelling was to reproduce the geometry of the syn- and postVariscan and before the Upper Cretaceous situation, various palaeo-surfaces and objects were included in the model as passive markers. These include the late Variscan granitoid plutons of Leuchtenberg, Falkenberg and Friedenfels (Fig. 2), the Erbendorf Lineament and the ZEV-Moldanubian s.s. boundary, and representative down-plunge projections of Variscan foliation planes. Of interest for future modelling, locations of radiometric dating and ascertained metamorphic P - T - t paths were also retrodeformed. To visualize the retro-deformation the current topography was also added to the footwall and hanging wall.

Deformation kinematics and parameters. Retrodeformation was accomplished using the reverse movement sense to that of the Upper Cretaceous event, i.e. north-south oriented, sinistral (with respect to the Frankonian Lineament) transtension. In this modelling, the hanging wall (the ZEV) was deformed according to a 3D inclined shear, constant heave, algorithm. The orientation of the shear vector during inclined shear is important, as it determines the internal structures of the hanging wall. Typically, antithetic oblique (30 ~ _+ 10~ from the vertical) shear has been found to model the structure of hanging-wall deformation most accurately (Hauge & Gray 1996). For this modelling we chose a 3D shear vector with azimuth 235 ~ and plunge 70~ for the following reasons: (1) this represents an oblique antithetic shear direction with respect to the Frankonian Lineament; (2) this is the average dip direction of the metamorphic foliation within the ZEV (Fig. 3) and the orientation in which the hanging wall was most likely to fail in a brittle fashion.

Results Figure 6b shows the model after retro-deformation with 7 km of sinistral transtensional hanging-wall heave. Clearly, the topographical surface of the ZEV has been down-thrown the

282

D.C. TANNER ETAL.

THREE-DIMENSIONAL RETRO-MODELLING OF TRANSPRESSION

283

Fig. 7. Map view of the ZEV model after retro-deformation, showing total finite strain magnitude (the value of es, as defined by Nadai (1963)) of the hanging wall. Faults and South German Basin are coloured as in Fig. 6a. m o s t directly along the F r a n k o n i a n L i n e a m e n t , in a classical rollover structure. T h e effects of this structure e x t e n d in the n o r t h e r n Z E V until the m i d d l e of the F a l k e n b e r g Granite, and in the south, up to the Z E V - M o l d a n u b i a n boundary. S u p e r i m p o s e d o n this structure is the effect of orthogonally pulling the hanging wall away f r o m the L u h e L i n e a m e n t . A zone of collapse occurs along this fault, the effects of which e x t e n d very locally into the M o l d a n u b i a n s . s . In Fig. 7, the latter is s h o w n as a zone of very high strain, as is, to a lesser extent, the eastern b o u n d a r y of the rollover structure. In a cross-section parallel to the F r a n k o n i a n L i n e a m e n t (Fig. 8), t h e effect of t h e L u h e

L i n e a m e n t can be clearly seen as a hanging-wall 'drag fold' that is rapidly d o w n - t h r o w n e v e n after m i n i m a l hanging-wall heave. O t h e r effects can also be seen in this section. T h e c. 30 ~ change in strike of the F r a n k o n i a n L i n e a m e n t ( m a r k e d by t h e b o l d a r r o w in Figs 6b a n d 8), f r o m N N W - S S E in the s o u t h e r n Z E V to N W - S E in the n o r t h e r n ZEV, causes releasing-confining b e n d effects (e.g. C r o m w e l l 1974). C o n s e quently, the area 5-10 k m south of the change of fault strike suffers only <1.5 k m vertical disp l a c e m e n t after 7 k m h a n g i n g - w a l l h e a v e , w h e r e a s in t h e area a r o u n d t h e K T B it is b e t w e e n 4 and 6 kin. In addition, the hanging wall is displaced in a series of 'folds' (c. 5 k m

Fig. 6. (a) The 3D model before deformation. Topographies of the South German Basin and Moldanubian s.s., but not the ZEV, are shown as brown translucent surfaces. For all the elements shown, only the outcrop trace is known and the third dimension has been projected with the help of geophysical information (see text for details). It should be noted that jagged edges of surfaces are artefacts and do not affect the modelling. Perspective view from south. (b) The model after removal of the Upper Cretaceous deformation. The orange, dark green and yellow-green surfaces represent the present-day topographical surfaces of the Moldanubian, ZEV and South German Basin, respectively. Faults have the same colour as in (a). Kinematic parameters: movement vector directly north, shear vector with azimuth 235~ and dip 70~ hanging-wall heave 7 km. The bold arrow illustrates the position of the 30~ change in strike of the Frankonian Lineament (see discussion in text).

D.C. TANNER ETAL.

284 NW

Current

KTB

/

~

i ~~__

- -

- ------~-'6000m-S- /~- ) ~ . - ' - - - -~' "" -

..........

_

1

. ....

i

~ .....

KTB,]~-~

_

".....

_ -- - i i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

........

"-- .........

__..jr ~

SE

Current

v

.... Detachment

%'. _,<

....

= .... km

~.e~

,,0 /

Fig. 8. NW-SE section through the ZEV after various increments of retro-deformation (see Fig. 2 for location). Bold lines represent the traces of the active faults; the projected position of the detachment (although not intersected in this section) is shown as a dashed line. The position of the present topography is displayed after various amounts of hanging-wall heave (alternating finely dashed and continuous lines). Numbers give the amount of hanging-wall heave. The lines KTB and KTB' represent the section of the KTB borehole above the Frankonian Lineament before and after 7 km hanging-wall heave, respectively. The bold arrow illustrates the position of the 30~ change in strike of the Frankonian Lineament (see discussion in text).

wavelength, <0.5 km amplitude; Fig. 8) which migrate northwards during the retro-deformation. This indicates the migration of steep zones of high strain (Fig. 7), which may have been manifested as short thrusting events. In a north-south cross-section (Fig. 9), i.e. parallel to the movement vector, the irregular concave shape of the hanging-wall surface after retro-deformation can be seen. This is mainly caused by the along strike topography of the Frankonian Lineament. It should be noted that the KTB borehole (Figs 8 and 9) has not suffered more than 5 ~ body rotation in any direction (as required by palaeomagnetic data, see above; De Wall et al. 1995). This is a consequence of using antithetic inclined shear.

Discu s sio n o f the results The model will now be discussed in terms of the deformation event itself, i.e. as transpressive 'forward' deformation.

Tectonic thickening and uplift o f the Z E V Figure 9 shows that 7.0 km heave of the hanging wall caused between 2.9 and 3.9 km uplift of the KTB section during the Upper Cretaceous event. This is in good agreement with uplift values from the sphene fission-track analysis (see Fig. 5; Jacobs et al. 1993; Coyle & Wagner 1994). It should be noted, however, that these values are only for the KTB position in the ZEV; other areas suffered typically more or less uplift (Fig. 8). Unfortunately, apatite fission-track ages (see above), which were measured at various locations throughout the ZEV, are of limited use because they give apparent ages from 48 to

150 Ma (Wagner et al. 1991); that is, before and after the Upper Cretaceous event. The difference in uplift is caused by internal thickening of the hanging wall from 6 km to 7 km, a vertical elongation of c. 16%. In total, the whole KTB section is thickened from 6 to 9.1 km. In this model, this occurs by inclined shear sub-parallel to the foliation. This is probably the cause of the large abundance of cataclastic structures parallel to the foliation found in the KTB borehole (see above). The data for the KTB section from this model agree with those of Duyster et al. (1994) and Zulauf et al. (1994), but do not reproduce the complete 7 km thickening of the whole Z E V upper crust as suggested by those workers. This suggests that the model requires additional movement on faults parallel to the Frankonian Lineament such as the Fichtelnaab fault (Fig. 2), but with the same oblique movement vectors as used here and not the dip-slip model of Duyster et al. (1994). Our modelling also suggests major (of the order of several kilometres) movement on the Fichtelnaab Fault, because the hanging wall is offset at the position of the Fichtelnaab Fault (Fig. 6b). The fault is produced as a geometrical accommodation of hanging-wall deformation above the sharp bend at depth of the Frankonian Lineament where it meets the detachment.

Implications f o r the pre- Upper Cretaceous Z E V geometry Remarkably, the deformation of the hanging wall is generally limited to the ZEV, except for local deformation along the Luhe Lineament. The Moldanubian s.s. was virtually unaffected by the Upper Cretaceous event. Variscan fabrics

THREE-DIMENSIONAL RETRO-MODELLING OF TRANSPRESSION

285

Fig. 9. Section through the ZEV, parallel to the hanging-wall displacement vector, after 7 km northwardsdirected heave. The hanging wall is shaded grey. The hanging-wall section of the current KTB borehole profile (KTB) is displaced to KTB'. Figures on the right-hand side give the thicknesses of the KTB well section, above the hanging wall, before and after retro-deformation, and the calculated amounts of uplift at the base and top of the KTB hanging-wall profile as a result of the Upper Cretaceous deformation.

in the northern Z E V were not subjected to body rotation and were variably uplifted by between 2 and 6 kin. In contrast, fabrics at the very southern end of the Z E V were rotated through as much as 50 ~ (towards the SE; see Fig. 8) as a result of being effectively orthogonally thrust over the Luhe Lineament. As the Luhe Lineament is very steep (c. 70~ the result was very localized uplift of more than 6 km of rocks in this area. This may be the reason for the occurrence of the exotic Wetzeldorf Series (see above) next to the Luhe Lineament (Fig. 2), assuming these rocks once lay at the base of the Z E V (see Weber et al. 1993). Until now the position of these southern Z E V lithologies was though to be due to nappe e m p l a c e m e n t of the Z E V during the Variscan Orogeny (Weber et al. 1993; Lapp 1996). The difference in K - A r hornblende ages in the western and eastern Z E V (Scht~ssler et al. 1986) may also be due to the rollover effect of the Upper Cretaceous deformation.

Conclusions As a result of the U p p e r Cretaceous deformation, the Z E V was transpressively moved southwards over a linked fault system consisting of two orthogonal faults, the Luhe and Frankonian Lineaments, and a detachment at 9.5 km. In this modelling, the reverse kinematics was applied to reproduce the pre-deformation situation. Hanging-wall deformation was accommodated by antithetic (with respect to the Frankonian Lineament) inclined shear, Seven kilometres heave of the hanging wall fitted the fission-track, p a l a e o m a g n e t i c and petrological evidence from the KTB borehole best. Calculated uplifts for the U p p e r Cretaceous event are between 2.9 and 3.9 km close

to the KTB, without significant body rotation, a consequence of antithetic hanging-wall shearing. Hanging-wall deformation was complex, as rollover on the Frankonian Lineament was combined with drag-folding along the Luhe Lineament. Intense strain is recorded in this model around the latter structure. The single most important control on the vertical movement of the hanging wall was the 30 ~ change in the strike of the Frankonian Lineament. As a result of the hanging-wall deformation, the area west of the granites was uplifted between 2 and 6 km, but the area east of the granites was uplifted less than 2 km. U p p e r Cretaceous d e f o r m a t i o n along the western border of the Bohemian Massif was confined almost exclusively to the ZEV; the Moldanubian s.s. was affected only along the Luhe Lineament. The edge of the hanging-wall deformation is marked by the Fichtelnaab Fault and the Z E V - M o l d a n u b i a n boundary. Local, intense uplift along the Luhe Lineament may have brought deeper Z E V lithologies, such as the Wetzeldorf Series, to the surface. This work was funded by the Deutsche Forschungsgemeinschaft, Project Be 1041/9. Thanks are due to G. Zulauf and W. Franke for discussions, and to F. Neubauer and an anonymous reviewer for thorough reviews. Midland Valley Exploration Ltd provided excellent support for their software program 3DMove. We also thank R. Elsfisser, M. Persaud, S. Preiss, S. Stiasny and B. Zehner for their assistance and competence with 3D data.

References BAYERISCHES GEOLOGISCHESLANDESAMT1981. Geologische Karte yon Bayern. 1.'500 000. Bayersiches

Geologisches Landesamt, Miinchen.

286

D.C. T A N N E R E T A L .

BEHRMANN,J. H. & TANNER,D. C. 1997. Carboniferous tectonics of the Variscan basement collage in Eastern Bavaria and western Bohemia. Geologische Rundschau, 586, 15-27. BLOMEL, P. 1984. Mitteldruck- und Niederdruckmetamorphose in den auBeralpinen Varisziden Mitteleuropas. Fortschritte der Mineralogie, 62(1), 28-29. 1990. Metamorphism along the SW margin of the Bohemian Massif- an overview. In: FRANK~, W. (ed.) IGCP Terranes in the Circum-Atlantic Paleozoic Orogens, Conference on Paleozoic Orogens in Central Europe - Geology and Geophysics, GOttingen-Gieflen 1990. FieM Guide to Post Conference Excursion: Bohemian Massif. Giegen, 65-73. --, GRAFEN, B., KLEEMANN, U. & SCHREYER, W. 1988. Petrologic, Bohrkernmaterial, Oberpfalz. KTB Report, 88(10), 49. BOSUM, W., CARSTEN, U., FIEBERG, F., et al. 1993. Gravity and magnetic structural models of the KTB area. KTB Report, 93(2), 319-322. , CARSTEN, U., HEYDE, I. & Rorlq'GER, B. 1994. Integrated interpretation of gravity and magnetic data of the KTB main well and the surrounding area. KTB Report, 94(2), A143-A157. COYLE, D. A. & WAGNER,G. A. 1994. Fission-track investigations on sphene from the KTB Deep Drilling Project (Germany): post permian cooling history and in situ annealing. KTB Report, 94(2), A63-A70. CROMWELL,J. A. 1974. Origin of Late Cenozoic basins in Southern California. In: DICKENSON,W.R. (ed.) Tectonics and Sedimentation. Society of Economic Paleontologists' and Mineralogists, Special Publications, 22, 190-204. DAHLSTROM, C. O. A. 1969. Balanced cross sections. Canadian Journal of Earth Sciences, 6, 743-757. D E K O R P RESEARCH GROUP 1988. Results of the D E K O R P 4/KTB Oberpfalz deep seismic reflection investigations. Journal of Geophysics, 62, 69-101. DE WALL, H., WORM, H.-U. & KEYSSNER,S. 1995. Tectonic implications for the German Continental Deep Drilling Program (KTB) derived from paleomagnetic studies on lamprophyres. Zentralblatt fiir Geologic und Paliiontologie, Teil I, 1994(5-6), 523-527. DILL, H. 1990. Stratigraphie und Lithologie des Permokarbons im Weidener Becken (NOBayern, B. R. Deutschland). Zeitschrift der Deutschen Geologischen Gesellschaft, 141, 31-47. DURBAUM, H.-J., REICHERT,C. ~; BRAM, K. (eds) 1990. Dekorp Report. Integrated seismics Oberpfalz 1989. Longterm logging and testing programme of the KTB-Oberpfalz VB. KTB Report, 90(6b). --, SADOWIAK, P. • BRAM, K. (eds) 1992. Dekorp Report. Integrated seismics Oberpfalz 1989. Data evaluation and interpretation as of October 1992. KTB Report, 92(5). DUYSTER, J., DE WALL, H. & ZULAUF,G. 1994. 9101 m profile in the Superdeep Well KTB. Where are the depth dependent gradients? KTB Report, 94(2), A15-A24.

-

-

--,

GRAWINKEL, A. & KONTNY, A. 1995. Petrographic and structural characterization. KTB Report, 95(2), B1-B65. EMMERMANN, R. d~; WOHLENBERG, J. (eds) 1989. The German Continental Deep Drilling Program (KTB) Site-section Studies in the Oberpfalz and Schwarzwald. Springer-Verlag, Berlin. FRANKE, W. 1989. Tectonostratigraphic Units' in the Variscan Belt of Europe. Geological Society of America, Special Papers, 230, 67-90. 1990. Tectonostratigrapic units in the W part of the Bohemian M a s s i f - an introduction. In: FRANKE, W. (ed.) IGCP Terranes in the CircumAtlantic Paleozoic Orogens, Conference on Paleozoic Orogens in Central Europe - Geology and Geophysics, GOttingen-Gieflen 1990. Field Guide to Post Conference Excursion: Bohemian Massif. Giessen, 1-23. HARMS, U. & HOLZL, S. 1994. Lamprophyres from the KTB: petrogenetic implications fiom whole-rock geochemistry and Sr and Nd isotopes. KTB Report, 94(3), 165-177. HAUGE, T. A. & GRAY, G. G. 1996. A critique of techniques for modelling normal-fault and rollover geometries. In: BUCHANAN,P. G. & NIEUWLAND, D. A, (eds) Modern Developments in Structural Interpretation, Validation and Modelling. Geological Society, London, Special Publications, 99, 89-99. HEINICKE, E, FANGER, R., FRANKE, W., et al. 1986. Oberpfalz, Strukturgeologie IIL 2nd KTB Colloquium, Seeheim, Odenwald, 4. HENJES-KUNST, E, HOHNDORF, A. & KREUZER, H. 1994. K/Ar and Ar/Ar dating on minerals of the KTB. KTB Report, 94(2), B33. HIRSCHMANN, G. 1992. On the geological interpretation of the 3-d seismic data with special regard to the information from the KTB boreholes. KTB Report, 92(5), 351-373, - - , STETrNER, G. & ROHRMOLLER,J. 1994. The lithological units of the ZEV. KTB Report, 94(3), 63-74. HOSSACK, J. R. 1979. The use of balanced crosssections in the calculation of orogenic contraction: a review. Journal of the Geological Society, London, 136, 705-711. JACOBS, J., HEJL, E., VAN DEN HAUTE, P. & WAGNER, G. A. 1993. A preliminary tectono-thermal model for the KTB deduced from apatite, zircon und sphene fission-track analysis. KTB Report, 93(2), 125-128. KLARE, B. & SCHRODER,B. 1990. Extreme Randfazies der Oberkreide vor der ZEV. KTB Report, 90(4), 300-309. KLEEIVlANN,U. 1992. Petrologische Hinweise far einen tektonischen Versatz von Pal~iogeothermen am Westrand der BGhmischen Masse (NE Bayern). Frankfurter geowissenschafiliche Arbeiten, A l l , 82-85. KREUZER, H., HENJES-KUNST, E, SEIDEL, E., SCHUSSLER, U. & BUHN, B. 1993. A r - A r spectra on minerals from the KTB and related mediumpressure units. KTB Report, 93(2), 133-136. LAPP, M. 1996. Strukturelle Entwicklung im weiteren -

-

T H R E E - D I M E N S I O N A L R E T R O - M O D E L L I N G OF TRANSPRESSION

287

Umfeld der Bohrlokation der Kontinentalen Tiefbohrung (KTB). PhD thesis, University of

SIMON,M. & GEBRANDE, H. 1994a. Neue Seismographien aus dem KTB-Umfeld. KTB Report, 94(2),

GGttingen. MEISSNER, R. & BORTFELD, R. K. (eds) 1990. DEKORP-Atlas. Springer-Verlag, Berlin. NADAI, A. 1963. Theory of Flow and Fracture of Solids. McGraw-Hill, New York. PETEREK, A. 1994. Beitrag zur sp~it- und post-varistischen tektonischen Entwicklung am Westrand der BOhmischen Masse. KTB Report, 94(2), B l l . & SCIqRODER,B. 1997. Neogene fault activity and morphogenesis in the basement area of the KTB drill site (Fichtelgebirge and Steinwald). Geologische Rundschau, 86, 185-190. , RAUCHE, H., SCHRODER, g., FRANZKE, H. J., BANKWlTZ, P. & BANKWlTZ, E. 1994. Sp~itmesozoisch-k~inozoische tektonische Entwicklung am SW-Rand der BGhmischen Masse. GOttinger Arbeiten zur Geologie und Paliiontologie, Sbl, 174-176. REINHARDT, J. 1992. Klomplexe Element-Zonierungen in Granaten aus amphibolitfaziellen Gneisen der KTB-Bohrung. Berichte der Deutsche Miner-

B52. & -1994b. New seismic images of the Earth's crust: migration before stack. KTB Report, 94(2), A87-A96. STEIN, E. & KLEEMANN, U. 1992. Evidence for Late Variscan emplacement of the ZEV. KTB Report, 92(4), 133-146. STETI'NER, G. 1990. Geologische Karte des KTBUmfeldes Oberpfalz. 1:10 000. Bundesanstalt ftir Geowissenschaften und Rohstoffe und Bayerisches Geologisches Landesamt, Hannover. VOLL, G. 1960. Stoff, Bau und Alter in der Grenzzone Moldanubikum/Saxothuringikum in Bayern unter besonderer Berticksichtigung gabbroider, amphibolitischer und kalksilikatf0hrender Gesteine. Beihefte zum Geologischen Jahrbuch, 42, 382. WAGNER, G. A., HEJL, E., VAN DEN HAUTE, P. & VERCOUTERE, C. 1991. Spaltspurenuntersuchungen am Kern der KTB-Vorbohrung und an Umfeldgesteinen. KTB Report, 91(1), 259-268. WEBER, K. 1992. Die tektonische Position der KTBLokation. KTB Report, 92(4), 103-132. - - , LAPP, M. & ONCKEN, O. 1993. The basal units of the Erbendorf-Vohenstrauf3 Zone - attempted prognosis for the suspected lithology at the final depth of the KTB. KTB Report, 93(2), 87-88. WEMMER, K. • AHRENDT, H. 1992. Geochronologische Erfassung yon retrograden Prozessen in Gesteinen der KTB-Vorbohrung mit Hilfe der K-Ar-Methode. KTB Report, 92(4), 349-372. & -1994. Age determinations on retrograde processes and investigations on the blocking conditions on isotope systems of KTB rocks. KTB Report, 94(2), B32. WENDT, I. 1993. Geochemistry, radiometric ages and genetic relations of granites adjacent to KTB. KTB Report, 93-2, 407-410. WIEDERHOLT, H. 1992. Interpretation of envelopestacked 3-D seismic data and its migration another approach. KTB Report, 92(5), 67-113. WORM, H.-U. & ROLF, C. 1994. Remanent magnetisation of the KTB drill cores. Scientific Drilling, 4, 185-196. ZULAUF, G. 1993. Brittle deformation events at the western border of the Bohemian Massif (Germany). Geologische Rundschau, 82, 489-504. -~r DE WALL, H. 1994. Late to post-Variscan tectonometamorphic evolution of the KTB rock suite: depth dependent variation of fault and vein mineralisations. KTB Report, 94(2), B4. -DUYSTER, J. & DE WALL, H. 1994. Late to postVariscan tectonometamorphic evolution of the KTB rock site. The preliminary model: dramatic stacking within the supracrustal level. KTB Report, 94(2), B3.

alogischen Gesellschafi, Beihefl zum European Journal of Mineralogy, 4(1), 224. , KLEEMANN, U., BLI)MEL, P. & SCHREYER, W. 1989. Geothermobarometry of metapelites as a key to the pressure and temperature history of the Z E V (Zone von Erbendorf-Vohenstrauss) NE Bavaria. KTB Report, 89(3), 24-32. ROHR, C. & ZULAUF, G. 1992. Zur tektonometamorphen Entwicklung der Erbendorfer Grtinschiefer zone -abgeleitet aus petrologischen, geothermobarometrischen und strukturgeologischen Untersuchungen. KTB Report, 92(4), 181-212. SCHOPS, D. & FRIEDmCH, G. 1994. Hydrothermale Phasenbeziehungen in Scherzonen: die Paragenese Pyrit-Pyrrhotin-Graphit-Chlorit. KTB Report, 94(2), B20. SCURODER, B. 1975. Die geologische Entwicklung des Vorlandes der Oberpfalz. Der AufschluJ3, 26, 277-288. 1978. Friinkische Schweiz und Vorland. Borntraeger, Berlin. 1987. Inversion tectonics along the western margin of the Bohemian Massif. Tectonophysics, 137, 93-100. 1988. Outline of the Permo-Carboniferous basin at the western margin of the Bohemian Massif.

Zeitschrifi fiir geologische Wissenschaften, 16, 993-1001. 1992. Post-Hercynian fault block activities in the basement area near KTB-Drilling Site. KTB Report, 92(4), 287-294. SCHUSSLER,g., OPPERMANN,g., KREUZER, H., SEIDEL, E., OKRUSCH, M., LENZ, K.-L. & RASCHKA, H. 1986. Zur Altersstellung des ostbayerischen Kristallins, Ergebnisse neuer K-Ar-Datierungen. Geologica Bavarica, 89, 21-47.

-

-

Development of kinematic partitioning within a pure-shear dominated dextral transpression zone: the southern Ellsworth Mountains, Antarctica M. L. C U R T I S

Britbsh Antarctic Survey, N E R C , High Cross, Madingley Road, Cambridge CB3 0ET, UK

Abstract: The Heritage Range of the southern Ellsworth Mountains, West Antarctica, is composed of Cambrian to Permian sedimentary and volcanic rocks, which were deformed during the Permo-Triassic Gondwanian orogeny. The structural grain of the Heritage Range exhibits a previously unrecognized 18~swing from 333-153 ~in the north to 315-135 ~ in the south. A change in structural and kinematic style accompanies this strike-swing, with a structure consistent with near-orthogonal shortening present in the south, and a structural style consistent with dextral transpression within the central and northern Heritage Range. Kinematic partitioning is present within the central Heritage Range, where strikeparallel, contemporaneous domains of dextral and reverse shear have developed simultaneously with the regional cleavage. Comparison of the structure and kinematics within both structural domains suggests that the central and northern Heritage Range experienced pure-shear dominated dextral transpression, with an approximate angle of relative shortening (o0 of 65-70 ~ Results derived by integrating field data into a published kinematic partitioning model suggest relatively efficient kinematic partitioning has occurred. However, such efficient partitioning cannot be explained by strain partitioning models based purely on plate boundary conditions. Therefore, it is proposed that pre-existing weak structures were present within the unexposed basement facilitating the apparent high percentage of kinematic partitioning. Kinematic partitioning of the simple shear and pure shear components of oblique convergence is a c o m m o n p h e n o m e n o n in transpressive orogens. In both ancient and neotectonic orogenic belts, strain is invariably partitioned into strike-slip and thrust faults, striking parallel to the trend of the orogen (Harland 1971; Mount & Suppe 1987; Zoback et al. 1987; Jackson 1992; Strachan et al. 1992; Tikoff & Teyssier 1994; Jones & Tanner 1995; Teyssier et al. 1995; Goodwin & Williams 1996; among others). Partitioning of transpressive strains can result in complex strain patterns, with significant differences in finite and incremental strain directions (Robin & Cruden 1994; Kirkwood et aL 1995; Tikoff & Greene 1997). Modelling the complex results of kinematic partitioning has been tackled from two contrasting directions; either stress or strain is partitioned. Central to the argument of stress partitioning is the presence of a pre-existing weak fault zone (Mount & Suppe 1987; Zoback & Healy 1992) or weak lithological and rheological boundary (Jones & Tanner 1995) within a strong crust. It is argued that it is the properties of the fault zone or other rheological anisotropies which control the degree of kinematic partitioning. The alternative strain partitioned model proposes that the relative angle of oblique convergence controls the efficiency of kinematic partitioning and makes

no assumptions about crustal rheology (Tikoff & Teyssier 1994). This paper documents the structure and kinematics of a transpressive segment of the late Palaeozoic to early Mesozoic Gondwanian fold belt, exposed in the Ellsworth Mountains of West Antarctica (Curtis 1994, 1997). Detailed structural relationships are used to establish the transpressive nature of the deformation, and to examine the possible factors affecting kinematic partitioning within this segment of the orogen. The Ellsworth Mountains form a N N E - S S W trending range of mountains, over 500 km in length, centred at 79~ 84~ (Fig. 1). The mountains are geographically divided into two ranges; the northerly Sentinel Range, which is characterized by high alpine peaks (up to 4897 m) with large vertical relief, and the southerly Heritage Range, formed by lower relief, less rugged peaks up to 2640 m. The Heritage Range provides large, generally accessible, exposures over an area of 180 km by 60 km. Data presented in the following sections were collected during two field seasons in 1993-1994 and 1995-1996, and are primarily from the Heritage Range.

Geological setting The Ellsworth Mountains form part of the Ellsworth-Whitmore Mountains block (EWM),

CURTIS,M. L. 1998. Development of kinematic partitioning within a pure-shear dominated dextral transpression zone: the southern Ellsworth Mountains, Antarctica. In: HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. E (eds) 1998. Continental Transpressional and Transtensional Tectonics. Geological Society, London, Special Publications, 135, 289-306.

289

290

M.L. CURTIS a single orogenic event during the Gondwanian orogeny. Curtis (1994, 1997) identified a strikeparallel, contemporaneous component of dextral shear within the Gondwanian age deformation of the northern Heritage Range, and interpreted the Permo-Triassic Gondwanian structures as having been formed by pureshear dominated dextral transpression. Stratigraphy o f the Heritage R a n g e

Fig. 1. Geological map of the Heritage Range, southern Ellsworth Mountains, highlighting the basic stratigraphy distribution and structure of the range: W.H.F., Watlack Hills fault: M.D.F., Mount Dolence fault. Inset map: location of the Ellsworth Mountains within West Antarctica. one of five allochthonous crustal blocks forming West Antarctica (Dalziel & Elliot 1982; Storey et al. 1988). Stratigraphic, structural, and palaeomagnetic studies suggest the EWM block has been rotated c. 90 ~ anticlockwise, and tectonically transported from a position roughly adjacent to the Falklands Plateau and the Coats Land coast of East Antarctica, during the Mesozoic break-up of Gondwana (Schopf 1969; Clarkson & Brook 1977; Watts & Bramall 1981; Grunow et al. 1987; Dalziel & Grunow 1992; Lawver et aL 1992; Grunow 1993; Dalziel et al. 1994; Curtis & Storey 1996). Before supercontinent break-up, Andean-style convergence was inferred to have occurred along the palaeoPacific margin of Gondwana during the Permian to early Mesozoic (Smellie 1981; Johnson 1991; Collinson et al. 1994). The resultant Gondwanian orogenic belt was fragmented during breakup to form five fold belts dispersed across the present-day southern continents: the Sierra de la Ventana Fold Belt, and Falkland Islands, South America; the Cape Fold Belt, South Africa; and the Pensacola and Ellsworth-Whitmore Mountains, Antarctica. Previous geological studies (Craddock 1972; Craddock et al. 1992; SpOrli & Craddock 1992; Webers et al. 1992a, b; Curtis 1997) concluded that the Ellsworth Mountains were deformed by

The Heritage Range is composed of Middle-Upper Cambrian sedimentary and volcanic rocks of the Heritage Group, which are overlain with local disconformity (Goldstrand et al. 1994) by a dominantly quartzite sequence of Upper Cambrian to Permian age, the Crashsite Group (Webers et al. 1992b) (Fig. 2). The lowermost formations of the Heritage Group (the Union Glacier, Drake Icefall, Hyde Glacier, and Conglomerate Ridge formations), crop out in the Edson Hills and southern Soholt Peaks, where highly sheared tuffaceous diamictite, greywacke, conglomerate, marble, and argillite are exposed. Most rock exposure in the northern Heritage Range reveals intensely folded argillite and greywacke of the Springer Peak and Frazier Ridge formations, which are in turn overlain by the laterally discontinuous limestones of the Minaret Formation. The 3000 m thick Crashsite Group is predominantly formed of quartzites with some argillite of Late Cambrian to Devonian age (Sp6rli 1992). Structure o f Heritage R a n g e Sp6rli & Craddock (1992) and Curtis (1997) gave detailed accounts of the structure of the Heritage Range, and along with others recognized only a single phase of compressive deformation. However, following recent fieldwork (1995-1996) a locally developed set of structures was discovered forming an early deformation phase within a progressive Gondwanian orogenic event. In recognition of this fact, the main cleavage is now denoted $2. The following sections outline only the structure and kinematics of the main Gondwanian deformation event (D2) within the Heritage Range. Neither the early Gondwanian D1, nor post-Gondwanian D3 structures will be described. Folding. Several fold orders are present, with first-order folds possessing wavelengths of 20-25 km. Lower-order folds are generally upright to inclined, close to tight, with fold axes plunging gently about the horizontal, although a regional plunge toward the NNE is suggested in

KINEMATIC PARTITIONING IN SOUTHERN ELLSWORTH MOUNTAINS

Fig. 2. Stratigraphic section of rocks exposed in the Ellsworth Mountains (after Webers et al. 1992b). KPM, Kosco Peak Member; UGF, Union Glacier Formation; HGF, Hyde Glacier Formation; DIF, Drake Icefall Formation; CRF, Conglomerate Ridge Formation; FRF, Frazier Ridge Formation; SPF, Springer Peak Formation; LHF, Liberty Hills Formation; MF, Minaret Formation.

the data sets of Fig. 3, and the regional geological map of the Ellsworth Mountains (Craddock et al. 1986). Fold styles are largely lithology dependent, with Ramsay (1967) class 1C, 2, and 3 folds within the predominantly argillaceous Heritage Group, and chevron folding within the wellbedded quartzite succession of the Crashsite Group. Sheath folds and rootless intrafolial folds are common in areas of high strain, especially in the Minaret Formation of the southern Heritage Range and the carbonate and shale lithologies of Soholt Peaks. At the northern end of Soholt Peaks, steeply plunging third- and second-order folds possess an overall asymmetry consistent with their formation in a dextral shear domain (Curtis 1997). This zone of steep, NNW-plunging folds is expressed in Fig.

291

Fig. 3. Stereograms of contoured structural data sets from the Heritage Range.

3e, where these structures contribute to the asymmetrical, NNW-plunging distribution of F2 fold axes. Cleavage. A well-developed axial-planar cleavage (denoted $2) is associated with the folds (Fig. 3b), forming a slaty cleavage in the argillaceous sequences of the Heritage Group, and a spaced rough cleavage, with weak to strong microlithon orientation within the quartzites of the Crashsite Group. A well-developed shape fabric defines the cleavage within the intensely cleaved and sheared diamictites and conglomerates of the lower Heritage Group, exposed in the Edson Hills. Qualitative strain analysis reveals a generally oblate strain throughout the Heritage Range, with a mean k value of 0.59 (Curtis 1997). Lineations. Mineral stretching lineations, microboudinage, pressure shadows, and shape fabrics lie within the $2 cleavage plane defining the regional lineation, denoted L2. The mean L2 stretching lineation is oriented 71~

292

M.L. CURTIS

N

Kinematic partitioning within the southern Soholt Peaks Basic structure o f the southern SohoIt Peaks

Fig. 4. Hoeppener plot of predominantly mesoscale main Gondwanian age thrust faults from the central and northern Heritage Range. Mean slip direction of 240-O60~.

indicating roughly down-dip elongation within the cleavage plane. However, oblique to strikeparallel L2 lineations are common at southern Soholt Peaks, and commonly occur in discrete zones of strike-parallel, non-coaxial dextral shear, up to 450 m wide. The oblique to strikeparallel L2 lineations display an asymmetrical distribution plunging consistently toward the south or SW (Fig. 3d).

Faulting. Two significant faults striking NNW-SSE are present in the Heritage Range; the steep to sub-vertical Watlack Hills fault and the variably dipping Mount Dolence fault (Fig. 1). Stratigraphic relationships suggest that they were reverse faults before their folding during main Gondwanian deformation (Sp6rli & Craddock 1992). Dextral strike-slip reactivation along the Watlack Hills fault has occurred, although the lack of significant cataclastic or mineralized zones suggests the magnitude of dextral displacement is small. Thrust faulting, although present, is not particularly abundant within the Heritage Range. Figure 4 displays fault slip data for main Gondwanian age meso- and macroscale thrusts, which exhibit a mean slip direction oriented 240-060 ~ These data indicate that the minimum instantaneous strain direction, represented by the predominantly mesoscale thrust faults, is noncoaxial with the finite shortening direction across the range (243-063~

The southern Soholt Peaks expose a highly variable lithological succession of Middle Cambrian to lower Upper Cambrian age rocks, the oldest in the Ellsworth Mountains. The succession has been subdivided by Webers et al. (1992a, b) into four apparently conformable formations; the Union Glacier, Drake Icefall, Conglomerate Ridge, and Springer Peak. However, recent mapping reveals that the contacts between those formations are in fact tectonic, some of which display structural discontinuity (Fig. 5a and b). The oldest rocks, the Union Glacier Formation (UGF), are predominantly formed of highly sheared dark green, tuffaceous diamictites, with a subordinate component of conglomerates and sandstones. A strong L-S fabric, defined by stretched and flattened clasts within the diamictite, is present within the UGF. The U G F is locally overlain by 200-500 m of highly sheared, recrystallized white limestone possessing a strong L-S fabric and sheath folds. This recrystallized limestone marks the base of the Drake Icefall Formation (DIF), which is predominantly formed of black shales within interbedded highly sheared limestones. The L2 stretching lineation within the limestones and the underlying U G F is markedly oblique, indicating a reverse-dextral shear sense along this boundary (Fig. 5b). Within the immediate hanging wall of the sheared basal limestone body, black shales form a tight, NE-facing anticline, which is recumbent relative to the underlying limestone contact, indicating the presence of a tectonic detachment along this contact. The rest of the DIF is characterized by upright, tight to isoclinal, NE-verging folds. In contrast to the basal limestone, lineations developed along highly sheared limestone units within the shales display a dip-slip stretching direction (Fig. 5b). The top of the DIF is marked by a thin (<5 m) recrystallized limestone, dipping steeply toward the SW. The contact with the structurally overlying Conglomerate Ridge Formation (CRF) is tectonic, with a reverse fault lying immediately above, and parallel to, bedding within the DIF. The presence of this fault zone is betrayed by a localized, NE-verging, hanging-wall anticline within the coarse-grained siliciclastic rocks of the basal CRF. The hinge zone of this anticline is encountered c. 200 m SW of the fault, beyond which point bedding within the CRF assumes a consistent, steep, SW dip. Clast size rapidly

KINEMATIC PARTITIONING IN SOUTHERN ELLSWORTH MOUNTAINS increases up-section, with conglomerates dominating the 1100 m thickness of the formation. The top of the CRF is marked by a 90 m thick, highly sheared, recrystallized white limestone, displaying a pronounced L-S fabric. Strain is relatively low along the limestone-conglomerate contact, revealing their relationship to be conformable. The structural section through the southern Soholt Peaks is completed by generally NE-dipping interbedded argillite and sandstone of the Springer Peak Formation (SPF). Bedding becomes overturned and increasingly folded and sheared with proximity to the highly sheared limestone. This boundary marks a significant change in vergence direction, from NE vergence within the CRF and underlying formations, to SW vergence within the SPF. Therefore, the sheared limestones represent a significant tectonic discontinuity. Although tectonic contacts are present between the formations, there is little evidence to suggest that the stratigraphic order has been significantly disrupted or tectonically interleaved, as the dated formations, SPF and DIF (Buggisch et al. 1992; Jago & Webers 1992; Shergold et al. 1992; Webers et al. 1992a, b), are in the correct stratigraphic order. However, these tectonic contacts do offer an alternative explanation for the lateral wedging of the CRF and Hyde Glacier Formation (not exposed in the southern Soholt Peaks) outcrops, which are currently proposed to be rapid lateral facies changes (Webers et al. 1992a), but instead may be the result of tectonic juxtaposition. It is suggested that the presence of reverse shear zones along the lithological boundaries of an apparently undisrupted stratigraphic succession is the result of rheological contrasts within the heterolithic rock sequence. The deformation style of the individual formations reflects this contrast in theology. Bedding within the competent U G F and CRF generally forms steep, uniformly dipping sequences, whereas the DIF is intensely folded (Fig. 5b). Such disharmonic deformation must result in the development of ddcollements and shear along the Iithological boundaries. A similar explanation is proposed for the tectonic contact between the Springer Peak and Conglomerate Ridge formations. Kinematic partitioning

A steep to vertically inclined $2 cleavage is present in all lithologies within the southern Soholt Peaks, with an L2 stretching lineation commonly developed along it. This lineation is defined by many structures including shape fabrics, mineral alignments, pressure shadows,

293

plus meso- and microscale boudinage. The mean L2 lineation within the Heritage Range displays a slight obliquity to the structural grain, which is locally highly oblique, such as that developed along the contact between the Drake Icefall and Union Glacier formations. However, in the SW Soholt Peaks, L2 lineations display a complete spectrum of orientations relative to $2, from strike-parallel, to down-dip. The strike-parallel lineations generally possess a component of SE plunge (Fig. 3d) and are found in spatially discrete zones of dextral shear (Fig. 5a), whereas down-dip lineations are associated with either zones of reverse non-coaxial shear, or coaxial shortening orthogonal to strike. The zones of non-coaxial shear do not display an appreciable increase in strain relative to adjacent rocks, therefore they will be referred to as shear domains, as opposed to shear zones senso stricto (Ramsay & Huber 1983). Two dextral shear domains striking N N W - S S E (parallel to the structural grain of the Heritage Range) are exposed in the outcrop of the Conglomerate Ridge Formation (Fig. 5a). The domains vary in width from c. 250 to 450 m, and possibly anastomose slightly along strike, incorporating screens of rock exhibiting no strike-parallel shear. Stretching lineations are defined by the alignment of high-axial-ratio conglomerate cobbles, pressure shadows, mineral stretching lineations, and the micro-boudinage and elongation of competent clasts. A graph of foliation dip against lineation pitch (Fig. 6a) reveals that strike-parallel stretching lineations form exclusively within zones of steeply inclined foliation, whereas oblique and down-dip lineations may be present on cleavage planes of any inclination. The strike of the foliation plane containing the strike-parallel lineation shows a skewed distribution toward a more southerly strike, rather than a symmetrical distribution about the mean cleavage orientation of 153 ~ (Fig. 6b). This trend probably reflects the anastomosing nature of the cleavage within the conglomerates, and the development of C-S fabrics and shear bands within the domain. Shear sense criteria, such as asymmetrical 'o--style' tails and pressure shadows developed on competent porphyroclasts, shear bands, C-S fabrics, and asymmetrical boudinage, display a consistent dextral sense of shear (Fig. 7a-c). The most common shear sense indicator within the dextral domains is the asymmetrical boudinage of competent quartzite cobbles within the conglomerates. Although these structures can provide a potentially ambiguous sense D e x t r a l shear d o m a i n s .

294

M.L. CURTIS

Fig. 5. (a) Geological map of southwestern Soholt Peaks, displaying the orientation of dextral and reverse shear domains relative to the regional structure. (b) Composite structural cross-section through southern Soholt Peaks. Line of section indicated in (a). Legend for stereograms" O, poles to cleavage; Q, mineral stretching lineations.

KINEMATIC PARTITIONING IN SOUTHERN ELLSWORTH MOUNTAINS

a

strike-slip

.o~

70: . . . . " . . . . . .

=~

9"

Oblique-slip

i

9

'

Dip-slip

,~

" "" | '

'

cteavageinCline~

~ o ~ 60. _.:-... c: 5 0 ~.~-.

Moderately

inclined cleavage

40-

3o E3

20

Shallowly incllnea

cleavage

10

0

0

10

20

30

40

50

60

70

80

90

Pitch of (1_2)lineation

.. ~ 03

.~ ~ ~--~

90

.

8oi

"

6050

o.~

40

~- ~

30 20

~1

Mean foliation plane

'.---7.q "

ZS

~ 9

ot

l'st~k~-~,p

*

~ . ,,o,,

120

130

%= 9

....

%4"

1:

I

:'.

I*

I 9

:. 9

9

9

.

I 140

:-

' I'~ l~b 15o

9

.

.-

"

1-~0

180

Strike of fo/iation p/ane containing/ineation

Fig. 6. (a) Graph of $2 foliation dip against L2 lineation pitch for the northern and central Heritage Range. (b) Graph of L2 lineation pitch against the strike of foliation plane for the northern and central Heritage Range. of shear (Hanmer 1984; Goldstein 1988; Jordan 1991), detailed observations of the kinematics and evolution of the fractures within these asymmetrical boudinage systems have identified two distinct types which provide a consistent and complementary sense of shear to other locally developed kinematic indicators. Within the SW shear domain (Fig. 5a), shear fracture boudinage (Swanson 1992) is commonly developed. It is characterized by the development of an array of p r e d o m i n a n t l y sub-vertical shear fractures, which are at a clockwise oblique angle to the main cleavage and display dextral offsets. The boudin tiles or lozenges of quartzite formed by the bounding shear fractures are back-rotated in an anticlockwise sense, relative to the overall

295

shear regime. The distribution of fractures is skewed slightly from a mean fracture to cleavage angle of 38 ~ to a maximum of 55 ~ (Fig. 8). Fractures at these higher angles are pure extensional structures, commonly displaying quartz mineralization, which lie parallel to cross-cutting quartz veins in the same exposure. Such spatial and kinematic relationships within a fracture population are entirely consistent with dextral strikeslip deformation. The high angle (>45 ~) between the extensional structures and shear direction suggests the influence of a component of pure shear perpendicular to the shear direction within the domain. It is the presence and intimate relationship of the extensional fractures and veins within this fracture system which allows the shear fracture boudinage to be used confidently as a shear sense indicator. In contrast, the fracture arrays developed in aligned quartzite cobbles in the NE shear domain lie anticlockwise oblique to the regional cleavage, displaying sinistral offsets and clockwise rotation of the boudin tiles (Figs 7c and 8). The fractures form two types distinguished by the complexity of the fracture plane. Smooth, planar fractures, occasionally possessing a 'lazy-Z' sigmoidal curvature, traverse the entire clast, and are referred to as transverse fractures. Short, smooth fractures that do not traverse the clast display horsetail splays on the extensional side of the smooth fracture termination, or alternatively an array of micro-tension gashes emanating from the fracture tip (Fig. 7c). Both these fracture types are sporadically cross-cut by later calcitefilled extensional fractures formed perpendicular to the shear direction. The smooth, planar fractures are interpreted as having formed initially perpendicular to the shear direction, as Mode I tensile fractures, which were subsequently subjected to dextral shear resulting in 'forward-rotation', and antithetic reactivation of the fractures. The transverse tensile fractures were rotated and

296

M.L. CURTIS

Fig. 7. (a-c) Dextral shear sense indicators from the dextral shear domains in southern Soholt Peaks (plan view). (a) Competent quartzite clast displaying asymmetrical wings within northeastern dextral shear domain (note the presence of extensional veins perpendicular to the transport direction). Pencil is 14 cm long. (b) Dextral shear band within the northeastern dextral shear domain. Long edge of notebook is 16 cm. (c) Forward-rotated asymmetrical extensional fracture boudinage cross-cut by late tensile fracture perpendicular to the long axis of the clast (perpendicular to the shear direction). (Note extensional horsetails and micro tension gash arrays at the tips of non-transverse planar fractures). Clast is 18 cm long. (d) Asymmetrical winged rigid quartzite clasts with stair-step geometry indicating reverse shear. Profile view looking NNW; diameter of lens cap is 55 mm. (e) and (f) Superimposed non-coaxial fracture sets within the boundary between dextral and reverse shear domains. (e) Early sub-vertical dextral shear fractures cross-cut by extensional sub-horizontal fractures associated with reverse shear. Profile view looking SE, length of visible pencil is 6 cm. (f) Early extensional fractures associated with down-dip boudinage of quartzite clast, cross-cut by dextral shear fractures, view looking NE onto foliation plane; length of pencil is 14 cm. Examples (e) and (f) are from the southwestern dextral shear domain.

KINEMATIC PARTITIONING IN S O U T H E R N E L L S W O R T H MOUNTAINS

297

298

M.L. CURTIS

reactivated without damage to the fracture wall, whereas the non-transverse fractures display shear fracture propagation from the tip of smooth, originally tensile fractures into intact quartzite forming micro-tension gash arrays or terminal horsetail splays. Strikingly similar fracture relationships were described by Petit (1988) and Petit & Barquins (1988) from experimental shearing of a pre-existing flaw representing a tensile fracture. The late extensional fractures are perpendicular to the shear direction and cross-cut earlier rotated fractures, supporting the

interpretation that the incipient geometry of the rotated fractures was also perpendicular to the strike-parallel shear direction. Malavieille & Lacassin (1988) and to a greater extent Swanson (1992) described fracture systems of similar geometry and kinematic evolution. Swanson (1992) documented repeated tensile fracturing perpendicular to the shear direction during regional dextral shear, and related it to fault-related fluctuations in fluid pressure. A l t h o u g h t h e NE dextral shear domain lies in proximity to a significant reverse

KINEMATIC PARTITIONING IN SOUTHERN ELLSWORTH MOUNTAINS a

N

N=43

~-~

T

South-western dextral shear domain

b

299

Mean foliation

plane

N=85 Mean=114-294

North-eastern dextral shear domain

Fig. 8. Asymmetrical boudinage data from the dextral shear domains of southern Soholt Peaks. (a) Stereogram of poles to fractures within clasts deformed by asymmetrical shear fracture boudinage. (b) Rose diagram of fractures developed by forward-rotated extension fracture boudinage. 0, Poles to cobble fractures; 9 extension veins.

fault, the fracture system does not exhibit multiple generations of tensile fractures. Instead, a model similar to that of Malavieille & Lacassin (1988) is envisaged with initial stretching of the aligned clasts parallel to the shear direction, and formation of tensile fractures perpendicular to it. The tectonic alignment of elongate competent quartzite clasts parallel to the shear direction indicates that stretching along this axis prevailed before brittle deformation. Rotation of the tensile fractures initiated soon after their formation, thus preventing the development of mineralized veins. The sporadic nature of late cross-cutting tensile fractures and veins suggests that the process of tensile fracturing did not occur systematically, as clast elongation parallel to shear direction was accomplished by forward rotation of the boudin tiles. However, the observed examples of cross-cutting fractures may have resulted from the locking up of the antithetic shears after a finite amount of rotation (Nut et al. 1986). This domino style of asymmetrical boudinage was referred to as 'forwardrotated extension fracture boudinage' by Swanson (1992).

Down-dip shear domains. Outside the dextral shear domains, stretching lineations are downdip to slightly oblique within the regional cleavage plane. Again, competent conglomerate cobbles with high axial ratios are aligned parallel to the stretching direction, and frequently

exhibit extensive boudinage. The boudinage is generally characterized by the formation of lowangle extensional veins and fractures striking parallel to the regional structural grain, and perpendicular to the stretching lineation. Elongation of competent clasts parallel to the stretching lineation is less frequently achieved by the development of irrotational conjugate reverse shear fractures. The development of symmetrical pressure shadows adjacent to the conglomerate clasts, and the absence of boudin rotation, indicates that deformation occurred as a result of coaxial shortening across, and extension down the regional cleavage plane. The general coaxial nature of deformation within the regional cleavage is frequently disrupted by strike-parallel domains of non-coaxial reverse shear. Such domains are present adjacent to the boundaries of the dextral shear domains (Fig. 5a). The sense of shear within these domains is primarily indicated by cr-porphyroclasts (Fig. 7d), oblique extensional veining, and forward-rotated extensional fracture boudinage.

Relative chronology o f shear domains Brittle fracture of competent quartzite conglomerate cobbles, either by pure extensional boudinage or asymmetrical boudinage, is the primary deformation style within the Conglomerate Ridge Formation, and conglomerate

300

M.L. CURTIS

Fig. 9. Schematic block diagram summarizing the structure and kinematics of the partitioned shear domains in southern Soholt Peaks.

horizons within other stratigraphic units, and is present within all strain domains. Therefore, if the dextral shear domains represent a temporally distinct deformation event from the regionally predominant down-dip stretching domains, one would expect to find widespread and consistent superimposition and/or disruption of the fracture sets throughout the domains. However, this is clearly not the case, as the fracture populations occur in spatially distinct domains. Superimposition of the strain systems is present only within the boundary zones between dextral and reverse shear domains, where fracture systems from adjacent shear domains cross-cut one another without a consistent order of superimposition (Figs 7e and f, and 9). Therefore, the dextral and reverse shear domains must have been coeval with each other, and must have been contemporaneous with the coaxial deformation associated with the regional cleavage. No clear relationship is exposed between the dextral shear domains and the thrust faults bounding the main lithostratigraphic units of Soholt Peaks. The dextral shear domains are

clearly contemporaneous with main cleavage development, and are found developed within both NE- and SW-dipping fold limbs, suggesting that the partitioned shear domains form part of a progressive deformation event involving thrust initiation with associated folding, followed by cleavage development and kinematic partitioning.

Progressive strain partitioning within domains of distributed shear Variation in the orientation of strain axes is not confined to the boundaries of the partitioned shear domains, it is also present in many localities within areas of distributed deformation. Curved mineral fibres within pressure shadows developed adjacent to competent clasts preserve the progressive change in strain axis. These curved fibres consistently display a progressive change from sub-vertical down-dip extension to late sub-horizontal stretching. This late subhorizontal stretching appears to be associated with dextral shear, because slickenfibres on a

KINEMATIC PARTITIONING IN SOUTHERN ELLSWORTH MOUNTAINS reactivated steeply inclined bedding surface within the Meyer Hills display a dramatic change from reverse shear to dextral shear. Where developed, the progressive change in strain axes consistently indicates that initial shortening across areas of distributed deformation within the northern and central Heritage Range becomes superimposed by fold-axisparallel extension, locally associated with dextral shear. Similar progressive strain partitioning within a transpressive orogen has been documented by Kirkwood et al. (1995).

Transpressional model for the Heritage Range A comparison of the structural geology from the northern and central Heritage Range with that of the southern Heritage Range reveals a consistent and significant swing in the structural grain from a 333-153 ~ trend (northern and central areas) to 315-135 ~ (southern Horseshoe Valley), which coincides with a change in structural style and kinematics (Fig. 10). In the northern and central Heritage Range, folding is predominantly upright with a sub-vertical axial planar cleavage. Stretching lineation data display an asymmetrical distribution from down-dip to strike-parallel, with a consistent southerly directed plunge. The mean down-dip lineation (72~ forms a 09 ~ anticlockwise oblique angle to the finite shortening direction. Strike-parallel, contemporaneous, reverse and dextral shear zones are present within Soholt Peaks, with evidence of isolated strike-parallel dextral shear and associated sub-horizontal extension found in many localities throughout the area. In contrast, the southern Horseshoe Valley is dominated by moderately inclined, NE-verging, close to isoclinal folds with a mean axial planar cleavage of 135/54~ The mean stretching lineation (51~ present along the cleavage plane forms an anticlockwise oblique angle of only 04 ~ with the finite shortening direction. Curtis (1997) proposed a model of pure-shear dominated dextral transpression to explain the structure and kinematics of the northern and central Heritage Range, speculating an angle of relative external shortening (e0 of 70 ~ to the margin of deformation. The 18 ~ anticlockwise strike-swing recognized between the northern and southern Heritage Range presents an opportunity to test this transpressive model, and determine whether the swing in the structural grain results from a relative change in angle e~ because of an along-strike variation in the trend of the deformation zone margin, or

301

later oroclinal bending. The difference in structural relationships and kinematics between the two regions favours the strike-swing forming as an original feature of the orogenic belt and not as a result of orogenic bending, as relative structural relationships would be expected to be maintained within the latter scenario. If an original variation in the strike of the orogen is considered within the model of Curtis (1997), an approximate 20 ~ anticlockwise change in the trend of the deformation margin relative to the external shortening direction would result in the formation of a segment of orthogonal shortening (oL -- 90 ~ within the orogen, which should be reflected in the structure of the southern end of Horseshoe Valley. The structure of this region is indeed distinct from that to the north, with no evidence for along-strike dextral shear domains or sub-horizontal stretching observed. L2 lineations form exclusively down-dip within the moderately inclined cleavage plane and form a consistent low angle of obliquity (04 ~ relative to the finite shortening direction. The moderate inclination of structural fabrics, and consistent NE structural vergence, together with the kinematics are interpreted to reflect homogeneous deformation as a result of virtually orthogonal shortening. A minor component of distributed dextral shear is suggested by the slight obliquity of the L2 lineation, suggesting a _<90 ~ (Fig. l l a ) . The structure and kinematics of the southern Horseshoe Valley are therefore consistent with an original variation in strike of the orogen, forming a localized segment of virtually orthogonal shortening, relative to pure-shear dominated dextral transpression within the northern and central Heritage Range. The proposed external relative shortening direction (o0 of c. 70 ~ for the pure-shear dominated transpression within the northern and central Heritage Range (Curtis 1997) appears to be a good approximation, although the minor obliquity of L2 within the southern Horseshoe Valley suggests that shortening is not perfectly orthogonal within this segment of the orogen. Therefore, oL may be slightly less than 70 ~ within the main Heritage Range (Fig. 11a). As well as supporting the pure-shear dominated model of transpression, the addition of data from the 1995-1996 field season allows the model to be refined. The mean slip direction of the mesoscale main-Gondwanian thrust faults can be constrained more tightly, revealing that they form a mean 03 ~ anticlockwise relationship to the finite shortening direction. These smallscale fault displacements are interpreted to approximate the minimum instantaneous strain

302

M.L. CURTIS

Fig. 10. Swing in structural grain between the northern and southern Heritage Range. Stereograms of structural data from the northern and central Heritage Range, and the Patriot and Independence hills, highlight the change in structural style. direction (ISA), revealing that d e f o r m a t i o n within areas of distributed deformation is noncoaxial. For a transpressive deformation system

with boundary conditions of cx ~ 70, and a m i n i m u m ISA = 3 ~ the kinematic partitioning model of Tikoff & Teyssier (1994) suggests that

KINEMATIC PARTITIONING IN SOUTHERN ELLSWORTH MOUNTAINS

•

, , ~% , ~\ \

303

northernandcentral HeritageRange

~\

\\\ \

m -70 ~ 01 = 8 7 ~

\

\

\ X X

\

\\

~

l ~176 o0000 40

ge

(HorseshoeValley)

L

Fig. 11. (a) Kinematic model for the Heritage Range. c~,Angle of relative shortening; O, angle between the minimum instantaneous strain direction and the normal to deformation margin. (b) Relationship between the angle of oblique shortening and deformation boundary (c0, the orientation of the incremental strain axes in a horizontal plane (0), and the degree of strike-slip partitioning (Tikoff & Teyssier 1994). Values of al ~ 70~ and 0(90 - 01) =03~ from the northern and central Heritage Range suggest c. 67% kinematic partitioning.

67% of the strike-slip shear component has been partitioned along discrete strike-parallel dextral shear zones and faults (Fig. l l b ) . The figures derived from this model can only be considered approximations; nevertheless, they highlight the fact that the minimum finite and instantaneous strain axes are not parallel within the domains of distributed shear, as a significant degree of the external dextral simple shear component (33% predicted by Fig. l l b ) remains non-partitioned. This interpretation is supported by the oblique nature of L2 stretching lineations within the northern and central Heritage Range, and the presence of several examples of progressive strain partitioning from down-dip stretching to sub-horizontal stretching, in association with dextral shear.

Efficient partitioning within pure-shear transpression? Using known tectonic parameters such as the instantaneous strain directions within areas of distributed homogeneous deformation, and the angle of convergence (o0, Tikoff & Teyssier (1994) accurately modelled the approximate degree of partitioning within several modernday oblique convergent plate margins. Tikoff & Teyssier (1994) concluded from theoretical considerations and observation of present-day plate margins that pure-shear dominated transpressional settings (20 ~ < oL< 90 ~ are inefficient at accommodating strike-slip movement, and efficient partitioning of strike-slip shear will only occur within zones of wrench-dominated

304

M.L. CURTIS

transpression (~ < 20~ This is demonstrated in South Island, New Zealand, where the pureshear dominated transpressive (o~ = 30 ~ plate margin is non-partitioned, compared with the highly efficient wrench dominated transpressive (a = 5 ~ system of central California (Teyssier et al. 1995). However, structural data from the northern and central Heritage Range, Ellsworth Mountains, suggest that relatively efficient kinematic partitioning (c. 67 %) has occurred within a pureshear dominated transpression zone (cx 65-70~ Strain partitioning induced by relative plate motion alone cannot account for the apparent degree of partitioning; therefore, an additional influence must be operative. Jones & Tanner (1995) described a pure-shear dominated transpressive zone, with a similar high angle of relative convergence (~ = 60~ associated with the reactivated Highland Boundary Fault Zone (HBFZ) in central Scotland. The relationship of palaeostress directions adjacent to the fault zone is interpreted as indicating a high degree of strain partitioning, which is attributed to the pre-existence of the structurally weak HBFZ. Could pre-existing structures be the cause of partitioning within the Heritage Range? No evidence is present within the southern Soholt Peaks to suggest the existence of pre-existing structures, or rheological anisotropies along the sites of the dextral shear domains. Indeed, the obvious rheological contrasts formed by the heterolithic stratigraphy of the area, which might be expected to induce strain partitioning, show little or no evidence for strike-slip exploitation. This non-intuitive observation may be a result of the orientation of the faults or shear zones, for they dip <60 ~, which is below the observed lower limit for foliation planes displaying strike-parallel lineations (Fig. 6a), suggesting they are in a mechanically unfavourable orientation to facilitate strike-slip displacement. However, the presence of Cambrian age dykes and volcanic centres with associated extensional faulting trending parallel to the structural grain of the range (Curtis & Storey 1996), and pre-Gondwanian age faults also parallel to the orogen (SpOrli & Craddock 1992) indicate that the Heritage Range, and probably its basement, possessed a pre-existing structural architecture. It is proposed that weak structures are present within the unexposed basement of the Heritage Range, which have a positive influence on the degree of kinematic partitioning in the overlying cover rocks.

Conclusions Structural comparisons of the southern and northern structural domains of the Heritage Range support the pure-shear dominated dextral transpressive model for the northern and central Heritage Range proposed by Curtis (1997), and constrain the angle of external relative shortening (~) to c. 65-70 ~ Kinematic partitioning has occurred in southern Soholt Peaks, central Heritage Range, and is manifest as contemporaneous domains of strike-parallel, dextral and reverse shear, which are demonstrably coeval with regional cleavage development. Outside these domains of partitioned strain, deformation is dominated by the pure-shear component of transpression. However, the noncoaxial nature of finite and instantaneous strain directions (angle between the minimum instantaneous strain direction and the normal to the orogen (0) = 03 ~ indicates that not all the simple shear component of transpression has been partitioned into strike-parallel zones of dextral shear, which explains the oblique nature of the finite mineral stretching lineation throughout the range, and the examples of progressive strain partitioning in many areas of the range. Using the kinematic partitioning model of Tikoff & Teyssier (1994), the derived values of ot and 0 for the northern and central Heritage Range suggest that a high degree of kinematic partitioning has occurred (c. 67%). Such relatively efficient kinematic partitioning is not theoretically predicted, or demonstrated in modern oblique convergent margins (Tikoff & Teyssier 1994; Teyssier et al. 1995). Therefore, it is proposed that inherently weak zones, parallel to the strike of the Heritage Range, may be present within the unexposed basement of the Ellsworth Mountains, and may reasonably be expected to facilitate the apparently high degree of kinematic partitioning. The precedent for a high degree of kinematic partitioning within a pure-shear dominated transpressive system is provided by Jones & Tanner (1995), where the pre-existing Highland Boundary Fault Zone facilitates most of the strike-slip component of transpression, in a system where o~= 60 ~ B. C. Storey and M. R. A. Thomson are thanked for comments on the initial draft of this manuscript, with thorough, constructive reviews provided by E. Tavarnelli and an anonymous referee. I am indebted to B. Hull and S. Garrod for their assistance and companionship during the field seasons of 1993-1994 and 1995-1996, respectively. The Air Unit and support staff of BAS research station, Rothera, are gratefully acknowledged for logistical support.

KINEMATIC PARTITIONING IN SOUTHERN ELLSWORTH MOUNTAINS References

BUGGISCH,W.,WEBERS,G. E, REPETSKI,J. E. & GLENISTER, L. 1992. Cambrian conodonts from the Springer Peak and Minaret Formations, Ellsworth Mountains, West Antarctica. In: WEBERS, G. E, CRADDOCK,C. ~z SPLE'ITSTOESSER, J. E (eds) Geology and Paleontology of the Ellsworth Mountains, West Antarctica. Geological Society of America, Memoir, 170, 169-179. CLARKSON,R D. & BROOK, M. 1977. Age and position of the Ellsworth Mountains crustal fragment, Antarctica. Nature, 265, 615-616. COLLINSON,J .W., ISBELL,J. L., ELLIOT, D. H., MILLER, M. E, MILLER, J. M. G. c~z VEEVERS, J. J. 1994. Permian-Triassic Transantarctic Basin. In: VEEVERS, J. J. & POWELL, C. M. (eds) PermianTriassic Pangean Basins and Foldbelts along the Panthalassan margin of Gondwanaland. Geological Society of America, Memoir 184,173-222. CRADDOCK, C. 1972. Antarctic tectonics. In: ADIE, R. J. (ed.) Antarctic Geology and Geophysics. Universitetsforlaget, Oslo, 449-455. - - , SPORLI,K. B. d~ANDERSON,J. J. 1992. Structure of the Sentinel Range, Ellsworth Mountains, West Antarctica. In: WEBERS, G. E, CRADDOCK, C. & SPLETTSTOESSER,J. E (eds) Geology and Paleontology of the Ellsworth Mountains, West Antarctica, Geological Society of America, Memoir, 170, 393-402. - - , WEBERS, G .E, RUTFORD, R. H., SPORLI, K. B. & ANDERSON, J. J. 1986. Geological Map of the Ellsworth Mountains, Antarctica, Map and Chart Series MC-57, scale I :250 000. Geological Society of America, Boulder, CO. CURTIS, M. L. 1994. Evidence for vertical extrusion within a dextral transpression zone: the Heritage Range, Ellsworth Mountains, Antarctica. Geological Society of America, Abstracts with Programs, 26, A-504. 1997. Gondwanian age dextral transpression and spatial kinematic partitioning within the Heritage Range, Ellsworth Mountains, West Antarctica. Tectonics, 16, 172-181. - & STOREY,B. C. 1996. A review of the geological constraints on the pre-break-up position of the Ellsworth Mountains within Gondwana: implications for Weddell Sea evolution. In: STOREY,B. C., KING,E. C. & LIVERMORE,R.A. (eds) Weddell Sea Tectonics and Gondwana Break-up. Geological Society, London, Special Publications, 108, 11-30. DALZIEL, I. W. D. & ELLIOT,D. H. 1982. West Antarctica: problem child of Gondwanaland. Tectonics, 1, 3-19. -& GRUNOW,A. M. 1992. Late Gondwanide tectonic rotations within Gondwanaland. Tectonics, 11, 603-6O6. , DALLA SALDA, L. H. & GAHAGAN,L. M. 1994. Palaeozoic Laurentia-Gondwana interaction and the origin of the Appalachian-Andean mountain system. Geological Society of America Bulletin, 106, 243-252.

305

GOLDSTEIN,A.G. 1988. Factors affecting the kinematic interpretation of asymmetric boudinage in shear zones. Journal of Structural Geology, 10, 707-715. GOLDSTRAND, P. M., FITZGERALD,P .G,, REDFIELD,T. E, STUMP,E. & HOBBS, C. 1994. Stratigraphic evidence for the Ross orogeny in the Ellsworth Mountains, West Antarctica: implications for the evolution of the paleo-Pacific margin of Gondwana. Geology, 22, 427-430. GOODWIN, L. B. & WILLIAMS,P. E 1996. Deformation partitioning within a transpressive shear zone, Marble Cove, Newfoundland. Journal of Structural Geology, 18, 975-990. GRUNOW, A. M. 1993. Creation and destruction of Weddell Sea floor in the Jurassic. Geology, 21, 647-650. - - , KENT, D. V. & DALZIEL,I. W. D. 1987. Mesozoic evolution of West Antarctica and the Weddell Sea Basin: new paleomagnetic constraints. Earth and Planetary Science Letters, 86, 16-26. HANMER,S. 1984. The potential use of planar and elliptical structures as indicators of shear regime and kinematics of tectonic flow. Current Research Geological Survey of Canada, Paper, 84-1B, 133-142. HARLANO,W. B. 1971. Tectonic transpression in Caledonian Spitsbergen. Geological Magazine, 108, 27-42. JACKSON, J. 1992. Partitioning of strike-slip and convergent motion between Eurasia and Arabia in East Turkey and the Caucasus. Journal of Geophysical Research, 97, 12471-12479. JAGO, J. B. & WEBERS, G. E 1992. Middle Cambrian trilobites from the Ellsworth Mountains, West Antarctica. In: WEBERS, G. E, CRADDOCK,C. SPLETTSTOESSER,J. E (eds) Geology and Paleontology of the Ellsworth Mountains, West Antarctica. Geological Society of America, Memoir, 170, 101-124. JOHNSON,M. R. 1991. Sandstone petrography, provenance and plate tectonic setting in Gondwana context of the southeastern Cape-Karoo Basin. South African Journal of Geology, 94, 137-154. JONES, R. R. & TANNER,P. W. G. 1995. Strain partitioning in transpression zones. Journal of Structural Geology, 17, 793-802. JORDAN, P. G. 1991. Development of asymmetric shale pull-aparts in evaporite shear zones. Journal of Structural Geology, 13, 399-409. KIRKWOOD,D., MALO,M., ST-JULIEN,P. & THERRIEN,P. 1995. Vertical and fold-axis parallel extension within a slate belt in a transpressive setting, northern Appalachians. Journal of Structural Geology, 17, 329-343. LAWYER,L. A., GAHAGAN,L. M. & COFFIN,M. E 1992. The development of paleoseaways around Antarctica. Antarctic Research Series, 56, 7-30. MALAVIEILLE,J. & LACASSIN,R. 1988. 'Bone-shaped' boudins in progressive shearing. Journal of Structural Geology, 10, 335-345. MOUNT, V.S. & SUPPE,J. 1987. State of stress near the San Andreas fault: implications for wrench tectonics. Geology, 15, 1143-1146.

306

M.L. CURTIS

NUR, A., ROE, H. & Scorrl, O. 1986. Fault mechanics and the kinematics of block rotations. Geology, 14, 746-749. PETIT, J.-P. 1988. Normal stress dependent rupture morphology in direct shear tests on sandstone with applications to some natural fault surface features. International Journal ofRock Mechanics Mineral Sciences & Geomechanics Abstracts, 25, 411-419. & BARQUINS,M. 1988. Can natural faults propagate under Mode II conditions? Tectonics, 7, 1243-1256. RAMSAY, J. G. 1967. Folding and Fracturing of Rocks. McGraw-Hill, New York. -& HUBER, M. I. 1983. The Techniques of Modern -

-

Structural Geology Volume 1: Strain Analysis. Academic Press, London, 53. ROme, R-Y. E & CRUDEN,A. R. 1994. Strain and vorticity patterns in ideally ductile transpression zones. Journal of Structural Geology, 16, 447-466. SCHOPF, J. M. 1969. Ellsworth Mountains: position in West Antarctica due to sea-floor spreading. Science, 164, 63~56. SHERGOLD, J. H. & WEBERS, G. E 1992. Late Dresbachian (Idamean) and other trilobite faunas from the Heritage Range, Ellsworth Mountains, West Antarctica. In: WEBERS, G. E, CRADDOCK,C. & SPLETTSTOESSER,J. E (eds) Geology and Pale-

ontology of the Ellsworth Mountains, West Antarctica. Geological Society of America, Memoir, 170, 125-168. SNELLIE, J. L. 1981. A complete arc-trench system recognized in Gondwana sequences of the Antarctic Peninsula region. Geological Magazine, 118, 139-159. SPORH, K. B. 1992. Stratigraphy of the Crashsite Group, Ellsworth Mountains, West Antarctica. In: WEBERS, G. E, CRADDOCK, C. & SPLErrSTOESSER, J. E (eds) Geology and Paleon-

tology of the Ellsworth Mountains, West Antarctica. Geological Society of America, Memoir, 170, 21-36. --

& CRADDOCK, C. 1992. Structure of the Heritage

Range, Ellsworth Mountains, West Antarctica. //It: WEBERS, G. E, CRADDOCK, C. & SPLETTSTOESSER,J. E (eds) Geology and Paleon-

tology of the Ellsworth Mountains, West Antarctica, Geological Society of America, Memoir, 170, 375-392. STOREY, B. C., DALZIEL, I. W. D., GARRETF, S. W., GRUNOW, A. M., PANKHURST, R. J. & VENNUM,

W. R. 1988. West Antarctica in Gondwanaland: crustal blocks, reconstruction and break-up processes. Tectonophysics, 155, 381-390. STRACHAN, R. A., HOLDSWORTH,R. E., FRIDERICHSEN, J. D. & JEPSEN, H. E 1992. Regional Caledonian structure within an oblique convergence zone, Dronning Louise Land, NE Greenland. Journal of the Geological Society, London, 149, 359-371. SWANSON, M. T. 1992. Late Acadian-Alleghenian transpressional deformation: evidence from asymmetric boudinage in the Casco Bay area, coastal Maine. Journal of Structural Geology, 14, 323-341. TEYSSIER,C.,TIKOFF, B. & MARKLEY,M. 1995. Oblique plate motion and continental tectonics. Geology, 23, 447-450. T~KOVF,B. & GREENE, D. 1997. Stretching lineations in transpressional shear zones: an example from the Sierra Nevada Batholith, California. Journal of Structural Geology, 19, 29-39. & TEYSSlER, C. 1994. Strain modelling of displacement-field partitioning in transpressional orogens. Journal of Structural Geology, 16, 1575-1588. WArtS, D. R. & BRAMALL,A. M. 1981. Palaeomagnetic evidence for a displaced terrane in Western Antarctica. Nature, 293, 638-641. WEBERS, G .E, BAUER, R. L., ANDERSON, J. J., BUGGISCH,W., OJAKANGAS,R. W. • SPORLI,K. B. 1992a. The Heritage Group of the Ellsworth Mountains, West Antarctica. In: WEBERS, G. E, CRADDOCK, C. & SPLErrSTOESSER, J. E (eds) -

-

Geology and Paleontology of the Ellsworth Mountains, West Antarctica. Geological Society of --,

America, Memoir, 170, 9-19. CRADDOCK, C. & SPLETTSTOESSER,J. E 1992b. Geological history of the Ellsworth Mountains, West Antarctica. In: WEBERS,G. E, CRADDOCK,C. & SPLETrSTOESSER,J. E (eds) Geology and Pale-

ontology of the Ellsworth Mountains, West Antarctica. Geological Society of America, Memoir, 170, 1-8. ZOBACK, M. D. & HEALY, J. H. 1992. In situ stress measurements to 3.5 km depth in the Cajon Pass Scientific Research Borehole: implications for the mechanics of crustal faulting. Journal of Geophysical Research, 97(B4), 5039-5057. - - , ZOBACK,M. L., MOUNT,V S., et al. 1987. New evidence on the state of stress of the San Andreas fault system. Science, 238, 1105-1111.

Transpressional tectonics along the Karakoram fault zone, northern Ladakh: constraints on Tibetan extrusion M. P. S E A R L E 1, R. F. W E I N B E R G

2,a & W. J. D U N L A P 2

1Department o f Earth Sciences, Oxford University, Parks Road, Oxford 0)(1 3PR, UK 2Research School o f Earth Sciences, Australian National University, Canberra, A. C. T. 0200, Australia Abstract: The Tibetan plateau north of the Himalaya has approximately double normal

crustal thickness (60-75 km) and has been homogeneously shortened since the India-Asia collision at 60--50 Ma ago, yet, with minimal erosion rates, has almost no middle or deep crustal rocks exposed at the surface. In the Karakoram range, west of Tibet, early Tertiary crustal thickening and regional metamorphism resulted in Miocene crustal melting producing the Baltoro monzogranite-leucogranite batholith. Late Tertiary transpression along the western margin of the Tibetan plateau, caused by the continued northward penetration of India into Asia, led to exhumation of migmatites in the Pangong range and Baltoro-type granites along the Karakoram fault. Detailed studies of the Karakoram fault zone in northern Ladakh, India, show that Baltoro-type two-mica _+ garnet leucogranites, intruded 21-18.0 Ma ago, have been offset a maximum of 150 km right-laterally. Average slip rates since 18.0 +_0.6 Ma (2or) are 8.3 mm/a. 4~ mica cooling ages are 11.3 Ma on both sides of the main (southwestern) strand of the fault, suggesting that most of the exhumation of the Pangong migmatites and leucogranites must have occurred between 18.0 and 11.3 Ma. During this time, at slip rates of 8.3 mm/a the rocks would have moved horizontally right-laterally for c. 56 km and been exhumed by c. 20 km vertically during transpression, using the measured 20~ plunge of lineations. The high exhumation rate (3.0 mm/a) and amount of erosion (20 km) inferred between 18.0 and 11.3 Ma may also reflect the partitioning between an early transpressional strain associated with crustal thickening and exhumation of the Pangong deep crustal migmatites and leucogranites, and a later dominantly dextral strike-slip phase of fault motion along the central part of the Karakoram fault from c. 11 to 0 Ma. This timing may also coincide with the initiation of the N-S aligned normal faults and E-W extension in southern Tibet. We suggest that the relatively minor dextral offset (150 km) and the young age of initiation on this bounding fault do not support the model of large-scale extrusion of Tibetan crust, but they suggest instead that deformation of Tibet was taken up predominantly by crustal thickening.

One of the major questions concerning the continental tectonics of Asia is to what extent has the double normal thickness crust of the Tibetan plateau region been extruded eastwards, out of the way of the Indian indentor. The pattern of active strike-slip faults across Asia (Fig. 1) led Molnar & Tapponnier (1975) to suggest that Tibet was being forced eastwards out of the way of the northward indenting Indian crust. The continental extrusion hypothesis has subsequently been greatly expanded upon by Tapponnier & Molnar (1976, 1977), Tapponnier et al. (1982, 1986), Peltzer & Tapponnier (1988), Armijo et al. (1989), Peltzer et al. (1989) and Avouac & Tapponnier (1993). The boundaries of the extruding crust have been defined as the 2500 km long sinistral Altyn Tagh fault along the northern boundary of Tibet (Peltzer & Tapponnier 1988; Peltzer et al. 1989; Meyer et al. 1996) and the 1000 km long dextral Karakoram fault

along the southwestern margin of Tibet (Fig. 1). The Karakoram fault actually merges with the Indus suture zone in the region of Mt Kailas, although most proponents of the continental extrusion hypothesis link the Karakoram fault in the west to the Jiali fault in southeast Tibet via a series of N W - S E aligned small-scale strike-slip faults across southern Tibet (Armijo et al. 1989; Avouac & Tapponnier 1993). Estimates of geological offsets by Peltzer & Tapponnier (1988, plate 2, p. 15 358) suggested c. 500 km of left-lateral offset along the Altyn Tagh fault and c. 1000 km of right-lateral offset along the Karakoram fault. Holocene slip rates along the Altyn Tagh fault were estimated as being c. 20 mm/a in the west, 30 mm/a in the central part and 4 _+ 2 mm/a in the far east (Avouac & Tapp o n n i e r 1993; M e y e r et al. 1996). For the Karakoram fault, Holocene slip rates of c. 32 mm/a were proposed by Avouac & Tapponnier

SEARLE,M. P., WEINBERG,R. E & DUNLAP,W. J. 1998. Transpressional tectonics along the Karakoram fault zone, northern Ladakh: constraints on Tibetan extrusion. In: HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. E (eds) 1998. Continental Transpressionaland TranstensionalTectonics. Geological Society, London, Special Publications, 135, 307-326.

307

308

M.P. SEARLE E T A L .

7~"

\

Muji

Pamir,

&~

&"

V~&Kongur &MuztaghAta X

- 38"

\

J

TAR M BASIN]

38"

-..,,.._ /,%~," - 36 ~

36 ~

~aAng/Ladakh~

Nl

Parldat batholith

34~

~~

;el

34 = -

Pangong

~ k .

I TIBET i I PLA_TEAU i

32"

32 = -

f~[f~ii:[ batholithbatholith ~ ~ - 30"

[ + + Jbatholith ~

.~ [

Indus suture zone

. . . .

\

~

~YA--~,,,~_

~,~

/

&

44=.'>...)

~

+ Gangdese 3 0 a --

"~/a

~t-UreZo.,.

~'o

-28 ~

28"

o,

17o km

200 78" 1

. 81o I

,a,,.o.0o. 84" I

Fig. 1. Map of the Karakoram fault, western Himalaya and western part of the Tibetan plateau showing the major structures. ISZ, Indus suture zone; STD, South Tibetan Detachment system of normal faults which bounds the upper part of the High Himalayan slab. (1993) and Liu et al. (1992) based on offset geomorphological features. However, as noted by Searle (1996), none of these Holocene slip rates are constrained by absolute ages on the displaced morphological features; all are assumed to have formed at the end of the last glaciation and the beginning of the Holocene warming

period, which happened around 12 + 2 ka, and, as pointed out by Gillespie & Molnar (1995), this age could be older by a factor of two to four. Searle (1996) challenged the notion of c. 1000 km offset along the K a r a k o r a m fault by matching geological markers on either side of the fault. Using the well-dated Baltoro granites

TRANSPRESSION ALONG THE KARAKORAM FAULT (U-Pb zircon ages of 21 + 0.5 Ma (Parrish & Tirrul 1989) and U - P b monazite ages of 25.5 _+ 0.3/-0.8 Ma (Sch~irer et al. 1990)) he demonstrated that c. 120 km of dextral offset had occurred since the early to mid-Miocene from the Bilafond glacier region of north Pakistan to the Shyok region of northern Ladakh (Fig. 2). Similar offsets of the Ladakh batholith, Shyok-Pangong suture and the offset course of the Indus river suggest that the fault initiated after formation of all of these geological features. By determining the age of the youngest of these features we can determine the maximum age of initiation of the fault. In this paper we briefly review the zones of transpression and transtension along the Karakoram fault before discussing the geology of the northern Ladakh segment of the fault, where deep crustal metamorphic, migmatitic and leucogranitic rocks have been exhumed by transpression. We present new U - P b SHRIMP and 4~ geochronological data, which are used to constrain dextral slip rates along the fault, and we discuss the evolution of the Karakoram fault, especially with regard to the major debate concerning the amount and timing of eastward extrusion of Tibetan crust.

Transpression and transtension along the Karakoram fault The Karakoram fault stretches for some 1000 km from the Pamir mountains in the northwest to the Mt Kailas region in SW Tibet in the southeast (Fig. 1). Both ends of the fault terminate in a series of splays where transtensional basins are bounded by active normal faults. The northwestern termination of the fault in the central Pamirs shows active normal faults bounding the Muji graben around the Kongur and Muztagh Ata gneiss domes (Brunel et al. 1994; Searle 1996). The Tashkurgan graben south of Muztagh Ata is a transtensional pullapart basin with the bounding faults showing both dextral strike-slip and dip-slip fabrics. The southeastern termination of the Karakoram fault in SW Tibet similarly shows a large-scale dextral transtensional pull-apart basin, the Gar basin, bounded by faults which show both strike-slip and dip-slip components (Armijo et al. 1989). Southeast of Gar, the Karakoram fault merges into the Indus suture zone in the region of Mt Kailas, where east-west normal faults parallel to the strike of the suture zone predominate. The main dextral strike-slip motion occurs along the central segment of the fault, where

309

several transpressionally uplifted mountain ranges are bounded by late Neogene or Quaternary faults which have both strike-slip and thrust-related kinematics associated with them. Examples are the K2-Gasherbrum range in northern Pakistan (Fig. 1) and the Pangong range in northern Ladakh, India (Figs 2 and 3). K2 - Gasherbrum

range

North of the Karakoram batholith in Pakistan and along the border of Ladakh (India) with western Tibet, a well-bedded sequence of Carboniferous black slates and Permian-Triassic massive carbonates form the Gasherbrum range (Fig. 2). These sediments have been intruded by Jurassic to early Cretaceous quartz diorites. Structural culminations of mid-crustal amphibolite-facies rocks, the K2 gneiss, have Cretaceous crystallization ages (U-Pb zircon ages 115-120 Ma), and K - A r cooling ages of 111-94 Ma (Searle et al. 1990). The K2-Gasherbruin range forms a very high (8000-8600 m) barrier of mountains parallel to, and immediately SW of the Karakoram fault, which runs along the Shaksgam valley. North of K2, mylonites exposed along the fault in the Shaksgam valley separate amphibolite-facies rocks of the K2 gneiss to the southwest from unmetamorphosed carbonates of the Aghil range to the northeast (Searle 1991). Fission-track zircon ages from samples collected over the altitude range 4900-7150 m on Gasherbrum IV have a range from 125 to 31 Ma (Cerveny et al. 1989), indicating that only about 6 km of material has been eroded since the Cretaceous. Because of the very high present-day relief and deep erosion, most of this erosion must have occurred in very recent times. Fission-track apatite ages from K2 support this assumption, with samples collected over the altitude range 5300 -8611 m giving ages of 4.3 + 1.4 to 2.1 + 0.6 Ma (Foster et al. 1994). Combined with the U - P b zircon and K - A r hornblende, muscovite and biotite ages (Searle et al. 1989, 1990), the cooling history of the K2 gneiss shows long slow cooling from 110 to c. 5 Ma and then extremely rapid erosion of about 6-7 km of material in the last 5 Ma. The deep crustal metamorphic rocks of K2 are bounded by faults on either side, which show both dextral strike-slip and normal fault kinematic indicators (Searle 1991). The higher exhumation rates, greater erosion and high present-day topography in the K2 range support the contention that it was uplifted by transpression during dextral shear along the Karakoram fault in the period from 5 to 0 Ma.

310

M.P. S E A R L E E T A L .

0

9

E U_

g ~_

E

0 ~_

o

'-0

-

~

.~

C1 ~_

~

~

~

~

~

E

E

~

0

-~_
~

~ ~

0

o

~ ~

>,

~, ~

,--

--

r-

o

~

~ ~

0

~

o

0 ,<

O

9" . "

<

0

,,<

~.

~l~

(k

ii

~d

rr

~_.

.~

"I"

~

v <

o

o

Z

-=

-,I

i:::

0

~

Q

o

E

09

9

,~-~

_

I..I_ ,-.

e

2

o

=

'=

_I

0

N

OO

/

~.

.,..~

,% .... ,

.

, g,:

...." ,

1~ [[

+~

4 .

,

-b~

Y,,

0

4-

~ I//

......

' / ,.4:,.:?

...... . .....,~:.

:7::~'~{~7i :B:,' .=~c-;~,>;;.J,

CO

+J

+ \

~

~.o

A

~o

. ,,,,~

~'~

~

~

~_s~,

,

TRANSPRESSION ALONG THE KARAKORAM FAULT

80o

311

90"

100o

,

40=_

"/L

~..@': Kangri

!~!i~ ~.

~s - ia"

"<"r ''`'- ~ <%'•215215215215 w.~" x•215215215 %.•215215215215215 •215215215 ~ • .. ~

~ vvvv /

iiiiiiii

?

-/

iiii:

/,

/'" q-'" t / Khardung-la )( q,

~:~.:::. ~7.7..<: %. ::..-....;. ;;~.: ..;...::::"-;.

,

/

-1-

§ F ~

+/

~176176 .... ;:Z:D ,@:;o~

-Key

-~vw ~vv~v ~ , ' , k N a n g t s e ~ ~ -P Jr ~vv arul~ '-;-'."5""!.' - ; , (+ Chang-la \~v ~ru-*~ -:'"..........

+ ~/'

Indus molasse

+

+

I~:ii i ~

Ladakh granites

+

Karakoram granites

/f /

/+ J§

~< ~ ~

l -

~.<.%/%/

,.,

~

~.-~- 7rd,gOO,-.

%~o~~ , ~

/

+/

c.____ ,.' /

. ~~~-~%~~ ~~o~ "o. ~

10

20

.~

"'"'--~'"';;:~.... ) # ~ 7 !

,.~_

+

W:~1767,,
+ granodiorites complexKarak~ metarnorlohic

~ , , ,~,%~. <,._,.,_::):

~%

~ <~o!~;; ~ [ ~ % ' ~ ' ~ t

O

'.vi

+

k~/

[~

/

'-\

§

30

40

S"

Ci

~

#..~L + ~ / (

50km ~176 ~

~\

Fig. 3. Geological map of the Karakoram fault in northern Ladakh, India. Inset shows the area of approximately double thickness continental crust (shaded area), which, in the extrusion model, is moving eastward bounded by the dextral Karakoram fault in the southwest and the sinistral Altyn Tagh fault along the north.

Pangong range In the Shyok region of Ladakh, a major restraining bend is present where the Karakoram fault splays into two branches (Figs 3 and 4). The northeastern Pangong strand of the Karakoram fault runs along the Shyok valley, and cuts across to the north end of the Tangtse gorge (Fig. 5) to the southern part of Pangong lake, where it cuts across the margin of the Pangong range. The

Pangong lake is a drowned river valley which appears to have been dammed by recent movement along this branch of the Karakoram fault (Fig. 6). The southwestern branch, the Tangtse strand, runs through Darbuk and Tangtse villages towards Chusul and is the major fault which separates Karakoram terrane rocks to the northeast from pre-collisional Ladakh granites and Khardung volcanic rocks to the southwest. Southeast of Chusul b o t h branches of the

312

M.P. SEARLE E T A L . Darbuk

Karakoram fault

|

v ~ v v vv

$hyok

::1 ~ :~ <~ : 4 x

:<~x< i<~ ~

55.~4

:l:!l:: :::::::::::::::::::

leucogranite

/p intrusive contact

-.

strongly foliated bi - diorite intruded by leucogranite dykes

x I'1 " l fault

Tangtse ~

Karakoram

fault

@

gar + ms + bi leucogranite ~ ~

~

~t~ ~',-~

|

strain aureole

calc - silicate amphibolites + leucogranite dykes

bi I nr

/_.~z~=(]

"~ " A

I

I,:,0 ~marbles

+

+

+

+

+ 9

+++ + +

+

+ +

+ + +

+

+

+ +

+ +

+ +

+

Ill,

+

+ kadakh + + granite+ + + + -~ + +

+ +

+ .

+ +

+ +

+ +

+

+

X:i:!l

, i,, ii/i

* +

+ +

+

a;~ghj~lltg~ a

bi. ,

,.gr.

leucogranite dykes

Tangtse mylonites 0

1

I

|

granodiorite+ dykes

2

3

I

I

T migmatites

gar-- + -ms Igr.

51

Karakoram metamorphic complex

km

Fig. 4. Simplified cross-sections across the two strands of the Karakoram fault at Darbuk and Tangtse (see Fig. 3 for location). Karakoram fault merge to form one dominant strand (Fig. 3). The Pangong range is a transpressionally uplifted mountain belt rising to over 6500 m between these two strands of the Karakoram fault. Relative exhumation is shown by the high metamorphic grade of rocks within the range compared with lower-grade (greenschist-facies) volcano-sedimentary rocks to the northeast and virtually unmetamorphosed volcanic rocks of the Khardung Formation to the southwest. Preferential uplift of the Pangong range is also indicated by its greater surface topography compared with the flatter region northeast of the Pangong strand. Recent uplift of the Pangong range is synchronous with motion along the Pangong strand of the Karakoram fault. Hanging glaciers draining northeastwards off the

Pangong range, south of Pangong lake, have been truncated by the fault, which also shows down-to-the-northeast movement. Between the two strands, a variety of migmatites and high-grade metamorphic rocks including amphibolites, orthogneisses, calc-silicates and rare pelites are exposed. Amphibolites and granodioritic gneisses in the Pangong range containing biotite and hornblende show in situ partial melt textures forming migmatites. These rocks have been intruded by a network of quartzo-feldspathic pegmatites as well as biotite + muscovite _+garnet _+tourmaline leucogranite dykes and sheets (the Tangtse injection complex; Fig. 7), which locally come together to form kilometre-sized plutons (Weinberg & Searle in press). Staurolite + g a r n e t metapelitic schists occur

TRANSPRESSION ALONG THE KARAKORAM FAULT

313

Fig. 5. Kilometre-sized pluton of a two-mica + garnet leucogranite adjacent to the northeastern strand of the Karakoram fault, north of Tangtse. Foliation in the surroundings was deformed into concordance with the pluton after emplacement. along the road leading to Pangong lake, and less deformed, more massive bedded meta-carbonates and pelites occur northeast of the Pangong strand of the Karakoram fault. Whereas the Karakoram metamorphic complex crops out south of the batholith in the Pakistan sector, in northern Ladakh the metamorphic rocks have been inter-sheared with the Baltoro-type granites and Pangong migmatites along the Karakoram fault zone. All rocks show strong deformation fabrics and mylonites outcrop in a zone 800-1200 m wide along both branches of the Karakoram fault (Fig. 4). Dextral shear fabrics have been superimposed on previously undeformed Ladakh granite lithologies. At Tangtse a strong mylonitic lineation is ubiquitous, plunging at 20 ~ towards the NW. This is consistent with the observation that deeper crustal rocks have been exhumed along the Karakoram fault in this area.

S H R I M P U - P b zircon dating from the migmatites reveals core ages of 106.3 _+2.3 Ma (see below). Based on the lithogical and age similarities, the Pangong migmatites are correlated with similar, pre-collisional, deformed granodioritic rocks in northern Pakistan, notably the K2 gneiss (Searle et al. 1990), the Muztagh Tower gneiss and the Hushe gneiss, which are also intruded by garnet leucogranite dykes (Searle 1991; Crawford & Searle 1992).

Offset Karakoram granites One of the youngest features displaced by the Karakoram fault is the Karakoram batholith. The youngest phase of granite magmatism within the Karakoram batholith is the middle Miocene Baltoro plutonic unit. Searle (1996) originally proposed 120 km of dextral offset of the Baltoro-type granites along the Karakoram

314

M.P. SEARLE E T A L .

Fig. 6. The northeastern strand of the Karakoram fault runs along the Pangong Lake, which is a drowned river valley, dammed by transpressional uplift at its western end. fault. Following our more detailed regional mapping in north Ladakh, we now refine this value to 150 km based on the offset points shown in Fig. 2. Baltoro granite

The Baltoro granite forms a large component of the Karakoram batholith in northern Pakistan, where it consists of co-magmatic monzogranites and leucogranites with the assemblage qtz + Kfs+ pl + bt _+ms _+grt (Searle et al. 1989, 1992; Searle 1991). The Baltoro granites are mildly peraluminous with SYSr/86Sr ratios of 0.7072-0.7128, and have geochemical characteristics compatible with dehydration melting of a biotite-rich pelite in a lower-crustal source (Rex et al. 1988). The Baltoro granite intrudes sillimanite grade gneisses along the southern margin and unmetamorphosed Carboniferous-Permian and Mesozoic sediments along the n o r t h e r n boundary, which have been contact metamorphosed along the margin to form andalusite hornfels (Searle et aL 1989, 1992). The age of crustal melting is known from U - P b zircon ages of 21.0 + 0.5 Ma from samples collected along the Baltoro glacier (Parrish & Tirrul 1989) and U - P b monazite ages of 25.5 +~ s Ma from the

Latok peaks, west of the Baltoro (Sch~irer et al. 1990). A suite of lamprophyre dykes (hornblende vogesites and biotite minettes) intrude the country rocks north of the Baltoro granite and a few occur south of the batholith in Hushe. Two of these lamprophyres from Broad Peak and the Abruzzi glacier have K - A t biotite ages of 22.0 + 0.7 Ma and 22.4 + 0.7 Ma, and one from the Masherbrum glacier south of the batholith has a K - A r biotite age of 24.0 + 1.0 Ma (Searle et al. 1989). These volatile-rich, mantle-derived melts were therefore intruded at the same time as the lower-crustal melts forming the Baltoro granite. Because of the volume of lower-crustal melting along the central Karakoram, which is far greater than the volume of crustal melt granites along the High Himalaya, and the synchroneity of the lamprophyre dyke intrusion, Searle et al. (1992) proposed a model involving simultaneous lower-crustal and upper-mantle melting at c. 21 Ma beneath the Karakoram, with mantle melts providing an additional source of heat to promote widespread lowercrustal melting. In the Baltoro and northern Hushe regions of north Pakistan, the Baltoro granites are undeformed with intrusive contacts along both north and south margins. These monzogranites and

TRANSPRESSION ALONG THE KARAKORAM FAULT

Fig. 7. Two-mica leucogranite dykes (Pangong Injection Complex) intruding amphibolites, calcsilicates and pelites within the Karakoram fault zone, north of Tangtse village.

leucogranites continue eastward to the Saltoro Kangri-K12 range (Fig. 2). This area is extremely high and largely inaccessible, and lies right on the disputed border between India and Pakistan. The position marked 'a' in Fig. 2 marks the location where the upper Miocene Baltorotype granites have been cut by the Karakoram fault. South of this point there are no more leucogranites of similar mineralogy or composition to the Baltoro unit. The Karakoram fault cuts the batholith at an acute angle and the trend of the batholith has been rotated clockwise NE of the fault into parallel alignment with it in the Siachen glacier region (Searle 1996). Solid-state deformation in the form of mylonitic S-C fabrics has been superimposed on the Baltoro granites along the Karakoram fault. The fault must therefore have been active after crystallization of the granites at c. 25-21 Ma.

Tangtse granite Larger leucogranite intrusions within the Karakoram fault zone at Darbuk and Tangtse

315

(Figs 3 and 4) contain biotite, muscovite and garnet, and were also intruded before movement along the Karakoram fault. Spectacular dextral S-C fabrics (Fig. 8) have been superimposed on the igneous texture. These leucogranites are compositionally similar to the Baltoro leucogranite in Pakistan and we correlate the two as originally belonging to the same batholith. Their similar mineralogy, trace element contents, and Rb/Sr and Rb/Ba ratios (normalized to primordial mantle values from Taylor & McLennan (1985)) strongly contrast with those of the neighbouring Ladakh granite. Furthermore, as we will show below, the ages of the Baltoro and the Tangtse granites are both early to mid-Miocene, and contrast with the much older, pre-collisional Ladakh granites (which span the range 103-50 Ma). The leucogranites have been offset dextrally along the Karakoram fault for a maximum of 150 km from the Saltoro Kangri-K12 range southwest of the Karakoram fault in Pakistan to the Darbuk-Tangtse area in north Ladakh (Fig. 2). The fact that no Baltoro-Tangtse-type granites occur southwest of the Karakoram fault in Ladakh strongly suggests that the Karakoram fault initiated after 18.0 _+0.6 Ma (20-), the crystallization age of the youngest offset granite (see below). The Tangtse granite intruded earlier, foliated, biotite-rich diorite, migmatites, calc-silicates and pelites, and has been intejasely deformed by non-coaxial dextral shear r a t h e greenschist facies along the Karakorarmmult after crystallization. Solid-state d e f o r l i l l o n is defined by strong S-C fabrics indicanve of dextral shear. Although the entire granite is deformed, the S-C fabrics are best developed along the southwestern margin of the Karakoram fault zone. Microstructures suggest that deformation of the Tangtse leucogranite occurred under greenschist facies conditions (450-300~ and there is no evidence of a high-temperature deformation history. The quartz exhibits ductile deformation features only, with the grains having accommodated strain by both grain boundary migration recrystallization and sub-grain rotation. The Kfeldspar in the leucogranite has deformed mainly by fracturing and cataclastic flow, which is characteristic of granites deformed under greenschist-facies conditions. U - P b and 4~ geochronology was undertaken on key samples of Tangtse granite and rocks on either side of the Karakoram fault, to determine the age of strike-slip faulting and the cooling history of rocks within the Karakoram fault and to the southwest in the Ladakh batholith.

316

M.P. SEARLE E T A L .

Fig. 8. Polished slab of Tangtse granite from the southwestern strand of the Karakoram fault near Tangtse village, showing spectacular dextral S-C fabrics.

U-Pb SHRIMP dating Two samples were prepared for U - P b dating using the Sensitive High-Resolution Microprobe (SHRIMP II) at the Australian National University. A sample of the granite from the in situ Pangong migmatite was collected in the Tangtse gorge (sample 022), and a sample of the mylonitic Tangtse leucogranite was collected between the villages of Tangtse and Darbuk (sample 215). The leucogranite sample is particularly important because its crystallization age provides a maximum age for the initiation of the dextral shear along the Karakoram fault. The Tangtse granite, sample 215, has prismatic, 150-400 txm long zircons. Cathodoluminescence (CL) images show that these zircons generally have luminescent (low U) cores with euhedral fine-scale oscillatory igneous zoning, truncated and overgrown by weakly luminescent (high U) finely zoned prismatic mantles. Primary ion beams were focused to spots of c. 30 Ixm in diameter, and using the CL images we dated the cores and mantles and avoided overlapping the two distinct domains (overlap occurred, however, in spot 9.2 (Table 1) as found in CL images produced after analysis; this spot was not included in the plots of Fig. 9). Details of the procedure for zircon U - P b isotopic analysis using the S H R I M P have been given by

Compston et al. (1984) and Williams & Claesson (1987), and the standard used was SL13 (Claou6-Long et al. 1995). During analysis, for every four spots of unknown sample, two standards were analysed. For each analysis, seven cycles of secondary ion yields were measured on the following ions: Zr2 O+, the four lead isotopes (2~ 2~ 2~ 2~ U + and U O +, Th + and ThO +. Because neither 2~176 nor 2~ chronology by SHRIMP provides a fine age resolution for the Phanerozoic, we relied entirely on the 2~ and when possible on 2~ geochronometers. The procedure to determine the errors associated with the unknown ages has been described by Claou6-Long et al. (1992), and includes the error of the UO/U v. 2~ calibration line, and the counting statistics error of the unknowns. The capability of the SHRIMP in resolving the age of young rocks has been shown by Zeitler et aL (1989), who successfully dated thin zircon rims younger than 11 Ma. In total we analysed 39 spots in 23 zircons (Table 1). We corrected the results for common lead by assuming the isotopic composition of lead from Broken Hill, Australia. The results may be divided into two age groups and a few scattered values (Fig. 9a, Table 1). The older group represents inherited zircons from the magma source and the younger group

T R A N S P R E S S I O N A L O N G T H E K A R A K O R A M FAULT

317

Table 1, SHRIMP analyses of zircons from the Tangtse leucogranite sample 215

Spot Date 1.1 1.2 2.1 2.2 3.1 4.1 4.2 5.1 5.2 6.1 7.1 7.2 8.1 8.2 9.1 9.2 10.1 10.2 11.1 11.2 11.3 12.1 12.2 13.1 14.1 14.2 15.1 16.1 16.2 17.1 18.1 19.1 19.2 20.1 20.2 20.3 21.1 22.1 22.2 23.1

23.06 23.09 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.09 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.09 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.06 23.09 23.09 23.09 23.09 23.09

Core (1) or U rim (2) (ppm) 1 2 2 1 1 1 2 2 1 1 1 2 1 2 1 2 1 2 1 2 1 1 2 1 1 2 1 1 2 1 1 1 2 1 0 2 2 2 1 2

4845.0 15782.0 1456.0 10562.0 1821.0 448.0 11310.0 3967.0 4092.0 183.0 655.0 13625.0 103.0 6098.0 381.0 561.0 1109.0 7487.0 60.0 7070.0 88.0 945.0 14136.0 938.0 631.0 13486.0 1387.0 328.0 7958.0 3512.0 841.0 1365.0 5176.0 1308.0 728.0 9221.0 12207.0 15982.0 398.0 13641.0

Th (ppm) %206c %208c 2~176 1143 317 837 257 578 42 158 652 169 116 858 438 26 264 239 333 545 231 78 979 70 333 261 33 122 217 277 12 987 179 33 569 736 568 13 190 221 317 456 3172

-0.01 0.13 0.21 0.09 -0.02 0.40 0.06 0.21 0.17 0.84 1.06 0.10 3.53 0.16 0.01 0.29 -0.01 0.27 0.69 0.11 0.13 1.31 0.25 0.55 0.05 0.03 0.69 1.83 0.15 0.24 0.07 0.38 0.14 0.15 0.17 0.13 0.00 0.21 0.41 0.09

-0.4 45.9 2.4 23.5 -0.5 14.2 24.2 8.0 24.7 7.9 5.2 21.5 35.2 22.3 0.2 2.9 -0.2 49.8 3.7 5.5 44.3 8.6 2.7 88.1 1.8 11.6 19.2 105.0 7.3 23.5 9.4 5.2 6.5 2.6 22.9 37.5 0.0 59.7 2.4 2.5

0.0463 0.0477 0.0483 0.0473 0.0462 0.0502 0.0470 0.0483 0.0488 0.0548 0.0570 0.0474 0.0797 0.0478 0.0474 0.0497 0.0472 0.0489 0.0872 0.0474 0.0815 0.0487 0.0476 0.0515 0.0482 0.0467 0.0528 0.0633 0.0478 0.0495 0.0471 0.0507 0.0477 0.0487 0.0479 0.0476 0.0462 0.0483 0.0512 0.0473

23sU/2~ + 7 __ 10 __+12 _ 8 + 14 + 29 + 5 ___10 + 5 + 28 + 14 + 7 + 47 + 12 + 20 + 15 _+ 13 _+ 12 _+ 13 +_ 8 _+ 17 _ 28 + 6 + 26 _ 12 + 7 + 20 + 78 _+ 5 _+ 7 _+ 21 + 11 + 7 _+ 10 _+ 31 _+ 12 + 10 + 7 _+ 25 _+ 8

307.4 326.5 346.1 335.5 330.6 325.4 383.6 398.5 103.8 121.4 111.9 370.2 148.9 424.3 99.9 133.2 101.9 359.3 4.8 357.1 6.4 336.4 354.1 353.3 71.6 342.8 408.8 507.7 352.3 100.5 353.5 115.2 381.3 101.9 369.7 378.2 365.1 369.6 99.6 320.0

2~ age (lo-)

2~ age (10-)

_ 2.7 18.7 _+ 0.5 20.9 __+0.2 _ 4.6 9.3 _+ 6.4 19.7 _ 0.3 _ 3.4 18.2 _+ 0.5 18.6 _ 0.2 _ 5.9 13.2 + 3.7 19.2 __ 0.3 ___4.2 17.2 + 0.9 19.5 _ 0.3 _+ 6.6 33.1 _+ 5.2 19.7 _+ 0.4 _+ 2.7 15.1 + 4.6 16.8 + 0.1 +_ 4.1 15.6 + 0.8 16.1 + 0.2 _+ 0.6 51.6 + 5.4 61.7 + 0.3 _+ 3.4 56.6 + 3.1 52.5 + 1.5 _+ 1.0 58.8 -+ 1.2 56.7 __ 0.5 _+ 5.1 12.0 -+ 2.4 17.4 _+ 0.2 _+ 4.3 69.3 -+ 6.5 41.6 + 1.2 + 3.4 12.3 _+ 0.3 15.2 -+ 0.1 -+ 1.5 60.0 + 2.3 64.2 + 0.9 + 1.8 50.2 _+ 2.1 48.1 -+ 0.6 _+ 0.9 65.0 + 1.6 63.0 + 0.5 _+ 3.8 11.3 _+ 5.5 17.9 + 0.2 _+ 0.1 1261.5 _+ 22.1 1221.0 +_ 16.2 _ 3.0 14.6 _+ 0.7 18.0 -_ 0.2 +- 0.1 1327.4 + 33.0 918.2 __ 14.6 _+ 5.2 36.1 __+1.7 19.1 __ 0.3 _+ 4.5 9.6 + 3.5 18.2 _+ 0.2 _+ 7.0 2.8 + 10.3 18.1 +_ 0.4 _+ 0.6 83.0 +- 4.8 89.4 _+ 0.7 __+3.6 19.0 -+ 5.6 18.8 _+ 0.2 ___6.0 16.6 -- 1.5 15.6 + 0.2 _+ 20.9 12.5 _+ 0.5 _+ 3.2 17.0 -+ 0.7 18.2 _+ 0.2 _+ 0.6 61.0 -+ 6.2 63.7 + 0.4 + 5.2 23.3 _+ 7.8 18.2 _+ 0.3 + 1.0 63.7 + 2.0 55.5 _+ 0.5 _+ 2.2 16.3 + 0.8 16.9 _+ 0.1 _+ 0.9 56.5 + 1.7 62.9 _+ 0.5 +_ 5.4 38.5 _+ 24.2 17.4 _+ 0.3 _+ 4.1 14.4 _+ 8.6 17.0 _+ 0.2 + 2.5 21.1 + 7.5 17.6 _+ 0.1 + 3.0 8.4 + 4.1 17.4 _+ 0.1 _+ 1.8 66.6 -+ 1.7 64.4 + 1.1 +_ 3.1 20.8 + 0.5 20.1 _+ 0.2

%206c indicates the percentage of c o m m o n 2~ correction in the total measured 2~ %208c indicates the percentage of common 2~ correction in the total measured 2~ Gaps in the 2~ age column correspond to poorly resolved Th ages because of low Th content. All errors correpond to l a .

represents the age of crystallization of the T a n g t s e g r a n i t e . Six d a t a p o i n t s d e f i n e t h e o l d e r g r o u p w i t h a w e i g h t e d m e a n o f 63.0 _+ 0.8 M a ( 2 a ) a n d f o u r d a t a p o i n t s s h o w a n age r a n g e d e c r e a s i n g f r o m c. 60 to 40 M a , s u g g e s t i n g t h e s e g r a i n s w e r e a f f e c t e d b y P b loss. T w o s p o t s in t h e s a m e z i r c o n y i e l d e d m u c h o l d e r ages at c. 1220 a n d 920 M a ( T a b l e 1), s u g g e s t i n g i n v o l v e m e n t o f o l d e r c r u s t a l rocks. O n e s p o t y i e l d e d c. 89 M a ( T a b l e 1, Fig. 9a), w h i c h c o u l d e i t h e r b e a n

o u t l i e r or a n i n h e r i t e d z i r c o n f r o m a n o l d e r m a g matic phase. T h e y o u n g e r g r o u p , c o m p r i s i n g 25 a n a l y s e s , m a i n l y o f z i r c o n m a n t l e s b u t also o f a f e w z i r c o n c o r e s (Fig. 9b, T a b l e 1), y i e l d s a w e i g h t e d m e a n a g e o f 18.0 M a , a n d a s t a n d a r d d e v i a t i o n o f 1.4 M a , a n d a s t a n d a r d e r r o r in t h e m e a n o f 0.3 M a (Io-). ( s p o t 16.1 o f 12.5 _+ 0.5 M a w a s c o n s i d e r e d a n o u t l i e r a n d n o t i n c l u d e d in this g r o u p . ) T h e r e l a t i v e l y h i g h s p r e a d o f ages ( h i g h s t a n d a r d

318

M.P. SEARLE E T A L .

Fig. 9. Sample 215. (a) Stack histogram of the 2~ age distribution of 37 spots on zircons from the mylonitic Tangtse leucogranite determined using SHRIMP II. (b) Detailed histogram of the 25 young ages of mantles and cores.

deviation) compared with the relatively small individual errors (of the order of 0.3 Ma) suggests one of three possibilities: (a) the age spread is caused by either Pb loss or underestimation of U content in high-U zircons; (b) individual errors are underestimated; (c) the zircons have a true age spread. A potential source for the high standard deviation could be the high U content of some analyses (up to 1.6% U), suggesting the possibility of calibration problems, but, in contrast to sample 022 (below), there is no clear relation between U content and age. Spots with relatively high Th contents yielded high-precision 2~ ages. These were nearly identical within error to U/Pb ages (e.g. spots 1.1, 2.1, 3.1, 5.1, 11.2, 16.2, 19.2, 23.1 in Table 1), but with a consistent tendency towards lower values (spots 1.1, 3.1, 11.2, 16.2). This last observation suggests that either the high U content may in fact have caused small unwanted analytical effects or the U - T h - P b system was opened after zircon crystallization. Either of these effects may explain the relatively large age spread, but neither was sufficiently important to have fundamentally changed the apparent age obtained. The 18.0 _+ 0.6 Ma (2o-) Tangtse granite is younger than the compositionally similar 21.0 + 0.5 Baltoro granite (dated by conventional zircon U - P b methods, (Parrish & Tirrul 1989)). The relatively short crystallization age difference between the two rocks and their similar geochemistry suggest they could have resulted

from the same crustal melting event; such events typically last a few million years. The older 63.0 + 0.8 Ma (20-) cores are of clear igneous origin and are comparable in age with Rb-Sr ages on granitoids studied by Debon (1995) and Debon et al (1996) in the Hunza Karakoram in Pakistan. Sample 022 has well-formed, prismatic zircons, 100-300 txm long. Under the microscope, these zircons can be divided into two groups: yellowish and clear transparent zircons, with some zircons showing yellowish cores overgrown by clear, transparent rims. Cathodoluminescence images show similar patterns to sample 215, with luminescent (low-U) cores with euhedral fine-scale oscillatory igneous zoning, truncated and overgrown by weakly luminescent (high-U) finely zoned prismatic mantles. Dating defined two main age groups (Table 2 and Fig. 10): the yellowish zircons and cores represent an older phase with a weighted mean age of c. 106.3 _+2.3 Ma (2o.), interpreted as the crystallization age of the orthogneiss. The transparent zircons and transparent rims represent a younger group ranging in age from 22 to 15 Ma, interpreted as dating zircon growth during partial melting of the orthogneiss. U/Pb ages of the older group are confirmed by similar Th/Pb ages (Table 2). A few analyses, however, yielded U/Pb ages intermediate between the two main age groups suggesting Pb loss. Pb loss is confirmed by comparison to Th/Pb ages of these same spots, which are much

TRANSPRESSION ALONG THE KARAKORAM FAULT Table 2.

SHRIMP analyses of zircons from Pongong migmatite sample 022

Spot D a t e 1.1 1.2 1.3 2.1 2.2 3.1 3.2 4.1 4.2 5.1 5.2 6.1 6.2 7.1 8.1 8.2 9.1 9.2 10.1 11.1 11.2 11.3 12.1 13.1 14.1 15.1 16.1 16.2 17.1 18.1 19.1 19.2 20.1 21.1 22.1 23.1 23.2 24.1 24.2 25.1 25.2 26.1 26.2 27.1 28.1 29.1 30.1 31.1 32.1 33.1 33.2

319

18.06 18.06 23.09 18.06 23.09 18.06 18.06 18.06 18.06 18.06 18.06 18.06 18.06 18.06 18.06 18.06 18.06 18.09 18.06 18.06 23.09 23.09 18.06 18.06 18.06 18.06 18.06 18.06 18.06 18.06 18.06 23.09 18.06 23.06 23.06 23.06 23.09 23.06 23.09 23.06 23.09 23.06 23.09 23.06 23.06 23.06 23.06 23.06 23.06 23.09 23.09

Core (1) or U rim (2) (ppm) 1 2 2 1 1 2 1 2 1 2 1 2 1 1 2 1 2 1 1 2 2 1 1 2 1 2 1 1 1 1 2 1 1 1 1 1 2 1 1 1 2 1 2 1 1 1 1 1 2

101.8 363.0 265.3 1107.9 1274.5 911.1 134.3 233.0 824.0 283.4 312.7 296.8 113.7 9384.1 1023.8 286.2 280.6 187.8 391.7 215.3 210.1 97.7 5889.9 1078.8 419.3 292.9 622.5 6123.5 826.2 2187.8 353.4 729.1 3758.0 144.6 3158.8 3068.8 6421.8 496.2 1605.2 3726.0 3356.1 87.0 730.4 99.8 936.0 116.3 118.1 692.7 358.5 12674.0 1902.8

Th (ppm) %206c %208c 2~176 101 3.29 15 7.02 17 3.23 18 2.03 19 0.60 108 0.97 96 4.74 14 14.44 273 1.73 16 14.33 227 1.24 46 3.01 118 3.13 31 0.22 1777 0.39 14 8.89 127 2.18 111 2.42 262 1.17 18 16.78 8 36.72 60 35.86 97 0.53 16 8.27 17 5.06 189 1.93 18 8.01 101 1.59 304 0.64 37 2.02 292 1.53 142 1.05 134 0.94 66 0.65 18 0.38 7 0.37 47 0.17 210 0.41 36 0.86 60 0.47 64 0.22 76 1.64 20 1.24 73 1.00 52 0.94 76 1.10 109 1.15 449 - 0 . 0 2 268 1.34 311 0.14 30 0.20

19.7 85.0 85.8 86.0 89.9 19.1 31.6 86.5 20.1 114.7 11.0 46.2 18.7 103.3 1.6 66.5 24.3 22.6 10.5 79.1 4.0 3.3 80.2 134.1 82.9 16.9 96.0 79.7 10.0 97.7 11.9 13.1 68.5 7.9 98.2 124.3 88.5 4.1 66.4 102.3 60.6 12.6 70.4 9.0 25.2 10.6 8.4 -0.2 10.4 28.7 35.0

0.0788 __ 46 0.1117 __ 85 0.0764 __+77 0.0653 _+ 33 0.0520 _+ 23 0.0558 _+ 21 0.920 _+ 41 0.1807 _+ 271 0.0634 _+ 20 0.1795 _+ 313 0.0597 - 27 0.0751 _+ 34 0.0773 _+ 48 0.0485 _+ 8 0.0519 _+ 8 0.1291 _+ 217 0.0682 _+ 36 0.0706 _+ 30 0.0589 _+ 24 0.2025 _+ 234 0.3887 _+ 92 0.3814 _+ 100 0.0514 _+ 9 0.1232 _+ 75 0.0935 _+ 56 0.0659 + 26 0.1209 + 71 0.0612 _+ 12 0.0540 _+ 12 0.0653 _+ 24 0.0624 _+ 24 0.0566 _+ 33 0.0552 _+ 12 0.0540 _+ 40 0.0499 _+ 14 0.0498 + 12 0.0480 _+ 10 0.0512 _+ 16 0.0544 _+ 24 0.0508 _+ 10 0.0484 _+ 12 0.0633 _+ 43 0.0579 _+ 37 0.0575 _+ 32 0.0551 _+ 31 0.0583 _+ 34 0.0589 __ 25 0.04800 _+ 9 0.0605 _+ 22 0.04780 _+ 8 0.0482 _+ 16

238U/Z~ 55.3 _+ 1.1 429.6 _+ 13.3 464.0 _+ 17.5 374.3 __ 7.4 407.2 _+ 5.1 170.8 _+ 3.1 62.4 _+ 2.0 401.1 + 19.2 97.2 _+ 0.9 635.9 _+ 102.1 58.4 _+ 0.7 115.0 _+ 2.4 55.9 _+ 1.1 302.6 _+ 4.3 56.8 _+ 0.6 350.5 _ 26.5 63.6 _+ 1.5 58.2 _+ 1.0 60.6 _+ 0.8 321.9 + 32.7 63.8 _+ 2.2 33.9 _+ 0.6 306.2 _+ 4.7 503.1 _+ 16.7 291.8 _ 7.4 61.8 _+ 0.8 350.3 _+ 12.3 298.3 _+ 3.0 61.6 + 0.8 316.7 _+ 4.5 57.1 _+ 0.9 155.6 + 3.3 319.0 _+ 2.8 67.6 _+ 1.5 341.5 _+ 5.0 339.3 _+ 2.9 338.4 + 3.1 95.2 _+ 1.5 399.1 +_ 7.7 389.7 + 4.1 391.0 _+ 4.1 62.1 _+ 1.4 390.9 _+ 6.3 57.5 _+ 1.4 310.2 _+ 6.4 61.8 _+ 1.2 57.7 -+ 2.0 59.0 __ 0.4 62.0 + 0.9 307.6 + 1.6 388.6 __+3.9

2~ age (lo-) 109.8 29.8 8.6 24.9 5.3 92.0 102.3 39.6 95.7

_+ 5.3 _+ 22.6 -+ 13.5 _+ 27.1 _+ 19.9 _+ 5.9 _+ 5.9 _+ 51.3 -+ 3.6

106.5 + 4.6 87.3 -+ 11.7 104.6 _+ 6.4 111.7 111.9 107.5 108.5 110.1 71.4

+_ 1.3 _+ 59.8 _+ 8.2 _+ 6.7 _ 4.5 _+ 41.4

12.0 _+ 8.6 39.9 106.6 14.4 37.5 115.9 3.9 107.5 96.4 17.1 112.3 1.6

-+ 23.1 _+ 4.1 + 32.8 -+ 11.9 _+ 3.8 _+ 21.5 _+ 3.4 _+ 5.8 _+ 5.0 _+ 8.7 _+ 35.0

4.2 _+ 21.5 108.3 _+ 3.6 20.2 _+ 12.1 9.0 98.4 23.3 110.9 77.6 104.3 107.9 109.8 125.5 22.3 29.6

_+ 7.9 _+ 6.7 _+ 17.8 _+ 6.8 -+ 10.2 _+ 5.3 _+ 5.9 __ 2.3 _+ 4.2 _+ 5.2 _+ 14.1

2~ age (lo-) 111.7 _+ 2.3 13.9 _+ 0.5 13.4 _+ 0.5 16.8 __ 0.3 15.8 _+ 0.2 37.3 + 0.7 97.6 _+ 3.1 13.7 _+ 0.8 64.8 _+ 0.6 8.7 _+ 1.4 108.1 _+ 1.3 54.1 _+ 1.1 110.7 _+ 2.2 21.2 _+ 0.3 112.1 _+ 1.1 16.7 _+ 1.3 98.4 _+ 2.3 107.3 + 1.9 104.3 _+ 1.3 16.6 -+ 1.8 58.5 _+ 2.3 111.9 _+ 2.9 20.9 _+ 0.3 11.7 + 0.4 20.9 _+ 0.5 101.4 _+ 1.3 16.9 _+ 0.6 21.2 _+ 0.2 103.2 + 1.4 19.9 + 0.3 110.1 + 1.7 40.9 _+ 0.9 20.0 _+ 0.2 94.0 _+ 2.2 18.8 _+ 0.3 18.9 _ 0.2 19.5 + 0.2 67.1 _+ 1.0 16.0 _+ 0.3 16.4 _+ 0.2 16.6 _ 0.2 101.3 _+ 2.3 16.3 + 0.3 110.1 _+ 2.7 20.6 _+ 0.4 102.3 _+ 2.0 109.5 _+ 3.7 108.4 __ 0.7 101.8 __ 1.6 22.5 _ 0.1 16.6 _ 0.2

%206c indicates the percentage of c o m m o n 2~ correction in the total measured 2~ %208c indicates the percentage of c o m m o n 2~ correction in the total measured 2~ Gaps in the 2~ age column correspond to poorly resolved Th ages because of low T h content. All errors c orre pond to lo-.

320

M.P. SEARLE E T A L .

Fig. 10. Sample 022. (a) Stack histogram of the 2~ age distribution of 51 spots on zircons from the in situ granite of the Pangong migmatite. (b) Detailed histogram of the 26 spots analysed belonging to the young age group showing two distinct subgroups. Inset shows age plotted against U content, suggesting a positive relation between the two for spots with >6000 ppm of U.

less affected than U/Pb ages, yielding ages similar to those of unaffected zircons. The younger age group, represented by 21 spot analyses, has a wide spread of ages that cannot be explained as resulting from a Gaussian distribution around a single mean age. The histogram in Fig. 10b suggests that this group may in fact represent two subgroups, one with weighted mean centred at 20.5 _ 0.7 Ma (20-) and the other at 16.4 _+0.2 Ma (20-). A X2 test (a standard statistical test used to determine the likelihood that a group of measurements is derived from a parental distribution of measurements) for each subgroup suggests that whereas the younger subgroup is well described by a single age Gaussian distribution, the older subgroup is not and would need to be further divided. However, a closer look at the analyses of this subgroup (20.5 _ 0.7 Ma) shows that a few points have very high U content (up to 1.2%) and that the older ages within the subgroup are associated with the highest U content (insert in Fig. 10b; Table 2). This suggests that the age spread and the mean age of this subgroup might have been artificially raised by unwanted analytical effects (i.e. the instrument underestimated U content for high values). A n o t h e r point of concern regarding these two young subgroups is that their U/Pb ages could not be confirmed by Th/Pb ages, because of the large uncertainties owing to low Th content. A

string of ungrouped younger ages (down to 8 Ma old) suggest that at least some zircons or parts of zircons may have undergone Pb loss. Thus, whereas high U seems to explain the spread towards older ages of the 20.6 Ma subgroup, Pb loss may explain the spread towards lower ages of the younger subgroup. Despite these two effects, the strong grouping at c. 17 Ma and the group of low-U spots centred at c. 20 Ma still suggest two separate zircon growth phases. However, in view of the uncertainties, we choose to take the broader and safer interpretation that the partial melting of the 106 Ma orthogneiss occurred between 20 and 17 Ma, broadly contemporaneous with the age of crystallization of the intruding leucogranite (sample 215).

4~

dating

The 4~ method of isotopic dating has been applied to micas from the Tangtse leucogranite sample dated above and to the Ladakh granodiorite at Taruk. This method, which relies on the decay of 4~ to 4~ and the subsequent retention of this argon in the mica lattice, commonly gives the age of cooling. Micas undergo closure to diffusion loss of 4~ from the lattice at about 300-350~ for cooling rates of tens of ~ (+ 100~ depending on a number of kinetic factors (Dodson & McClelland-Brown

TRANSPRESSION ALONG THE K A R A K O R A M FAULT

321

Table 3. 4~ step-heating data for the Taruk granite biotite sample 450 and muscovite from the Tangtse leucogranite, sample 128

4~

37Ar/39Ar 36Ar/39Ar 39Ar Cumulative (10 -3) (10 -3) (10 -16 mol) 39Ar(%)

%4~

4~

Calculated age (Ma) _+1 SD

K/Ca

32.09 _+27.75 12.80 _+21.20 17.13 _+9.91 12.13 _+2.75 8.65 _+1.96 7.26 + 1.11 10.79 + 0.62 10.76 + 0.34 11.01 _+0.13 11.21 _+0.15 11.38 _ 0.08 11.22 + 0.05 11.35 _+0.04 11.32 _+0.07 11.31 _+0.03 11.28 _+0.11

3.3 9.3 60.2 7.7 9.3 12.6 22.6 58.1 138 244 141 184 225 140 41.2 101

Sample 450 biotite, Taruk granite, ANU No. 95450, 100-420 ~m, J = 0.0005449 • 0.4% 437.1 361.4 193.2 82.09 52.12 44.70 35.86 26.59 19.85 16.02 14.26 13.12 12.83 13.66 14.78

157.6 56.43 8.721 68.38 57.10 41.70 23.31 9.055 3.817 2.157 3.723 2.864 2.341 3.757 12.81

1367 1177 593.6 234.6 145.3 125.1 82.89 51.66 27.93 14.26 7.748 4.416 3.022 5.920 9.728

0.1456 0.1561 0.2517 0.4276 0.7582 1.453 2.782 4.971 8.307 12.05 16.12 34.76 33.65 21.20 166.0

0.05 0.10 0.18 0.32 0.57 1.05 1.97 3.61 6.35 10.3 15.7 27.1 38.2 45.2 100

7.5 3.6 9.1 15.1 16.9 16.6 30.7 41.3 56.6 71.4 81.4 87.3 90.2 84.6 78.1 Total

32.94 13.06 17.51 12.38 8.817 7.398 11.01 10.98 11.23 11.44 11.61 11.45 11.58 11.55 11.54 11.51

Sample 128 muscovite, Tangste leucogranite, ANU No. 96216,180-700 ~m, J = 0.0004896 • 0.4% 350.6 93.63 44.19 40.54 35.66 27.67 18.51 16.61 17.45 18.87 19.79 19.57 17.93 16.44 16.01 16.45 26.89

21.42 19.53 10.82 3.054 2.383 1.057 0.4508 0.1209 0.3936 0.5729 0.1314 0.0669 0.5363 0.9771 0.2069 0.8371 2.649

1124 269.1 102.5 91.80 75.71 48.54 17.92 11.63 14.12 18.81 22.38 20.98 15.69 10.65 9.029 10.12 45.65

0.8712 1.963 3.517 8.105 17.70 46.58 50.60 64.60 40.02 29.08 26.55 27.61 31.10 33.00 29.65 21.24 16.59

0.2 0.6 1.4 3.2 7.2 17.6 28.8 43.2 52.1 58.6 64.5 70.7 77.6 85.0 91.6 96.3 100

1985). P u r e separates of b o t h muscovite (>700 Ixm d i a m e t e r ) f r o m the Tangtse l e u c o g r a n i t e and biotite (180-350 fxm) f r o m the T a r u k granodiorite, SW of the K a r a k o r a m fault, w e r e i r r a d i a t e d a n d s u b s e q u e n t l y s t e p - h e a t e d in a v a c u u m f u r n a c e at progressively increasing t e m p e r a t u r e s (for analytical details, see D u n l a p et al. (1995)). T h e m u s c o v i t e y i e l d e d an age s p e c t r u m with a p l a t e a u at 11.4 + 0.1 M a and the T a r u k biotite yielded a similar p l a t e a u at 11.3 _+ 0.1 M a (Table 3, Fig. 11). D e s p i t e the small d i f f e r e n c e in closure t e m p e r a t u r e s b e t w e e n m u s c o v i t e and biotite, a

5.1 14.6 30.5 32.1 36.2 46.7 69.2 76.9 73.8 68.4 64.5 66.3 71.9 78.4 80.8 79.4 48.3 Total

17.96 13.71 13.49 13.01 12.89 12.94 12.81 12.77 12.88 12.91 12.77 12.97 12.89 12.89 12.94 13.05 13.00 12.90

15.79 + 7.52 12.07 + 1.19 11.88 + 0.42 11.46 + 0.29 11.35 + 0.15 11.39 + 0.07 11.28 ___0.05 11.24_ 0.03 11.34 + 0.04 11.36 ___0.06 11.25 _+0.08 11.42 _+0.08 11.35 _+0.06 11.35 _+0.05 11.39_+0.08 11.49 _+0.07 11.44 _+0.11 11.36 +_0.09

24.6 27.0 48.5 172 221 498 1167 4348 1339 917 4000 7874 980 1901 2545 629 199 2134

conservative i n t e r p r e t a t i o n is that b o t h samples c o o l e d t h r o u g h 300-350~ and w e r e at similar depths in the interval 11.4-11.3 M a ago.

Geological offsets and timing of initiation of the Karakoram fault Peltzcr & T a p p o n n i e r (1988) p r o p o s e d c. 1000 k m of dextral m o t i o n along the K a r a k o r a m fault, b u t S e a r l e (1996) d e m o n s t r a t e d t h a t t h e s e w o r k e r s w e r e linking up different granite belts that were never connected. The timing of

322

M.P. SEARLE E T A L . 12.o

12.0

Darbuk-Tangtse Muscovite 11.4_+0.1 Ma

11.5

11.5 Age (Ma) 11.0

11.0 10.5

10.5

10.0 0.0

0.2

0.4

0.6

0.8

4. 10.0 1.0

Fraction 39Ar Released

12.0

12.o

Taruk Biotite 11.3_+0.1 Ma 11.5

11.5

Age (Ma) 11.0

11.0

10.5

10.5

10.0 0.0

0.2

0.4

0.6

0.8

10.0 1.0

Fraction 39Ar Released

Fig. 11. Results from 4~ step-heating experiments on muscovites from the mylonitic Tangtse leucogranite (Karakoram terrane) and on biotites from the Taruk granodiorite (Ladakh batholith). initiation of the Karakoram fault is crucial to determining if the fault can be related to the indentation of India into Asia and crustal extrusion of the thickened Tibetan plateau. If the fault was a major boundary of the eastwards extruding Tibetan crust, formed as a result of the India-Asia collision, then it would be expected that the fault initiated around the time of collision some 50 Ma ago. However, the Karakoram fault clearly cuts several geological features of widely ranging ages by a similar amount (c. 150 km). These include the late Cretaceous to early Eocene Ladakh granite batholith, the Miocene Baltoro and Tangtse granites of the Karakoram batholith, and the course of the Indus river (Searle 1996). Granites that were formed only c. 21 Ma ago in the Baltoro region (Parrish & Tirrul 1989) and c. 18.0 + 0.6 Ma at Tangtse in northern Ladakh (this paper) have clearly been truncated by the later Karakoram fault. No two-mica garnet-bearing leucogranites occur SW of the Karakoram fault south of point (a) in Fig. 2.

These granites are very distinct mineralogically and geochemically from the biotite + hornblende-bearing granodiorites of the Ladakh range to the south. North of the Baltoro granite plutons in northern Karakoram and Pamir there are no similar leucogranites to the Baltoro pluton exposed. Correlation of the Tangtse granite at point 'a" in Fig. 2 with the Baltoro granite is therefore almost impossible to deny on geological evidence. The post-magmatic, dextral mylonitic S-C fabrics superimposed on the Tangtse granite along the southwestern strand of the Karakoram fault indicate that motion along the Karakoram fault in this area must have occurred after intrusion and crystallization of the Tangtse leucogranite. We suggest that the fault was initiated soon after the intrusion of the Tangtse leucogranite, and dextral slip associated with movement along the fault imposed S-C fabrics on the granite in the greenschist facies. What is clear is that the fault was not initiated as a direct result of the Indian plate collision at c. 50 Ma; it is much more likely that it was initiated as a result of late Neogene crustal shortening in the Pamir indenter to the west (Searle 1996). This late-stage shortening in the Pamir has resulted in clockwise rotations in western Tibet, which are also reflected in the block rotations along the Karakoram fault (Armijo et al. 1989; Searle 1991).

Rates of dextral shear, cooling and exhumation Maximum offsets of pre-Holocene geological features along the Karakoram fault are here determined to be 150 km right-lateral. Distinctive garnet-two-mica leucogranites of the Baltoro plutonic unit have been offset by this amount from the K12-Saltoro Kangri range in Pakistan to the Darbuk-Tangtse area in northern Ladakh. The Baltoro leucogranites and the Tangtse leucogranites have yielded early to midMiocene U-Pb zircon ages and have similar mineralogy and geochemistry, suggesting that they belong to the same plutonic unit and were subsequently separated by movement along the Karakoram fault. The average dextral strike-slip rate since 18.0 Ma was therefore 8.3 mm/a. 4~ results for the micas indicate that by 11.3 + 0.1 Ma rocks on both sides of the Darbuk-Tangtse strand of the Karakoram fault were at similar temperatures, suggesting that uplift across the fault at this time was not rapid enough to induce significant thermal perturbations (e.g. differential cooling), and that thermal perturbation produced by rapid uplift since 18

TRANSPRESSION A L O N G THE K A R A K O R A M FAULT

Ma had largely relaxed. Exhumation of the deep crustal migmatites and leucogranites within the Karakoram fault zone must therefore have occurred between 18.0 and 11.3 Ma, when these rocks cooled from >700 to 350~ The cooling rate from 18.0 to 11.3 Ma was therefore c. 50~ We suggest that exhumation, erosion and unroofing of these migmatites and leucogranites within the fault zone occurred during this time; 11.3 Ma probably reflects the latest time at which the central part of the Karakoram fault changed from dominantly transpressional to dominantly strike-slip motion. Along the entire Karakoram fault, maximum exhumation of deep crustal metamorphic and migmatitic rocks occurs in the D a r b u k Tangtse-Pangong area. To the NW and SE of this area, only upper-crustal shallower level rocks are exposed along the fault. At Tangtse, lineations consistently plunge at 20 ~ towards the NW. Elsewhere in the Darbuk to Pangong Lake sector, lineation plunge varies from 0 to 30 ~ Given the measured 20 ~ plunge in lineations along the Darbuk-Tangtse sector of the Karakoram fault, we can calculate the amount of uplift of the rock. At an average slip rate of 8.3 mm/a the dextral offset along the fault from 18.0 to 11.3 Ma would be c. 56 km. Assuming 56 km of dextral slip in 6.7 Ma at slip rates of 8.3 mm/a and a 20 ~ plunge in lineations, the calculated uplift of the migmatites and leucogranite at Tangtse (tan 20 ~ • 56) is 20.4 km at a mean rate of 3.0 mm/a. Transpression and transtension Although we argue here that the major phase of transpressive exhumation occurred in the Tangtse region between 18.0 and 11.3 Ma, there is evidence from along the Karakoram fault elsewhere of some later Pliocene-Quaternary transpression. Fission-track thermochronology from the K2 gneiss indicates that the K2-Gasherbrum range in north Pakistan has been uplifted by transpression within the last 5-7 Ma (Foster et al. 1994). The Pangong range in northeast Ladakh has also apparently been uplifted by transpression in Quaternary times, as shown by the morphology of hanging glaciers truncated by the fault scarp. At the same time, further SE along the Karakoram fault, the Gar basin opened by transtension (Armijo et aL 1989). Whereas the central part of the Karakoram fault accommodated most of the dextral slip, the northwest (Pamir) and southeast (Kailas) segments show dominantly dip-slip motion with Quaternary transtensional basins accumulating sediment eroded from the adjacent mountain ranges. As India continued to penetrate northwards

323

into Asia, Tibetan crust continued to thicken until the change from north-south compression to east-west extension. A series of ten N-S aligned graben systems in south Tibet are connected to N W - S E aligned dextral strike-slip faults, indicating that the two sets of faults may be temporally as well as spatially connected. Only two of the north-south normal faults have been dated by 4~ on muscovites, at c. 14 Ma in the Thakhola graben, northeast of Annapurna in north Nepal (Coleman & Hodges 1995) or c. 7-8 Ma in central Tibet (Harrison et al. 1992, 1995; Pan & Kidd 1992). We suggest that the timing of transpressional exhumation in the Tangtse sector of the Karakoram fault (18.0-11.3 Ma) immediately preceded initiation of the N-S normal faults in south Tibet and the beginning of E - W extension of the Tibetan plateau. Tibet: to extrude or n o t to extrude - that is the q u e s t i o n Although strike-slip faults in and around the Tibetan plateau are some of the most impressive geological features, they do not show the very large amounts of offset that have previously been proposed (e.g. Peltzer & Tapponnier 1988). The Karakoram fault and the Altyn Tagh fault are undoubtedly extremely active faults today, cutting glacial features, truncating river terraces, changing the course of rivers and bounding rapidly uplifting mountain ranges. However, the estimation of Holocene slip rates cannot be determined with accuracy without precise dating of the offset geomorphological features, and present-day slip rates can certainly not be extrapolated back in the past beyond the Holocene. The pattern of strike-slip faults in Asia does suggest that some eastward motion of thickened Tibetan crust has occurred, but the amount of lateral extrusion can only be of the order of 100-200 km and not of the order of >1000 km demanded by the models of Tapponnier and co-workers. If it was of this order of magnitude the bounding strike-slip faults should show similar amounts of geological displacements. Our work along the Karakoram fault, combined with detailed mapping and geochronology in the Pakistan sector (summarized by Searle (1991)) shows that, at least for the Karakoram fault, geological offsets since 18.0 Ma are no more than 150 km at mean slip rates of 8.3 mm/a, suggesting that crustal thickening of the Tibetan plateau was more important than lateral extrusion. Strain along the Karakoram fault was partitioned into predominantly dip-slip and

324

M.P. SEARLE ETAL.

transtensional tectonics in the N W (Pamir) and SE (Gar to Mt Kailas) terminations and predominantly strike-slip with transpressionally uplifted mountain ranges in the central part of the fault in the K 2 - G a s h e r b r u m range, Siachen glacier region and Tangtse area of northern Ladakh.

Conclusions (1) The initiation of the Karakoram fault cannot have been older than 18.0 Ma, because it cuts and offsets granites of that age. The Karakoram fault cannot therefore be directly related to the 60-50 Ma India-Asia collision, which is much older. It is probably more likely to be a result of crustal shortening in the Pamir indentor and clockwise rotations of western Tibet. (2) The maximum dextral offset along the Karakoram fault is 150 km based on offsets of the Baltoro-type granites in the Karakoram and the Ladakh batholith, and the offset course of the antecedent Indus river. (3) Average dextral slip rates along the Karakoram fault since 18.0 Ma were 8.3 mm/a, approximately one-quarter of the proposed Holocene dextral slip rates proposed by Liu et al. (1992) and Avouac & Tapponnier (1993). (4) Crustal melt garnet-two-mica leucogranites and migmatites were formed at 20-18 Ma at depths of >20 km in the Pangong range between the two strands of the Karakoram fault and were then exhumed by transpressional uplift from 18.0 to 11.3 Ma at rates of 3.0 mm/a. (5) Maximum exhumation along the Karakoram fault occurred in the Tangtse area, where deep crustal rocks (Pangong m e t a m o r p h i c complex) are exposed within the fault zone. Northwest (Pamir and Shaksgam) and southeast (SW Tibet) of this area only upper-crustal rocks are exposed along the fault zone. (6) 4~ dating of micas shows that by 11.3 + 0.1 Ma ago temperatures of rocks on opposite sides of the Karakoram fault (but now at the surface 2 km apart) were similar, indicating an end of relative tectonic e x h u m a t i o n (uplift of the rock) by that time. (7) Deformation was partitioned into a dominantly transpressional phase from 18.0 to 11.3 Ma, w h e n e x h u m a t i o n of the P a n g o n g migmatites and crustal melt granites within the Karakoram fault zone occurred, and a dominantly strike-slip phase starting from 11.3 Ma. The northwestern segment of the fault in the Pamir and the southeastern termination near Mt Kailas in southwest Tibet show dominantly dipslip motion and Quaternary transtensional pullapart basins.

(8) The relatively minor total offset along the K a r a k o r a m fault, combined with the age of initiation of dextral slip and the 8.3 mm/a slip rates suggest that the Karakoram fault has not accommodated the amount of eastward extrusion of the Tibetan plateau required by the extrusion model. The timings and amounts of exhumation, both in the Baltoro region and the Tangtse region, suggest that Asian crust was being actively t h i c k e n e d during the middle Miocene. This work was supported by NERC Grant GT5/96/13/E to M. P. Searle. M. P. S. acknowledges with gratitude many discussions with P. Molnar, P. Tapponnier and J.-P. Avouac, even though they may not agree with our model. R. F. W. is very grateful to W. Compston for assistance with the SHRIMP analyses at the ANU and data reduction. R. F. W. and W. J. D. thank D. Green, R. Griffiths and I. McDougall for support. We thank the Australian Nuclear Science and Technology Organisation and the Australian Institute for Nuclear Science and Engineering for irradiations. We are extremely grateful to R. Parrish for a detailed review on the manuscript and discussions. We are also grateful to Fida Hussein Mitoo of Leh, D. Willis and J. Bradley for assistance in the field.

References ARMIJO,R.,TAPPONNIER,P. & HANTONGLIN1989. Late Cenozoic right-lateral strike-slip faulting across southern Tibet Journal of Geophysical Research, 94, 2787-2838. AVOUAC, J.-P. & TAPPONNIER, P. 1993. Kinematic model of active deformation in Central Asia. Geophysical Research Letters, 20, 895-898. BRUNEL, M., ARNAUD, N., TAPPONNIER, P., PAN,Y. & WANG,Y. 1994. Kongur Shan normal fault: type example of mountain building assisted by extension (Karakoram Fault, eastern Pamir). Geology, 22, 707-710. CERVENY,P. F., NAESER,C. W., KELEMEN,P. B. LIEBERMAN, J. E. & ZEITLER, P. K. 1989. Zircon fissiontrack ages from the Gasherbrum Diorite, Karakoram range, northern Pakistan. Geology, 17, 1044-1048. CLAOUt~-LONG, J. C., COMPSTON,W., ROBERTS, J. & FANNING,C. M. 1995. Two Carboniferous ages: a comparison of SHRIMP zircon dating with conventional zircon ages and 4~ analysis. In: BERGGREN, W. A. ET AL. (eds) Geochronology Time Scales and Global Stratigraphic Correlation. Society for Sedimentary Geology Special Publication, 4, 3-21. , JONES, P. J., ROBERTS,J. & MAXWELL,S. 1992. The numerical age of the Devonian-Carboniferous boundary. Geological Magazine, 129, 281-291. COLEMAN, M. & HODGES, K. 1995. Evidence for Tibetan plateau uplift before 14 Ma ago from a new minimum estimate for east-west extension. Nature, 374, 49-52.

TRANSPRESSION ALONG THE K A R A K O R A M FAULT

COMPSTON,W., WILLIAMS, I. S. & MEYER, C. 1984. Geochronology of zircons from lunar breccia 73217 using a sensitive high mass-resolution ion microprobe. Journal of Geophysical Research, 8 9 , B525-B534. CRAWFORD, M. B. • SEARLE, M. P. 1992. Field relations and geochemistry of pre-collisional (India-Asia) granitoid magmatism in the central Karakoram, northern Pakistan. Tectonophysics, 206, 171-192. DEBON, F. 1995. Incipient India-F.urasia collision and plutonism: the Lower Cenozoic Batura granites (Hunza Karakorum, North Pakistan). Journal of the Geological Society, London, 152, 785-795. , ZIMMERMAN,J. L. ~; LEFORT, P. 1996. Upper Hunza granites (North Karakorum, Pakistan): a syn-collision bimodal plutonism of mid-Cretaceous age. Comptes Rendus de l'Acaddmie des Sciences, 323, 381-388. DODSON, M. H. & MCCLELLAND-BROWN,E. 1985. Isotopic and palaeomagnetic evidence for rates of cooling, uplift and erosion. In: SNELLING,N. J. (ed.) Geochronology and the Geologic Record Geological Society, London, Memoirs, 10, 315-325. DUNLAP, W. J., TEYSSIER, C., McDOUGALL, I. & BALDWIN, S. 1995. Thermal and structural evolution of the intracratonic Arltunga nappe complex, central Australia. Tectonics, 14, 1182 1204. FOSTER, D. A., GLEADOW,A. J. W. 8r MORTIMER, G. 1994. Rapid Pliocene exhumation in the Karakoram (Pakistan), revealed by fission track thermochronology of the K2 gneiss. Geology, 22, 19-22. GILLESPIE, A. & MOLNAR, P. 1995. Asynchronous maximum advances of mountain and continental glaciers. Reviews of Geophysics, 33(3) 311-364. HARRISON, T. M., COPELAND,P., KIDD,W. S. F. & AN YIN 1992. Raising Tibet. Science, 255,1663-1670. & LOVERA, O. M. 1995. Activation of ;he Nyainqentanghla Shear Zone: implications for uplift of southern Tibetan Plateau. Tectonics, 14, 658-676. LIu, Q., AvOuAC, J.-P., TAPPONNIER, P. 8r ZHANG, Q. 1992. Holocene movement along the southern part of the Karakoram Fault. Abstracts, Inter-

national Symposium on the Karakoram and KunLun Mountains, Kashgar, China, 91. MEYER, B., TAPPONNIER,P., GAUDEMER,Y., PELTZER, G., SHUNMIN,G. & ZHITAL, C. 1996. Rate of leftlateral movement along the easternmost segment of the Altyn Tagh fault, east of 96~ (China). Geophysical Journal International, 124, 29-44. MOLNAR,P. & TAPPONNIER,P. 1975. Cenozoic tectonics of Asia, effects of a continental collision. Science, 189, 419-426. PAN, Y. & KIDD, W. S. F. 1992. Nyainqentanghla Shear Zone: a late Miocene extensional detachment in southern Tibet plateau. Geology, 20, 775-778. PARRISH, R. R. & TIRRUL, R. 1989. U-Pb age of the Baltoro granite, northwest Himalaya, and implications for zircon inheritance and monazite U-Pb systematics. Geology 17, 1076-1079.

325

PELTZER, G. Be:TAPPONNIER,P. 1988. Formation and evolution of strike-slip faults, rifts and basins during the India-Asia collision: an experimental approach. Journal of Geophysical Research, 93, 15085-15117. , & ARMIJO, R. 1989. Magnitude of Late Quaternary left-lateral displacements along the north edge of Tibet. Science, 246, 1285-1289. REx,A. J., SEARLE,M. P., TIRRUL,R., CRAWFORD,M. B., PRIOR, D. J., REX, D. C. & BARNICOAT,A. 1988. The geochemical and tectonic evolution of the central Karakoram, North Pakistan. Philosophical Transactions of the Royal Society, London, A326, 229-255. SCHARER, U., COPELAND, P., HARRISON, T. M. & SEARLE,M. R 1990. Age, cooling history and origin of post-collisional leucogranites in the Karakoram batholith; a multi-system isotope study N. Pakistan. Journal of Geology, 9 8 , 233-251. SEARLE, M. P. 1991. Geology and Tectonics of the Karakoram Mountains. John Wiley, Chichester. 1996. Geological evidence against large scale preHolocene offsets along the Karakoram Fault: implications for the limited extrusion of the Tibetan Plateau. Tectonics, 15, 171-186. --, CRAWFORD, M. B. & REX, A. J. 1992. Field relations, geochemistry, origin and emplacement of the Baltoro granite, Central Karakoram.

Transactions of the Royal Society, Edinburgh, Earth Sciences, 83, 519-538. --,

PARRISH, R. R., TIRRUL, R. & REX, D. C. 1990. Age of crystallization and cooling of the K2 gneiss in the Baltoro Karakoram. Journal of the Geological Society, London, 147, 603-606. - - , REX, A. J., TIRRUL, R., REX, D. C., BARNICOAT,A. & WINDLEY,B. E 1989. Metamorphic, magmatic and tectonic evolution of the Central Karakoram in the Biafo-Baltoro-Hushe regions of North Pakistan. Geological Society of America, Special Paper, 232, 47-73. TAPPONNIER, P. & MOLNAR, P. 1976. Slip-line field theory and large scale continental tectonics. Nature, 264, 319-324. - & -1977. Active faulting and tectonics in China. Journal of Geophysical Research, 82, 2905-2930. - - , PELTZER,G. & ARMIJO, R. 1986. On the mechanics of the collision between India and Asia. In: COWARD,M. P. & RIES, A. C. (eds) Collision Tectonics. Geological Society, London, Special Publications, 19, 115-157. - - , LEDAIN,A. Y., ARMIJO,R. & COBBOLD,P. R. ]982. Propagating extrusion tectonics in Asia: new insights from simple experiments with plasticine. Geology, 10, 611-616. TAYLOR, S. R. & MCLENNAN,S. M. 1985. The Conti-

nental Crust: It's Composition and Evolution. Blackwell Scientific Publications, Oxford. WEINBERG, R. E & SEAm-E, M. P. 1998. The Tangtse Injection Pangong Complex, Indian Karakoram: a case of pervasive granite flow through hot viscous crust. Journal of the Geological Society, London, in press. WILLIAMS,I. S. & CLAESSON,S. 1987. Isotopic evidence for the Precambrian provenance and Caledonian

326

M.P. SEARLE E T A L .

metamorphism of high grade paragneisses from the Seve Nappes, Scandinavian Caledonides, II, Ion microprobe zircon U-Th-Pb. Contributions to Mineralogy and Petrology, 97, 205-217. ZEITLER, P. K., SUTFER,J. E, WILLIAMS,I. S., ZARTMAN,

R. & TAHIRKHELI,R. A. K. 1989. Geochronology and temperature history of the Nanga ParbatHaramosh Massif, Pakistan. Geological Society of America Special Publications, 232, 1-22.

Transpressional kinematics and magmatic arcs M I C H E L D E S A I N T B L A N Q U A T 1, B A S I L T I K O F F 2, C H R I S T I A N T E Y S S I E R 3 & JEAN LOUIS VIGNERESSE 4

1 C N R S - UMR5563, Laboratoire de POtrophysique et Tectonique, UniversitO PaulSabatier, 38 rue des 36-Ponts, 31400 Toulouse, France (e-mail: michel@lucid, ups-tlse.fr) 2Department o f Geology and Geophysics, Rice University, Houston, T X 77005, USA (e-mail: btikoff@geophysics, rice. edu) 3Department of Geology and Geophysics, University of Minnesota, Minneapolis, M N 55455, USA (e-mail: [email protected]) 4CREGU, 54501 Vandoeuvre, Nancy Cedex, France (e-mail: [email protected]) Abstract: Most continental magmatic arcs occur in obliquely convergent settings and display strike-slip movement within, or adjacent to, the magmatic arc, and contractional structures in the forearc and backarc regions. Thus, three-dimensional transpressional kinematics typifies many arc settings, both modern and ancient. Intrusions cause magma-facilitated strike-slip partitioning, even in cases where the relative angle of plate convergence is almost normal to the plate boundary. Transpressional systems are preferentially intruded by magmas because of the steep pressure gradients in vertical strike-slip shear zones and their ability to force magma upward. Both buoyancy and transpressional dynamics cause a component of magma overpressuring, which in turn expels granitic magma upward following the vertical pressure gradient. The tectonic and magmatic processes are linked in a positive feedback loop which facilitates the upward movement of magma. We propose a lithospheric-scale, three-dimensional model of transpressional arc settings. Strike-slip motion is partitioned into the magmatic arc settings because of the linear and margin-parallel trend of the vertical, lithospheric-scale weakness caused by ascending magma. The parallelism of contraction structures in the forearc and backarc regions is caused by mechanical coupling through the lower crust and upper lithospheric mantle. The displacement field of the basal layer of the arc system provides the boundary condition for the upper-crustal, strike-slip partitioned deformation.

In oblique subduction, a common observation is that some transcurrent motion is accommodated within the magmatic arc (e.g. Fitch 1972; Jarrard 1986). Strike-slip faulting occurs in the volcanic edifice of many magmatic arcs (Beck 1983). For example, the central A n d e s (Soulas 1977; Megard 1987), Sumatra (Fitch 1972; McCaffrey 1988; Bellier & Sebrier 1994), Japan (Matsuda et al. 1967; Lallemand & Jolivet 1986), and New Zealand (Cashman et al. 1992) are all active magmatic arcs where strike-slip deformation is observed at the surface of the Earth. Deeper crustal levels in eroded magmatic arcs also demonstrate strike-slip tectonism, despite the tendency for magma to obscure structural relations. These ancient magmatic arcs include the Andes (Petford & A t h e r t o n 1992), north Cascades (Brown & Talbot 1989), I d a h o Batholith (Lund & Snee 1988), and the Sierra Nevada (Busby-Spera & Saleeby 1990; Tikoff & Teyssier 1992). Strike-slip tectonism associated with oblique subduction is facilitated by two obvious factors: (1) the linear, margin-parallel trend of the magmatic arc, and (2) the high heat

flow in the arc because of magma transport, which induces a zone of weakness across the whole continental lithosphere (e.g. Jarrard 1986). In addition to strike-slip tectonism, a comp o n e n t of contraction is often seen across continental magmatic arcs. An arc-normal contractional component is well documented in the Andes (Suarez et aL 1983; Megard 1987) and Sumatra (McCaffrey 1991; M o u n t & Suppe 1992); as well as inferred for many Mesozoic arcs, the Coast Belt (Crawford et al. 1987; Rusmore & Woodworth 1991), north Cascades (Brown & Walker 1993), Idaho Batholith (Manduca el al. 1993), Sierra Nevada (Tikoff & Saint B l a n q u a t 1997), and A n d e s (Megard 1987). The combination of strike-slip and arcnormal contractional tectonism suggests that a transpressional setting is a common setting for magmatic arcs, although other possibilities clearly exist (Grocott et al. 1995; Tobisch et al. 1995). Despite the evidence for three-dimensional d e f o r m a t i o n within obliquely convergent

SAINTBLANQUAT,M., TIKOFF,B., TEYSSIER,C. d~;VIGNERESSE,J. L. 1998. Transpressional kinematics and magmatic arcs. In: HOLDSWORTH,R. E., STRACHAN,R. A. • DEWZY,J. E (eds) 1998. Continental Transpressionaland TranstensionalTectonics.Oeological Society, London, Special Publications, 135, 327-340.

327

328

M. DE SAINT BLANQUAT E T A L .

settings, the conventional view of continental magmatic arcs remains overwhelmingly two dimensional. The aim of this paper is two-fold: (1) to illustrate the connection between transpressional d e f o r m a t i o n and magmatic processes, and (2) to propose a lithospheric model that addresses the inherent three-dimensional nature of continental magmatic arcs.

Transpression and exhumation The three-dimensional kinematics of magmatic arcs may explain some long-standing problems concerning the wallrock fabrics and upward ascent of magmas within arc settings. For instance, transpressional kinematics may create crustal thickening and lead to granite generation (Hutton & Reavy 1992). In a magmatic arc setting, granites are generated by the geodynamic processes related to oceanic plate subduction and partial melting of the upper mantle and lower continental crust. Thus, magma generation presumably occurs regardless of the particular kinematic framework. However, crustal thickening associated with transpressional deformation would influence the depth at which magmas react with the base of the crust. Perhaps more importantly, transpressional tectonics may explain the vertical movement of the crust which accompanies arc magmatism. Let us consider, for example, a traverse across the central section of the Cretaceous Sierra Nevada batholith (Fig. 1; e.g. Bateman 1992). Ague & Brimhall (1988), in a barometric study, found that the oldest (120 Ma) part of the batholith, along the eastern edge, intruded at depths of c. 4 kbar. The youngest part (90 Ma) of the batholith intruded at c. 1 kbar. Thus, within the Sierra Nevada magmatic arc setting, there was e x h u m a t i o n during arc magmatism, a common observation in other magmatic arcs (Read 1957). The lack of any later structure to cause this pattern of uplift suggests that it occurred during arc magmatism. A similar conclusion of uplift during magmatism was reached by Pickett & Saleeby (1993) for the southern Sierra Nevada batholith. The efficiency of transpressional deformation at causing vertical movement is shown by the flow lines of homogeneous transpression (Fig. 2). Upward movement of particles is due to the pure shear component of deformation, itself a consequence of the fixed lower-boundary and free upper-boundary condition. In the horizontal plane of transpression, material points move in straight lines toward the plate margins but at any angle from 0~ to 90 ~ (Fossen et al. 1994). A more sophisticated model of 'bulk'

Fig. 1. Simplified geological map of the central Sierra Nevada batholith. The youngest plutons are located on the eastern margin of the batholith. The axis of magmatism of the Cretaceous batholith progresses from west (c. 120 Ma) to east (c. 80 Ma) and corresponds to a change in emplacement depth (from 4 kbar to 1 kbar). TIS, Tuolumne intrusive suite; MPIS, Mono Pass intrusive suite; MWIS, Mount Whitney intrusive suite. transpression, which fulfils compatibility constraints of rigid walls (i.e. no-slip boundary conditions), has been given by Robin & Cruden (1994). Their model predicts essentially the same deformation pattern: rapid uplift and distributed horizontal shearing. Transpressional kinematics may explain the common observation of vertically foliated and vertically lineated wallrocks in magmatic arcs (e.g. Sams & Saleeby 1988; Tikoff & Greene 1997). The existence of vertical foliation and vertical lineation in wallrocks at lower batholithic levels is commonly cited as evidence for vertical return-flow during magma migration (e.g. Saleeby 1990). However, these fabrics are also consistent with numerical models and geological observations in other transpressional settings (Fig. 3; e.g. Hudleston et al. 1988; Fossen et al. 1994; Robin & Cruden 1994). Most kinematic

MAGMATIC ARCS

329

Fig. 2. Flow lines of homogeneous transpression showing efficiency of vertical movement. (a) Flow lines in the horizontal plane. The axes are the flow apophyses of movement (fa), which are not mutually orthogonal. (b) Flow lines in the vertical plane. Upward movement of the particles is due to the pure shear component of deformation. (c) Three-dimensional flow lines resulting from the fixed lower-boundary and free upperboundary conditions. analyses in the last 20 years have focused on the so-called movement plane, parallel to lineation and perpendicular to foliation. This two-dimensional approach has failed to recognize the potential role of transpression in the development of regional deformation patterns. Bulk transpressional deformation is accomplished by moving the segregated magma to higher crustal levels, providing vertical growth required by transpressional kinematics. This type of emplacement is inferred for the Ox Mountain igneous complex (McCaffrey 1992). The melt is moved to a higher crustal level during deformation, resulting in a 'magmatic' partitioning of the transpressional flow field (Fig. 3). This model relies on the observation that the flow pattern of homogeneous transpression in the horizontal plane is identical to the pattern of anisotropic area loss (Fossen & Tikoff 1993; Fossen et al. 1994). For homogeneous transpression, the area loss in the horizontal plane is only apparent as material is moved into the vertical direction by the flow (Figs 2 and 3). In a 'magmatic' partitioning case, at deep levels within the arc, the flow may approximate

anisotropic volume loss as magma segregates. This type of 'magmatic'-partitioned transpression may commonly occur in magmatic arcs, and

Fig. 3. Homogeneous and magmatic-partitioned transpression. In magmatic-partitioned transpression, the vertical growth required by transpressional kinematics is provided by the segregation and the upward movement of magma. The fabrics exhibited by wallrocks are consistent with transpression (vertical foliations, vertical to horizontal lineations).

330

M. DE SAINT BLANQUAT E T A L .

would tend to accentuate the vertical foliation and lineation of the wallrocks at lower crustal levels (Fig. 3).

Magma-facilitated strike-slip partitioning Case e x a m p l e : S u m a t r a

Sumatra is probably the best example of an active, obliquely convergent magmatic arc system (Fig. 4). The present rate of convergence is c. 70 mm/a between the Australian and Eurasian plates (DeMets et al. 1990). The Great Sumatran fault is parallel to and generally lies within the volcanic arc (Bellier & Sebrier 1994). Displacement on this fault increases from southeast Sumatra (50-70 km, 6 mm/a) (Lassal et al. 1989: Bellier et al. 1991) to the A n d a m a n Sea (450 km, 40 mm/a) (Curray et al. 1979), possibly because of a component of arc-parallel stretching (McCaffrey 1991). The forearc region displays SW-directed thrust faults and associated folds with axes oriented c. 15-20 ~ counter-clockwise to the plate margin (Karig et al. 1980: Mount & Suppe 1992). A kinematic model of strike-slip partitioning was applied to Sumatra (Tikoff & Teyssier 1994), showing that approximately one-third of the transcurrent motion is accommodated in the Great Sumatran fault,

Fig. 4. Map of the obliquely convergent magmatic arc system of Sumatra. The Great Sumatran and Mentawai faults record dextral motion. Orientations of borehole elongations, earthquake slip vectors, and fold axes in young sediments in both forearc and backarc show a consistent 15~ (or 75~ obliquity to the Great Sumatran fault, indicating mechanical coupling across the magmatic arc. From Tikoff & Teyssier (1994).

consistent with geodetic measurements in the region (Sieh 1994). The remaining transcurrent motion is accommodated by diffuse deformation or margin-parallel strike-slip zones (e.g. Mentawai fault; Diament et al. 1992). The shallow slip vectors in the forearc and backarc regions of Sumatra are neither parallel to the plate motion nor perpendicular to the plate boundary (Mount & Suppe 1992). Rather, folds and faults tend to have an en 6chelon orientation indicating a maximum horizontal infinitesimal contraction (commonly interpreted as stress) direction oriented at c. 75 ~ from the Great Sumatran fault (assumed parallel to the boundary of the deforming zone). This obliquity reflects a wrench component of deformation (see Teyssier & Tikoff this volume), and thus the forearc and backarc regions of Sumatra are accommodating wrench motion. The consistent orientation and obliquity of contractional structures in a deformation zone several hundred kilometres wide encompassing both sides of the Great Sumatran fault strongly suggests that the entire transpressional system is linked kinematically and forms a single dynamic system. Strike-slip partitioning

Fitch (1972) first recognized the three-dimensional structure of a continental magmatic arc above an obliquely subducting oceanic plate: a strike-slip fault system in or adjacent to the magmatic arc and a forearc region dominated by contractional deformation. However, within the last 25 years, two other aspects have been found to be characteristic of oblique subduction: deformation in the backarc as well as the forearc and a component of wrench deformation in both the forearc and backarc. The obliquity of these structures to the plate margin indicates distributed wrench deformation, as observed throughout Sumatra (Mount & Suppe 1992; Tikoff & Teyssier 1994). Therefore, to a first approximation, the kinematics of magmatic arcs is described by strike-slip partitioning (Tikoff & Teyssier 1994; Teyssier et al. 1995). In the absence of magmatism, strike-slip partitioning appears to cease at relatively low angles of convergence. Tikoff & Teyssier (1994) attributed the process of strike-slip partitioning to wrench-dominated transpression, which requires angles of convergence less than 20~. The physical experiments of Richard & Cobbold (1990) and the numerical models of Braun & Beaumont (1995) likewise suggested that strikeslip partitioning takes place at low angles of convergence. The tectonics of the South Island of New Z e a l a n d supports this interpretation.

MAGMATIC ARCS There, with a relative angle of convergence of 16--29~ (Walcott 1979; DeMets et al. 1990), no strike-slip partitioning occurs (Norris et al. 1990; Teyssier et al. 1995). The requirement of a low angle of plate convergence is not applicable to magmatic arcs. Sumatra, with a convergence angle of c. 50 ~ partitions approximately one-third of the translation component of plate motion onto slip on the Great Sumatran fault (Tikoff & Teyssier 1994). Slip within the Sierra Crest shear zone system occurs with a relative angle of convergence of 75-90 ~ (Kelley 1993; Kelley & Engebretson 1994). Strike-slip tectonism in the present Cascade arc of the western United States is another example. Weaver et al. (1987) interpreted magmatism under Mt Saint Helens to occur in a dilational jog within a strike-slip fault system. However, oblique subduction between the Juan de Fuca and North American plates is approximately c. 85~ (DeMets et al. 1990). Strike-slip partitioning occurs in other arcs also at higher angles of plate convergence: North Island of New Zealand (o~ = 70~ Walcott 1979) and Andes (c~ = 60-90~ Dewey & Lamb 1992). A vertical zone of lithospheric weakening (particularly of the upper mantle), caused by the rise of magmas, is well oriented to accommodate the tangential component of plate motion and forces strike-slip partitioning. In summary, in a magmatic arc setting, very little component of tangential plate motion is required for strike-slip tectonism to occur. These observations support the idea that magmatism facilitates partitioning of the plate motion into components of tangential and normal motion, even for almost normal subduction. Thus, in a transpressional magmatic arc, the kinematics and the magmatism appear to be intricately linked. As more magma is transferred and emplaced within the magmatic arc, localization of the transcurrent component of plate motion occurs into strike-slip fault zones within the arc. This strike-slip tectonism, in turn, provides room for the ascending magma.

Magma overpressuring: buoyancy and tectonic forces The exact reason why melt moves upward, particularly in an overall contractional arc setting, is a currently debated topic. Magma movement is governed by the magma driving pressure, which results from a combination of (Hogan & Gilbert 1995; Hutton 1997): (1) The density-contrast driven buoyancy, because of the lower density of granitic melt compared with average crustal densities; this component of the driving pressure

331

is always available, regardless of crustal connectivity; (2) the confining pressure, or lithostatic load, which acts as an effective driving force only if crustal connectivity exists; (3) the volume change during melting, acquired in the source region; (4) the viscous resistance to magma flow, which is strongly dependent on temperature; (5) the increase in pressure because of the increase of vapour pressure during the late stages of ascent. We are more concerned here with the first and second components, which result in a positive driving pressure as a result of gravitydriven buoyancy. Studies of melt movement in migmatites indicates that gravity-driven buoyancy forces are an important component of the driving forces (Burg & Vanderhaeghe 1993). The observation that magma reaches the surface, whatever the composition or the geodynamic context, has led many workers to propose that most magma is overpressured as a consequence of buoyancy (Lister & Kerr 1991; Parson & Thompson 1991; Clemens & Mawer 1992; Petford et al. 1993; Hutton 1997). Magma overpressuring is simply the idea that the magma pressure exceeds the lithostatic load. The overpressured magma will move toward the lowestpressure region by opening its own conduits, either by dyking or intruding pre-existing pathways, such as shear zones (strike-slip, normal, or thrust). This idea of magma overpressuring also explains the common observation of plutons emplaced in contractional settings, such as magmatic arcs (e.g. Guglielmo 1993; Ingram & Hutton 1994; Hutton 1997). It is routinely assumed that the cause of the magma overpressuring is primarily a buoyancy effect. However, an overlooked aspect of magmatic overpressuring is the role of a differential tectonic stress or 'tectonic overpressuring'. Tectonic overpressuring is simply the idea that the pressure induced by a tectonic deformation locally exceeds the lithostatic load and causes extrusion of material. This type of tectonic overpressuring has been invoked in subduction zone dynamics (Mancktelow 1995). If tectonic overpressuring acts on a magmatic system, it will strongly accentuate the magma overpressuring. Thus, magma overpressuring has two main components: buoyancy and tectonic pressures. As suggested by Robin & Cruden (1994), a dynamic effect induced by the boundary conditions of transpression is to cause a component of tectonic overpressuring. At shallow crustal levels, tectonic overpressuring in transpression results in extrusion in the form of flower structures (Mancktelow 1993). In a magmatic arc setting, this tectonic overpressuring may contribute directly to magmatic overpressuring. As

332

M. DE SAINT BLANQUAT ETAL.

the wall rocks push together, as a result of the 'press'-part of the transpression (Robin & Cruden 1994), they exert a force on the magma, which increases its pressure significantly: the horizontal tectonic load is partly transferred into a vertical driving force. Because this overpressuring is related to the 'press' rather than 'trans' part of transpression, the ability of the magma to withstand a shear stress is not of primarly importance. The magma, which is potentially already overpressured by buoyancy forces, becomes more so. Robin & Cruden (1994) noted that the transpression-induced overpressuring increases with the square of depth. Thus, a pronounced vertical gradient in tectonic overpressuring is created in the crust. This effect, combined with vertical lithostatic load, acts to effectively move granites upward in the crust (Fig. 5). The relative magnitudes of buoyancy forces v. tectonic forces leading to magmatic overpressuring is complicated by rheology, viscosity of melt, compressibility of country rock, drainage rate, and length/thickness ratio of the deformed zone. Thus, we emphasize that ours is only a

qualitative model. However, recognition that tectonic stress could be as high as 500 MPa (Bott & Kuznir 1984) suggests that the tectonic part of the magma overpressure could be significant. Both buoyancy and differential stress act to facilitate upward movement of granitic melt. Therefore, we envision a positive feedback loop between the two processes (Fig. 6). Whereas the buoyancy pressure is active in any tectonic setting, the tectonic pressure is limited to the appropriate tectonic setting (e.g. compression, transpression). The necessity of vertical pressure gradients moving overpressurized magma may also explain the common observation of magmatism associated with strike-slip, rather than thrust, movement. Strike-slip faults, simply by virtue of their vertical orientation, provide a higher pressure gradient (shortest way to a free boundary at the surface) and are therefore preferentially exploited by rising magmas. It is worth noting as well that, although the effect of transpression-induced overpressuring increases with the square of depth (Robin & Cruden 1994), the effect of buoyancy-induced magmatic overpressuring does not. Therefore, at progressively higher crustal levels, magmas have less upward force caused by transpression dynamics (Fig. 5). Rather, structural weaknesses, ballooning, or magma-induced tectonism (Tommasi et al. 1994) may control upper-crustal emplacement, even in transpressional zones. This reasoning may explain why large plutons are often associated with large shear zones, but not emplaced directly in them, such as the south Armorican shear zone (Vigneresse 1983) and sections of the Great Sumatran fault (Bellier & Sebrier 1994).

Implications for pluton emplacement

Fig. 5. Magma overpressuring resulting from both buoyancy and tectonic forces. The transpressional dynamics induce a tectonic overpressuring which is proportional to the viscosity, strain rate, and the square of the depth/thickness ratio of the deformed zone (Robin & Cruden 1994). This tectonic pressure, in addition to buoyancy forces, causes magma overpressuring and upward-directed magma transport. The buoyancy forces are relatively constant with depth, although the transpressioninduced forces decrease toward the free, upper surface.

The old concept of upward movement of granitic magmas as diapirs originated the idea of emplacement. At some point in their ascent, these diapirs stop because of neutral buoyancy, or became trapped or 'emplaced' at a particular place in the upper crust. The increased recognition of dyking as a viable mechanism for magma ascent (e.g. Petford et al. 1993) and the interaction of tectonism and magmatism (e.g. Karlstrom 1989; Hutton 1997) suggests that 'emplacement', i.e. upward or lateral movement of a plutonic-sized volume of coherent magma, may be rare. Rather, pulses of magma aggregate into coherent plutonic bodies during pluton construction. In other cases, what we call 'emplacement', for small granitic bodies (<1 km 2) or dykes, consists of magmas that 'froze' at some point in their upward ascent (see, e.g. Castro et al. 1995).

MAGMATIC ARCS .....................

Dsg~m!.~176

333

L L.~176 .................................

[

.~t':'. ~.m.a..t!.~176176176

Magmatransport (ascent/emplacement) Ik ~

Steeppressure 1 . . Magmatic Idl.~--...-_ graaients overpressure]" ~ J Horizontaltectonicload t causesverticaldrivingforcef magmainduced fi de~ormat,on,~,,~ Tectonic / i ~ " ~ I

........

J ~ - thermallyweakened verticalzone,

tectonc"space" . . ' ~..[ ~_StriKe-slip :~.~.___.._~ partitioning] ~ " A i i

/

] /

mranspressional kinematics J Fig. 6. Interactions between tectonic and magmatic processes in a magmatic arc setting. The first feedback loop describes the dynamic interaction between the two main components of magma overpressuring, the tectonic-induced and the buoyancy-induced forces, both acting to facilitate upward movement of magma. The second feedback loop relates the kinematics. Magma is transferred and emplaced within the magmatic arc, thereby localizing the transcurrent component of plate motion into strike-slip fault zones. The strike-slip tectonism, in turn, provides room for the ascending magma and causes a steep vertical pressure gradient. If the concept of magma overpressuring is correct (e.g. Hutton 1997), buoyancy is not necessarily the primary reason for the cessation of upward movement. Rather than magmas rising to a point of neutral buoyancy, magmas rise until they are capable of deforming or moving the wallrocks. Magmas may simply push rocks out of their way, using the free-boundary condition at the Earth's surface (e.g. roof raising) or lateral ballooning. At this point, pluton construction begins by aggregation of overpressurized magmas. This idea of pluton construction explains why granites commonly appear as 'tectonically controlled'. If local emplacement of overpressurized magma is distributed into a regional deformation (Hutton, 1997), no distinction can be made between an emplacement rate and a tectonic rate; yet it is critical to realize that this tectonism may not occur without magmatism. Using an automotive analogy, a pre-existing void in the crust is similar to having a reserved parking space in Toulouse: no such thing exists (to our knowledge). Parking places do not exist in Toulouse, they are created by the person who needs to park. Likewise, tectonic voids do not exist in the crust before granites intrude. Rather, granitic bodies are opportunistic: they assist regional tectonics in creating space, pushing

away their margins and moving to lower-pressure regions. This is the concept of magmaenhanced regional deformation (Tommasi et al. 1994; Neves et al. 1996): the magmas actually create a regional deformation (e.g. Hutton 1997). Thus, the effects of tectonically controlled magmatism and magma-controlled deformation are not only indistinguishable, but ultimately the same phenomenon.

Application to the Sierra Nevada batholith The Late Cretaceous Sierra Nevada is an ancient magmatic arc, which preserves a record of both strike-slip and contractional tectonism (Fig. 1). In this paper, we limit our discussion to the latest magmatic events in the Cretaceous Sierra Nevada batholith, which resulted in nested 'Intrusive Suites' (Tuolumne, Whitney, Mono Pass) in the eastern part of the batholith. Plutons emplaced between 92 and 83 Ma indicate intrusion at upper-crustal depths (a few kbar) and account for a large volume of magmatism (about 105 kin3), c. 20% of the currently exposed batholith. Several lines of indirect evidence suggest that large, intrabatholith, strike-slip displacements affected the Sierra Nevada. The Sr i = 0.706 line (isopleth of 87Sr/86Sr initial ratio) is offset by

334

M. DE SAINT BLANQUAT ETAL.

several arc-parallel, NW-SE trending faults within the Sierran block, amounting to c. 200 km of dextral displacement during the Early to Late Cretaceous (Kistler 1990). Lahren et al. (1990) suggested 400 km of dextral offset across the Sierra Nevada on the basis of stratigraphic correlation of the Snow Lake pendant. In addition, direct observation of shear zones occurring along the eastern crest of the Sierra Nevada also suggests significant strike-slip displacement. The Sierra Crest shear system (Fig. 1) is a 220 km long continuous shear zone, composed of a series of shear zones that were all active during the emplacement of the youngest plutons (Greene & Schweickert 1995; Tikoff & Saint Blanquat 1997). In the Kern Canyon area, southern Sierra Nevada, the proto-Kern Canyon fault exhibits dextral, ductile motion and shallowly plunging mineral lineations. This shear zone system may be connected to the deformation zone (White Wolf-Breckenridge fault) in the southernmost Sierra Nevada adjacent to the Rand Schist (Busby-Spera & Saleeby 1990; Pickett & Saleeby 1993). Further north, the Rosy Finch shear zone is 80 km long, and displays ductile, dextral deformation (Tikoff & Teyssier 1992; Tikoff & Saint Blanquat 1997). The Rosy Finch shear zone is in structural continuity with the Gem Lake shear zone (Greene & Schweickert 1995) and Cascade Lake shear zone (Davis 1996; Tikoff & Greene 1997) to the north (Fig. 1). The plutons themselves show good evidence for a combination of strike-slip and contractional deformation. Generally, these plutons of the eastern Sierra Nevada have a well-defined NW-SE trending, vertical foliation. Relation to plate motion Plate reconstructions for Late Cretaceous California indicate a high angle of plate convergence (75-90 ~) between North America and the offshore Farallon plate (Kelley 1993; D. C. Engebretson, pets. comm. 1995). Thus, large volumes of magma were emplaced in the Late Cretaceous Sierra Nevada batholith, associated with strikeslip tectonism, during a time of nearly normal subduction (Tikoff & Saint Blanquat 1997). M a g m a t i s m and deformation: the M o n o Creek granite A case example provides a basis for the concept of a magma-enhanced regional deformation and the feedback between transpression and magma ascent. A particularly well-studied example is the Mono Creek granite (Mono Pass Intrusive

Series) of the Late Cretaceous Sierra Nevada batholith (Fig. 7; Saint Blanquat & Tikoff 1997; Tikoff & St Blanquat 1997). Fabrics within the Mono Creek granite are in good agreement with transpressional deformation models. The foliation (flattening plane) is systematically vertical, indicating a sub-horizontal axis of compression throughout the construction of the pluton, and the shape of the fabric ellipsoid is oblate. We observe a continuous evolution between two geometries: the inferred earliest fabrics are characterized by east-west sub-vertical foliations and sub-vertical lineations, and the latest are characterized by north-south sub-vertical foliations and sub-horizontal lineations. The transition between these two fabrics is always progressive. We propose this fabric evolution characterizes a progressive switch from emplacement-dominated strain to regional-deformation-dominated strain, although both are transpressional (Saint Blanquat & Tikoff 1995) (Fig. 7). The E-W foliation orientation of the first batches of magma occurs at a high angle to the inferred regional strain, and is due to overpressuring magma pushing to open its own space (Fig. 7a). The first batches of melt may actually initiate, rather than being passively emplaced between, the shear zones that parallel the pluton. This is similar to the tectonic situation suggested for plutons intruding transcurrent shear zones in Brazil (Tomassi et al. 1994; Neves et al. 1996). For the case of the Mono Creek granite, it is also a good example of how magmatically induced strikeslip partitioning may occur. Thus, strike-slip movement occurs even for a high angle of relative plate motion (75-90~ Kelley & Engebretson 1994). Further, magma-induced overpressuring may partially explain the lack of offset of pluton contacts seen in this setting (Tikoff & Saint Blanquat 1997). Magma overpressuring is also clearly seen in the second stage in the evolution of the Mono Creek granite (Fig. 7b). In the main part of the pluton, lineations are shallowly plunging and foliations are rotated parallel to the inferred regional strain. However, along the NE side of the pluton, the magmatic overpressuring creates a 'bulge' in the otherwise elongate shape of the pluton. This bulge is the result of forceful emplacement of the Mono Creek granite (Bateman 1992; Tikoff et al. 1996), as it deflects the surrounding foliation and displaces metamorphic wall-rocks from their regional strike. The final stage of emplacement involves tectonically assisted magmatic intrusion (Fig. 7c). The Rosy Finch shear zone is activated at this time, and by shearing provides tectonic space for

MAGMATIC ARCS

335

pluton, it constitutes a zone of weakness and localizes a component of strike-slip motion in the N-S shear zone developed during the latter stages of pluton construction. However, once strike-slip tectonism occurs, it further facilitates magma emplacement. Thus, a positive feedback between magmatism and deformation induces partitioning of strike-slip motion within the magmatic arc during transpression (Fig. 6).

Three-dimensional view of arcs and strikeslip partitioning

Fig. 7. Simplified emplacement model of the Mono Creek granite (Sierra Nevada, California). (a) The first stage of magma intrusion is due to overpressurized magma pushing to open its own space, leading to a foliation (flattening plane) which is E-W, at high angle to the inferred regional strain orientation. The overpressured magma initiates the strike-slip movement and the pull-apart setting. (b) Magmatic overpressuring causes emplacement of a 'bulge' in the otherwise elongate shape of the pluton. (e) The final stage of emplacement involves tectonically assisted magma intrusion, through activation of the Rosy Finch shear zone, which provides tectonic space for the latest stages of magma intrusion. the latest stages of magma intrusion (Saint Blanquat & Tikoff 1997; Tikoff & Saint Blanquat 1997). We envision the following pattern in the ascent and construction of the Mono Creek granite, based on structural fieldwork, AMS analysis, and microstructural studies. The magma initiates a pull-apart geometry, indicating that magma is facilitating strike-slip partitioning. As more magma is emplaced within the

A clear relationship exists between m a g m a emplacement and strike-slip tectonism in neotectonic settings (e.g. Sumatra) and can also be discerned in ancient magmatic arcs (e.g. Sierra Nevada). Further, in both cases, the upper crust deforms over a wide region, in a consistent orientation, in response to the plate motion. We propose that deformation of the upper crust is driven by the bulk lithospheric deformation induced by oblique plate interaction. Strike-slip partitioning, or the relative efficiency for strikeslip faults or shear zones to take up the wrench component of plate motion, is facilitated by magmatism which tends to localize shear in the volcanic arc region (Sumatra), but also in the underlying plutons (Sierra Nevada). The interaction between magmatic and kinematic processes has broad implications for the dynamics of continental arcs. We propose the following three-dimensional model for magmatic arc construction, based on an approximate rheological lithospheric layering in an arc setting (Fig. 8). According to lithospheric strength profiles determined experimentally (e.g. Brace & Kohlstedt 1980), the upper mantle is the strongest layer in the lithosphere, although it may be thermally weakened below the arc. Oblique subduction transfers motion to the overriding continental plate presumably through the strong upper mantle, or possibly the lower crust. Deformation of the upper mantle is poorly known in modern, obliquely converging arc settings. However, studies of seismic anisotropy in the San Andreas region (e.g. Ozalaybey & Savage 1995), which is also characterized by oblique convergence and strike-slip partitioning, indicate that a strongly anisotropic layer more than 100 km thick and probably over 100 km wide underlies California (e.g. Molnar 1992). In a magmatic arc, one may also expect an upper mantle accommodating the imposed plate motion by distributed shearing. Because of heat advection from magmas, the thinned sub-arc mantle becomes the weakest lithospheric

336

M. DE SAINT BLANQUAT E T A L .

Fig. 8, Three-dimensional geometry of magmatic arcs. Oblique motion imposed by the oceanic lithosphere is transferred to continental lithosphere through the 'strong' upper mantle. Vertical coupling between the mantle and upper crust is provided by distributed shearing in the lower crust. This explains the consistency in orientation of structures from the forearc to the backarc. Because of its relative weakness caused by rising magma, the continental lithospheric mantle beneath the forearc region undergoes margin-parallel tFanslation ('forearc mantle sliver').

segment of the subduction system. However, this zone probably remains stronger than the lower or upper crust, which is also thermally weakened. As a result of these two effects, the continental lithospheric mantle beneath the forearc region moves tangentially with the obliquely subducting oceanic plate as a forearc (mantle) 'sliver' (e.g. Beck 1983; Jarrard 1986), in which the forearc region urldergoes margin-parallel translation (although discrete offset is not necessary). The lower crust undergoes distributed shearing, which acts to transmit the deformation to both the upper-crustal forearc and backarc regions, in response to the velocities field imposed by the upper mantle (basal forcing). This type of deformation is inferred by shallow (<40 kin) crustal seismic anisotropy, recorded for example in the Hikurangi subduction zone, New Zealand (Gledhill & Stuart 1996), where a strong seismic anisotropy orientation is found, with lineation parallel to the trend of the overlying geological structures. This type of anisotropy is consistent with a bulk transpressional shearing of the lower crust. In this manner, the forearc, the

active arc, and the backarc are mechanically linked through distributed shearing in the lower crust. Further, this mechanical explanation provides a rationale for why the deformation in the upper crust is systematically consistent with plate motion (Teyssier & Tikoff this volume). The influx of magmas into the lower lithospheric levels is expected to affect the distribution of shearing in the upper mantle, lower crust, or both. Thus heat advection from rising magmas has the effect of thinning the upper mantle and weakening the lower crust. This vertically oriented zone of high heat flow is ideally oriented to accommodate the tangential component of the plate motion. For both the lower crust and upper mantle, heat transfer is expected to result in a zone of enhanced wrenching deformation below the arc (Fig. 8). This increased wrench zone occurs within a larger transpressional regime, resulting in heterogeneous transpression of the lower layers. Despite the gradient of simple shear and the heterogeneous displacement field, the system is mechanically continuous.

MAGMATIC ARCS Deformation in the upper crust adjacent to the magmatic arc is best considered as a doublewedge: two wedges or prisms which verge in opposite directions and may or may not be separated by a strike-slip fault system (Fig. 8). This model of a double-wedge is inferred from geological studies (Norris et al. 1990) and physical experiments (Richard & Cobbold 1990), and is supported by three-dimensional numerical studies (e.g. Koons 1994; Braun & Beaumont 1995). This double-wedge is similar to the classic geometry of a flower structure, insofar as thrust faults emerge from a central strike-slip fault (e.g. Harding 1985), except the scale is much larger (hundreds v. tens of kilometres). The strike-slip partitioning implied by the double-wedge is an upper-crustal effect of transpressional deformation (e.g. Teyssier et al. 1995). The parallelism of structures on the forearc and backarc implies kinematic connection across the arc, provided by basal coupling (e.g. Molnar 1992; Teyssier & Tikoff this volume). This idea of lithospheric layers explains a recurring problem in magmatic arcs: the active strike-slip fault does not always correlate with the surface expression of volcanism (Fig. 8; Fitch 1972). For example, the Sumatran fault does not always lie within the Sumatran volcanic arc (Bellier & Sebrier 1994). On a lithospheric scale, the location of the strike-slip fault occurs over the area of maximum displacement in the lower crust, controlled by thermal weakening or offset in the lithospheric mantle. In contrast, overpressured magmas in the uppermost crust exploit structural weaknesses and their location may not correlate exactly with the surface expression of the main strike-slip fault. In our model, the two main interfaces (upper crust-lower crust, and lower crust-upper mantle) are theological boundaries. They necessarily constitute zones of deformation maintaining the mechanical continuity of the whole system. A vertically weakened zone exists in the lithosphere because of the upward movement of magmas. The upper mantle, as the strongest lithospheric layer, controls partitioning of the plate motion. The lower crust tends to widen the zone affected by transpressional deformation and homogenizes the movement. Strike-slip partitioning describes the response of the upper crust. Therefore, even on a lithospheric scale, there is a clear relationship between magmatism and tectonism.

Conclusion A positive feedback loop exists between magmatism and transpressional processes; the result of each is to cause magmatic overpressuring.

337

Transpressional kinematics results in tectonic overpressuring, seen in upward movement of material from transpressional zones. In a magmatic arc setting, the horizontal tectonic load of transpression is partly transferred into a vertical driving force, which facilitates the upward movement of granitic magma along a vertical pressure gradient. Overpressured magmas can cause regional deformation, to facilitate magma movement, and thereby initiate strike-slip movement. Thus, transpression-induced and buoyancy-induced overpressuring of magma are linked in a positive feedback loop, which provides the force for upward magma movement. The same feedback cycle explains how a magmatic arc facilitates strike-slip partitioning, even in cases where the relative angle of plate convergence is almost normal to the plate boundary. We propose that transpressional kinematics typifies many arc settings, both modern and ancient. Strike~slip motion is partitioned into the magmatic arc settings because of the linear and margin-parallel trend of the vertical, lithospheric-scale weakness caused by ascending magma. Offset occurs within the mantle forearc sliver (e.g. Beck 1983), creating a strike-slip fault that broadly correlates with the magmatic arc. The lower crust couples the upper mantle and upper crust, explaining the parallelism of contraction structures in the forearc and backarc regions. The displacement field of the basal layer of the arc system provides the boundary condition for the upper-crustal, strike-slip partitioned deformation. However, strike-slip motion is ultimately related to shearing at lower lithospheric levels and may not correlate exactly with upper-crustal magmatism. We would like to thank A. Cruden, D. Hutton, and K. McCaffrey for critical and helpful reviews. Thanks are also due to R. E. Holdsworth for his encouragement and editorial patience. This work was supported by CNRS (INSU/DBT, MDRI and UMR 5563) (M. S. B.) and NSF grants (EAR 9305262) (B. T. and C. T.).

References A~uE, R. & BRIMHALL,G. H. 1988. Magmatic arc asymmetry and distribution of anomalous plutonic belts in the batholiths of California: effects of assimilation, crustal thickness, and depth of crystallization. Geological Society of America Bulletin, 100, 912-927. BATEMAN,P. C. 1992. Plutonism in the Central Part of the Sierra Nevada Batholith, California. US Geological Survey Professional Papers, 1483. BECK,M. E. 1983. On the mechanism of tectonic transport in zones of oblique subduction. Tectonophysics, 93, 1-11. BELLmR, O. & SEBRIER, M. 1994. Relationship between tectonism and volcanism along the Great

338

M. DE SAINT B L A N Q U A T E T A L .

Sumatran Fault Zone deduced by SPOT image analyses. Tectonophysics, 233, 215-231. -& PRAMUMIJOYO, S. 1991. La grande faille ~le Sumatra: gGometrie, cindmatique et quantit6 de ddplacement mises en 6vidence par l'imagerie satellitaire. Comptes Rendus de l'AcadOmie des Sciences, SOrieH, 312, 1219-1226. Borr, M. R & KUZNm, N. J. 1984. Origin of tectonic stress in the lithosphere. Tectonophysics, 105, 1-14. BRACE, W. E & KOHLSTEDT, D. L. 1980. Limits on lithospheric stress imposed by laboratory experiments. Journal of Geophysical Research, 85, 6248-6252. BRAUN, J. & BEAUMONT, C. 1995. Three-dimensional numerical experiments of strain partitioning at oblique plate boundaries: implications for contrasting tectonic styles in the southern Coast Ranges, California and central South Island, New Zealand. Journal of Geophysical Research, 100, 1805%18074. BROWN, E. H. & TALBOT, J. L. 1989. Orogen-parallel extension in the North Cascades Crystalline Core. Tectonics, 8, 1105-1114. -& WALKER, N. W. 1993. A magma-loading model for Barrovian metamorphism in the southeast Coast Plutonic Complex, British Columbia and Washington. Geological Society of America Bulletin, 105, 479-500. BURG, J. P. ~z;VANDERHAEGHE, 0. 1993. Structures and way-up criteria in migmatites, with applications to the Velay Dome (French Massif Central)). Journal of Structural Geology, 15, 1293-1301. BUSBY-SPERA, C. J. & SAI_EEBY,J. 1990. Intra-arc strikeslip fault exposed at batholithic levels in the southern Sierra Nevada, California. Geology, 18, 255-259. CASHMAN, S. M., KELSEY, H. M., ERDMAN, C. E, CUTI?EN,H. N. C. & BERRYMAN,K. R. 1992. Strain partitioning between structural domains in the forearc of the Hikurangi subduction zone, New Zealand. Tectonics, 11, 242-257. CASTRO, A., DE LA ROSA, J. D., FERNANDEZ, C. & MORENO-VENTA, I. 1995. Unstable flow, magma mixing and magma-rock deformation in a deepseated conduit: the Gil-Marquez complex, southwest Spain. Geologishe Rundschau, 84, 359-374. CLEMENS, J. D. & MAWER, C. K. 1992. Granitic magma transport by fracture propagation. Tectonophysics, 204, 339-360. CRAWFORD, M. L., HOLLISTER, L. S. & WOODSWORTH, G. J. 1987. Crustal deformation and regional metamorphism across a terrane boundary, Coast Plutonic Complex, British Columbia. Tectonics, 6, 343-361. CURRAY, J. R., MOORE, D. G., LAWVER, L. A., EMMEL, E J., RAITT, R. W., HENRY, M. & KIECKHEFER, R. M. 1979. Tectonics of the Andaman Sea and Burma. In: WATKINS,J. S. et al. (eds) Geological

and Geophysical Investigations of Continental Margins. American Association of Petroleum Geologists, Memoir. 29, 189-198. DAVIS, M. 1996. Anisotropy of magnetic susceptibility

and fabric analysis in granitoids: the Cascade Lake

shear zone, Sierra Nevada, California. M.S. thesis, University of Minnesota, Minneapolis. DEWEY, J. E & LAMB,S. H. 1992. Active tectonics of the Andes. Tectonophysics, 205, 79-95. DEMETS, C., GORDON, R. G., ARGUS, D. E & STEIN, S. 1990. Current plate motions. Geophysical Journal International, 101, 425-478. DIAMENT, M., HARJONO, H., KARTA, K., ET AL. 1992. Mentawai fault zone off Sumatra: a new key to the geodynamics of western Indonesia. Geology, 20, 259-262. FITCH, J. T. 1972. Plate convergence, transcurrent faults, and internal deformation adjacent to Southeast Asia and the Western Pacific. Journal of Geophysical Research, 77(23), 4432-4460. FOSSEN, H. & TIKOFF,B. 1993. The deformation matrix for simultaneous pure shear, simple shear, and volume change, and its application to transpression/transtension tectonics. Journal of Structural Geology, 15, 413-422. -& TEYSSmR, C. 1994. Strain modeling of transpressional and transtensional deformation. Norsk Geologisk Tidsskrifi, 74, 134-145. GLEIgHmL,K. & STUART,G. 1996. Seismic anisotropy in the fore-arc region of the Hikurangi subduction zone, New Zealand. Physics of the Earth and Planetary Interiors, 95, 211-225. GREENE, D. C. & SCIqWEICI,:ERT,R. A. 1995. The Gem Lake shear zone: Cretaceous dextral transpression in the northern Ritter Range pendant, eastern Sierra Nevada, California. Tectonics, 14, 945-961. GROCOTT, J., BROWN, M., DALLMEYER, R. D., TAYLOR, G. & TRELOAR, P. J. 1995. Mechanisms of continental growth in extensional arcs: an example from the Andean plate-boundary zone. Geology, 22, 391-394. GUGLIELMO, G. 1993. Magmatic strains and foliation triple points of the Merrimac plutons, northern Sierra Nevada, California: implications for pluton emplacement and timing of subduction. Journal of Structural Geology, 15(2), 177-189. HARDING, T. P. 1985. Seismic characteristics and identification of negative flower structures, positive flower structures, and positive structural inversion. American Association of Petroleum Geologists, Bulletin, 69, 582-600. HOGAN, J. P. & GILBERT,M. C. 1995. The A-type Mount Scott granite sheet: importance of crustal magma traps. Journal of Geophysical Research, 100, 15779-15792. HUDLESTON, P. J., SCHULTZ-ELA,D. & SOUTHWICK,D. L. 1988. Transpression in an Archean greenstone belt, northern Minnesota. Canadian Journal of Earth Sciences, 25, 1060-1068. HUTTON, D. H. W. 1997. Syntectonic granites and the principle of effective stress: a general solution to the space problem? In: BOUCHEZ,J.-L., STEVHENS, W. E. & HUTrON, D. H. W. (eds) Granite: From Melt Segregation to Emplacement Fabrics, Kluwer Academic Publishers, 189-197. & REAVY, R. J. 1992. Strike-slip tectonics and granite petrogenesis. Tectonics, 11, 960-967. INGRAM, G. M. & HUTrON, D. H. W. 1994. The Great Tonalite Sill: emplacement into a contractional -

-

MAGMATIC ARCS shear zone and implications for late Cretaceous to early Eocene tectonics in southeastern Alaska and British Columbia. Geological Society of America Bulletin, 106, 715-728. JARRARD, R. D. 1986. Terrane motion by strikeslip faulting of forearc slivers. Geology, 14, 780-783. KARIG, D. E., LAWRENCE, M. B., MOORE, G. E & CURRAY, J. R. 1980. Structural framework of the fore-arc basin, NW Sumatra. Journal of the Geological Society, London, 137, 77-91. KARLSTROM, K. E. 1989. Toward a syntectonic paradigm for granitoid. Eos Transactions, American Geophysical Union, 70, 762. KELLEY, K. 1993. Relative motions between North

America and oceanic plates of the Pacific basin during the last 130 million years. M.S. thesis, Western Washington University, Bellingham. & ENGEB~TSON, D. C. 1994. Updated relative motions and terrane trajectories for North America and oceanic plates: Cretaceous to present. Geological Society of America, Abstracts with Programs, 26, 459. KISTLER,R.W. 1990. Two different lithosphere types in the Sierra Nevada, California. In: ANDERSON,J. L. (ed.) The Nature and Origin of Cordilleran Magmatism. Geological Society of America Memoir, 174, 271-281. KOONS, P. O. 1994. Three-dimensional critical wedges: tectonics and topography in oblique collisional orogens. Journal of Geophysical Research, 99, 12301-12315. LAHREN,M. M., SCHWEICKERT,R. A., MATTINSON,J. M. & WALKER, J. G. 1990. Evidence of Uppermost Proterozoic to Lower Cambrian miogeoclinal rocks and the Mohave-Snow Lake fault: Snow Lake Pendant, Central Sierra Nevada, California. Tectonics, 9, 1585-1608. LALLEMAND,S. & JOLIVET,L. 1986. Japan Sea: a pullapart basin? Earth and Planetary Sciences Letters, 76, 375-389. LASSAL, O., HUCHON, P. & HARJONO, H. 1989. Extension crustale dans le drtroit de la Sonde (Indonesie). Donnres de la sismique reflrxion (campagne Krakatoa). Comptes Rendus de l'AcadOmie des Sciences, Paris, Serie II, 309, 205-212. LISTER, J. R. & KERR, R. C. 1991. Fluid-mechanical models of crack propagation and their application to magma transport in dykes. Journal of Geophysical Research, 96, 10049-10077. LUND, K. & SNEE,L. W. 1988. Metamorphism, structural development, and age of the continentisland arc juncture in west-central Idaho. In: ERNST, W. G. (ed.) Metamorphism and Crustal --

Evolution, Western Coterminous United States, Rubey Vol. VII. Prentice-Hall, Englewood Cliffs, N J, 296-331. MANCKTELOW, N. S. 1993. Tectonic overpressure in competent mafic layers and the development of isolated eclogites. Journal of Metamorphic Geology, 11, 801-812. - 1995. Nonlithostatic pressure during sediment subduction and the development and exhumation

339

of high pressure metamorphic rocks. Journal of Geophysical Research, 1 0 0 , 571-583. MANDUCA, C. A., KUNTZ, M. A. & SHNER, L. T. 1993. Emplacement and deformation of the western margin of the Idaho batholith near McCall, Idaho: influence of a major terrane boundary. Geological Society of America Bulletin, 105, 749-765. MATSUDA, T., NAKAMURA, K. & SIGIMURA, A. 1967. Late Cenozoic orogeny in Japan. Tectonophysics, 4, 349-366. MCCAFFREY, K. J. W. 1992. Igneous emplacement in a transpressive shear zone; Ox Mountains igneous complex. Journal of the Geological Society, London, 149, 221-235. MCCAFFREY, R. 1988. Active tectonics of the eastern Sunda and Banda arcs. Journal of Geophysical Research, 93, 15163-15182. - 1991. Slip vectors and stretching of the Sumatran forearc. Geolog); 19, 881-884. MEGARD,F. 1987. Structure and evolution of the Peruvian Andes. In: SCHAER,J.-E & RODGERS,J. (eds) The Anatomy of Mountain Ranges. Princeton University Press, 179-210. MOLNAR, P. 1992. Brace-Goetze strength-profiles, the partitioning of strike-slip and thrust faulting at zones of oblique convergence, and the stress-heat flow paradox of the San Andreas fault. In: EVANS, B. & WONG,T.-E (eds) Fault Mechanics and Transport Properties of Rocks. Academic Press, New York, 435-459. MOUNT,V. S. & SUPPE,J. 1992. Present-day stress orientations adjacent to active strike-slip faults: California and Sumatra. Journal of Geophysical Research, 97, 11995-12013. NEVES, S. P., VAUCHEZ, A. & ARCHANJO, C. J. 1996. Shear zone-controlled magma emplacement of magma-assisted nucleation of shear zones? Insights from northern Brazil. Tectonophysics, 205, 171-185. NORRIS, R. J., KOONS,E O. & COOPER,A. E 1990. The obliquely-convergent plate boundary in the South Island of New Zealand. Journal of Structural Geology, 12, 715-725. OZALAYBEY, S. ~; SAVAGE, M. K. 1995. Shear-wave splitting beneath western United States in relation to plate tectonics. Journal of Geophysical Research, 100, 18135-18149. PARSON, T. & THOMPSON, G. A. 1991. The role of magma overpressure in suppressing earthquakes and topography: worldwide examples. Science, 253, 1399-1402. PETFORD, N. & ATHERTON, M. P. 1992. Granitoid emplacement and deformation along a major crustal lineament: the Cordillera Blanca, Peru. Tectonophysics, 205, 171-185. - - - , KERR, R. C. & LISTER,J. R. 1993. Dike transport of granitoid magmas. Geology, 21, 845-848. PICKEtt, D. A. & SALEEBY,J. B. 1993. Thermobarometric contraints on the depth of exposure and conditions of plutonism and metamorphism at deep levels of the Sierra Nevada batholith, Tehachapi Mountains, California. Journal of Geophysical Research, 98, 60%629.

340

M. DE SAINT B L A N Q U A T E T A L .

READ, H. H. 1957. The Granite Controversy. Thomas Murby. RICHARD, E & COBBOLD, P. 1990. Experimental insights into partitioning of fault motions in continental convergent wrench zones. Annales Tectonicae, IV, 35-44. ROBIN, R-Y. E & CRUDEN,A. R. 1994. Strain and vorticity patterns in ideally ductile transpression zones. Journal of Structural Geology, 16, 447-466. RUSMORE, M. E. & WOODWORTH,G. J. 1991. Coast Plutonic Complex: a mid-Cretaceous contractional orogen. Geology, 19, 941-944. SAINT BLANQUAT,M. & TIKOFF, B. 1995. Passive and forceful emplacement of the Mono Creek pluton (Sierra Nevada, California), during regional transpressive deformation. Terra Nova, 7, 138. & -1997. Development of magmatic to solidstate fabrics during syntectonic emplacement of the Mono Creek granite, Sierra Nevada batholith. In: BOUCHEZ,J.-L., STEPHENS,W. E. (eds) Granite: -

-

From Melt Segregation to Emplacement Fabrics. Kluwer Academic Publishers, 231-252. SALEEBY, J. 1990. Progress and tectonic and petrogenetic studies in an exposed cross-section of young (-100 Ma) continental crust, southern Sierra Nevada, California. In: SALISBURY,M. H. & FOUNTAIN, D. M. (eds) Exposed Cross-sections of the Continental Crust. D. Reidel, Dordrecht, 137-158. SAMS,D. B. & SALEEBY,J. B. 1988. Geology and petrotectonic significance of crystalline rocks of the southernmost Sierra Nevada, California. In: ERNST, W. G. (ed.) Metamorphism and Crustal

Evolution of the Western United States (Rubey Vol. 7). Prentice-Hall, Englewood Cliffs, New Jersey, 865-893. SIEH, K. 1994. Active tectonics of Sumatra. Geological Society of America, Abstracts with Programs, 26, 382. SOULAS, J. E 1977. Les phases tectoniques andines du Tertiaire Superieur, r6sultats d'une transversale Pisco-Ayacucho (Peru Central). Comptes Rendus de l'Acaddmie des Sciences, Paris, Serie H, 284, 2207-2210. SUAREZ,G., MOLNAR,P. & BURCHFIEL,B. C. 1983. Seismicity, fault plane solutions, depth of faulting, and active tectonics of the Andes of Peru, Ecuador,

and Southern Colombia. Journal of Geophysical Research, 88, 10403-10428. TEYSSIER, C. & TIKOFF, B. 1998. Strike-slip partitioned transpression of the San Andreas fault system: a lithospheric-scale approach. This volume. & MARKLEY, M. 1995. Oblique plate motion and continental tectonics. Geology, 23, 447-450. TIKOFF,B. & GREENE, D. 1997. Stretching lineations in transpressional shear zones. Journal of Structural Geology, 19, 29-39. -& SAINT BLANQUAT, M. 1997. Transpressional shearing and strike-slip partitioning in the Late Cretaceous Sierra Nevada magmatic arc, California. Tectonics, 16, 442-459. - - & TEYSSIER,C. 1992. Crustal-scale, en echelon 'Pshear' tensional bridges: a possible solution to the batholithic room problem. Geology, 20, 927-930. & -1994. Strain modeling of displacementfield partitioning in transpressional orogens. Journal of Structural Geology, 16, 1575-1588. , BECK, M., OSWALD, B., SAINT BLANQUAT, M., KELSO, P. & HALVORSEN, M. 1996. Forceful emplacement of the Mono Creek 'bulge', Sierra Nevada, California. Geological Society of America, Abstracts with Programs, 28, 42. TOBISCH, 0. T., SALEEBY,J. B., RENNE, P. R., MCNULTY, B. & TONG, W. 1995. Variations in deformation fields of a large volume magmatic arc, Central Sierra Nevada, California. Geological Society of America, Bulletin, 107, 148-166. TOMMASI, A., VAUCHEZ, A., FERNANDES, L. A. D. & PORCHER, C. C. 1994. Magma-assisted strain localization in an orogen-parallel transcurrent zone of southern Brazil. Tectonics, 13, 421-437. VIGNERESSE, J. L. 1983. Enracinement des granites armoricains estim6 d'apr6s la gravim6trie. Bull-

-

etin de la SociOtO GOologique et MinOralogique de Bretagne, 15C, 1-15. WALCOTT, R. I. 1979. Plate motion and shear strain rates in the vicinity of the Southern Alps. Royal Society of New Zealand Bulletin, 18, 5-12. WEAVER, C. S., GRANT, W. C. d~ SHEMETA, J. E. 1987. Local crustal extension at Mount St. Helens, Washington. Journal of Geophysical Research, 92, 10170-10178.

Quaternary tectonics of the Pollino Ridge, Calabria-Lucania boundary, southern Italy MARCELLO

SCHIATTARELLA

Centro di Geodinamica, Universitd della Basilicata, Via Anzio, 1-85100 Potenza, Italy (e-mail." [email protected] Abstract: The Pollino Ridge is a N120 ~ trending morpho-structure, formed by Meso-

Cenozoic carbonate rocks, that marks the boundary between Calabria and Lucania (southern Italy). It is bordered by Quaternary basins filled by both marine and continental sediments. A detailed geological survey and structural analysis in an area elongated perpendicular to the axis of the chain, ranging from the Morano Calabro basin to Mount Madonna de1 Pollino, was carried out. Data show that the entire mountain ridge and nearby Quaternary basins experienced several deformations in recent times, after the MioPliocene thrust tectonics. Moreover, data are also consistent with numerous observations along the Calabria-Lucania boundary, from the Lauria Mountains to the Castrovillari basin. The present structure derives from reorganization of the pre-existing tectonogenetic configuration by Plio-Quaternary brittle tectonics. Mapping and kinematic analysis of a thrust along the northern (Lucanian) side of the Pollino Chain was performed for the first time. This thrust brings both the Cretaceous platform limestones and the Cerchiara Formation (lower Miocene age) over the Bifurto Formation (middle Miocene age). Low-angle transtensional faults along the southern (Calabrian) side were also mapped and analysed. These faults have created 'younger-on-older' type tectonic contacts. The present structure of the ridge is due to the effects of two different Quaternary tectonic stages. The first, lower Pleistocene in age, was characterized by strike-slip tectonics. It truncated ramp folds, producing new contractional and extensional features on the opposite sides of the carbonate ridge. The second stage occurred during middle Pleistocene times and was a purely extensional regime with a counter-Apenninic tensional axis, which reactivated the pre-existing structural pattern with different kinematics. However, both tectonic stages should be interpreted as a consequence of the continuous reorganization of the local stress fields as a result of a rotational field acting along the Pollino shear zone. The best explanation of this phenomenon seems to be a counter-clockwise block rotation of the carbonate ridge in an unchanging regional stress field. This study defines the basic tectonic units outcropping in the area of the Pollino Ridge at the Calabria-Lucania boundary (southern Apennines, Italy), establishes their mutual relationships and analyses the Q u a t e r n a r y brittle deformation. This work presents a kinematic picture of the tectonic evolution of the ridge and nearby basins during Pleistocene times. The Quaternary strike-slip and extensional tectonics has strongly modified the thrust belt formed by Meso-Cenozoic platform carbonates and oceanderived 'internal' units, often changing their original geometric relationships. The Pollino Ridge is a N120 ~ trending morpho-structure made of Meso-Cenozoic carbonate rocks. It is classically interpreted as a simple monoclinal structure dipping N E under ophiolitic nappes which were emplaced in early Miocene time. The ridge is bordered by Quaternary basins filled by both marine and continental sediments (Fig. 1). A n important fault (the 'Pollino Line' after Bousquet & Gueremy, 1969), showing a

Quaternary offset of about 4 km with a strain rate of 3-4 mm/a (Colella & Cappadona 1988), borders the southern slope of the carbonate ridge. The kinematic behaviour and the geodynamic meaning of this fault have been interpreted in m a n y different ways (Bousquet & Gueremy 1969; Ghisetti & Vezzani 1982; Knott & Turco 1991; Van Dijk & Okkes 1991; Russo & Schiattarella 1992; Cinque et al. 1993). The Pollino Fault System probably is still active, as revealed by recent palaeoseismological studies (Ferreli et al. 1996). Therefore, a kinematic study of this brittle shear zone is necessary to understand the seismotectonic picture of the Calabria-Lucania boundary. The complex pattern of the Quaternary deformation has been examined by surveying largescale tectonic features and by carrying out a structural analysis of minor fault and joint arrays in Meso-Cenozoic and Plio-Pleistocene units. Fault sets have been measured in several localities and grouped according to their orientations and kinematics. The study area is a transverse

SCHIATrARELLA,M. 1998. Quaternary tectonics of the Pollino Ridge, Calabria-Lucania boundary, southern Italy. In: HOLDSWORTH,R. E., STRACHAN,R. A. & DEWEY,J. F. (eds) 1998. Continental Transpressional and Transtensional Tectonics. Geological Society, London, Special Publications, 135, 341-354.

341

342

M. SCHIATTARELLA

,!iiiiNiiii imi

"Lit

l10 O k m

[illio~anI

~ t ~ . : . : . : . : . : . ~ ~

....

,i oon ,s ~ ~

' Qu emarydepo'

~

Sicilideunits

~

Lagonegrounits

Tortonianto Messinianclastics

Fig. 1. Geological sketch map of the Calabria-Lucania boundary. Black frames indicate the study areas. sector of the ridge between the Morano Calabro basin, to the south, and Mount Madonna del Pollino to the north, and the adjoining Mercure, Morano Calabro and Castrovillari basins. The new data show that the entire mountain ridge and nearby Quaternary basins underwent several deformational stages after Mio-Pliocene thrust tectonics and clarify the role of strike-slip tectonics.

Geological outlines of the southern Apennines The southern Apennines are an E-verging accretionary wedge which developed on the western border of the Adriatic promontory from early Miocene to Pleistocene. The wedge is mainly composed of sedimentary cover, of platform and deep-water environments, which was scraped off from the Mesozoic Ligurian oceanic crust (Ogniben 1969; Knott 1987; Bonardi et al. 1988; Mauro & Schiattarella 1988), from the western passive margin of the Adriatic plate (D'Argenio et al. 1973), and from the Neogene-Pleistocene foredeep deposits of the active margin (Casnedi 1988; Pescatore 1988; Pieri et al. 1996). The Apenninic chain developed in the hanging wall of

a W-directed and E-retreating subduction zone (the strike of the average axis of the chain is about N150~ The thrusting in the frontal eastern part of the accretionary wedge is followed by back-arc extension to the rear (Malinverno & Ryan 1986). Recent shortening occurred at the belt front deforming Plio-Pleistocene sediments (Prosser et al. 1996) and volcanic rocks (Beneduce & Schiattarella 1997) whereas widely documented extension is still active along the Apennines axis (Brancaccio et al. 1991; Santangelo 1991; Di Niro et al. 1992; Ortolani et al. 1992; Amato & Selvaggi 1993). The belt is also affected by Plio-Quaternary strike-slip faults mainly oriented on N120 ~ _ 10~ and N50-60 ~ trends (Knott & Turco 1991; Ortolani et al. 1992). In a representative regional cross-section moving from W to E, i.e. from the Tyrrhenian Sea to the Adriatic (Apulian) foreland, and from the top to the bottom of the accretionary wedge (Fig. 2), the following tectonic units are observed (Mostardini & Merlini 1986; Prosser et al. 1996; Schiattarella et al. 1997): (1) a JurassicCretaceous ophiolitic unit, covered by syntectonic deposits of early Miocene age; (2) a carbonate platform unit (Apenninic or Campania-Lucania platform), whose age ranges from

TECTONICS OF THE POLLINO RIDGE

343

Fig. 2. Tectono-stratigraphic scheme of the southern Apennines (west to east).

late Triassic to early Miocene; (3) a unit mainly composed of deep-sea sediments, ranging from early Triassic to lower-middle Miocene (Lagonegro basin); (4) a frontal imbricate fan made up of Cretaceous to lower Miocene deepsea clays covered by middle to upper Miocene flysch deposits; (5) the Apulian carbonate platform, which has been incorporated by underplating at the base of the accretionary wedge. The units forming the frontal imbricate fan (unit 4) are likely to be derived from an 'Apulian' domain which underlies the chain units. Such an Apulian domain could be a Mesozoic pelagic basin, which may represent a timeequivalent of the Lagonegro basin (unit 3), or a carbonate platform evolving into a basinal environment during the early Cretaceous. Pliocene to Pleistocene deposits are tectonically sandwiched between the Apulian platform and unit 3 or 4 (Patacca & Scandone 1989). The basement is not directly known in the southern Apennines. Based on the available well data and on a simple superimposition of the above-mentioned tectonic units, the thickness of the sedimentary cover involved in the Apenninic chain should be about 15 km on the Tyrrhenian side. No more than 5-10 km thick crystalline basement is then expected in the lower part of the

crust if an average crustal thickness of 20-25 km (Scarascia et al. 1994) is assumed. Because of shortening (average value c. 70-80%; up to 100% after Schiattarella et al. (1997)) and the estimated time span for the Apenninic orogeny (about 20 Ma), a very high strain rate can be inferred (from 6-7 mm/a to 1 cm/a).

Geological framework of the Calabria-Lucania boundary The Calabria-Lucania boundary separates the sedimentary terranes of the southern Apennines and the metamorphic units of the Calabrian Arc (Fig. 1). The Pollino Ridge, one of the widest W N W - E S E striking geological structures of this region (Fig. 3), has been considered a simple NE-dipping monocline (see, for example, Carta Geologica d'Italia, scale 1:100 000, Foglio 221, Castrovillari). It is made up of a Meso-Cenozoic shallow-water carbonate succession ('Complesso Panormide' after Ogniben (1969), corresponding either to the Alburno-Cervati Unit after D'Argenio et al. (1973) or to the Pollino Unit after Amodio Morelli et al. (1976)). This unit results from the Campania-Lucania platform deformation (D'Argenio et al. 1973). The Alburno-Cervati Unit is formed, starting

344

M. SCHIATTARELLA

Fig. 3. Main structures of the Calabria-Lucania boundary. Bold lines indicate faults and other tectonic contacts; fine lines indicate stratigraphic contacts. from the bottom, by the following units: (i) Upper Triassic dolomite; (ii) dolostone and limestone ranging from Jurassic to Cretaceous; (iii) neritic grainstone, early Miocene in age (Cerchiara Formation); (iv) calcarenites, quartzarenites, marls and shales, early-middle Miocene in age (Bifurto Formation). The last terrigenous formation outcrops north of the carbonate ridge and is tectonically covered by ophiolite-bearing units ('Complesso Liguride', sensu Bonardi et aL (1988); see also Mauro & Schiattarella (1988) for the Cilento region, and Knott (1987), Cello et al. (1990) and Monaco et al. (1995) for the southern Lucania region). Therefore the Ligurian units occupy the highest geometric position of the thrust belt (Fig. 2). Along the southern side of the ridge a strikeslip brittle shear zone (Pollino Fault Zone), juxtaposing tectonic units derived from different palaeogeographical domains, has been recognized by several workers (Ghisetti & Vezzani 1982; Colella & Cappadona 1988; Knott 1988; Turco et aL 1990; Knott & Turco 1991; Russo & Schiattarella 1992; Schiattarella 1996). South of this line, the Verbicaro Unit (margin-to-slope limestones and marls ranging from upper Triassic to lower Miocene) and the S. Donato Unit (Triassic marbles and phyllites) widely cropout (Amodio Morelli et aL 1976; Ietto et al. 1992).

Results and discussion Tectonic setting o f the Pollino R i d g e

The Pollino Ridge can be divided into three axis-parallel sectors that show different structural features (Fig. 4). For this reason, a crosssection may be divided into a central sector, a northern (frontal) sector, and a southern one (Fig. 5). The central sector includes the major (up to 2200 m high) NE-dipping carbonate massifs and is bordered to the south by the 'Pollino Line' (Bousquet 1973), the major N120 ~ trending fault of the Calabria-Lucania boundary. This sector is separated to the north from the frontal sector by a fault showing the same strike as the Pollino Line. The major faults of the central sector strike N E - S W (the 'counter-Apenninic' trend) and show normal dip-slip kinematics. The ridge was progressively downthrown to the NW by these faults. As a consequence, the Cretaceous limestones cropout more widely in the NW. The most downthrown block forms the bedrock of the Quaternary Mercure basin, which is a transverse half-graben breaking off the lateral continuity of the ridge. The Mercure basin is bordered to the SE by the N50-60 ~ striking Viggianello-Mormanno fault system,

TECTONICS OF THE POLLINO RIDGE

Fig. 4. Tectonic map of the Pollino Ridge area between Timpone S. Angelo and Mt Madonna del Pollino. Detailed legend: (A) carbonate platform units (Pollino Unit and Verbicaro Unit, upper Triassic-lower Miocene); (B) Bifurto Formation (middle Miocene); (C) undifferentiated Ligurian units (Eocene?); (D) Frido Unit (Cretaceous-Oligocene). Pollino and Serra del Prete lines separate three different structural domains (see text). A-A', trace of the cross-section shown in Fig. 5. which shows a left-lateral transtensional kinematics as the last identified movements on the fault planes. In the northern sector high-angle N E - S W and N-S trending faults are recognizable. The N-S trending system exhibits right-lateral strike-slip kinematics. The western slope of Mount Madonna del Pollino is bordered by normal faults that have partially obliterated a high-angle transpressional fault. The latter has accommodated back-thrusting of the Mesozoic carbonates onto the ophiolite-bearing Frido Unit. On the top of Mt Madonna del Pollino a reverse fault brings the Cretaceous limestone onto the Bifurto Formation (middle Miocene in age). The southern part of Mt Madonna del Pollino is affected by a N120 ~ trending fault system showing a left-lateral sense of slip (see Fig. 4 and stereoplot 2a in Fig. 7, below). Back-thrusts in the eastern part of the Pollino Ridge have been recognized by Catalano et al. (1993). These

345

workers believe that back-thrusting of N-S trending carbonate structures is genetically linked to Pleistocene strike-slip tectonics. Along the frontal part of the northern sector Meso-Cenozoic limestones are thrust onto the Bifurto Formation (Figs 4 and 5). A comparison between the attitude of the NW-dipping carbonates in the northern sector and the attitude of the NE-dipping limestones outcropping in the central area (Fig. 4) suggests that the frontal thrust developed by counter-clockwise rotation of blocks bordered by left-lateral strike-slip N120 ~ trending faults and right-lateral strike-slip N-S trending faults, N W - S E being the basic orientation of the ridge. Owing to the transverse streams cutting the hanging wall, it has been possible to estimate a thrust motion of 1.5 km at least. However, it is likely that the moderate thrusting because of block rotation reactivated inherited faults. Towards the NW, the Cretaceous limestones of the Lauria Mounts and Mt Coccovello are thrust on 'internal' oceanderived Ligurian terranes. A similar situation is observed at Mt Bulgheria, sited in the Cilento region. On this basis, the entire northern front of the arc formed by the PoUino Ridge-Lauria Mounts-Mt Coccovello-Mt Bulgheria (from east to west) is likely to be generated by one of the widest out-of-sequence thrusts of the southern Apennines (Fig. 6), related to the earlier Pliocene contractional tectonics. Further, the propagation of a deeper thrust under the Pollino structure may be responsible for the growth strata and general architecture of the southern zone of the Sant'Arcangelo basin (Hippolyte et al. 1994; Pieri et al. 1994), located not far from the ridge towards the northeast. Growth structures expanding towards the northeast in the southern flank of the basin and N W - S E trending transpressional structures have both been observed in the Plio-Pleistocene clastic deposits. In the southern sector the tectonic setting is complicated by N W - S E trending low-angle oblique faults (Figs 4 and 5), which have created 'younger-on-older' type tectonic contacts (Ietto & D'Argenio 1990; Oldow et al. 1993; Schiattarella 1996). The fault planes show transtensional indicators that are kinematically compatible with the left-lateral motion along the N120 ~ master faults. The lateral continuity of the low-angle faults is interrupted by N-S faults which exhibit both normal and right-lateral displacements. An analysis of the structural data collected within the area provides an understanding of the kinematics and chronology of the faults. The data related to the fault kinematics are presented in Fig. 7. Joint orientations are shown in

346

M. SCHIATTARELLA

SW

1

NE

2

3

4

5

6

i'", Fig. 5. Geological cross-section of the Pollino Ridge from Timpone S. Angelo to Mt Madonna del Pollino (the trace of the section is shown in Fig. 4). Legend: 1, Bifurto Formation (middle Miocene, Pollino Unit); 2, grainstones and rudites of the 'Calcari cristallini' Formation (Maastrichtian, Verbicaro Unit); 3, well-bedded limestones and dolomitic limestones (Lias-upper Cretaceous, Pollino Unit); 4, massive white dolomite (upper Triassic-Infralias, Pollino Unit); 5, dark dolomite, locally stromatolites (Norian, Verbicaro Unit); 6, faults.

Fig. 6. Tectonic sketch map of the Meso-Cenozoic carbonate structures of the Campania-Lucania Apennine included in the left-lateral megashear zone. The line with filled triangles indicates the out-of-sequence thrust discussed in the text; bold lines correspond to the high-angle faults; fine lines indicate the outcrop limits of the Meso-Cenozoic shallow-water carbonates. Toponyms, structures and geological units: Pi, Picentini Mts; VS, High Sele Valley graben; M, Mt Marzano; PS, Sele Plain; S, Mt Soprano; A, Alburni Mts; VD, Vallo di Diano graben; Ma, Maddalena Mrs; Ce, Mt Cervati; B, Mt Bulgheria; C, Mt Coccovello; L, Lauria Mts; BM, Mercure basin; P, Pollino Ridge; Se, Mt Sellaro push up (after Knott & Turco, 1991); Cc, Northern Calabria Coastal Range; BC, Castrovillari basin; CU, Cilento Unit and ocean-derived 'internal' units; LU, Lagonegro units and, north of Pollino Ridge, Ligurian units (ocean-derived 'internal' units). Fig. 8. Figure 9 shows azimuthal diagrams for the high-angle fracture systems (a) of the northern

sector (i.e. the frontal part of the carbonate ridge), major faults of the same sector (b) and

TECTONICS OF THE POLLINO RIDGE

347

Fig. 7. Stereograms of fault systems (Wulff net, lower hemisphere) from the Pollino Ridge, separated on the basis of different kinematics (all the dip-slip components of the oblique faults are normal). (1) Southern sector: A, NW-SE strike-slip or oblique faults; B, NE-SW normal faults; C, NW-SE normal faults; D, NE-SW strike-slip or oblique faults. A and B represent synkinematic structures related to strikeslip tectonics acting along N120~ master faults (about N150~ extension axis). C and D are referred to structural association acting during extensional tectonics (about N50~ extension axis). (2) Northern sector: A, NW-SE strike-slip or oblique faults; B, NE-SW normal faults; C, NW-SE normal faults; D, NE-SW strike-slip or oblique faults. A and B represent synkinematic structures related to strikeslip tectonics acting along N120~ master faults. C and D are referred to structural association acting during extensional tectonics. (3) Entire area: N-S and E-W trending faults showing several kinematics. 1t

N

N

morpho-tectonic lineaments (c) from the entire area shown in the tectonic map (Fig. 4). Fracture patterns and kinematics of the faults suggest a strain field produced by strike-slip tectonics acting along N120 ~ trending master faults with left-lateral sense of shear; yet it is also clear that the fault pattern produced by strike-slip tectonics was reactivated by a N E - S W extension. Cross-cutting relationships among the fault systems and superimposition of distinct

generations of kinematic indicators on fault planes reveal that strike-slip pre-dates dip-slip on the N120 ~ trending faults whereas dip-slip pre-dates strike- or oblique-slip on the N50-60 ~ trending faults. In this way the N W - S E trending planes were reactivated as n o r m a l faults whereas the N E - S W and N-S trending planes, which originated respectively as normal and dextral wrench faults, were reactivated as transfer faults by extensional tectonics. Conjugate normal faulting is then observable for both N W - S E and N E - S W trending systems, and is related to chronologically different extension axes (about N150 ~ extension axis during strikeslip tectonics and about N50 ~ extension axis during extensional tectonics). Most of the jointing in carbonate rocks is linked to strike-slip tectonics, being orthogonal to the N150 ~ extension axis, although N W - S E striking high-angle fractures have also been observed. The latter represent the extensional joints produced during the successive normal block faulting. The N-S trending dextral wrench faults seem to be the most ancient system, for they pre-date the tilting of the beds. In fact, if the attitudes of the beds from the single measurement stations are rotated to the horizontal, the poles of these faults appear strongly clustered in the diagrams. Uplift of the ridge occurred and near presentday configurations of the adjacent basins were reached during the extensional stage. A subsequent stage created N150 ~ _+ 20 ~ trending normal faults affecting Quaternary deposits or reactivated N - S trending planes as normal faults in the Mesozoic limestones. A clockwise rotation of the extension axis is thought to have caused this event. E - W trending faults often cut all the other structures and are also observed in

348

M. SCHIATrARELLA

Fig. 8. Density diagrams of the fracture systems (Schmidt net, lower hemisphere) from the Pollino Ridge. (a) Station 1 (Serra del Prete); (b) Station 2 (Timpone Capanna); (r Station 3 (Timpone Viggianello); (d) Station 4 (Timpone S. Angelo); (e) Station 5 (northern frontal sector of the ridge); (f) total data diagram. upper Pleistocene slope deposits. Their kinematic role is not completely clear but their genesis seems to be associated with the last tectonic event. The age of deformation can be deduced by correlating the sedimentary sequences of the surrounding Quaternary basins. The strike-slip

stage was active from late Pliocene to early Pleistocene times (Russo & Schiattarella 1992; Cinque et al. 1993). The extensional tectonics started during the passage from lower to middle Pleistocene (late Villaffanchian stage according to Bousquet & Gueremy (1969); Sicilian stage according to Russo & Schiattarella (1992)).

TECTONICS OF THE POLLINO RIDGE

349

FRACTURE SYSTEMS FROM THE POLLINO RIDGE

A) MESOSCOPIC FRACTURES

N

N

B) MAJOR FAULTS

C) TECTONIC LINEAMENTS

Fig. 9. Azimuthal diagrams of mesoscopic fracture systems (a) collected in the northern sector of the Pollino Ridge, major faults surveyed in the same sector (b) and tectonic lineaments (c) from the entire area of the tectonic map (Fig. 3).

Mercure

basin

The Mercure basin is located west of the Pollino Ridge and to the east of the Lauria Mountains, which represent the extension of the Pollino carbonate structure to the northwest. Its trend is oblique to the axis of the ridge (Figs 1 and 3). The basin is filled by middle-upper Pleistocene fluvial and lacustrine sediments. Geomorphological features and Quaternary brittle structures of the basin have been recently surveyed (Schiattarella et al. 1994). Orientation measurements were collected in both Mesozoic carbonate bedrock and Quaternary lacustrine and slope deposits. The existence of two tectonic stages responsible for the genesis and the evolution of the basin were documented. Both stages pre-date the deposition of the middle-upper Pleistocene lacustrine sediments. The correlation between ages and unconformities of the sedimentary sequences of the Mercure, S a n t ' A r c a n g e l o and Castrovillari basins suggests that we can ascribe those tectonic events to the Emilian and Sicilian stages (lower Pleistocene). The first tectonic stage occurred in a transcurrent regime related to the kinematics of the Pollino left-lateral strike-slip N120 ~ trending fault zone (Knott 1988; Knott & Turco 1991; Russo & Schiattarella 1992; Schiattarella et al. 1994). It created an obliquely trending half-graben which breaks off the lateral continuity of the chain (Fig. 10). The second stage took place into an extensional strain field characterized by a N E - S W extension axis rotating clockwise through time. This stage occurred between the end of the lower Pleistocene and the beginning of the middle Pleistocene. It

resulted in the development of a lacustrine basin because of the uplift of the Nl10-120 ~ trending La Fagosa Ridge. Both tectonic stages are interpreted as a predictable consequence of a continuous reorganization of the local stress fields because of a rotational field acting along the Pollino Fault System. A weak tectonic event is also recorded by the lacustrine deposits. The presence of N150 ~ _+20 ~ trending joint sets, besides N-S and E - W trending fault systems with little normal offset, is here interpreted as the last effect of the extension axis clockwise rotation. Clays and marls of the lake sequences are also affected by synsedimentary listric normal faults, which indicate the persistence of an extensional tectonic regime during middle-upper Pleistocene times.

Fig. 10. Block-diagram related to the genesis of the Mercure half-graben (after Schiattarella et al. 1994), which occurred during the early Pleistocene strikeslip tectonics (ML, Lauria Mountains block; CP, Pollino Ridge block).

350

M. SCHIATTARELLA C

.

I4 \.

D

N /

Fault systems from Morano basin area (lower hemisphere)

Fig. 11. Stereograms of fault systems (Wulff net, lower hemisphere) from the Mesozoic carbonates surrounding the Morano basin, separated on the basis of different orientations: (a) N40qS0~ trend; (b) N45-70~ trend; (e) N10-30~ trend; (d) N-S trend; (e) E-W trend. Note the polyphase hystory of fault systems, except for the N10-30~ trend, which represents the most recent tectonic event.

Morano Calabro basin

Castrovillari basin

The M o r a n o Calabro basin (Fig. 1), which developed in Quaternary times along the southern side of the Pollino Ridge ( n o r t h e r n Calabria), represents a good example of a tectonic-controlled depression. It was generated within a strike-slip shear zone but strongly modified by extensional tectonics. Stratigraphical, structural and geomorphological data shown by Perri & Schiattarella (1997) indicate a polyphase deformation history characterized by changes in orientation of the tensional axis during the Pleistocene (Figs 11 and 12). In particular, evidence for a transition from leftlateral strike-slip along N120 ~ trending faults to normal dip-slip along the same faults is very clear. Most of the fault planes show several generations of mechanical striae and/or syntectonic fibrous calcite with a progressive increase of the pitch value. Indeed, transtensional faults are very common in the Mesozoic carbonate rocks surrounding the basin. Furthermore, a subsequent tectonic event produced N150-170 ~ trending normal faults (Fig. 11, stereoplot c), and in the same stress field the N120 ~ trending faults were reactivated with a dextral component of movement (Fig. 11, several faults in stereoplot b).

The Castrovillari-Cassano basin is located immediately south of the Pollino Ridge (Figs 1 and 3). It is filled by three sedimentary cycles separated by unconformities (Russo & Schiattarella 1992). The first cycle consists of marine clay and sand, lower Pliocene in age, outcropping along the western side of the Crati Valley in n o r t h e r n Calabria. The second is formed by marine fine sediments and conglomerates, ranging from the upper part of late Pliocene to lower Pleistocene (Emilian). The third cycle is represented by marine, transitional and continental deposits ascribed to the lower Pleistocene (Emilian - Sicilian). The succession is completed by upper Pleistocene alluvial fan deposits and slope breccias. Plio-Quaternary clastic sediments, geomorphological features and recent brittle deformations of the Castrovillari basin have been studied (Russo & Schiattarella 1992). The existence of two tectonic stages responsible for the genesis and subsequent evolution of the basin has been demonstrated. The reconstruction of sedimentary sequences separated by unconformities allowed the assignment of those tectonic events to the Emilian and Sicilian (lower Pleistocene). The first tectonic stage occurred

TECTONICS OF THE POLLINO RIDGE

351

Fig. 12. Schematic tectonic evolution of the Morano Calabro basin (after Perri & Schiattarella 1997). Bold lines represent the major faults acting during the single tectonic stages (plus and minus signs represent respectively the uplifted and downthrown blocks).

Fig. 13. Azimuthal diagrams of high-angle fractures and minor faults from Castrovillari-Cassano basin and adjacent carbonate massifs. (a) Serra del Prete Mesozoic carbonates; (b) Cycle II elastic deposits (lower Pleistocene); (e) Cycle III elastic deposits (lower-middle Pleistocene).

under transtensional conditions, related to the kinematics of the Pollino left-lateral strike-slip fault, whereas the second took place into an extensional strain field characterized by a N E - S W extension axis; yet, in this case also, both tectonic stages should be interpreted as a consequence of the continuous reorganization of the local stress fields owing to a rotational field acting along the Pollino brittle shear zone. Cycle II clastic deposits (Cassano area), in fact, are affected by extensional features which show a preferred N30-40 ~ trend (Fig. 13b). The extensional axis related to this fracture pattern essentially coincides with that predicted for strike-slip along the N120 ~ trending master

faults. The extensional fracture systems linked to strike-slip tectonics surveyed in the Mesozoic limestone are, however, mainly arranged along a N50-60 ~ trend (Fig. 13a). This fact suggests a counter-clockwise rotation of the strain field during Plio-Pleistocene transcurrent motion. Such tectonics delineates the shape of the basin in which the Cycle III sediments were deposited. These successions were deformed by N120-130 ~ and N150-160 ~ striking conjugate n o r m a l faults and hybrid joint sets (sensu Hancock 1985). Of the two systems the latter is the youngest, being observed in the Sicilian deposits (Figs 13c and 14). E - W striking minor faults and joints represent the most recent

352

M. SCHIATI'ARELLA

Fig. 14. Tectonic sketch map of the Castrovillari basin (modified after Russo & Schiattarella (1992)). Bold lines indicate the most important faults. (Note the set of three N150--170~ trending faults inside the basinal area and the right-lateral strike-slip fault, which border to the NE the Cassano Horst (both represent the effects of the last tectonic stage).)

features, being recorded also in the upper Pleistocene alluvial fan.

Conclusions This study of the Calabria-Lucania boundary, from the Lauria Mountains to the Castrovillari basin, has revealed the characteristics of the deformation experienced by the Meso-Cenozoic and Plio-Quaternary of the Pollino Ridge and adjacent basins. The present-day structure of the area is a consequence of the Quaternary brittle tectonics, which modified the pre-existing structures of the fold-and-thrust belt. Two Quaternary tectonic stages were recognized. The first, lower Pleistocene in age, was characterized by strike-slip tectonics with N120 ~ trending master faults. This truncated ramp folds and produced the new contractional and extensional features on opposite sides of the carbonate ridge. The second stage occurred during middle Pleistocene times, and was a purely extensional regime with a N E - S W tensional axis. This reactivated the pre-existing structural pattern with different kinematics. Both tectonic stages have to be interpreted as a consequence of the continuous reorganization

of the local stress fields as a result of a rotational field acting along the Pollino shear zone. The observed p h e n o m e n o n could be explained as a counter-clockwise block rotation of the carbonate ridge in an unchanging regional stress field. It is interesting to note that palaeomagnetic data from Sant'Arcangelo basin indicate that the area experienced a counter-clockwise rotation of 22 ~ during Pleistocene times (Sagnotti 1992). Left-lateral strike-slip faults trending N120 ~ are spread along a wide axial portion of the southern Apennines (Fig. 6) affecting also Mesozoic platform carbonates of the Alburni Mts (Cinque et al. 1993) and Mt M a r z a n o (Caiazzo et al. 1992). The entire belt between northern Calabria and the Picentini Mts forms a megashear zone and appears to be a rotated segment with respect to N150 ~ trending morpho-structures of the northern part of the south Apenninic chain. P. Di Leo kindly revised the English form of the text. E. Tavarnelli strongly encouraged the draft of this paper. S. Knott and T. Needham provided helpful comments. Special thanks are due to I. Giano for help in redrawing some figures, and to other colleagues from the Centro di Geodinamica for their support and criticism.

TECTONICS OF THE POLLINO R I D G E References

AMATO, A. & SELVAGGI,G. 1993. Aflershock location and P-velocity structure in the epicentral region of the 1980 Irpinia earthquake. Annali di Geofisica, 36, 3-15. AMODIO MORELLI, t., BONARDI, G., COLONNA,V., ET AL. 1976. L'arco calabro-peloritano nell'orogene appenninico-maghrebide. Memorie della Societh Geologica Italiana, 17, 1-60. BENEDUCE, P. & SCHIATTARELLA,M. 1997. Relazioni tra tettonica regionale quaternaria e deformazione vulcanogenica nelle aree dei Campi Flegrei, Isola di Ustica e Monte Vulture (Italia meridionale). II Quaternario, 10, 585-590. BONARDI, G., AMORE, F. O., CIAMPO,G., DE CAPOA, P., MICONNET, P. & PERRONE,V. 1988. Il Complesso Liguride Auct.: stato delle conoscenze e problemi aperti sulla sua evoluzione pre-appenninica ed i suoi rapporti con l'Arco Calabro. Memorie della Societgt Geologica Italiana, 41, 17-35. BOUSQUET, J. C. 1973. La tectonique r6cente de l'Apennin calabro-lucanien dans son cadre g6ologique et g6ophysique. Geologica Romana, 12, 1-103. -& GUEREMY,P. 1969. Quelques ph6nom6nes de n6otectonique dans l'Apennin calabro-lucanien et leurs cons6quences morphologiques. II L'escarpement mdridional du Pollino et son pi6mont. Revue de Gdographie Physique et de Gdologie Dynamique, 11, 223-236. BRANCACCIO, L., CINQUE,A., ROMANO, P., ROSSKOPF, C., Russo, E, SANTANGELO,N. & SANTO,A. 1991. Geomorphology and neotectonic evolution of a sector of the Tyrrhenian flank of the Southern Apennines (Region of Naples, Italy). Zeitschrifi far Geomorphologie, Neue Folge, 82, 47-58. CAIAZZO, C., GIOVINE,G., ORTOLANI,E, PAGLIUCA,S., SCHIATTARELLA,M. & VITALE,C. 1992. Genesi ed evoluzione strutturale della depressione tettonica dell'alta valle del Fiume Sele (Appennino campano-lucano). Studi Geologici Camerti, special issue, 1992(1), 245-255. CASNEDI, R. 1988. La Fossa Bradanica: origine, sedimentazione e migrazione. Memorie della Societd Geologica Italiana, 41, 439-448. CATALANO,S., MONACO, C., TORTORICI, L. & TANSI, C. 1993. Pleistocene strike-slip tectonics in the Lucanian Apennine (Southern Italy). Tectonics, 12, 656-665. CELLO, G., LENTINI,E & TORTORICI,L. 1990. La struttura del settore calabro-lucano e suo significato nel quadro dell'evoluzione tettonica del sistema a thrust sudappenninico. Studi Geologici Camerti, special issue, 1990, 27-34. CINQUE, A., PATACCA,E., SCANDONE,P. & TOZZI, M. 1993. Quaternary kinematic evolution of the Southern Apennines. Relationships between surface geological features and deep lithospheric structures. Annali di Geofisica, 36, 249-260. COLELLA, A. & CAPPADONA,P. 1988. Evidenze stratigrafiche del carattere trascorrente sinistro della zona di faglie del Pollino. Valutazioni sulla velocith di trascorrenza. Pre-stampe Atti del 74 ~

353

Congresso Nazionale della Societd Geologica Italiana, Sorrento, Vol. B, 147-150. D'ARGENIO, B., PESCATORE,T. & SCANDONE,P. 1973. Schema geologico dell'Appennino Meridionale (Campania, Lucania). Quaderni Accademia Nazionale dei Lincei, 183, 49-72. D~ NIRO,A., GIANO,S. I. & SANTANGELO,N. 1992. Primi dati sull'evoluzione geomorfologica e sedimentaria del bacino dell'alta val d'Agri (Basilicata). Studi Geologici Camerti, special issue, 1992(1), 257-263. FERRELI, L., M1CHETTI,A. M., SERVA,L., V~TTORI,E. & ZAMBONELLI, E. 1996. Tettonica recente ed evidenze di fagliazione superficiale nella Catena del Pollino (Calabria settentrionale). Memorie della Societd Geologica Italiana, 51, 451-466. GHISETTI,E & VEZZANI,L. 1982. Strutture tensionali e compressive indotte da meccanismi profondi lungo la linea del Pollino (Appennino meridionale). Bollettino della Societ~ Geologica Italiana, 101, 385-440. HANCOCK,P .L. 1985. Brittle microtectonics: principles and practice. Journal of Structural Geology, 7, 437-457. HIPPOLYTE,J. C., ANGELIER,J., ROURE, E & CASERO,E 1994. Piggyback basin development and thrust belt evolution: structural and palaeostress analyses of Plio-Quaternary basins in the Southern Apennines. Journal of Structural Geology, 16, 159-173. IETTO, A. & D'ARGENIO, B. 1990. Some accounts on thrust and subsequent extensional tectonics in the Pollino Mountains, Southern Apennines. Rendiconti della Societgl Geologica Italiana, 13,121-124. - - , BARILARO,A. M., CALLIGARO,G. & MANCUSO,C. 1992. Elementi per una revisione dei rapporti Appennino - Arco Calabro. Bollettino della Societa Geologica Italiana, 111, 193-215. KNOTT, S. D. 1987. The Liguride Complex of Southern Italy - a Cretaceous to Paleogene accretionary wedge. Tectonophysics, 142, 217-226. -1988. Structure, sedimentology and petrology of an ophiolitic flysch terrain in Calabria, Italy. PhD thesis, Oxford University. - & TURCO, E. 1991. Late Cenozoic kinematics of the Calabrian Arc, Southern Italy. Tectonics, 10, 1164-1172. MALINVERNO,g . & RYAN,W. B. E 1986. Extension in the Tyrrhenian Sea and shortening in the Apennines as a result of arc migration driven by sinking of the lithosphere. Tectonics, 5, 227-245. MAURO, A. & SCHIATTARELLA,M. 1988. L'Unitg Silentina di Base: assetto strutturale, metamorfismo e significato tettonico nel quadro geologico dell'Appennino meridionale. Memorie della Societd Geologica Italiana, 41, 1201-1213. MONACO, C., TORTORICI, L., MORTEN, L., CRITELLI, S. & TANSI, C. 1995. Geologia del versante nord-orientale del massiccio del Pollino (confine calabro-lucano): nota illustrativa sintetica della Carta geologica alla scala 1:50 000. Bollettino della Societd Geologica Italiana, 114, 277-291. MOSTARDINI,E & MERLIM, S. 1986. Appennino centro meridionale. Sezioni geologiche e proposta di

354

M. SCHIATTARELLA

modello strutturale. Memorie della Societd Geologica Italiana, 35, 177-202. OGNIBEN, L. 1969. Schema indroduttivo alia geologia del confine calabro-lucano. Memorie della Societgt Geologica Italiana, 8, 453-763. OLDOW, J. S,, D'ARGENIO, B., FERRANTI, L., PAPPONE, G., MARSELLA,E. & SACCHI,M. 1993. Large-scale longitudinal extension in the southern Apennines contractional belt, Italy. Geology, 21, 1123-1126. ORTOLANI, E, PAGLIUCA,S., PEPE, E., SCHIATTARELLA, M. & TOCCACELI,R. M. 1992. Active tectonics in the Southern Apennines: relationships between cover geometries and basement structure. A hypothesis for a geodynamic model. IGCP No. 276, Newsletter, 5, 413-419. PATACCA, E. & SCANDONE, P. 1989. Post-Tortonian mountain building in the Apennines. The role of the passive sinking of a relic lithospheric slab. In: BOR1ANI, A., BONAFEDE, M., PICCARDO, G. B. & VAI, G. B. (eds) The Lithosphere in Italy. Mid-term Conference, Rome, 1987. Atti Accademia Nazionale dei Lincei, 80, 157-176. PERRI, E. & SCHIATFARELLA,M. 1997. Evoluzione tettonica quaternaria del bacino di Morano Calabro (Catena del Pollino, Calabria settentrionale). Bollettino della Societgl Geologica Italiana, 116, 3-15. PESCATORE,T. 1988. La sedimentazione miocenica nell'Appennino campano-lucano. Memorie della Societd Geologica ItaIiana, 41, 37-46. PIERI, P., SABATO,L., LOIACONO,E d~;MARINO,M. 1994. I1 bacino di piggyback di Sant'Arcangelo: evoluzione tettonico-sedimentaria. Bollettino della Societ~ Geologica Italiana, 113, 465-481. , & TROPEANO, M. 1996. Significato geodinamico dei caratteri deposizionali e strutturali della Fossa bradanica nel Pleistocene. Memorie della Societgt Geologica Italiana, 51, 501-515. PROSSER, G., SCHIATrARELLA, M., TRAMUTOLI, M., DOGLIONI, C., HARABAGLIA, E & BIGOZZI, A. 1996. Una sezione rappresentativa dell'Appennino

meridionale. Conferenza sulla Ricerca Scientifica in Basilicata, Universit~t della Basilicata-Regione Basilicata, Potenza, 1996. Russo, E & SCHIATTARELLA,M. 1992. Osservazioni preliminari sull'evoluzione morfostrutturale del bacino di Castrovillari (Calabria settentrionale). Studi Geologici Camerti, special issue, 1992(1), 271-278. SAGNOTn, L. 1992. Paleomagnetic evidence for a Pleistocene counterclockwise rotation of the Sant'Arcangelo basin, southern Italy. Geophysical Research Letters, 19, 135-138. SANTANGELO, N. 1991. Evoluzione stratigrafica, geomorfologica e neotettonica di alcuni bacini lacustri del confine campano lucano (Italia meridionale). PhD thesis, Naples University. SCARASCIA, S., LOZEJ, A. & CASSINIS,R. 1994. Crustal structures of Ligurian, Tyrrhenian and Ionian seas and adjacent onshore areas interpreted from wide-angle seismic profiles. Bollettino di Geofisica Teorica ed Applicata, 36, 5-19. SCHIATFARELLA,M. 1996. Tettonica della Catena del Pollino (Confine Calabro-Lucano). Memorie della Societd Geologica Italiana, 51, 543-566. ~, DOGLIONI, C., PROSSER, G. & TRAMUTOLI, M. 1997. Large-scale geometry and kinematics of the Southern Apennines. Terra Nova, 9, Abstract Supplement 1,109. ~,TORRENTE, M. M. & Russo, E 1994. Analisi strutturale ed osservazioni morfostratigrafiche nel bacino del Mercure (confine calabro-lucano). II Quaternario, 7, 613-626. TURCO, E., MARESCA, R. & CAPPADONA,P. 1990. La tettonica plio-pleistocenica del confine calabrolucano: modello cinematico. Memorie della Societgt Geologica Italiana, 45, 519-529. VAN DIJK, J. 19.& OKKES,M. 1991. Neogene tectonostratigraphy and kinematics of Calabrian basins; implications for the geodynamics of the Central Mediterranean. Tectonophysics, 196, 23-60.

Index Aegean Sea, North, map 77 Africa, SE, Gondwana break-up 203-14 Agulhas Plateau 210, 211 Alps, Central, tectonic map 4 Altai, Jargalant and Sutai Domains 75 Mongolia 75 Alterosa Suture Zone 232, 235 analogue modelling brittle-ductile regimes, SE Africa 209-10 comparison with natural examples 74-7 faulting in transpression/transtension zones 59-80 materials and scaling 60-1 summary and discussion 69-4 transpression experiments 64-6 transtension experiments 66-9 Andes, slip vector ENE 5 Andorra, Mont Louis-Andorra pluton, map 269 anisotropy of magnetic susceptibility (AMS) 267-73 Antarctica Ellsworth Mountains, kinematic partitioning, dextral transpression zone 28%306 and Gondwana breakup 211-12 Apennines s e e Pollino Ridge, Calabria-Lucania boundary 40Ar/39Ar dating 320-1 Asia, East, Tertiary extensional basins, contrasting models 220 Atacama Fault System 127-142 Cenozoic/Mesozoic 129 Avalon Zone, NE Newfoundland Appalachians, ductile transpressional structures 8, 42 Baja-British Columbia terrane, Canadian Cordillera 28 Baltica and Laurentia, transtension and oblique plate divergence 176-7 Baltoro granites, Karakoram Fault Zone 308-9, 314-15 blueschist belts, exhumation, terrane collision 30 Bocono Fault, dextral faulting 76 Bohai Basin, N China, transtensional deformation 215-29 initial Tertiary deformation 219-21 kinematics of Tertiary transtension 224-7 Mesozoic compressional structures, map 221 Middle Eocene to Oligocene deformation 221-3 other Tertiary extensional basins 216 pre-Tertiary structure and stratigraphy 218-19 regional setting 217-18 Tertiary structures 217 Bohemian Massif, Germany 3D retromodelling of transpression 281 deep continental borehole (KTB) 275,277-81 geological setting 275-80 Mesozoic and Cenozoic crustal movements 279-80 Upper Cretaceous deformation event 280-1 Variscan units, map 276 Zone of Erbendorf-Vohenstrauss (ZEV) 277, 283 Bolsa Point Syncline 110, 115 bounding faults and shear zones 27 Bozhong depression 222-3

Brasilia Plate, map 235 Brazil, Precambrian transpression, strain partitioning 231-52 Campo de Meio strike-slip shear belt (CMSB) 231-6, 238-40 map 233 evolution of regional strain regime 247 Guaxup6 Syntaxis 233,241,248 physical experiments 244-7 regional geological structural map 233 regional strain regime 242-3 regional tectonic maps 232, 234 Rio Paraiba do Sul shear belt (RPSSB) 231-48 map 232 structural framework 234-42 tectonic setting of plutonism 243 British Variscides, SW, transpression rotation in external orogenic zones 253-66 California Confidence Hills, Death Valley 74-5 Eastern California shear zone (ECSZ) 185, 188 Franciscan Complex 30 W, kinetic deformation model 116 s e e a l s o San Andreas Fault; Sierra Nevada batholith; Trans Mojave-Sierran shear zone Campania-Lucania Apennines s e e Pollino Ridge, Calabria-Lucania boundary Campo de Meio strike-slip shear belt (CMSB), Brazil 231-6, 238-40 Canadian Appalachians Avalon Zone 8 Cape Breton Island, map 42 Roper Lake shear zone 41-57 Canadian Cordillera Baja-BC terrane 28-9 Coast shear zone 28 Cape Breton Island, Canadian Appalachians 41-57 map 42 carbonate structures, Campania-Lucania Apennines 341-54 Carpathians, Western, transpression rotation in external orogenic zones 255-6, 262-4 regional geological maps 255,256 summary and discussion 262-4 Castrovillari basin, Italy, map 352 Cauchy-Green tensor 49 Cauterets-Panticosa, Pyrenees, plutonic complex 269 map 270 Cerro Chuculay Fault 135, 136 Chile s e e Salar Grande China Tertiary extensional basins 216 s e e a l s o Bohai Basin cluster analysis, stress tensors calculated from fault groups 261 coal-bearing foreland basins, Variscan northern margin in SW Britain 259 computerized tomography (CT), in analogue modelling 63 Confidence Hills, Death Valley 74-5

356

INDEX

Conglomerate Ridge Formation, Ellsworth Mountains, Antarctica 292-300 Cordillera de la Costa, Chile 127-9 Cordillera, Southern s e e Trans Mojave-Sierran shear zone Cretaceous deformation event, West Bohemian Massif, transpression retro-modelling 275-88 Salar Grande pull-apart basin, Atacama Fault System 138 Sierra Nevada batholith 328-40 Dead Sea Transform, Lebanon, restraining bend 81-106 Hasbaya and Rachaiya Faults, maps 93-4 Horns Basalt, K-Ar ages 92 kinematic evolution 102 linking evolution to plate tectonics 103-4 Niha syncline area, map 100 rotation pole 82 search for transcurrent faulting 92-5 seismicity 87-8 tectonic setting 83-7 tectonics and landscape evolution 95-9 Tyre-Nabatiy6 plateau 88, 94-5, 98 Yammouneh Fault 89-92 maps 88, 97-8 Death Valley, Confidence Hills 74-5 dextral faulting Bocono Fault 76 Nazca Plate subduction 137 dextral shear Hercynian orogeny, Pyrenees 268-9 Conglomerate Ridge Formation, Ellsworth Mountains, Antarctica 293-300 rates, cooling and exhumation, Karakoram Fault Zone, North Ladakh 322-3 dextral transpression coalescence pattern 74 kinematic partitioning, Antarctica 289-306 Pyrenees 267-74 Early Miocene collapse, Trans-Mojave-Sierran shear zone, North America 183-202 Eastern California, shear zone, Pacific-North America plate boundary, map 185 Einardsdalen D6collement Zone (EDZ) 175 Ellsworth Mountains, Antarctica Conglomerate Ridge Formation 292-3 dextral shear domains 293-300 Drake Icefall Formation 292-3 geological setting 289-92 Heritage Range 290-2, 303-4 progressive strain partitioning within domains of distributed shear 300-1 Sentinel Range 289 Southern Soholt Peaks kinematic partitioning 292-300 map 294 transpressional model for Heritage Range 301-3 Union glacier Formation 292-3 Watlack Hills and Mount Dolence Faults, map 290 Erbendorf Lineament 276-7

Europe, W, Variscan northern margin, coal-bearing foreland basins, map 258 exhumation of deep crustal rocks by transtension, Western Gneiss Region, Scandinavian Caledonides 159-82 mechanism, oblique plate interactions 177 migmatites and leucogranites, Karakoram Fault Zone, North Ladakh 323 Sierra Nevada batholith 328-30 terrane accretion 29-30 Falklands-Agulhas Fracture Zone 210-11 Fichtelnaab Fault 277-8 finite strain 19-22 finite strain ellipsoid principal axes, stereonet 39 shapes 52 finite strain geometry 19-22 finite strain-related lineations 50-1 Fitzgerald Marine Reserve 111,113 Flinn space and diagram 20, 37-8, 52 flow apophyses 22 foreland basins, Sandersonian strain computation 259-64 Franciscan Complex, California 30 Frankonian Lineament 276-85 Garlock fault, junction with San Andreas Fault 75, 120, 122, 152, 186 Gastre-Agulhas Fault Zone, Triassic-Jurassic 210-11 German Continental Deep Drilling Project (KTB) 277-81 Germany s e e Bohemian Massif gneiss, Western Gneiss Region 159-81 Gondwana break-up Antarctica-Africa fit 210 fold belts of southern continents 290 transtension in SE Africa 203-14 Karoo Supergroup 206 Gondwana break-up transtension in SE Africa (continued) KwaZulu-Natal coastal faulting 206-9 Lebombo monocline 205-6 map 204, 211 Margate and Mzumbe terranes, analogue model 208-10 palaeogeology at time of faulting 210-12 Renken Fault 207,209 Guaxup6 Syntaxis, Brazil Precambrian transpression 233,241,248 Gulf of California, opening 119, 121 Har Us Naar Fault 75 Hasteinen Basin, Standal Fault (NSD) 165-8 Hercynian orogeny, Pyrenees 267-73 evidence for compression during D2 269-71 evidence for dextral shear component of D2 268-9 map 268 syntectonic origin of granitoids 268 Heritage Range, Ellsworth Mountains, Antarctica 290-2, 3034 Highland Boundary Fault Zone, Scotland 304 Hoeppener plot, Gondwana thrust faults 292

INDEX Hornelen Detachment, Standal Fault (NSD) 16%70, 175 hydrocarbons, Tertiary basins, China 215-16 India-Asia collision Tibetan plateau 322 s e e a l s o Karakoram Fault Zone, North Ladakh infinitesimal stretching axes (ISAs) 18-19 Italy s e e Pollino Ridge, Calabria-Lucania boundary Jargalant structural domain 75 Jizhong and Liaohe Depressions, China, maps 222 Juan de Fuca and North American plates, subduction angle 331 Kaapvaal Craton 204 Rooi Rand dyke swarm 206 Karakoram Fault Zone, North Ladakh 307-26 4~ dating 320-1 Baltoro granites 308-10, 313-15 cross-sections 312 Darbuk-Tangste area 322 exhumation of migmatites and leucogranites 323 geological offsets and timing of initiation 321-2 K2-Gasherbrum range 309 maps 308, 310, 311 offset Karakoram granites 313-15 Pangong Injection complex 315 Pangong Range 311-13,323 Karakoram Fault Zone, North Ladakh (continued) Pliocene-Quaternary transpression 323 rates of dextral shear, cooling and exhumation 322-3 Tangtse granite, U-Pb dating 315-20 Tashkurgan graben 309 Tibetan extrusion 323-4 transpression and transtension 309-13 U-Pb SHRIMP dating 316-20 Karoo Supergroup, SE Africa 206-7, 211 kinematic partitioning, dextral transpression zone, South Ellsworth Mountains, Antarctica 289-306 kinematic vorticity 18, 47-8 kinematics, transpressional s e e transpressional kinematics KTB (German Continental Deep Drilling Project) 277-81 Kvamhesten Basin, Standal Fault (NSD) 171 KwaZulu-Natal coastal faulting 206-9 La Negra Formation, Salar Grande 129-31 Ladakh batholith 309 Laurentia and Baltica, transtension and oblique plate divergence 176-7 Lebanon s e e Dead Sea Transform Lebombo monocline, SE Africa 205-6 Liaohe depression, structural map 222 London Brabant Massif 256-7 Luhe Lineament 276-85 magma overpressuring, buoyancy and tectonic forces 331-3 magma-facilitated strike-slip partitioning 330-1 Sumatra 330, 337

357

magmatic arcs 3D geometry 336 interactions between tectonic and magmatic forces 333 and transpressional kinematics 327-40 Mercure Basin, S Italy 344, 349 Merida Andes, Venezuela, structural map and block diagram 76 Meso-Cenozoic platform carbonates, Campania-Lucania Apennines 341-54 migmatites, and leucogranites, Karakoram Fault Zone, North Ladakh 323 modelling s e e analogue modelling; transpression/transtension models Mojave Desert Eastern California shear zone (ECSZ) 185, 188 Kramer Hills, multiple vertical axis rotations 194 s e e a l s o Trans Mojave-Sierran shear zone (TMSSZ) Mojave Extensional Belt 187 Moldanubian Zone, Bohemian Massif 275-87 Mongolia, Altai, Jargalant and Sutai Domains, maps 75 Mono Creek granite, Sierra Nevada batholith 334-5 Mont Louis-Andorra pluton 270 map 269 Montara Mountains La Honda Block 110-14 maps 113 More-Trondelag Fault Zone (MFTZ) 160-1,168, 171,173-7 Moro Calabro Basin, S Italy 350 Mount Lebanon uplift, Laklouk, Qartaba horst, map 99-103 Yammouneh Fault and Homs Basalt relationship, map 90 Nazca Plate and South American Plate, ENE slip vector 5 subduction, dextral faulting 137 New Zealand, transpression/transtension 26 Newfoundland Appalachians, Avalon Zone, ductile transpressional structures 8 Nordfjord-Sogn Detachment (NSD) s e e Scandinavian Caledonides North America, Trans-Mojave-Sierran shear zone, Early Miocene collapse 183-202 North American plate and Juan de Fuca plate, subduction angle 331 s e e a l s o Pacific-North America plate boundary North Salinian s e e Salinian Block, Northern Norway, W, tectonigraphic map 160 oblique plate interactions convergence basal boundary condition, transpressional kinematics 148-50 five reference models 17, 24 partitioning 5-6 divergence, as exhumation mechanism 177 flow apophyses and angle of convergence 17 simple shear, transpression zones 35-40 transpression/transtension models 15-33 Orogenic collapse, 159-77, 183-212.

358

INDEX

orogenic zones, transpression rotation, British Variscides 253-66 Ox Mountain igneous complex 329 Pacific Ocean (proto), subduction beneath Patagonia Block 210-12 Pacific-Asia plate boundary, Tertiary reconstruction 226 Pacific-North America plate boundary 152 California, USA, map 197 Salinian Block, tectonic significance of Palaeogene to Recent shortening 115-16 strike-slip partitioning 115-16 Trans Mojave-Sierran shear zone 183-202 palaeostress determination and analysis 254-6 Patagonia Block, subduction of proto-Pacific Ocean 210-12 Pigeon Point Block, California, map 110 Pilarcitos Fault 113 pluton emplacement, magma overpressuring, buoyancy and tectonic forces 332-3 Point Reyes Peninsula, California, map 114 Pollino Ridge, Calabria-Lucania boundary 341-54 Bifurto Formation 345,346 Castrovillari basin 350-2 cross-section 346 density diagrams of fracture systems 348 geological framework of Calabria-Lucania boundary 343-4 geological setting of S Apennines 342-3 Lauria Mountains 342, 345 maps 342, 344, 345 Mercure Basin 342, 344, 349 Morano Calabro Basin 342, 350-1 Mt Bulgheria 345 stereograms of fault systems 347, 350 tectonic maps 345,346 polydimethylsiloxane(PDMS), analogue modelling 60-3 Precambrian transpressional tectonics, SE Brazil 231-52 Precordillera Fault System 136 Pyrenees dextral transpression 267-74 Hercynian plutons, map 268 see also Cauterets-Panticosa; Trois Seigneurs Quaternary tectonics, Pollino Ridge, Calabria-Lucania boundary 341-54 restraining bends Dead Sea Transform, Lebanon 81-106 San Andreas Fault Big Bend 119-26 Rio Paraiba do Sul shear belt (RPSSB), Brazil 231-48 Rio Seco Fault 135, 137 Roper Lake shear zone, Canadian Appalachians 41-57 strain geometry 53~5 theoretical modelling 46-9 Ross diagram, San Andreas Fault 124 Roum and Yammouneh Faults, Lebanon, map 88 Salar Grande Fault 133, 135

Salar Grande pull-apart basin, Atacama Fault System 127-142 maps 128, 131,133 reconstruction, Early Cretaceous 138 regional geology and structure 127-35 structural evolution 139 summary and discussion 135-8 Salinian Block, Northern Bodega Head Peninsula 114 Bolsa Point Syncline 110,115 Salinian Block, Northern (continued) geological setting and structure 107-10 La Honda Block 110-15 map 108 Pigeon Point Block 110 Point Reyes Peninsula 114 tectonic significance of Palaeogene to Recent shortening 115-16 San Andreas Fault contractional and extensional structures (Big Bend) 119-126 transpressional tectonics 119-20 cross-section and schema 154-5 geological setting 107-8 junction with Garlock fault 75, 120, 122, 152,186 maps 108, 113, 120-1,184 Northern Salinian Block 107-18 Point Arena 115 Ross diagram 124 San Andreas discrepancy 143, 151-2 strike-slip partitioning 143-158 application of model to central California 150-5 simulation 144-50 see also strain partitioning and oblique plate interactions San Gregorio Fault 110 Sandersonian strain computation 259-64 Sao Paulo Plate, map 235 Scandinavian Caledonides, Western Gneiss Region 159-81 Caledonian P - T - t evolution 163-4 Devonian basins 161 east-west trending folds 168-71 geological setting 159-62 late orogenic extension 164-5 Standal Fault (NSD) 165-8 strain estimation 171 Vestranden Gneiss Complex 175 Scotland, Highland Boundary Fault Zone 304 Sensitive High-Resolution Microprobe (SHRIMP) dating 4~ 316-20 U-Pb 316-20 Sentinel Range, Ellsworth Mountains, Antarctica 289 shear strain analysis 35-9 see also dextral shear Sierra Crest shear system 334 Sierra Nevada batholith emplacement model, Mono Creek granite 334-5 interactions between tectonic and magmatic forces 333-5 Kern canyon 334

INDEX magma overpressuring, buoyancy and tectonic forces 331-3 magma-facilitated strike-slip partitioning 330-1 map 328 transpression and exhumation 328-30 s e e a l s o Salinian Block, Northern Silurian, Laurentia and Baltica, oblique plate divergence 176-7 sinistral transpression, coalescence pattern 74 sinistral transtension, strain and displacement indicators 172-4 Soledad Formation 129-31 South America Andes, slip vector ENE 5 Merida Andes, Venezuela, structural map and block diagram 76 proto-Pacific Ocean, subduction beneath Patagonia Block 210-12 s e e a l s o Salar Grande South American Plate, and Nazca Plate, ENE slip vector 5 South Wales coalfield 257 stress tensors calculated from fault groups 261 Southern Soholt Peaks, Ellsworth Mountains, Antarctica 292-300 Standal, Western Gneiss Region, Scandinavian Caledonides, map 166 stereonet, principal axes, finite strain ellipsoid 39 strain analysis, simple shear 35-40 and displacement indicators, sinistral transtension 172-4 ellipsoid principal axes of stereonet 39 variation in shape 21, 36-8, 52 finite 19-22 geometry 48-53 finite strain-related lineations 50-1 infinitesimal stretching axes (ISAs) 18-19 patterns 2-5 strain partitioning 5-6 Precambrian transpression, Brazil 231-52 stress, palaeostress determination and analysis 254-6 stress tensors, calculated from fault groups, S Wales coalfield 261 strike-slip basins s e e Salar Grande strike-slip deformations s e e transpression/transtension zones strike-slip partitioning bulk kinematics simulation, San Andreas Fault 143-158 calculation 47 magma-facilitated 327-40, 330, 337 Sumatra, magmatic arc system 330, 337 terrane accretion 27-30 Baja-British Columbia terrane 28-9 exhumation of deep crustal rocks 29-30 speeds 28-9 terrane collision exhumation of blueschist belts 30 India-Asia 322 Tertiary extensional basins, China 215-16, 219-27

359

Pacific-Asia plate boundary, map 226 Tibet s e e Karakoram Fault Zone, N Ladakh 307-26 Trans Mojave-Sierran shear zone (TMSSZ) 183-202 Early Miocene markers 189-90 Early Miocene palinspastic reconstruction 185-7, 191,195 maps 186, 189-91 Pacific-N America transform boundary, maps 184, 185 shear strain profile 193 structure 187-96 summary and discussion 196-200 transform fault-bounded continental fragment (Salinian Block), Palaeogene to Recent shortening 107-18 transpression analogue modelling experiments 61-6 basic definitions 2, 15 dextral 74, 267-74, 289-306 four plane strain components 37 general model 54 non-partitioned 6 passive planar markers 22-3 rotation, external orogenic zones, British Variscides 253-66 sinistral, coalescence pattern 74 various boundary wall orientations 9 s e e a l s o exhumation transpression/transtension models 15-33 analogue modelling 59-80 five reference models 17, 24 movement and slip rates 25-6 retro-modelling, West Bohemian Massif, Late Cretaceous deformation 275-88 tectonic applications 24-7 transpression/transtension zone properties 1-14 basic definitions 2, 15 constant-volume, with vertical stretch 3, 4 deformations, spectrum 16 fabric patterns and structural style 6 lateral stretch, extrusion and escape 10 modelling 16-33 oblique simple shear 35-40 partitioning styles 7 strain partitioning 5-6, 231-52 strain patterns 2-5, 12 termination modes 11 triclinic symmetry Roper Lake shear zone 42-6 theoretical modelling 46-53 transpressional kinematics basal boundary con0ition, oblique convergence 148-50 Central California 150-5 horizontal extension 146-8 magmatic arcs 327-40 mantle and crustal relationship 153 rotation, W Carpathians and SW British Variscides 253-66 transpressional tectonics Baja-British Columbia terrane 28-9 North America terrane accretion 27 Precambrian, Brazil 231-52 Pyrenees, Hercynian 267-73

360

INDEX

transtension analogue modelling experiments 63, 66-9 examples 59 sinestral, strain and displacement indicators 172--4 Trois Seigneurs, Pyrenees, maps 268, 271 Tyre-Nabatiy6 plateau, Miocene marine terrace 88 U-Pb SHRIMP dating 316-20 UHP rocks, exhumation by transtension, Scandinavian Caledonides, Western Gneiss Region 159-81 Variscan northern margin in SW Britain 256454 coal-bearing foreland basins, maps 259 Variscan tectonometamorphic unit, Zone of Erbendorf-Vohenstrauss (ZEV) 275-87

Venezuela, structural map and block diagram 76 Vestranden Gneiss Complex 175 Vit6ria Plate, map 235 vorticity-normal section (VNS) 47-9 Wales, S Wales coalfield, stress tensors 257, 261 Weddell Sea, Antarctica, and Gondwana breakup 211-12 Western Gneiss Region, Scandinavian Caledonides 159-81 Zone of Erbendoff-Vohenstrauss (ZEV) Bohemian Massif, Germany, map 277,283 NW-SE section 284 tectonic thickening and uplift 284-5

L'. i J

9 .

_

.

.

_

_ __

,

.

r _

_

- ,

_

-

Continental Transpressionai and Transtensional Tectonics edited by R. E. Holdsworth (University of Durham, UK) R. A. St-rachan, (Oxford Brookes University, UK) .and J.~ F. Dewey (Oxford University, UK) Many ipresent'day and. ancient continental deformation systems appear to have formed due: to_. s,ign!ificant!y Oblique retat[ve plate motions, Transpressi(~n a n d tranStension z o n ~ are. formed: Where .the obl~ique motions .invo,l~ve,eom.pone.nts .of:co.mpressiort and exten.sion, respectively, This book covers t:he .:recent .adva,nces ~n, ,o,ur understanding: o,r transpressional, and :tranten.siona.I; deformation tones both ;in,theory and; in rreal geo]og=ica.I;.settings from around thee, world. The volume opens with an up-to-date overview of the topic that sets: the scene for the more detailed papers which fol,low. The papers are grouped into four sections, The first;. Modelling Transpression and Transtension, includes a series of papers that discuss theoretical. strain models in the context of field examples and analogue experiments,. The. second section details the tectonic evolution of Continenta/Transform: Zones and includes papers on the Dead Sea Transform, western USA and Chile, The third section, Oblique Divergence Zones,. has papers on gravitational collapse in the = Norwegian Cal.edonides and in SW North America, the break-up of Gondwana and a pull-apart basin in northern China. T.he~final section, on..Oblique .Convergence Zones, has case. studies from Brazil, European, Varisci,des, Antar~.ica, the Himalayas, the Sierra Nevada batholith :and Italy. 9 9 9 9 9 9

international field of contributors all the main workers in the field are included 370 pages 22 chapters 230 illustrations index

I S B N

1

-

8

6

2

3

9

-

0

0

7

-

X

Cover illustration: WNW-verging folds and overthrust

unit from the transpressional Caledonian orogen, Durham Kli.pp.e, Dronning Louise Land, NE Greenland.. Height of cliff is 400m. Photo by R. E. Holdsworth.

>