Flow Processes in Faults And Shear Zones (Geological Society Special Publication No. 224)

Flow Processes in Faults and Shear Zones

Geological Society Special Publications Society Book Editors R. J. PANKHURST (CHIEF EDITOR) P. DOYLE F. J. GREGORY J. S. GRIFFITHS A. J. HARTLEY R. E. HOLDSWORTH

J. A. HOWE P. T. LEAT A. C. MORTON N. S. ROBINS I P. TURNER

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

It is recommended that reference to all or part of this book should be made in one of the following ways: ALSOP, G. L, HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224. CASCIELLO, E., CESARANO, M. & COSGROVE, J. 2004. Shear deformation of pelitic rocks in a largescale natural fault. In: ALSOP, G. L, HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224,113-125.

GEOLOGICAL SOCIETY SPECIAL PUBLICATION NO. 224

Flow Processes in Faults and Shear Zones

EDITED BY

G. I. ALSOP

University of St Andrews, UK

R. E. HOLDSWORTH University of Durham, UK

K. j. w. MCCAFFREY University of Durham, UK and

M. HAND University of Adelaide, Australia

2004 Published by The Geological Society London

THE GEOLOGICAL SOCIETY

The Geological Society of London (GSL) was founded in 1807. It is the oldest national geological society in the world and the largest in Europe. It was incorporated under Royal Charter in 1825 and is Registered Charity 210161. The Society is the UK national learned and professional society for geology with a worldwide Fellowship (FGS) of 9000. The Society has the power to confer Chartered status on suitably qualified Fellows, and about 2000 of the Fellowship carry the title (CGeol). Chartered Geologists may also obtain the equivalent European title, European Geologist (EurGeol). One fifth of the Society's fellowship resides outside the UK. To find out more about the Society, log on to www.geolsoc.org.uk. The Geological Society Publishing House (Bath, UK) produces the Society's international journals and books, and acts as European distributor for selected publications of the American Association of Petroleum Geologists (AAPG), the American Geological Institute (AGI), the Indonesian Petroleum Association (IPA), the Geological Society of America (GSA), the Society for Sedimentary Geology (SEPM) and the Geologists' Association (GA). Joint marketing agreements ensure that GSL Fellows may purchase these societies' publications at a discount. The Society's online bookshop (accessible from www.geolsoc.org.uk) offers secure book purchasing with your credit or debit card. To find out about joining the Society and benefiting from substantial discounts on publications of GSL and other societies worldwide, consult www.geolsoc.org.uk, or contact the Fellowship Department at: The Geological Society, Burlington House, Piccadilly, London W1J OBG: Tel. +44 (0)20 7434 9944; Fax +44 (0)20 7439 8975; E-mail: [email protected]. For information about the Society's meetings, consult Events on www.geolsoc.org.uk. To find out more about the Society's Corporate Affiliates Scheme, write to [email protected]. Published by The Geological Society from: The Geological Society Publishing House Unit 7, Brassmill Enterprise Centre Brassmill Lane Bath BA1 3JN, UK (Orders: Tel. +44 (0)1225 445046 Fax +44 (0)1225 442836) Online bookshop: http://bookshop.geolsoc.org.uk The publishers make no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility for any errors or omissions that may be made. © The Geological Society of London 2004. 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 itemfee code for this publication is 0305-8719/04/$15.00. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. ISBN 1-86239-153-X

Typeset by Type Study, Scarborough, UK Printed by Cromwell Press, Trowbridge, UK

Distributors USA AAPG Bookstore PO Box 979 Tulsa OK 74101-0979 USA Orders: Tel. +1 918 584-2555 Fax +1 918 560-2652 E-mail [email protected] India Affiliated East-West Press PVT Ltd G-l/16 Ansari Road, Daryaganj, New Delhi 110 002 India Orders: Tel. +91 112 327-9113 Fax+91 112326-0538 E-mail [email protected] Japan Kanda Book Trading Co. Cityhouse Tama 204 Tsurumaki 1-3-10 Tama-shi Tokyo 206-0034 Japan Orders: Tel. +81 (0)423 57-7650 Fax +81 (0)423 57-7651 E-mail [email protected]

Contents ALSOP, G. I. & HOLDSWORTH, R. E. Shear zones - an introduction and overview

1

DIJKSTRA, A. H., DRURY, M. R., VISSERS, R. L. M., NEWMAN, J. & VAN ROERMUND, H. L. M. Shear zones in the upper mantle: evidence from alpine- and ophiolite-type peridotite massifs

11

WHITE, J. C. Instability and localization of deformation in lower crust granulites, Minas fault zone, Nova Scotia, Canada

25

LLOYD, G. E. Microstructural evolution in a mylonitic quartz simple shear zone: the significant roles of dauphine twinning and misorientation

39

PIAZOLO, S., ALSOP, G. L, M0LLER NIELSEN, B. & VAN GOOL, J. A. M. The application of GIS to unravel patterns of deformation in high grade terrains: a case study of indentor tectonics from west Greenland

63

VIGNERESSE, J. L. Rheology of a two-phase material with applications to partially molten rocks, plastic deformation and saturated soils

79

COLLETTINI, C. & BARCHI, M. R. A comparison of structural data and seismic images for low-angle normal faults in the Northern Apennines (Central Italy): constraints on activity

95

CASCIELLO, E., CESARANO, M. & COSGROVE, J. Shear deformation of pelitic rocks in a large-scale natural fault

113

MALTMAN, A. & VANNUCCHI, P. Insights from the Ocean Drilling Program on shear and fluid-flow at the mega-faults between actively converging plates

127

JANSSEN, C, LUDERS, V. & HOFFMANN-ROTHE, A. Contrasting styles of fluid-rock 141 interaction within the West Fissure Zone in northern Chile GUPTA, S. & BICKLE, M. J. Ductile shearing, hydrous fluid channelling and high-pressure metamorphism along the basement-cover contact on Sikinos, Cyclades, Greece

161

ALSOP, G. I. & HOLDSWORTH, R. E. Shear zone folds: records of flow perturbation or structural inheritance?

177

MAC!NNES, E. A. & WHITE, J. C. Geometric and kinematic analysis of a transpression terrane boundary: Minas fault system, Nova Scotia, Canada

201

PEREIRA, M. F. & SILVA, J. B. Development of local orthorhombic fabrics within a simple-shear dominated sinistral transpression zone: the Arronches sheared gneisses (Iberian Massif, Portugal)

215

OCCHIPINTI, S. A. & REDDY, S. M. Deformation in a complex crustal-scale shear zone: Errabiddy Shear Zone, Western Australia

229

BAILEY, C. M., FRANCIS, B. E. & FAHRNEY, E. E. Strain and vorticity analysis of transpressional high-strain zones from the Virginia Piedmont, USA

249

GIORGIS, S. & TIKOFF, B. Constraints on kinematics and strain from feldspar porphyroclast populations

265

GUMIAUX, C., BRUN, J. P. & GAPAIS, D. Strain removal within the Hercynian Shear Belt of Central Brittany (western France): methodology and tectonic implications

287

vi

CONTENTS

CARRERAS, J., DRUGUET, E., GRIERA, A. & SOLDEVILA, J. Strain and deformation history in a syntectonic pluton. The case of the Roses granodiorite (Cap de Creus, Eastern Pyrenees)

307

MOLLI, G. & TRIBUZIO, R. Shear zones and metamorphic signature of subducted continental crust as tracers of the evolution of the Corsica/Northern Apennine orogenic system

321

CHEW, D. M., DALY, J. S., FLOWERDEW, M. I, KENNEDY, M. J. & PAGE, L. M. Crenulation-slip development in a Caledonian shear zone in NW Ireland: evidence for a multi-stage movement history

337

MAZZOLI, S., INVERNIZZI, G, MARCHEGIANI, L., MATTIONI, L. & CELLO, G. Brittle-ductile shear zone evolution and fault initiation in limestones, Monte Cugnone (Lucania), southern Apennines, Italy

353

Index

375

Referees The editors are very grateful to the following people for their help in refereeing papers for this volume. C. M. Bailey, College of William and Mary, USA W. Bailey, CSIRO Petroleum, Australia M. Barchi, Universita di Perugia, Italy M. Bestmann, Liverpool University, UK M. Bickle, Cambridge University, UK M. Brown, University of Maryland, USA J. Carreras, Universitat Autonoma de Barcelona, Spain C. Ehlers, Abo Akademi University, Finland D. Gapais, Universite de Rennes 1, France J. Imber, University of Durham, UK R. Jamieson, Dalhousie University, Canada D. Jiang, University of Maryland, USA S. Johnson, University of Maine, USA R. R. Jones, CognIT, Norway M. Krabbendam, British Geological Survey, UK R. Law, Virginia Tech, USA S. Lin, University of Waterloo, Canada G. E. Lloyd, University of Leeds, UK A. Maltman, Aberystwyth University, UK A. McCaig, University of Leeds, UK B. Miller, San Jose State University, USA T. Needham, University of Leeds, UK G. J. H. Oliver, University of St Andrews, UK C. W. Passchier, Johannes Gutenberg Universitat, Germany

N. Petford, Kingston University, UK G. Potts, Liverpool University, UK G. Roberts, University College London, UK A. Robertson, Edinburgh University, UK C. Rosenberg, Freie Universitat, Germany M. Sandiford, University of Melbourne, Australia S. Sherlock, The Open University, UK R. Sibson, University of Otago, New Zealand C. Simpson, Boston University, USA C. J. Spiers, Utrecht University, The Netherlands R. Strachan, Portsmouth University, UK E. Tavarnelli, Universita di Siena, Italy B. Tikoff, University of Wisconsin-Madison, USA S. Treagus, University of Manchester, UK J. Turner, University of Birmingham, UK A. Vauchez, Universite de Montpellier II, France J.-L. Vigneresse, CREGU, France R. Weinberg, University of Western Australia, Australia M. Wells, University of Las Vegas, USA B. Yardley, University of Leeds, UK

This page intentionally left blank

Shear zones - an introduction and overview G. I. ALSOP1 & R. E. HOLDSWORTH2 Crustal Geodynamics Group, School of Geography &. Geosciences, University of St Andrews, St Andrews, Fife, Scotland, KY16 9AL, UK (e-mail [email protected]) ^Reactivation Research Group, Department of'Earth Sciences, University of Durham, Durham, DH1 3LE, UK l

Faults and their deeper-level equivalents, shear zones are localized regions of higher strain which effectively accommodate differential movement in the Earth's crust and mantle during deformation of the lithosphere. Shear zones may be more precisely defined as approximately tabular regions of concentrated deformation and flow across which adjacent relatively undeformed rock units are offset. They are recognized at all sizes from micro to plate boundary scale (Ramsay 1980; Sornette et al. 1990) (Figs 1 & 2). Faults and shear zones are therefore important examples of the heterogeneous nature of deformation in natural rocks, and profoundly influence the location, architecture and evolution of a broad range of geological phenomena (e.g. Rutter et al. 2001). The topography and bathymetry of the Earth's surface is marked by mountain belts and sedimentary basins which are controlled by faults and shear zones. In addition, faults and shear zones control fluid migration and transport, including hydrothermal fluids and hydrocarbons of economic significance (e.g. McCaig 1997). Magma transport, emplacement and eruption are also frequently controlled by faults and shear zones, as are earthquakes. Once faults and shear zones are established, they are often long-lived features prone to multiple reactivation over very large time-scales (e.g. Holdsworth et al. 1997). Faults and shear zones are typically arranged into complex interlinked networks that permit 3D strain in response to plate tectonic forces (Dewey et al. 1986). However, analysis of ductile shear zones is complicated as they are only directly accessible to geoscientists after exhumation to the Earth's surface. In such cases, the relationships between the observed finite deformation patterns, the preserved microstructures at any given location, and the deformation path and strain rate history are potentially difficult to resolve (e.g. Knipe 1989). It is therefore necessary to first consider the bulk deformation behaviour of the lithosphere and the nature and strength of deformed rocks at depth within shear zones.

Fig. 1. Banded orthogneiss with darker amphibolite layers displaying dextral offset across minor shear zones in West Greenland. Note the attenuation of layering, variable displacement and minor melt component along the shears. Pocket knife for scale.

Strength, strain-rate histories and fault rocks at depth Laboratory-based rock deformation experiments, together with geophysical studies and field observations of natural examples suggest that fault and shear zone deformation products, processes and rheology change with depth

From: ALSOP, G. L, HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224,1-9. 0305-8719/$15.00 © The Geological Society of London 2004.

Fig. 2. Aeromagnetic map of part of the 180 km long Nordre Str0mfjord shear zone in the Nagssugtoqidian belt of West Greenland. Magnetic patterns clearly define the ENE-WSW-trending shear zones which are marked by pronounced swings in the trend of the regional magnetic signature (see Rasmussen & Van Gool 2000 and references therein).

INTRODUCTION TO SHEAR ZONES

Fig. 3. General graphs (partly modified from Knipe 1989) and schematic sketches illustrating variations in shear resistance, deformation regime, fault rocks and strain rate with depth in the lithosphere. The lower diagram represents a schematic profile through crustal faults and shear zones illustrating the frictional to viscous regime together with the typical fault rocks and products of deformation.

(Fig. 3; Sibson 1977, 1983). For a crustal-scale structure, most models suggest that an upper crustal network of brittle faults and cataclastic deformation products will connect directly at depth with a generally broader system of anastomosing shear zones where deformation products are mylonitic in character (Fig. 3). In the upper crustal region, deformation mechanisms involve brittle failure and frictional sliding, with fault

3

strength generally rising with increasing depth due to increasing effective pressure (Fig. 3) (Byerlee 1978; Paterson 1978; Sibson 1983). This frictional deformation regime is often seismogenic. In the shear zones at greater depths, the regime changes to one of viscous flow where deformation is generally considered to be aseismic with dominant operative mechanisms exemplified by crystal plasticity and diffusional creep (Fig. 3) (Sibson 1977; Tullis & Yund 1977; Schmid & Handy 1991). Here, the main controls on strength are temperature, strain rate and grain size. Due to increasing temperature, strength is believed to decrease with increasing depth. The transition between frictional and viscous regimes is typically thought to lie at depths between 10 to 15 km in continental regions and to coincide with the strength maxima of the crust and - in some cases - perhaps the lithosphere (Figs 3 & 4) (Sibson 1983). The character and deformation behaviour within this transition zone is complicated by the pre-existing compositional heterogeneity of rocks on different scales (e.g. Handy 1990) and by changes in composition, deformation mechanisms and rheology that occur during fault rock deformation and associated, often retrograde, metamorphism (e.g. Schmid & Handy 1991; Wintsch et al 1995; Holds worth et al 2001). A number of recent experimental and field-based studies (e.g. Bos & Spiers 2001,2002; Stewart et al 2000; Imber et al 1997) have highlighted the importance of fluidrock interactions and how these may lead to marked fault zone weakening in the region of the frictional viscous transition along crustalscale faults.

4

G. I. ALSOP & R. E. HOLDSWORTH

plete understanding of fault and shear zone processes and their controls, it is necessary to consider the relationships and interactions of processes across all scales.

Lithospheric-scale controls

Fig. 4. Schematic strength profile through the crust and upper mantle illustrating the principal loadbearing regions of the lithosphere. The strength of the crust and mantle are based on quartz-feldspar and olivine rheologies respectively.

Based on what we know regarding strain-rate evolution under different deformation regimes, it seems likely that strain-rate histories will also show fundamental variations with increasing depth (Fig. 3) (Knipe 1989). Although the shallow frictional regime is dominated by cataclasites associated with punctuated high strainrate episodes, the frictional-viscous transition is marked by a greater component of aseismic flow between such short-lived high strain-rate events. The deeper viscous regime is dominated by ductile mylonites in which deformation is accommodated by more continuous and relatively steady flow (Fig. 3). Once again, the strainrate behaviour in the main load-bearing frictional-viscous transition zone will be of considerable importance, with the possibility that periods of unstable viscous flow in mylonites could trigger seismic episodes in the overlying brittle crust (Figs 3 & 4) (e.g. Hobbs et al 1986).

Lithospheric-scale controls on faults and shear zones include the general tectonic regime (shortening, extension and strike-slip) and plate boundary conditions (e.g. Sibson 1983; Teyssier & Tikoff 1998). The thickness of the quartzofeldspathic crust and age of the (mantle) lithosphere, which reflects its thermal history, also provides large-scale controls on lithosphere deformation and therefore fault and shear zone development (e.g. see Tommasi et al. 1995). Experimental data suggest that the primary load-bearing region of the lithosphere lies in the upper mantle, whereas a secondary load-bearing region is developed at the frictional-viscous transition in the crust (Fig. 4). The general view then is that large-scale deformation response is determined by development of shear zones in the upper mantle (e.g. Molnar 1992; RegenauerLieb & Yuen 2004). If the crust and mantle are coupled, then flow within the mantle will ultimately control deformation in the crust (e.g. Tikoff etal. 2002).

Grain-scale controls Grain-scale controls on faults and shear zones include lithological controls such as the composition, mineralogy and grain size of the host, together with the presence of pre-existing fabrics and introduction of a fluid or melt. Local environmental conditions of pressure, temperature and strain rate will also affect faulting and shear zone processes. The influx of hydrous fluids into active deformation zones at depth appears to be particularly important and will lead to profound changes in the chemistry and mineralogy of the host, grain size and dominant deformation mechanisms and ultimately therefore the rheology and strength of the fault/shear zone rock (e.g. Bos & Spiers 2002; Holdsworth et al. 2001).

Controls on fault and shear zone development

Network geometry-scale processes

Three sets of controls on fault and shear zone development exist: lithospheric-scale, network geometry-scale, and grain-scale. Each scale range tends to be the focus of different research communities, e.g. geodynamicists, structural geologists and microstructural geologists, respectively. However, in order to gain a com-

Geometric controls are central to the transfer of displacement between mantle shear zones driving lithosphere deformation and overlying systems of crustal faults and shear zones (e.g. see Tommasi et al. 1995; Tikoff et al. 2002). More fundamentally, they form the key mechanical link between grain-scale and lithospheric-scale

INTRODUCTION TO SHEAR ZONES

5

Fig. 5. Schematic 3D sketch illustrating shear zones anastomosing around low-strain augen. The relatively small amount of mechanically weak but interconnected shear zone rock forms a network which effectively controls the bulk strength of the overall rock volume.

Fig. 6. Diffuse zone of shearing between quartzofeldspathic orthogneiss (right-hand side of photograph) and amphibolite-rich gneisses (left-hand side of photograph) in West Greenland. Note how variable rheology results in boudinage and folding of amphibolites. Hammer for scale (top centre).

processes. Such controls include the size and interconnectivity of adjacent shear zones, the orientation and dip of shear zones in relation to far-field stress, and displacement compatibility. In general, larger and more interconnected shear zones will require smaller volumes of mechanically weak fault rocks in order to form an interconnected weak network that controls the overall strength of the system on crustal or lithospheric scales (Fig. 5). Relatively low strain units may be entirely surrounded by high strain (and weak) interconnected shear zones which effectively mechanically isolate these regions of relatively strong rock as augen (Fig. 5). Displacement compatibility also needs to be considered in relation to how stick-slip faulting at shallow depths relates to steady-state flow at depth. In order to understand faults and shear zones better, we must consider the interaction of all of these factors on lithosphere, grain and networkscale controls. We shall now consider some of the main points arising from this volume.

timescales (e.g. see Tommasi etal. 1995; Vauchez & Tommasi 2003). The present volume opens with seven papers which discuss these topics from deeper to shallower levels using a variety of techniques and scales of observation. Dijkstra et al. describe localization mechanisms of deformation in alpine- and ophiolite-type mantle massifs. They suggest that softening processes responsible for localization are associated with melt-related weakening, together with a change in the dominant deformation mechanism from dislocation to grain size-sensitive creep. Such reactions occur over a broad range of P-T conditions in the upper mantle, suggesting that mantle shear zones are widespread and may significantly reduce the bulk strength of the lithosphere. White describes localization of deformation in lower crust granulites in Nova Scotia. Localization may occur at several different length scales resulting in finer-grained material associated with dislocation creep microstructures. The finest-grained mylonites are marked by the introduction of partial melts and pseudotachylites associated with transient frictional events. Microstructures which formed during an event which initiated an instability can be obliterated by subsequent ductile flow. The role of microstructure in the evolution of an amphibolite facies shear zone is described by Lloyd. This contribution explores the role twinning plays in both the microstructural and petrofabric evolution of shear zones by assisting in the initial grain size comminution processes. Twinning may thus help to accommodate high shear strains whilst maintaining a stable microstructure and constant (single crystal) petrofabric

Lithosphere deformation and rheology of shear zones As noted above, shear zones processes and products change with depth and the type of material being deformed (Fig. 6). The nature of these processes and how they may be linked kinematically and mechanically, particularly across major geological and geophysical interfaces such as the Moho, will profoundly affect the overall rheological behaviour of the lithosphere during deformation on geological

6

G. I. ALSOP & R. E. HOLDSWORTH

which may be detected by seismic investigation of deep shear zones. The testing of remote geophysical techniques via the application of Geographical Information Systems (GIS) to granulite and amphibolite facies shear zones in West Greenland is presented by Piazolo et al. Multidisciplinary geophysical and geological datasets are compared in terms of subject, space and scale and suggest that Theologically weak amphibolite facies rocks are moulded and sheared around a cooled granulite facies block which effectively behaves as a rigid indentor on a regional scale. The rheology of deforming material is also addressed by Vigneresse in terms of two phase materials such as crystallizing magma. The general behaviour of viscosity is taken as a function of the strain rate and the amount of solid phase present. At high strain rates, the viscosity contrast between the two phases at its lowest, whilst at low strain rates, the viscosity contrast between the phases is at its highest. These relationships will clearly control the structures that are developed. The role of large-scale extensional faults in crustal deformation in the Northern Appenines of Central Italy is studied by Colletini & Barchi. Recorded microseismicity suggests that faults are presently active with a vertical o^ and that fluid movement along the gently-dipping fault planes has resulted in the faults being mechanically weak. Casciello et al. also investigate the role of fluids in upper crustal faults via analysis of the hydraulic characteristics and hence frictional properties of sheared clays. Shear strain can induce mineralogical changes in smectite resulting in their replacement with anhydrous illite minerals. Localization of the illitization process along the shear zone may generate water, leading to fluid overpressure and hydraulic circulation. Partitioning processes in shear zones The central section of the book contains a series of eight papers on partitioning and localization processes, which are an increasingly recognized phenomena in shear zones across a range of scales from microstructural to outcrop to regional (Fig. 7). Partitioning may be analysed in terms of the role of fluid flow, together with the localization and focusing of fluids along faults and shear zones. Mega-faults between actively converging plates are described by Maltman & Vannucchi. Lithological influence on localization of fault propagation seems absent with faults responding to continuing deformation by intensifying and focusing strain inwards rather than propagating outward splays. Faults penetrated by the Ocean Drilling Programme may be

Fig. 7. Outcrop-scale dextral shear zone developed in pelite at Cap de Creus, Spain (see Carreras 2001 for further details). Regional deformation is focused into the retrogressive shear zone, which further partitions and localizes deformation. Note the sheared pegmatite and pronounced strike-swing of regional foliation into the shear zone.

just tens of metres in thickness and yet can efficiently channel fluids for distances of tens of kilometres whilst at the same time inhibiting fluid flow across them. Styles of fluid rock interaction within the West Fissure Zone in Chile are examined by Janssen et al. Their results demonstrate considerable variation in the degree of fluid interaction with limited ascending hydrothermal fluids in some traverses, whereas in other cases fluid-enhanced weakening mechanisms are dominant, reflecting the heterogeneity and partitioning of fluid flow in largescale continental fault zones. The partitioning and localization of fluid flow in the Cyclades of Greece is described by Gupta & Bickle. Their work shows an association of higher strain with increased hydration in the footwall to a basement-cover shear zone along which fluids have been focused. The restricted availability of water outside of this zone allows the preservation of earlier metamorphic assemblages which are otherwise destroyed. Partitioning processes in shear zones can also be investigated in terms of deformation

INTRODUCTION TO SHEAR ZONES

localization and partitioning, which results in the multiple generation of a range of structures displaying overprinting relationships at a variety of scales. Alsop & Holdsworth describe deformation within shear zones in which folding displays predictable geometric patterns analysed on fabric topology plots. Multiple generations of folds may be explained in terms of a fold evolution model, where early sheath folds represent a more highly deformed and evolved variety of synshearing folds originally generated during perturbations in ductile flow. Alternatively, the fold inheritance model suggests that the structural architecture generated during sheath folding may subsequently control the geometry and govern the orientation of later synshearing folds. Machines & White describe the largescale partitioning of deformation into distinct external (contractional structures) and internal (shear) zones along the Minas Fault System of Nova Scotia. Ductile structures are composite features derived from the multiple transposition of pre-existing layers, with strain accommodated within progressively narrower volumes of rock marked by contrasting deformation micromechanisms. Pereira & Silva investigate the partitioning of deformation in a transpressive sinistral shear zone in the Iberian Massif of Portugal. Although the shear zone is dominated by widespread monoclinic fabrics, mechanical conditions locally exist in which deformation has been partitioned into orthorhombic fabrics highlighting the importance of scale when investigating deformation processes. Occhipinti & Reddy undertake a detailed investigation of the crustal-scale Errabiddy Shear Zone of Western Australia which shows a temporal evolution from generation of ductile fabrics, to folding to a brittle overprint. Deformation is locally partitioned into simple shear dominated displacement zones that separate regions of flattening and pure shear. Analyses of small segments of shear zones may thus not give the complete history of an evolving shear zone because of strain localization and partitioning over time. Bailey et al. analyse transpressional high strain zones from the Virginia Piedmont of the southern Appalachians. Estimates of strain based on quartz grain shapes, porphyroclasts and folded/boudinaged pegmatites suggest significant strike-slip displacement is associated with orogen-parallel material elongation together with 40-70% contraction normal to this zone.

Unravelling shear zone histories Deformation associated with faults and shear zones may be both spatially and temporally variable resulting in a complex interplay of

7

Fig. 8. Minor aplite intrusion rotated and attenuated by sinistral shears formed within the Roses Granodiorite (Cap de Creus, Spain). Minor intrusions may provide effective markers in the unravelling of shear zone histories (see paper by Carreras et al.).

structures developed across a range of scales. This final section of the book comprises six papers concerned with unravelling shear zone histories through a variety of techniques and approaches. Estimates of strain, metamorphic history, isotopic age dating, fluid influx and minor igneous intrusions acting as 'timemarkers' may all be particularly useful in interpreting and unravelling such histories (Fig. 8). Giorgis & Tikoff constrain the kinematics and finite strain of deformation in shear zones based on 3D numerical models of the rotation populations of rigid clasts. This numerical model suggests that there is no consistent relationship between the asymmetrical orientation of a population of rigid markers and the simple shear component of deformation. This suggests that the asymmetrical alignment of a population of porphyroclasts is not a reliable shear sense indicator. They apply the technique to the Western Idaho Shear Zone and suggest significant oblique convergence during transpressional deformation. Gumiaux et al. undertake a statistical analysis of cleavage orientation data in

8

G. I. ALSOP & R. E. HOLDSWORTH

order to restore major deformation associated with the Hercynian shear belt of central Brittany. Strain patterns reflect domains on different scales and the occurrence of local superimposed deformations highlighting spatial and temporal shear zone evolution on a variety of scales. Carreras et al. investigate the strain and deformation history of a syntectonic pluton in the Eastern Pyrenees. This pluton displays a continuous deformation history from magmatic structures developed during cooling to latestage (solid state) mylonitic fabrics developed along shear zones when the pluton had crystallized. Quantitative strain analysis of magmatic and solid state fabrics suggest that both record bulk finite strains of similar magnitudes, but with marked differences in the distribution of strain which becomes increasingly localized during solid state deformation. Molli & Tribuzio examine the metamorphic and structural history of major shear zones in the Tenda massif of Corsica. Contractional shear zones display a HP/LT metamorphic signature which is overprinted by retrogressive (greenschist facies) exhumation-related extensional structures indicating a protracted metamorphic and structural history. Chew et al. also describe a protracted structural evolution within Caledonian shear zones of NW Ireland. Early recumbent fold nappes are overprinted by dextral strike-slip crenulations which are followed by a subsequent phase of sinistral deformation. Major crustal scale shear zones may thus be reactivated over significant time scales. Mazzoli et al. investigate the structural evolution of brittle-ductile shear zones developed in low grade settings of the Southern Appenines of Italy. Analysis of shear zone related vein arrays suggests that they are controlled by displacement accumulation and shear strain developed at approximately constant temperature. Detailed analysis of such vein fills may therefore provide important constraints on the factors which control fluids and fault development during brittle-ductile deformation. Summary Lithospheric-scale controls on the development of faults and shear zones are clearly important with current work concentrating on crust-mantle linkages, together with the attributes and processes in mantle shear zones. Improved remote sensing of deep shear zones will aid in the interpretation of large-scale deformation localization and magma transport. Grain-scale controls relate to deformation mechanisms and recognition of diffusive creep

in shear zones associated with studies of microrheology. Further work may concentrate on chemical, strain and energy/heat partitioning processes in polymineralic rocks which will undoubtedly be aided by modelling flow and grain-scale deformation mechanisms. Geometric controls on faults and shear zones are reflected in their kinematic and flow characteristics, which may be marked by patterns of vorticity and spin together with strain partitioning and localization. Future work may in general attempt to quantify and model fault and shear zones processes across this range of controls. This may involve the development of new analytical and geochronological techniques, together with the application and integration of analogue/numerical modelling to studies of natural faults and shear zones. Thus, although significant advances have undoubtedly been made in the understanding of fault and shear zone processes, new work may attempt to integrate studies further encompassing a range of techniques varying from fieldbased to remote geophysical, and a range of scales from lithospheric to microscopic. Such a mixed, multidisciplinary approach will provide a more holistic view of fault and shear zone development and the controls and evolution of lithospheric deformation. Most of the papers in this volume were presented at a joint international meeting of the Tectonic Studies Group of the Geological Society, London, the Structural Geology and Tectonics Division of the Geological Society of America and the Geological Society of Australia on the 2-3 September 2002. We would like to thank all contributors, and participants at the Transport and Flow Processes in Shear Zones meeting and field trip. We would also like to thank Lorna Stewart for her help with this Special Publication.

References Bos, B. & SPIERS, C.J. 2001. Experimental investigation into the microstructural and mechanical evolution of phyllosilicate-bearing fault rock under conditions favouring pressure solution. Journal of Structural Geology, 23,1187-1202. Bos, B. & SPIERS, CJ. 2002. Frictional-viscous flow of phyllosilicate-bearing fault rock: Microphysical model and implications for crustal strength profiles. Journal of Geophysical Research, 107, (B2). BYERLEE, J.D. 1978. Friction of rocks. Pure and Applied Geophysics, 116, 615-626. CARRERAS, J. 2001. Zooming on Northern Cap de Creus shear zones. Journal of Structural Geology, 23,1457-1486. DEWEY, J.F., HEMPTON, M.R., KIDD, W.S.F., SAROGLU, F. & SENGOR,A.M.C. 1986. Shortening of continental lithosphere: the neotectonics of Eastern Anatolia- a young collision zone. In: COWARD, M.P &

INTRODUCTION TO SHEAR ZONES RIES, A.C (eds) Collision tectonics. Geological Society, London, Special Publications, 19, 3-36. HANDY, M.R. 1990. The solid-state flow of polymineralic rocks. Journal of Geophysical Research, 95, 8647-8661. HOBBS, B.E., ORD, A. & TEYSSIER, C. 1986. Earthquakes in the ductile regime. Pure and Applied Geophysics, 124, 310-336. HOLDSWORTH, R.E., BUTLER, CA. & ROBERTS, A.M.

1997. The recognition of reactivation during continental deformation. Journal of the Geological Society, London, 154, 73-78. HOLDSWORTH, R.E., STEWART, M., IMBER, J. & STRACHAN, R.A. 2001. The structure and rheological evolution of reactivated continental fault zones: a review and case study. In: MILLER, J.A., HOLDSWORTH, R.E., BUICK, I.S. & HAND, M. (eds) Continental reactivation and reworking. Geological Society, London, Special Publications, 184, 115-137. IMBER, I, HOLDSWORTH, R.E., BUTLER, C.A. & LLOYD, G.E. 1997. Fault-zone weakening processes along the reactivated Outer Hebrides Fault Zone, Scotland. Journal of the Geological Society, London, 154,105-110. KNIFE, R.J. 1989. Deformation mechanisms - recognition from natural tectonites. Journal of Structural Geology, 11,127-146. McCAiG, A.M. 1997. The geochemistry of volatile fluid flow in shear zones. In: HOLNESS, M. (ed.) Deformation enhanced melt segregation and metamorphic fluid transport. Chapman and Hall, 227-260. 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, London, 435-459. PATERSON, M.S. 1978. Experimental rock deformation - the brittlefield.Springer-Verlag, Berlin. RAMSAY, J.G. 1980. Shear zone geometry: a review. Journal of Structural Geology, 2, 83-99. RASMUSSEN, T.M. & VAN GOOL, J.A.M. 2000. Aeromagnetic survey in southern West Greenland: project Aeromag 1999. Geology of Greenland Survey Bulletin, 186, 73-77. REGENAUER-LIEB, K. & YUEN, D.A. 2004. Modeling shear zones in geological and planetary sciences: solid- andfluid-thermal- mechanical approaches. Earth Science Reviews, 63, 295-349. RUTTER, E.H., HOLDSWORTH, R.E. & KNIPE, R.J. 2001. The nature and tectonic significance of fault zone weakening: an introduction. In: HOLDSWORTH, R.E., STRACHAN, R.A., MAGLOUGHLIN, J.F. & KNIPE, R.J. (eds) The nature and tectonic significance of fault zone weakening. Geological Society, London, Special Publications, 186,1-11.

9

SCHMID, S.M. & HANDY, M.R. 1991. Towards a genetic classification of fault rocks: geological usage and tectonophysical implications. In: MULLER, D.W., MCKENZIE, J.A. & WEISSERT, H. (eds) Controversies in modern geology: evolution of geological theories in sedimentology, earth history and tectonics. Academic, San Diego, California, USA, 339-361. SIBSON, R.H. 1977. Fault rocks and fault mechanisms. Journal of Geological Society, London, 133, 191-213. SIBSON, R.H. 1983. Continental fault structure and the shallow earthquake source. Journal of Geological Society of London, 140, 741-767. SORNETTE, D., DAVY, P. & SORNETTE, A. 1990. Structuration of the lithosphere in plate-tectonics as a self-organized critical phenomenon. Journal of Geophysical Research, 95,17353-17361. STEWART, M., HOLDSWORTH, R.E. & STRACHAN, R.A. 2000. Deformation processes and weakening mechanisms within the frictional-viscous transition zone of major crustal-scale faults: insights from the Great Glen Fault Zone, Scotland. Journal of Structural Geology, 22, 543-560. 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. & DEWEY, J.F. (eds) Continental transpressional and transtensional tectonics. Geological Society, London, Special Publications, 135,143-158. TIKOFF, B., TEYSSIER, C. & WATERS, C. 2002. Clutch tectonics and the partial attachment of lithospheric layers. In: BERTOTTI, G, SCHULMANN, K. & CLOETINGH, S.A.P.L. (eds) Continental Collision and the Tectono-sedimentary Evolution of Forelands. EUG Stephan Mueller Special Publication Series, 1, 57-73. TOMMASI, A., VAUCHEZ, A. & DAUDRE, B. 1995. Initiation and propogation of shear zones in a heterogeneous continental lithosphere. Journal of Geophysical Research, 100, 22083-22101. TULLIS, J. & YUND, R.A. 1977. The brittle-ductile transition in feldspathic rocks. Transactions American Geophysical Union, 68,1464. VAUCHEZ, A. & TOMMASI, A. 2003. Wrench faults down to the asphenosphere: geological and geophysical evidence and thermo-mechanical effects. In: STORTI, F, HOLDSWORTH, R.E. & SALVINI,F. (eds) Intraplate strike-slip deformation belts. Geological Society, London, Special Publications, 210,15-34. WINTSCH, R.P, CHRISTOFFERSON, R. & KRONENBERG, A.K. 1995. Fluid-rock reaction weakening of fault zones. Journal of Geophysical Research, 100, 13021-13032.

Shear zones in the upper mantle: evidence from alpine- and ophiolite-type peridotite massifs ARJAN H. DIJKSTRA1, MARTYN R. DRURY2, REINOUD L. M. VISSERS2, JULIE NEWMAN3 & HERMAN L. M. VAN ROERMUND2 1 Department of Applied Geology, Curtin University of Technology, Perth, Western Australia Postal address: GPO Box U1987, Perth, WA 6845, Australia (e-mail: a. dijkstra@curtin. edu. au) 2 Vening Meinesz School of Geodynamics, Utrecht University, Utrecht, The Netherlands 3 Department of Geology and Geophysics, Texas A &M University, College Station, Texas,

USA Abstract: There is abundant field and microstructural evidence for localization of deformation in alpine- and ophiolite-type mantle massifs. On the basis of field relationships and microstructures we recognize two types of tectonite shear zones (medium- to coarse- and fine-grained), as well as two types of mylonitic shear zones (anhydrous and hydrous peridotite mylonites). In tectonite shear zones, softening processes responsible for localization are probably melt-related weakening in the medium to coarse tectonites and a change in limiting slip system in the fine-grained tectonites. In peridotite mylonites, the most likely cause for softening and localization is a change in dominant deformation mechanism from dislocation to grain size sensitive creep. Microstructural and petrological study of mylonite rocks reveals that reactions, either continuous net-transfer reactions (anhydrous and hydrous) or melt-rock reactions, play a key role in the formation of fine-grained material that promotes grain size sensitive creep. These reactions occur over a broad range of pressure-temperature conditions encompassing a large part of the lithospheric upper mantle. We conclude that mantle shear zones are widespread and that they reduce the (bulk) strength of the lithosphere significantly.

Attempts to understand the dynamics of tectonic processes depend critically on quantitative knowledge concerning the mechanical properties of the lithosphere and the relative strength of the crust and upper mantle (e.g. Brun 2002). For many years, it has been accepted that the upper part of the mantle lithosphere is strong compared to the crust (e.g. Kirby 1983; Kuznir & Park 1986). The notion of a strong uppermost mantle is based on extrapolation of laboratory flow laws (Kohlstedt et al 1995) and the depth distribution of seismicity in the lithosphere (Chen & Molnar 1983). The inferred high strength of the shallow mantle implicitly refers to homogeneous flow. However, at large strains, deformation of the mantle lithosphere may become localized within weak shear zones (Fig. 1) so the application of low-strain flow laws may seriously overestimate the strength of the mantle (Rutter & Brodie 1988; Handy 1989; Vissers et al. 1995). In a recent study on the seismicity and elastic thickness of continental lithosphere, Magi et al. (2000) have suggested that the crust is in fact the strongest layer. Thus, it is

emerging that the strength of lithospheric upper mantle may vary considerably. Over the last decade we have studied a number of peridotite massifs, both of orogenic ('alpine-type') and ophiolitic origin, in order to establish under which conditions, and to what extent, deformation in mantle rocks is localized. In this contribution we present an overview of shear localization features in these peridotite massifs. From our data as well as published data from other workers it emerges that, under lithospheric conditions, localization of deformation in mantle rocks is a common phenomenon (Fig. 1; Table I). In addition, shear localization seems to occur over a wide range of pressure-temperature conditions, producing a number of distinct shear zone types. The range of shear zone types is comparable to types of deformation structures found in crustal rocks (e.g. Carreras et al. 1980). Microstructural work has highlighted the importance of mineral reactions as a softening mechanism in the development of mantle mylonite shear zones (Handy 1989; Drury et al.

From: ALSOP, G. L, HOLDSWORTH, R. E., MCCAFFREY, K. IW. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224,11-24. 0305-8719/$15.00 © The Geological Society of London 2004.

12

A. H. DIJKSTRA ET AL.

Fig. 1. Shear zone features in selected mantle massifs (see references in figure). Tectonites in Voltri (Italy), Othris (Greece) and Ronda (Spain) are fine-grained tectonites. Tectonites in Oman are mostly coarse-grained, 'asthenospheric' in character, except for some relatively fine-grained tectonites in the 'sub-Moho shear zone' in the Hilti Massif (see text and Fig. 3).

1990; Newman et al. 1999; Furusho & Kanagawa 1999; Handy & Stunitz 2002). Because of the key role of these reactions, we predict that the mechanical strength of the mantle lithosphere may vary in different tectonic settings, and that the greatest reduction of the bulk strength may be expected in systems in which mantle rocks move quickly through pressure-temperature space, such as in lithosphere-scale extensional detachment and/or subduction zone systems. We also conclude that reaction-derived mantle mylonite zones are relatively stable features, that may be easily reactivated, and that mechanical models of the mantle lithosphere should take the possible presence of weak shear zones into account.

Structures and microstructures in mantle rocks Based on mapping and microstructural analysis of different peridotite massifs, we recognize three main types of deformation structures and microstructures in mantle rocks: medium- to coarse-grained tectonites, fine-grained tectonites and peridotite mylonites. We use the term 'tectonite' to describe strongly deformed rocks with foliation and lineation defined by the shape of pyroxene, spinel and olivine grains. The recrystallized grain size varies from coarse (>4 mm), medium (1-4 mm) to fine (1-0.1 mm). The term mylonite is used to describe a cohesive fault (or shear zone) rock with strong foliation

Table 1. Peridotite massifs with evidence for localized deformation Massif

Type

Type of shear zone

References f

Anita Bay (New Zealand) Balmuccia (Italy)

a* a

Beni Bousera (Morocco) Hidaka (Japan) Horoman (Japan) Lanzo (Italy) Lizard (England) Oman

a t t t t o

Othris (Greece) Ronda (Spain) Table Mountain (Canada) Turon de Tecouere (France)

o a o a

Peridotite mylonites (tp) Peridotite mylonites (e) Hydrous peridotite mylonites (c) Peridotite mylonites (e)

Voltri (Italy)

t

Vourinos (Greece)

o

Raudhaugene (Norway) Zabargad (Red Sea)

a t

Fine grained tectonites (e) Hydrous peridotite mylonites grading into serpentinite mylonites (e) Peridotite gauge zones Hydrous peridotite mylonites grading into serpentinite mylonites (c) Peridotite mylonites (?) Hydrous peridotite mylonites (t)

Hydrous peridotite mylonites (e ) Fine grained tectonites (e) Pseudotachylytes (e?) Mylonites (e) Peridotite mylonites (c) Peridotite mylonites (c) Peridotite mylonites (e) Hydrous peridotite mylonites (e) Coarse grained tectonites (c?) Peridotite mylonites (c)

Wood (1972); Hill (1995) Boudier et al (1984); Brodie & Rutter (1987); Skrotzki et al. (1990) Jin et al. (1998) Reuberetal. (1982) Furusho & Kanagawa (1999) Sawaguchi (2002) Boudier (1978) Cook etal (2000) Ceuleneer et al. (1988); Ceuleneer & Rabinowicz (1992); Dijkstra et al. (20026) Boudier et al. (1988) Rassios & Konstantopoulou (1993); Dijkstra et al. (20020) Van der Wai & Vissers (1993,1996) Suhr (1993) Vissers etal. (1997); Fabrics et al. (1998); Newman et al. (1999) Drury et al. (1990); Vissers et al. (1991); Hoogerduijn Strating et al. (1993); Vissers et al. (1995,1998) Rassios et al. (1994); Rassios & Smith (2002); Personal observation (A.H.D.) Van Roermund & Drury (1998); Unpublished results (H.v. R.) Nicolas et al. (1987); Piccardo et al. (1993)

*)a = alpine-type (subcontinental mantle); o = ophiolitic (oceanic mantle); t = transitional (generally rifted margin) )e = extensional (including diapiric); c = contractional; tp = transpressional; t = transcurrent

f

14

A. H. DIJKSTRA ET AL.

SHEAR ZONES IN MANTLE ROCKS

and fine scale compositional banding resulting from tectonic grain size reduction. Usually, peridotite mylonites have a very fine (0.1-0.01 mm) to ultra-fine (<0.01 mm) grain size. Our terminology for grain size and mylonite definition is based on the recommendations of the IUGS sub-commission on Systematics of Metamorphic rocks (http//:www.bgs.ac.uk/SCMR/). Further detailed descriptions of peridotite microstructures can be found in Mercier & Nicolas (1975); Nicolas (1986); Ceuleneer et al. (1988); Drury et al. (1991); Hoogerduijn Strating et al. (1993); Ildefonse et al. (19980,6); Vissers et al. (1998); Newman et al. (1999); and Dijkstra et al. (20020,6). For references to the peridotite massifs mentioned below, the reader is referred to Table 1.

Medium- to coarse-grained tectonites Medium- to coarse-grained tectonites, equivalent to the so-called 'asthenospheric' microstructures of Nicolas (1986), Ceuleneer et al. (1988) and Ildefonse et al. (19980), have medium- to coarse-grained (often >1 mm) olivine crystals showing extensive evidence for recovery and dynamic recrystallization by subgrain rotation and fast grain boundary migration. There is generally a strong lattice preferred orientation present, with the dominant slip system [a] (010) (where [a] is the slip direction and (010) the slip plane). Medium- to coarse-grained tectonites often contain evidence for the presence of small amounts of interstitial melt during deformation. It is generally assumed that these tectonites form by high temperature (>1200 °C, i.e. near- to supersolidus temperatures), low stress (of the order of a few MPa) deformation, at conditions similar to those in the asthenosphere. These medium- to coarse-grained tectonites are mainly restricted to the mantle sections of ophiolite massifs and are usually non-localized at the scale

15

of the massifs. However, there is evidence for localization of high-temperature, 'asthenospheric' deformation in a 500-800 m wide meltrich, high-strain tectonite horizon immediately below the crust-mantle boundary in the Oman Ophiolite (Fig. 3; Dijkstra et al. 20020). The tectonites in this 'sub-Moho-shear zone' are medium- to fine-grained, with an average recrystallized grain size of about 0.5 mm, but otherwise strongly resemble highly recrystallized, 'asthenospheric' tectonites (Fig. 3).

Fine-grained tectonites Fine-grained tectonites (Fig. 2a), equivalent to the fine-porphyroclastic 'lithospheric' microstructures of Nicolas (1986), Ceuleneer et al. (1988) and Ildefonse et al. (19986), have olivine crystals generally smaller than 1 mm, often with a strongly developed dislocation substructure (undulatory extinction and well defined, closely spaced subgrain walls). Lattice preferred orientations are often consistent with [a] {Okl} slip. These are produced by relatively high stress (generally >10MPa) deformation, at temperatures at which recovery is not very efficient (950-1200 °C). Recrystallization is dominantly by subgrain rotation with moderate grain boundary migration. Fine-grained tectonites have been found in a number of different massifs, including the Voltri (Italy), Othris (Greece), and Ronda (Spain) massifs (Fig. 1). In Voltri, there is good evidence that fine-grained tectonites are essentially a form of localized deformation; in the Tugello and Tobbio areas, fine-grained tectonites occur in a kilometre-scale deformation zone within or adjacent to a kilometre-scale domain of granular peridotites (Fig. 1). Often, however, the limited dimensions of mantle massifs do not allow determination as to whether the deformation in the fine-grained tectonites was homogeneous, or localized in kilometre-scale shear zones.

Fig. 2. Photomicrographs of typical mantle deformation microstructures. Photo (e) in plane polarized light, all others in cross polarized light, (a) Fine-grained tectonite from Othris, Greece. Crystals in field of view are all olivine; (b) Mylonite from Turon de Tecouere, France, showing olivine porphyroclast in a fine-grained matrix consisting of olivine, pyroxene, spinel, plagioclase and minor amphibole; (c) Mylonite from Othris, showing olivine lenses and stretched orthopyroxene porphyroclast in fine-grained matrix of olivine and orthopyroxene; (d) Mylonitic band which formed in a peridotite which shows extensive development of fine-grained, reactionderived olivine and orthopyroxene. Also note the clast-like behaviour of plagioclase in the mylonitic band, suggesting that plagioclase was stronger than the fine-grained matrix during deformation. Sample from Othris; (e) Spinel mylonite from Voltri, Italy, with spinel porphyroclast in a fine-grained polyphase matrix consisting of bands of mostly olivine (light bands) and olivine, pyroxene, spinel and amphibole (dark bands); (f) Coarse amphibole (hornblende) fibres in the pressure shadows of a stretched orthopyroxene porphyroclast. The finegrained material mantling the pyroxene is also mainly amphibole. Sample from Voltri; (g) Fine-grained tectonite from Othris, showing abundant fine-grained olivine and orthopyroxene produced by the breakdown of orthopyroxene porphyroclasts, probably by a reaction involving a melt. The same reaction produced the fine-grained matrix of the Othris mylonites; (h) Peridotite gouge and possible mylonite precursor in dunite sample from Vourinos, Greece, consisting of olivine and minor (opaque) chromite.

16

A. H. DIJKSTRA ETAL.

Peridotite mylonites Peridotite mylonites (Figs 1, 2b-e) generally consist of a very fine grained matrix with olivine grain sizes as small as a few microns. The matrix often contains porphyroclasts of olivine (Fig. 2b) as well as other phases (e.g. spinel, pyroxene, plagioclase, garnet, amphibole; Fig. 2c-f), or lenticular domains of relatively coarse-grained olivine (Fig. 2c). Pyroxene porphyroclasts are often strongly stretched within the foliation plane (Fig. 2c); it is not uncommon to find pyroxene porphyroclasts with aspect ratios more than 10. Peridotite mylonite shear zones occur over a wide range of scales, from millimetrewide bands to kilometre-scale shear zones (e.g. in Othris; Fig. 1). Peridotite mylonites can be anhydrous (Othris; Ronda; Raudhaugene, Norway; and Oman), or weakly hydrous with the growth of small amounts of amphibole (Turon de Tecouere, N. Pyrenees, France). Strongly hydrous peridotites with the development of abundant amphibole (Fig. 2f) are found in Voltri (Hoogerduijn Strating et al. 1993) and in the mantle section of the Lizard ophiolite, England (Cook et al. 2000). The hydrous mylonites of the Voltri massif also contain chlorite, as well as various amounts of antigorite serpentine, and some of these mylonites grade into serpentinite mylonites. The formation of hydrous minerals requires fluid infiltration into the shear zones and the fluids may have been derived from adjacent crustal units during exhumation and after crustal emplacement of the peridotites. Relative softening mechanisms In the absence of large temperature, stress or fluid pressure gradients, shear localization in rocks deforming by crystal plastic or diffusion creep mechanisms requires a softening mechanism that weakens the material inside the shear zone with respect to the material in the wall rock (Bowden 1970; Poirier 1980; White et al. 1980; Hobbs etal. 1990; Drury etal 1991; Rutter 1999). Without a relative softening mechanism, the zone of localized deformation would widen rapidly resulting in relatively homogeneous deformation on the scale of the peridotite body. In many natural shear zones a history of progressive localization can be reconstructed (e.g. Vissers et al. 1991; Drury et al 1991; Jin et al. 1998). Softening mechanisms in general have been discussed in White & Knipe (1978), Poirier (1980) and White et al. (1980). Softening mechanisms in mantle rocks have been discussed in Drury et al. (1991) and Jin et al. (1998).

Below, we summarize the relevant mechanisms responsible for localization of deformation in tectonites and peridotite mylonites.

Relative softening in medium- to coarsegrained peridotite tectonites High temperature mantle flow producing medium- to coarse-grained tectonites appears to be homogeneous on the scale of individual peridotite bodies. Localization of high temperature, 'asthenospheric' mantle flow has so far only been observed in the mantle section of the Oman Ophiolite, in a melt-rich horizon immediately below the crust-mantle boundary (Ceuleneer etal. 1988; Ceuleneer & Rabinowicz 1992). Based on our microstructural study of such a melt-rich shear zone in the Hilti Massif (Fig. 3), we concluded that melt weakening was responsible for softening of the peridotites in the shallowmost 500-800 m of the mantle section (Dijkstra et al. 20020). Weakening of olivine aggregates deforming by dislocation creep containing an interstitial melt fraction is observed in laboratory deformation experiments (Hirth & Kohlstedt 1995; Bai et al. 1997). This weakening effect, which is proportional to the melt content but which only becomes significant when more than 4% melt is present, is probably related to an enhancement of grain boundary sliding and/or diffusion along melt-wetted grain boundaries (Hirth & Kohlstedt 1995). As this type of melt weakening requires high melt contents, higher than normally present in the upper mantle, shear localization of a result of melt weakening is only feasible in regions of melt accumulation (Dijkstra et al. 20020). Holtzmann et al. (2003) have recently shown that the compaction length of the melt flow system controls the spacing of localized shear zones produced by melt weakening.

Relative softening in fine-grained tectonites Drury et al. (1991) argued that shear heating could probably be ruled out as a softening mechanism in fine-grained tectonites. Recent experimental studies suggest a possible softening mechanism related to a change in the ratecontrolling slip system in olivine (Fig. 4). In olivine aggregates deforming by dislocation creep a transition may occur from creep controlled by the hard [c](010) system, to creep controlled by the weak [a]-slip system accommodated by grain boundary sliding and diffusion (see Hirth & Kohlstedt 1995 for details). Dislocation creep in fine-grained

SHEAR ZONES IN MANTLE ROCKS

17

Fig. 3. Cross-section across the Hilti mantle section, Oman, showing the presence of a melt-rich, tectonite shear zone immediately below the crust-mantle boundary. Also shown are tracings of representative microstructures in dunites from shear zone and wall rock (modified from Dijkstra et al. 20020).

Fig. 4. Olivine deformation mechanism map for the Othris fine grained tectonites, showing the fields for [c]- and [a]-slip limited creep. Deformation conditions for the Othris tectonites are shown by grey field. Weakening in fine-grained tectonites may be the result of switching off the hard [c]-slip system. For details on flow laws used see Dijkstra et al. (20026).

Fig. 5. Scanning electron microscope (SEM) backscatter image of grain boundary alignments in Turon de Tecouere mylonites.

controlled creep to [a]-slip controlled creep (Jin et al. 1998).

Relative softening in peridotite mylonites olivine limited by the weak [a]-slip system can be up to two orders of magnitude faster than creep limited by hard [c]-slip (Mackwell et al. 1990; Drury & Fitz Gerald 1998). Concentration of deformation in a tectonite shear zone, causing faster grain size reduction by dynamic recrystallization within the shear zone compared to the wall rock, may thus lead to strain softening of the shear zone tectonites by a change from [c]-slip

Peridotite mylonite zones are characterized by the small grain size of the matrix. There is often evidence that grain size sensitive (GSS) creep probably a combination of diffusion creep plus grain boundary sliding - has been the dominant deformation mechanism particularly in the polyphase layers (Drury et al. 1991; Newman et al. 1999; Jaroslow et al. 1996; Dijkstra et al. 20025). For instance, in the Turon de Tecouere

18

A. H. DIJKSTRA ET AL.

Fig. 6. SEM Backscatter images showing evidence for production of fine-grained material by reactions, (a) Clinopyroxene porphyroclast reacting to fine-grained plagioclase+spinel-lherzolite assemblage in Turon de Tecouere mylonites; (b) Spinel porphyroclast from coarse spinel-lherzolite assemblage reacting to finegrained plagioclase+spinel-lherzolite assemblage in Turon de Tecouere mylonites; (c) Garnet porphyroclast in garnet-lherzolite assemblage reacting to new garnet+spinel-lherzolite assemblage in Raudhaugene mylonites.

mylonites (Fig. 1), there is only a weak lattice preferred orientation in the matrix olivine and the matrix grains are virtually dislocation free (Newman et al. 1999). In addition, grain boundaries in the matrix are often aligned subparallel to the foliation (Fig. 5), over several grain lengths, and these alignments are interpreted as sliding surfaces (Newman et al. 1999). The Othris mylonites (Fig. 1) are characterized by the absence of a lattice preferred orientation in the matrix grains and by numerous grain boundary alignments (Dijkstra et al. 20026). This evidence suggests that softening and localization in mylonites are the result of a change of deformation mechanism, from dislocation to grain size sensitive creep, brought about by extreme grain size reduction. Detailed microstructural and petrological work in peridotite mylonites in Turon de Tecouere, Othris, and Raudhaugene has further shown that reactions play a key role in produc-

ing the very fine grain size of the matrix. Newman et al. (1999) demonstrated that in Turon de Tecouere small, micron-sized grains were produced by a continuous net transfer reaction associated with the transformation of spinel-lherzolite assemblage into a plagioclaseIherzolite assemblage (Fig. 6a,b). This reaction took place over a broad pressure range of 5-11 kbar (Fig. 7). The same reaction causing grain size reduction was also responsible for the development of the Hidaka mylonites (Furusho & Kanagawa 1999). Similarly, grain size reduction associated with the formation of mylonites in the Raudhaugene peridotite body in western Norway (Van Roermund & Drury 1998; Van Roermund et al. 2001) involved a continuous reaction in which a garnet-lherzolite assemblage was transformed into a spinel-lherzolite assemblage (Fig. 6c). In the Othris peridotites, a reaction between harzburgites and a percolating melt led to the formation of a mixture of

SHEAR ZONES IN MANTLE ROCKS

Fig. 7. Pressure-temperature grid with exhumation paths for Pyrenean ('Py'; e.g. Turon de Tecouere) and Norwegian ('N'; e.g. Raudhaugene) peridotites, showing the possible reaction lines that may be responsible for reaction-softening. Note that, as phases involved are solid solutions, reactions occur continuously over a range of pressures and temperatures (shaded fields).

fine-grained material which later formed the matrix for the peridotite mylonites (Fig. 2d,g; Dijkstra et al. 20026). A similar sort of reaction is observed in Lizard (Cook et al. 2000). Finally, in Voltri, amphibole-, chlorite-, and antigoriteproducing reactions played a role in the development of the hydrous mylonites (Fig. 2e,f; Hoogerduijn Strating et al. 1993). Clearly, these examples show how that reaction-enhanced softening, by both subsolidus and near-solidus reactions, as well as by hydrous and anhydrous reactions, is an important mechanism leading to grain size reduction and localization in mantle rocks during metamorphism. Metamorphic reactions may also induce softening owing to transformational induced plasticity (White & Knipe 1978). Reactions do not always produce grain size reduction. In fact, grain size reduction should occur during retrograde reactions, whereas, prograde reactions may result in coarsening and strain hardening. Metamorphic reactions are often considered to occur over a limited range of P-T conditions. Many reactions in peridotites, however, are continuous, involving solid solution phases, so the reactions occur over a broad region of

19

pressure-temperature space spanning a large part of the lithospheric mantle (Fig. 7). Moreover, because of the polyphase nature of finegrained reaction products, subsequent grain growth after the reaction is inhibited (Olgaard 1990). This means that once formed, reactionderived mylonites are relatively stable features, which can be reactivated during subsequent deformation events. Another consequence of the key role of reactions in the development of peridotite mylonites is that fine grained rocks, such asfine-grainedtectonites - if polyphase might be better candidates for future mylonites than coarse-grained, undeformed rocks because reaction kinetics will tend to be faster in finergrained rocks. The growth of hydrous minerals in mylonites requires fluid-infiltration and probably a component of brittle-dilatant deformation. It is therefore possible that hydrous peridotite mylonites form by localization of the brittle component of deformation (Hobbs et al. 1990; Drury et al. 1991; Jin et al 1998; Handy and Stiinitz 2002). Cataclasis may also contribute to grain size reduction (Jaroslow et al. 1996; Handy & Stiinitz 2002). In the mantle section of the Vourinos Ophiolite, Greece, there is a close spatial and temporal association of brittle gouge zones, hydrous peridotite mylonites and serpentinite mylonites, which are often the locus of high-temperature mineralizing fluid flow (Rassios et al. 1994). Some of the peridotite mylonites in Vourinos may have started off as fine grained brittle gouge zones such as shown in Fig. 2h. Gouge zones and serpentinite mylonites may be formed after the peridotite bodies are emplaced in the upper crust, but low-temperature fault gouges and mylonites like this should also occur in the shallow mantle beneath oceanic crust and thinned continental margins. Discussion There is now a wealth of field and detailed microstructural observations suggesting that localization of deformation in tectonite and mylonite shear zones may be the rule rather than the exception in mantle rocks. Evidence for localized deformation is found in alpine, ophiolite and transitional type mantle massifs. Similar evidence for tectonite and mylonite mantle structures is also found in peridotite xenoliths from alkali basalts (e.g. Mercier & Nicolas 1975; Hunter et al. 1984; Cabanes & Brique 1986; Downes 1987,1990; Cabanes & Mercier 1988; Xu et al. 1993). There remains, however, the question as to how representative such structures are for true mantle deformation. The structures

20

A. H. DIJKSTRA ET AL.

Fig. 8. Possible tectonite and mylonite shear zones in extensional and transcurrent systems under hydrous conditions. In transcurrent settings, the formation of reaction-derived mylonites is much less likely than in extensional (or contractional) systems, as mantle rocks do not move rapidly through pressure-temperature space. Lithospheric weakening can be caused by the formation of tectonite shear zones in such settings. The formation of hydrous mylonites as the result of ingress of fluids can lead to the formation of mylonites in the roots of transcurrent systems.

present in exposed upper mantle rocks may have been formed when the peridotites were part of the crust, or in-situ within the upper mantle. Insitu temperature (T) and pressure (P) conditions vary widely within the upper mantle, ranging from very low T and P conditions beneath oceanic crust through to ultra-high P and Tin the convecting upper mantle. In the ophiolitic massifs discussed, relations between deformation and magmatism (Othris, Oman) and/or high-temperature fluid infiltration (Vourinos) demonstrate that deformation structures are related to on-ridge or early, intra-oceanic, deformation events taking place well before the rocks became emplaced onto continental margins. There is little doubt that such structures were produced within the shallow oceanic mantle. In continental settings, there is considerable overlap between P—T conditions within the shallow upper mantle and crust, depending on crustal type and thickness, making it difficult to identify mantle or crustal structures in alpinetype peridotite bodies solely on the basis of P-T conditions of deformation. Even so, we argue that there is abundant evidence for localization of deformation in rocks with mantle compositions under conditions that are at least equivalent to those in the mantle lithosphere. Experimental studies show that quartzo-feldspathic crustal rocks are significantly weaker than dry olivine-pyroxene mantle rocks (Kohlstedt et al. 1995). As a result, dry mantle rocks will behave as strong bodies within the crust after incorporation into that crust. The common discordance of structures in peridotite bodies compared to crustal country rock structure is consistent with limited deformation of the peridotites after crustal emplacement. A weak serpentinite rim and internal serpentinite shear zones usually accommodate any crustal defor-

mation in peridotite bodies. Some examples of intense coherent deformation of crust and mantle rocks occur in hydrated peridotite bodies included in high-pressure crustal terrains (Wood 1972; Cordellier et al. 1981). The extensive deformation of these mantle rock bodies within the crust can be explained by water weakening of olivine (Chopra & Paterson 1984). Shear zone structures in ophiolitic massifs are generally contractional, related to the ophiolite emplacement, whereas shear zones in alpinetype peridotite massifs are often related to extensional exhumation of the massifs. This is essentially a sampling bias because we tend to observe the structures that have brought the mantle massifs to the earth surface. In order to extrapolate our observations to the unexposed part of the mantle it is crucial to understand the processes that are responsible for shear zone development. As discussed above, we have found that during exhumation of the mantle footwall in an extensional detachment system (Fig. 8), drastic weakening occurs through the development of peridotite mylonites by deformation and reactions in mantle mineral assemblages under changing pressure and temperature conditions. The same processes may produce similar mylonites during burial in a thickening lithospheric mountain root, or in a subducting slab. In contrast, we expect that peridotite tectonites rather than mylonites are the dominant mode of localized deformation in the root zones of lithospheric scale transcurrent shear zones. In such systems driving forces for reactions are absent, as mantle rocks are essentially stationary in pressure-temperature space. Only as soon as fluid infiltration and/or brittle processes set in may very fine-grained mylonites be produced, leading to significant weakening of the lithospheric mantle (Fig. 8).

SHEAR ZONES IN MANTLE ROCKS

Implications for the strength of the mantle The upper part of the mantle is potentially the strongest part of the subcontinental lithosphere (e.g. Kirby 1983; Kuznir & Park 1986), however, the development of mantle shear zones may reduce the strength of the mantle lithosphere significantly (Rutter & Brodie 1988). Under some conditions, mantle lithosphere containing mylonite shear zones can be shown to be weaker than dry lower crustal, feldspar-dominated rocks (Fig. 2d; Dijkstra et al 20026). Mantle shear zones will, when present, play an important role during mountain building or during continental break-up (Vissers et al 1995; Vissers et al 1997; Handy & Stunitz 2002). The notion of a relatively weak continental mantle lithosphere is supported by recent assessments of lithosphere strength based on the depth distribution of seismicity (Magi et al 2000). Mylonite shear zones may be relatively stable structures that are not easily annealed (probably in contrast to tectonite shear zones) because of the polyphase fine-grained matrix in reactionderived peridotite mylonites (which suppresses grain growth). Mylonitic shear zones may therefore persist, and be easily reactivated during subsequent deformation events (e.g. Bailey et al 2000). Models of the strength of the mantle lithosphere need to take the presence of such shear zones, which act as pre-existing weaknesses, into account. In Turon de Tecouere it was found that strain rates in ultramylonitic and protomylonitic peridotites were equal, but that stresses were two times lower within the ultramylonites (Newman & Drury 2000). In Othris, estimated strain rates in peridotite mylonites were up to four orders of magnitude faster than strain rates in the wall rock (Dijkstra et al 20025). Conversely, lithospheric mantle containing only one volume percent of Othris-like, finegrained mylonitic material, for instance in an anastomosing network of shear zones, has a bulk strength that is a hundred times lower than mantle lithosphere without any mylonites. Conclusions Shear localization is a common process in mantle peridotite massifs. Reactions play a key role in the development of many peridotite mylonite zones, and these reactions occur over a large pressure and temperature range, both under hydrous/anhydrous and subsolidus/nearsolidus conditions. Shear zones are probably omnipresent in the lithospheric mantle and need to be taken into account in modelling studies. They may reduce the bulk strength of the mantle by up to two orders of magnitude.

21

This contribution is partly based on the work of many students and postgraduate students at Utrecht University. In particular, Eilard Hoogerduijn Strating and Dirk Van der Wai have made crucial contributions. Geoff Lloyd and Mike Bickle are thanked for numerous constructive comments and suggestions. This work has been supported by NWO Pioneer subsidy No. 030-75-346.

References BAI, Q., JIN, Z-M. & GREEN II, H. 1997. Experimental investigation of the rheology of partially molten peridotite at upper mantle pressures and temperatures. In: HOLNESS, M.B. (ed.) Deformationenhanced fluid transport in the Earth's crust and mantle. Chapman & Hall, London, 40-61. BAILEY, W.R., HOLDSWORTH, R.E. & SWARBRICK, R.E. 2000. Kinematic history of a reactivated oceanic suture: the Mamonia Complex Suture Zone. Journal of the Geological Society, London, 157, 1107-1126. BOUDIER, F. 1978. Structure and petrology of the Lanzo peridotite massif (Piedmont Alps). Geological Society of America Bulletin, 89,1574-1591. BOUDIER, E, JACKSON, M. & NICOLAS, A. 1984. Structural study of the Balmuccia Massif (Western Alps): a transition from mantle to lower crust. Geologic en Mijnbouw, 63,179-188. BOUDIER, E, CEULENEER, G. & NICOLAS, A. 1988. Shear zones, thrusts and related magmatism in the Oman ophiolite: initiation of thrusting on an oceanic ridge. Tectonophysics, 151, 275-296. BOWDEN, PB. 1970. A criterion for inhomogeneous plastic deformation. Philosophical Magazine, 22, 455-462. BRODIE, K. & RUTTER, E. 1987. Deep crustal extensional faulting in the Ivrea Zone of northern Italy. Tectonophysics, 140,193-212. BRUN, J.-P 2002. Deformation of the continental lithosphere: insights from brittle-ductile models. In: DE MEER, S., DRURY, M.R., DE BRESSER, J.H.P & PENNOCK, G.M. (eds) Deformation mechanisms, rheology and tectonics: current status and future perspectives. Geological Society, London, Special Publications, 200, 355-370. CABANES, N. & BRIQUE, L. 1986. Hydration of an active shear zone: Interactions between deformation, metasomatism and magmatism: the spinel Iherzolites from the Montferrier (Southern France) Oligocene basalts. Earth and Planetary Science Letters, 81, 233-244. CABANES, N. & MERCIER, J.-C.C. 1988. Insights into the upper mantle beneath an active extensional zone: the spinel-peridotite xenoliths from San Quintin (Baja California, Mexico). Contributions to Mineralogy and Petrology, 100, 374-382. CARRERAS, J., COBBOLD, PR., RAMSAY, J.G. & WHITE, S.H. 1980. Shear zones in rocks. Journal of Structural Geology, 2,1-287. CEULENEER, G. & RABINOWICZ, M. 1992. Mantle flow and melt migration beneath oceanic ridges: models derived from observations in ophiolites. In: MORGAN, J.P, BLACKMAN, D.K. & SINTON, J.M.

22

A. H. DIJKSTRA ET AL.

(eds) Mantle flow and melt generation at midoceanic ridges, Geophysical Monograph 71, American Geophysical Union, 123-154. CEULENEER, G, NICOLAS, A. & BOUDIER, F. 1988. Mantle flow patterns at an oceanic spreading center: the Oman peridotites record. Tectonophysics, 151,1-26. CHEN, W.-P. & MOLNAR, P. 1983. Focal depths of intracontinental and intraplate earthquakes and their implications for the thermal and mechanical properties of the lithosphere. Journal of Geophysical Research, 88,4183-4214. CHOPRA, P.N. & PATERSON, M.S. 1984. The role of water in the deformation of dunite. Journal of Geophysical Research, 89, 7861-7876. CORDELLIER, F., BOUDIER, F. & NlCOLAS, A.

1981.

Structural study of the Almklovdalen peridotite massif (Southern Norway). Tectonophysics, 77, 257-281. COOK, C.A., HOLDSWORTH, R.E., STYLES, M.T. & PEARCE, J.A. 2000. Pre-emplacement structural history recorded by mantle peridotites: an example from the Lizard Complex, SW England, Journal of the Geological Society, London, 157, 1049-1064. DIJKSTRA, A.H., DRURY, M.R. & FRIJHOFF, R.M. 20020. Microstructures and lattice fabrics in the Hilti mantle section (Oman Ophiolite): evidence for shear localization and melt weakening in the crust-mantle transition zone? Journal of Geophysical Research, 107, Bll, 2270, doi: 10.1029/2001JB000458. DIJKSTRA, A.H., DRURY, M.R., VISSERS, R.L.M. & NEWMAN, J. 2002£. On the role of melt-rock reaction in mantle shear zone formation in the Othris Peridotite Massif (Greece). Journal of Structural Geology, 24,1431-1450. DOWNES, H. 1987. Relationship between geochemistry and textural type in spinel Iherzolites, Massif Central and Languedoc, France. In: NIXON, PH. (ed.) Mantle xenoliths. John Wiley & Sons Ltd, New York, 125-133. DOWNES, H. 1990. Shear zones in the upper mantle: relation between geochemical enrichment and deformation in mantle peridotites. Geology, 18, 374-377. DRURY, M.R. & FITZ GERALD, J.D. 1998. Mantle rheology: insights from laboratory studies of deformation and phase transition. In: JACKSON, I.N.S. (ed.) The Earth's mantle - composition, structure and evolution. Cambridge University Press, 503-559. DRURY, M.R., HOOGERDUIJN STRATING, E.H. & VISSERS, R.L.M. 1990. Shear zone structures and microstructures in mantle peridotites from the Voltri Massif, Ligurian Alps, NW Italy. Geologie en Mijnbouw, 69, 3-17. DRURY, M.R., VISSERS, R.L.M., VAN DER WAL, D. & HOOGERDUIJN STRATING, E.H. 1991. Shear localization in upper mantle peridotites. Pure and Applied Geophysics, 137, 439-460. FABRIES, J., LORAND, J.-P & BODINIER, J.-L. 1998. Petrogenetic evolution of erogenic Iherzolite massifs in the central and western Pyrenees. Tectonophysics, 292,145-167.

FURUSHO, M. & KANAGAWA, K. 1999. Transformationinduced strain localization in a Iherzolite mylonite from the Hidaka metamorphic belt of central Hokkaido, Japan. Tectonophysics, 313, 411-432. HANDY, M.R. 1989. Deformation regimes and the rheological evolution of fault zones in the lithosphere: the effects of pressure, temperature, grain-size and time. Tectonophysics, 163,119-152. HANDY, M.R. & STUNITZ, H. 2002. Strain localization and reaction weakening: a mechanism for initiating exhumation of subcontinental mantle beneath rifted margins. In: DE MEER, S., DRURY, M.R., DE BRESSER, J.H.P. & PENNOCK, G.M. (eds) Deformation mechanisms, rheology and tectonics: current status and future perspectives. Geological Society, London, Special Publications, 200, 387-407. HILL, EJ. 1995. A deep crustal shear zone exposed in western Fiordland, New Zealand. Tectonics, 14, 1172-1181. HIRTH, G. & KOHLSTEDT,D.L. 1995. Experimental constraints on the dynamics of the partially molten upper mantle 2. Deformation in the dislocation creep regime. Journal of Geophysical Research, 100,15441-15449. HOBBS, B.E., MUHLHAUS, H.-B. & ORD, A. 1990. Instability, softening and localization of deformation. In: KNIPE, R.J. & RUTTER, E.H. (eds) Deformation mechanisms, rheology and tectonics. Geological Society, London, Special Publications, 54, 143-165. HOLTZMAN, B.K., GROEBNER, N.J., ZIMMERMAN, M.E., GINSBERG, S.B. & KOHLSTEDT, D.L. 2003. Deformation-driven melt segregation in partially molten rocks. Geochemistry, Geophysics and Geosystems, 4, doi: 10.1029/2001GC000258. HOOGERDUIJN STRATING, E.H., RAMPONE, E., PicCARDO, G.B., DRURY, M.R., VISSERS, R.L.M. 1993. Subsolidus emplacement of mantle peridotites during incipient oceanic rifting and opening of the Mesozoic Tethys (Voltri Massif, NW Italy). Journal of Petrology, 23, 901-927. HUNTER, R.H., UPTON, B.C. & ASPEN, P. 1984. Metaigneous granulite and ultramafic xenoliths from basalts of the Midland Valley of Scotland: petrology and mineralogy of the lower crust and upper mantle. Transactions of the Royal Society of Edinburgh: Earth Sciences, 75, 75-84. ILDEFONSE, B., BOUDIER, F. & NICOLAS, A. 19980. Asthenospheric deformation in ophiolitic peridotite. In: SNOKE, A.W., TULLIS, J. & TODD, V.R. (eds) Fault-related rocks - a photographic atlas. Princeton University Press, Princeton, New Jersey, 588-589. ILDEFONSE, B., BOUDIER, F. & NICOLAS, A. 19986. Lithospheric deformation in ophiolitic peridotite. In: SNOKE, A.W., TULLIS, J. & TODD, V.R. (eds) Fault-related rocks - a photographic atlas. Princeton University Press, Princeton, New Jersey, 590-591. JAROSLOW, G.E., HIRTH, G. & DICK, H.J.B. 1996. Abyssal peridotite mylonites: implications for grain-size sensitive flow and strain localization in the oceanic lithosphere. Tectonophysics, 256, 17-37.

SHEAR ZONES IN MANTLE ROCKS JIN, D., KARATO, S.-I. & OBATA, M. 1998. Mechanisms of shear localization in the continental lithosphere: inference from the deformation microstructures of peridotites from the Ivrea zone, northwestern Italy. Journal of Structural Geology, 20,195-209. KiRBY,S.H. 1983. Rheology of the lithosphere. Reviews of Geophysics and Space Physics, 21,1458-1487. KOHLSTEDT, D.L., EVANS, B. & MACKWELL, S.M.

1995.

Strength of the lithosphere: constraints imposed by laboratory experiments. Journal of Geophysical Research, 100,17587-17603. KUZNIR, N.J. & PARK, R.G. 1986. Continental lithosphere strength: the critical role of lower crustal deformation. In: DAWSON, IB., CARSWELL, D.A., HALL, J. & WEDEPOHL, K.H. (eds) The nature of the lower continental crust. Geological Society, London, Special Publications, 24, pp. 79-93. MACKWELL, S.J., BAI, Q. & KOHLSTEDT, D.L. 1990. Rheology of olivine and the strength of the lithosphere. Geophysics Research Letters, 17, 9-12. MAGI, A., JACKSON, J.A., MCKENZIE, D. & PRIESTLY, K. 2000. Earthquake focal depths, effective elastic thickness, and strength of the continental lithosphere. Geology, 28, 495-498. MERCIER, J.-CC. & NICOLAS, A. 1975. Textures and fabrics of upper-mantle peridotites as illustrated by xenoliths from basalts. Journal of Petrology, 16, 454-487. NEWMAN,! & DRURY,M.R. 2000. Deformation history along a peridotite shear zone with implications for dynamic processes within the upper mantle. Geological Society of America Abstracts with Programs, 32. NEWMAN, J., LAMB, W.M., DRURY, M.R. & VISSERS, R.L.M. 1999. Deformation processes in a peridotite shear zone: reaction-softening by an H2Odeficient, continuous net transfer reaction. Tectonophysics, 303,193-222. NICOLAS, A. 1986. Structure and petrology of peridotites: clues to their geodynamic environment. Reviews of Geophysics, 24, 875-895. NICOLAS, A., BOUDIER,F. & MONTIGNY, R. 1987. Structure of Zabargad Island and early rifting of the Red Sea. Journal of Geophysical Research, 92, 461-474. OLGAARD, D.L. 1990. The role of second phase in localizing deformation. In: KNIPE, R.J. & RUTTER, E.H. (eds) Deformation mechanisms, rheology and tectonics. Geological Society, London, Special Publications, 54,175-181. PICCARDO, G.B., RAMPONE, E., VANUCCI, R., SHIMIZU, N., OTTOLINI, L. & BOTTAZZI, P. 1993. Mantle processes in the sub-continental lithosphere: the case study of the rifted spirel-lherzolites from Zabargad (Red Sea). European Journal of Mineralogy, 5,1039-1056. POIRIER, J.P. 1980. Shear localization and shear instability in materials in the ductile field. Journal of Structural Geology, 2,135-142. RASSIOS, A. & KONSTANTOPOULOU, G. 1993. Emplacement tectonism and the position of chrome ores in the Mega Isoma peridotites, SW Othris, Greece. Bulletin of the Geological Society Greece, XXVIII/2, 463-474.

23

RASSIOS, A. & SMITH, A.G. 2000. Constraints on the formation and emplacement age of western Greek ophiolites (Vourinos, Pindos, and Othris) inferred from deformation structures in peridotites. In: DILEK, Y, MOORES, E.M., ELTHON, D. & NICOLAS, A. (eds) Ophiolites and oceanic crust: new insights from field studies and the ocean drilling program. Geological Society of America Special Paper 349, Boulder, Colorado, 473-483. RASSIOS, A., GRIVAS, E., KONSTANTOPOULOU, G. & VACONDIOS, I. 1994. The geometry of structures forming around the brittle-ductile transition in the Vourinos-Pindos-Othris oceanic slab. Bulletin of the Geological Society Greece, 30,109-121. REUBER, I., MICHARD, A., CHALOUAN, A. & JUTEAU,T. 1982. Structure and emplacement of the alpinetype peridotites from Beni Bousera, Rif, Morroco: a polyphase tectonic interpretation. Tectonophysics, 82, 231-251. VAN ROERMUND, H.L.M. & DRURY, M.R. 1998. Ultrahigh pressure (P>6GPa) garnet peridotites in Western Norway: exhumation of mantle rocks from more than 185 km. Terra Nova, 10,295-301. VAN ROERMUND, H.L.M., DRURY, M.R., BARNHOORN, A. & DE RONDE, A. 2001. Relict majoritic garnet microstructures from ultra-deep erogenic peridotites in Western Norway. Journal of Petrology, 42,117-130. RUTTER, E.H. 1999. On the relationship between the formation of shear zones and the form of the flow law for rocks undergoing dynamic recrystallization. Tectonophysics, 303,147-158. RUTTER, E.H. & BRODIE, K. 1988. The role of tectonic grain size reduction in the rheological stratification of the lithosphere. Geologische Rundschau, 77, 295-308. SAWAGUCHI, T. 2002. Structual petrology of the Horoman peridotite: deformation history and exhumation process of the Alpine-type peridotite massif. PhD Thesis, Waseda University, Japan. SKROTZKI, W, WEDEL, A., WEBER, K. & MULLER, WE 1990. Microstructure and texture in Iherzolites of the Balmuccia massif and their significance regarding the thermomechanical history. Tectonophysics, 179, 227-251. SUHR, G. 1993. Evaluation of upper mantle microstructures in the Table Mountain massif (Bay of Islands ophiolite). Journal of Structural Geology, 151,1273-1292. VAN DER WAL, D. & VISSERS, R.L.M. 1993. Uplift and emplacement of upper mantle rocks in the western Mediterranean. Geology, 21,1119-1122. VAN DER WAL, D. & VISSERS, R.L.M. 1996. Structural petrology of the Ronda Peridotite, SW Spain: deformation history. Journal of Petrology, 37, 23-43. VISSERS, R.L.M., DRURY, M.R., HOOGERDUIJN STRATiNG,E.H.& VANDERWAL,D. 1991. Shear zones in the upper mantle: a case study in an Alpine Iherzolite massif. Geology, 19, 990-993. VISSERS, R.L.M., DRURY, M.R., HOOGERDUIJN STRATING, E.H., SPIERS, C.J. & VAN DER WAL, D. 1995. Mantle shear zones and their effect on lithosphere strength during continental break-up. Tectonophysics, 249,155-171.

24

A. H. DIJKSTRA ET AL.

VISSERS, R.L.M., DRURY, M.R., NEWMAN, J. & FLIERVOET, T.F. 1997. Mylonitic deformation in upper mantle peridotites of the North Pyrenean Zone (France): implications for strength and strain localization in the lithosphere. Tectonophysics, 279, 303-325. VISSERS, R.L.M., HOOGERDUIJN STRATING, E.H., DRURY, M.R. & VAN DER WAL, D. 1998. Microstructures formed during shear localization in the upper mantle. In: SNOKE, A.W.,TuLLis, J. & TODD, V.R. (eds) Fault-related rocks - a photographic atlas, Princeton University Press, Princeton, New Jersey, 582-587. WHITE, S.H. & KNIPE, RJ. 1978. Transformation- and reaction-enhanced ductility in rocks. Journal of the Geological Society London, 135, 513-516.

WHITE, S.H., BURROWS, S.E., CARRERAS, J., SHAW, N.D. & HUMPHREYS, FJ. 1980. On mylonites in ductile shear zones. Journal of Structural Geology, 2, 175-187. WOOD, B.L. 1972. Metamorphosed ultramafites and associated formations near Milford Sound, New Zealand. New Zealand Journal of Geology and Geophysics, 15, 88-127. Xu, Y.-G., Ross, J.V & MERCIER, J.-C.C. 1993. The upper mantle beneath the continental rift of Tanlu, Eastern China: Evidence for the intralithospheric shear zones. Tectonophysics, 225, 337-360.

Instability and localization of deformation in lower crust granulites, Minas fault zone, Nova Scotia, Canada JOSEPH C. WHITE Department of Geology, University of New Brunswick, Fredericton, NB, Canada, E3B 5A3 (email: [email protected]) Abstract: Blocks of granulite from within the megabreccia at Clarke Head, Nova Scotia, Canada contain extremely well preserved mylonitic and ultramylonitic textures developed in mineral assemblages for which thermobarometic calculations have indicated temperatures and pressures between 700-860 °C and 750-950 MPa. Deformation within these rocks is characterized by localization at several discrete length scales associated with the development of new microstructures comprising finer-grained material. Mylonitized granulite exhibits dislocation creep microstructures, with development of intense S-C fabrics and shear bands during the transition to ultramylonite. Dynamically recrystallization of plagioclase can be followed through progressive grain size reduction to about 5 um, but there remain extensive zones with grains less than 1 um in diameter. Localization of the these finest-grained ultramylonites occurs by transient frictional events associated with the introduction of partial igneous melts and formation of pseudotachylyte which produces abrupt decreases in grain size that cannot arise during dislocation mediated grain size reduction. The heterogeneous response of these rocks demonstrates the importance of considering characteristic length scales when assigning evidence from the rock record (e.g. palaeopiezometry) to bulk behaviour of the lithosphere. Associated with the localization of strain and subsequent strain softening is the observation that microstructures formed during the event that initiated the instability can be an obliterated by ductile flow. In instances where critical components of the microstructural evolution are known to have been largely overprinted, it becomes possible to reconcile contradictions in the rock record, such as production of ultra-fine-grained superplastic aggregates in what otherwise appears to be a dominantly dislocation creep regime.

The Minas fault system (MFS) in northwestern Nova Scotia, Canada (Fig. 1) demarcates a longlived, crustal-scale Appalachian tectonic boundary that juxtaposes Avalon and Meguma tectono-stratigraphic terranes (Keppie 1982; Webster et al 1998). The temporal and spatial complexities of this fault system that reflect transpressional deformation (Waldron et al 1989) provide opportunities to examine deformation processes at a range of crustal conditions. Of specific interest to this study are occurrences of strongly deformed granulites at Clarke Head (Fig. 1) that, as samples of the deep crust, are unique within the MFS. The granulites are central to the geological history along the terrane boundary as they reflect the crustal conditions most remote from the current upper crustal, extensional basin setting. Exposed granulites (Gibbons & Murphy 1995) take the form of isolated mylonitized blocks within a carbonate-evaporite-clay matrix (Fig. 2a) and are an integral component of the Clarke Head 'megabreccia', itself an anomalous feature of the MFS. A multidisciplinary study of the granulites by Gibbons et al (1996) has reported

Fig. 1. Location map of Clarke Head megabreccia containing mylonitized granulite blocks as part of the Minas fault system (MFS).

geothermobarometric, isotopic, TEM and tectono-stratigraphic data used to establish the P-T-t trajectory of the Clarke Head megabreccia over approximately 55 million years of convergence between Laurentia and Gondwana. The latter study has demonstrated the occurrence of syntectonic metamorphic mineral

From: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 25-37. 0305-8719/$15.00 © The Geological Society of London 2004.

26

J. C. WHITE

Fig. 2. Field relationships of the granulites. (a) Megabreccia near Clarke Head, Nova Scotia. The large dark clasts are Carboniferous sedimentary rock. Granulite blocks are at the top of the cliff and eroded out on to beach, (b) Compositionally banded mylonitic granulite block with cross-cutting plagioclase-rich leucosome introduced as a melt, (c) Leucosome in mylonitic granulite in a stage of partial transposition, (d) The contact between granulite mylonite and ultramylonite comprising a change in fabric and bulk composition, although the mineral assemblage remains the same.

assemblages and microstructures indicative of lower crust conditions (700-860 °C and 750-950 MPa). Some retrogression of granulite blocks to amphibolite assemblages is observed, both by exhumation-related hydration and during introduction of late Cl-rich amphibole veins. Occurrence of the 'megabreccia' in its current configuration is distinct and unrelated to deformation of the granulites. This contribution pursues the nature of deformation microstructures in the granulites through detailed analytical electron microscopy with the aim of elucidating micromechanical records and their implication to the rheology of deep crustal rocks.

Clarke Head granulites: field observations The granulite blocks exhibit contrasting domainal layering (Fig. 2) defined by variations in mineralogy, grain size and deformation textures. Three domains are recognized within the blocks: host mylonite, ultramylonite and 'cherty' ultramylonite.

The host mylonite comprises the bulk of the granulites and although strongly deformed, it represents the earliest deformation and background strain. The mineralogical assemblage consists of clino- and orthopyroxene, plagioclase, K-feldspar and scattered garnet; ilmenite and magnetite are significant oxide phases. The principal foliation is a transposed, composite layering (Fig. 2b-d) defined by modal variations of feldspars and pyroxenes. Because the granulites occur as isolated blocks within the Clarke Head megabreccia, any reference of their internal structure (e.g. foliations) to an external geometric or kinematic reference frame is precluded. In this study, the compositional layering is taken as the primary reference frame (Cfoliation), in conjunction with a weak lineation within the compositional layering. A second foliation (S-foliation) oblique to, but within the compositional layers is defined by elongate plagioclase and pyroxene. Porphyroclasts of pyroxene and garnet within the compositional layering are typically 200-1000 um across, whereas feldspars dominate the finer-grained matrix. Plagioclase-rich

INSTABILITY AND DEFORMATION LOCALIZATION

layers (leucosomes) are observed both crosscutting this main compositional layering, and forming intrafolial isoclinal folds (Fig. 2b, c). The latter relationships record the introduction of anatectic melt (cross-cutting layers) and progressive high-strain transposition associated with formation of the compositional layering. The feldspathic layers commonly localize deformation as evidenced by their large attenuation and grain size reduction relative to pyroxenerich layers. Ultramylonite is recognized as zones of more intense deformation up to 40 cm thick that disrupt the predominant compositional layering of the mylonite at low angles and are recognized by a transition in colour, fabric and grain size (Fig. 2d). Feldspar-rich layers (leucosome selvages) are common at the mylonite-ultramylonite contacts. Shear foliation planes (C-surfaces) in the ultramylonite overprint the compositional layering (pre-exisiting C-surface) at a small angle. The compositional layering in the ultramylonite is much thinner, usually less than 1 mm thick. Ultramylonite S-foliation defined by elongate to curviplanar feldspar and pyroxene has the same orientation as that of the host mylonite and is pervasively developed. Shear bands, which do not occur in the host mylonite, are characteristic of the ultramylonite, commonly observed linking C-planes. Cherty ultramylonite comprises bands of cherty- or glassy-looking material up to 5 mm thick that occur sparsely throughout the granulite blocks. Although grossly parallel to the compositional layering and ultramylonite foliation, cherty mylonite bands have extremely sharp contacts that cut both of the latter domains. The distinctive appearance and crosscutting relationships in the field serve as the bases on which this domain is identified. Experimental procedures Thin sections were cut parallel and perpendicular to the lineation and compositional layering, respectively, observed in the host mylonite. All thin sections are doubly polished and mounted in Crystalbond™ adhesive in order to remove areas for transmission electron microscopy. Material was examined at progressively finer scales using light microscopy (LM), scanning electron microscopy (SEM), electron microprobe (EMP) and transmission electron microscopy (TEM). The SEM is JEOL 6400 operated at 15 kV and 2.5 nA. The analytical system combines a JEOL wavelength-dispsersive spectrometer for light elements with a LINK Pentafet™ energy-dispersive detector and analyser. The

27

EMP is a JEOL 733 operated at 15 kV and 10 nA. The TEM is a Philips EM400T operated at 120 kV, with a LINK Pentafet™ light element energy-dispersive detector and analyser. Specimens for TEM were mounted on copper grids directly from the thin sections, and ionmilled with argon. For compositional analyses by TEM, samples of coarser-grained granulites were first analysed in SEM and electron microprobe (EMP) and then used as empirical internal elemental standards for EDS analysis by TEM. All electron microscopy and the bulk of compositional analyses referred to herein were done in the UNB Electron Microscopy Unit in order to maintain strict spatial control of the analysis areas; samples analysed by EMP at the University of Wales and University of Calgary gave consistent results (Gibbons et al. 1996). Deformation microstructures

Host mylonite Microstructures (Fig. 3a) within the host (i.e. least deformed) granulite comprise orthopyroxeneclinopyroxene-garnet-feldspar assemblages, with both clinopyroxene and orthopyroxene forming equant, to slightly elongate porphyroclasts (100-1000 urn grains; aspect ratio typically < 2:1) in a finer-grained, dominantly feldspar matrix (25-150 urn grains). Pyroxene grains both contain and are mantled by ilmenite-magnetite aggregates. The plagioclase (An42-An48) grain size range is quasi-bimodal, with smaller grains being the recrystallization products (rotation recrystallization) of the larger porphyroclasts. A few primary grains with incomplete dynamic recrystallization are observed (Fig. 3b). Feldspar grain boundaries are irregular. Dislocations and dislocation networks indicative of intracrystalline creep occur in both pyroxene and plagioclase (Fig. 3c,d). Additionally, pyroxene-oxide aggregates would suggest concurrent diffusive mass transfer (Fig. 3a).

Ultramylonite As was observed in field exposures, the boundary (Fig. 4a) between mylonite and ultramylonite is commonly associated with feldspar-rich layers (leucosome) that reflect a broader compositional change. The plagioclase in these layers is compositionally distinct from that (An27-An39) in the mylonite and where pyroxenes are present, they occur in lower modal proportions than in the host mafic mylonite. The

28

J. C. WHITE

Fig. 3. Mylonite microstructures. (a) Typical pyroxene grains in plagioclase matrix. Pyroxene grains are mantled by ilmenite-magnetite (Fe-Ti oxide) aggregates. LM Crossed-polarizers with gypsum plate, (b) Relict primary plagioclase in leucosome exhibiting partial dynamic recrystallization (rotation recrystallization) to mylonitic texture. LM Crossed-polarizers. (c) Dislocation substructure in clinopyroxene. TEM Bright field, (d) Dislocations and twins in plagioclase. Subgrain boundaries indicated by arrows. TEM Bright field.

most common textural transition is an abrupt Zone H: plagioclase aggregates (5-25 jam) decrease in grain size and/or distortion of elonformed by dynamic recrystallization of grains from Zone I; these plagioclase grains form gate pyroxenes and feldspars into a new, finerelongate aggregates parallel to a new compograined foliation, concomitant with formation of sitional layering in the ultramylonite. a strong S-C fabric and shear bands Fig. 4a). Pyroxenes are associated with dramatic elongaZone III: discrete, optically isotropic, 100 urn thick seams with grain sizes routinely less tion of the mantling Ti-Fe oxide aggregates parallel to both S- and C-planes. In detail, the than 1 (im and rarely larger than 5 jim. ultramylonite has a very heterogeneous microstructure that enables a tripartite division The mineral assemblage remains effectively by grain size into three distinct textural zones unchanged throughout all three zones and no (Fig. 4b): systematic variation in mineral chemistries was observed among them. The most significant minZone I: plagioclase and pyroxene are com- eralogical change is the new occurrence of quartz parable in size to those in the host mylonite and a modal increase in K-feldspar, particularly (15-50 jam), but are more elongate; these in Zone III. The latter mineralogical changes, in elongate grains occur within enveloping conjunction with the cross-cutting field relationsurfaces that define zones parallel to the com- ships support the origin of the leucosomes as positionallayering in the host mylonite, but lower crustal anatectic melts. The black colour of are characterized by a more prominent Zone III arises primarily from the extremely S-foliation. small grain size. Semi-quantitative image

INSTABILITY AND DEFORMATION LOCALIZATION

29

Fig. 4. Ultramylonite microstructures. (a) Contact between mylonite and ultramylonite bands separated by plagioclase selvage. Compositional layering in the mylonite and C-planes (dark seams) in ultramylonite are subparallel. S-foliations in the mylonite become more strongly developed in the selvage, with a strong S-C fabric, plus shear bands in the ultramylonite. LM Crossed-polarizers with gypsum plate, (b) Details of ultramylonite bands. Zone I and Zone II both exhibit dynamic recrystallization, and are effectively a continuous deformation. Zone III are thin optically opaque seams parallel to C-planes. LM Crossed-polarizers with gypsum plate, (c) Zone III ultramylonite band parallel to C-plane demonstrating the very fined grained, polymineralic nature of the optically opaque seams. It is contained within Zone II ultramylonite. SEM Backscattered Image, (d) Dislocation-free K-feldspar grain typical of Zone III ultramylonite. Hole in the upper left is X-ray analysis spot induced during collection of mineral composition data. TEM Bright field.

processing of oxide phases showed there to be no significant variation in oxide proportions at length scales on the order of the thickness of the associated zonal foliation; for example, compositional layering vs. Zone I vs. Zone II. Grains in Zones I and II exhibit dislocation substructures comparable to those in the host mylonite (Fig. 3d). Zone III grains are distinctive by virtue of the absence of dislocations, irrespective of the mineral phase (Fig. 4d), except for occasional, larger feldspar grains. The most commonly observed transition from Zone II to Zone III textures is the apparent continuously ductile reorientation of S-foliations toward the C-plane (Fig. 5a) along which elongate Fe-Ti oxide aggregates and Zone III textures develop.

Close examination shows that Zone III seams (Fig. 5b) initiate parallel to both C-surfaces and shear band orientations as discrete brittle displacement and cataclastic zones (Fig. 5c,d). Zone III ultramylonite bands initially that form parallel to incipient shear bands (Fig. 5c) rotate into the C-plane orientation with increasing strain. The fractures are accompanied by finegrained aggregates (veins) of ilmenite/magnetite, plagioclase, K-feldspar, pyroxene and quartz that have substantively the same assemblage as seen throughout the mylonite and ultramylonite. The association of fracturing and veining leads to interpretation of these textures as fluid-assisted fracture events, where the fluid is related to anatectic melts.

30

J. C. WHITE

Fig. 5. Microstructures associated with the transition from Zone II to Zone III ultramylonite, from most to least common. The evolutionary sequence actually progresses in reverse as (d)-(c)-(b)-(a). (a) Typical, highly evolved contact between Zone I and II ultramylonites with well-defined S-fabric delineated by deformed orthopyroxene and truncated by Zone III ultramylonite parallel to C-plane. Backscattered electron contrast for the labelled minerals is the same in (b)-(d). SEM Backscattered Image, (b) Discrete Zone III seam (veinlet) with pyroxene, plagioclase, ilmenite and magnetite with no clear evidence of a transition. SEM Backscattered Image, (c) Elongated pyroxene and Fe-Ti oxides within a fracture cross-cutting the S-foliation along an incipient shear band. SEM Backscattered Image, (d) Fracturing and cataclasis along the shear band orientation associated with introduction of a pyroxene-plagioclase-oxide vein. The fracture and vein cross-cut orthopyroxene that has been previously deformed into the S-foliation. The rarity of these microstructures is interpreted to reflect their rapid transformation to textures such as in (a). SEM Backscattered Image.

Cherty ultramylonite microstructures The cherty ultramylonite bands are as distinctive in thin section as they are in the field. The sharpness of the macroscopic contacts arises from an abrupt and homogeneous reduction in grain size that is more extreme than the mylonite to ultramylonite transition (compare Fig. 6a with Fig. 4b). The contacts, although abrupt, are distinctly curviplanar, commonly have an ultra-fine-grained pyroxene selvage at contacts (Fig. 6a) with the host and truncate the host compositional layering, including individual grains. Flow folds (Fig. 6b) are common and restricted to these bands. Pyroxene porphyroclasts within the bands have two features not

seen in elsewhere: a shard-like morphology, particularly near the contacts and/or irregular grain boundaries suggestive of fluid corrosion (Fig. 6c-e). The absence of evidence for progressive grain size reduction would be consistent with, but not solely demonstrable of, catastrophic grain size reduction (Fig. 6d). Similarities with Zone III ultramylonite include matrix grain size of the order of 1 um (Fig. 6d) and the near complete absence of dislocations (see Fig. 4d), but the 'cherty' bands can be considerably thicker (> 1 mm). Garnet porphyroclasts likewise have distinctive rounded shapes not typical of typical of pseuodotachylyte (Fig. 6f and Lin 1999).

INSTABILITY AND DEFORMATION LOCALIZATION

31

Fig. 6. Cherty ultramylonite (deformed pseudotachylyte). (a) Cherty mylonite - mylonite contact. Unlike the common ultramylonite, there is no progressive transition from the host mylonite; the contact is both abrupt and eroded. A dark, pyroxene-rich selvage at the contact may be a chill margin. LM Crossed-polarizers with gypsum plate, (b) Typical flow folds that are common in these zones, and totally unobserved elsewhere. LM Plane-polarized light, (c) Highly elongated and distorted pyroxene at the contact with ultra-fine-grained cherty ultramylonite matrix. The corroded pyroxene grain boundary is indicative of interaction with a disequilibrium fluid. SEM Backscattered Image, (d) Pyroxene-matrix contact demonstrating attrition and grain size reduction by brittle processes. SEM Backscattered Image, (e) Overview of a cherty ultramylonite with deformed pyroxene and a foliation that are distinctive from those in the dominant ultramylonite. LM Crossed-polarizers with gypsum plate, (f) Elliptical garnet with extremely smooth contact with the matrix indicative of interaction with a melt. LM Plane-polarized light.

Interpretation of microstructures and deformation mechanisms Deformation environment The syndeformational granulite metamorphic minerals in each microstructural regime have effectively constant mineral chemistries that

correspond to those used in the thermobarometric calculations of Gibbons et al. (1996). The associated temperature range of 700-860 °C and 750-950 MPa is applicable to all the deformation states, with the possible exception of the cherty mylonite. Hence, the microstructures clearly evolved in the deep crust. Preservation of these microstructures can be explained by the

32

J. C. WHITE

small grain size and anhydrous mineral phases which will inhibit fluid ingress and also kinetically limit retrogression during exhumation of the rocks. Where there is imperfect preservation, the ability to capture a range of behaviour is potentially more useful than simply assigning some single set of parameters to the rock body (Means 1995).

ing and veining localized on pre-existing localized ductile structures provide an additionally mechanism for grain size reduction. The association of grain fragmentation, introduction of oxide/silicate veins and cataclasis (Fig. 5c,d) argues for operation of a pressuredependent brittle process induced by elevated fluid pressure. Based on the observed mineralogy, the fluid from which veins form is akin to that of the high temperature anatectic leucoDeformation mechanisms some melts in the host granulite. The fractures The bulk of the microstructures observed in and veins are concentrated at zones of ductile both mylonite and ultramylonite reflect disloca- strain localization such as shear bands, with subtion-mediated deformation-recovery-recrystal- sequent solid-state deformation of the crystallization cycles that reduce the grain size. lized melt obliterating any brittle textures. The However, the grain size reduction by demon- formation of elongate Fe-Ti oxide aggregates strable dynamic recrystallization contrasts with during ultramylonite deformation plays a the grain size reduction observed at the tran- central role in the brittle event. Fe-Ti oxides are sition from Zone II to Zone III ultramylonite anticipated to be involved with the vein as remowhere the latter relates to fracturing and vein bilized, extremely ductile phases with low visintroduction. Two aspects of the microstructural cosity behaviour. Similar, though not identical evolution are central to any interpretation of influences of oxides on fracturing has been these observations: (i) the increase in defor- reported for lower ocean crust shear zones mation intensity, for which decreasing grain size (Agar & Lloyd 1994). is a proxy, correlates with concomitant localizThe importance of fracturing, vein introducation of the deformation into narrower zones; tion and cataclasis is the reduction in grain size and (ii) microstructures that correlate to the beyond that obtainable in the dislocation creep initial stages of the localization process can be field by dynamic recrystallization (White 1982; masked by subsequent textural equilibration De Bresser 2001); this in turn allows a transition during ductile flow. to grain-size sensitive (GSS) flow (Figs 7 and 8) The host mylonitic texture is interpreted as and phenomenological superplasticity (Boullier having developed during deep-crustal flow of & Gueguen 1975; Edington etal 1976). Diffusive the granulite, wherein intracrystalline, disloca- control of the GSS flow in Zone III ultramylonite tion-mediated processes dominate, concomitant is inferred from the submicrometre grain size and with development of the macroscopic composi- dislocation-free grains that contrast with dislocational layering. These textures represent the tion dominated microstructures throughout the closest approach in the granulites to a bulk bulk of the granulite. The polymineralic nature quasi-steady-state deformation. Nevertheless, of Zone III ultramylonite favours mutual pinning heterogeneous flow occurs even under these of grain boundaries and suppression of grain conditions in response to the rheological con- growth (White 1982; Panda et al 1985; Chen & trast of pyroxene- and plagioclase-rich layers as Xue 1990; Olgaard 1990); hence, overprinting of shown by both more intense recrystallization the latter of the fine-grained material is not within the plagioclase layers and the associated anticipated even if readjustment to ambient localized development of S-C fabrics. stresses comparable to Zone II ultramylonite Mylonitic microstructures evolve progres- should occur. Thus it is possible to achieve consively to Zone I, then Zone II ultramylonite. trasting mixed-mode deformation in adjacent Incontrovertible evidence for dynamic recrystal- volumes of rocks while requiring only transient lization of plagioclase to grain sizes less than variations in stress and strain rate throughout the about 5 (im is not observed (Fig. 4b) even rock mass. though significantly smaller grain sizes are Cherty ultramylonite is interpreted to be an common in Zone III ultramylonite. Notably, extreme example of localization, whereby a fricboth the reduction in grain size and the ensuing tional melt (disequilibrium melt) develops (Fig. change in intracrystalline defect microstructure 7) with subsequent super- and sub-solidus flow. from Zone II to Zone III ultramylonites are This interpretation rests largely on the gross texabrupt (e.g. Fig. 5a), and there is no evidence for tural differences and obliquity of these zones to the progressive reduction in grain size observed the overall rock fabric, in conjunction with their between Zone I to Zone II ultramylonite. Tran- similarity to demonstrable pseudotachylytes. sient microstructures (Fig. 5) formed by fractur- Unlike the other inferred crystallized melts (e.g.

INSTABILITY AND DEFORMATION LOCALIZATION

33

Fig. 7. Schematic chart of microstructure evolution during deformation. The diagram illustrates the interplay between strain or strain-rate hardening of the system, the corresponding localization feature formed in response to hardening, and the subsequent strain or strain-rate softening processes within the localized zone that produces the stable microstructure observed in the rocks. Deformation mechanism regimes are demarcated in terms of how localization is controlled during the evolution of the ultramylonites.

leucosome, Zone III ultramylonite), cherty ultramylonite layers entrain pyroxene porphyroclasts that have scalloped contacts suggestive of corrosion by a fluid, and have irregular contacts with the host rock, as is observed in disequilibrium frictional melts (Magloughlin 1992; Spray 1992; White 1996). The flow folding that, within the granulite blocks, is unique to these zones is typical of ductile flow within pseudotachylyte (Sibson 1980; Passchier 1982; Koch & Masch 1992; White 1993, 1996). Upon crystallization of the frictional melt, the extremely fine-grained, polymineralic material moves directly into the grain-size sensitive flow regime, in much the same way as Zone III ultramylonite. The intensity of deformation immediately adjacent to these zones (e.g. very high aspect ratio pyroxene grains indicative of glide with insufficient time for recovery), in combination with the inherent strength of granulites in a pore fluid pressure deficient environment (White 1996; Kiister & Stockhert 1999) is consistent with melting initiated as a plastic instability; this in turn would explain the contrast in features between the 'cherty' and Zone III ultramylonites which initiated as transient high fluid pressure events. Similar textures have been observed in peridotite massifs (Vissers et al

1997) and are interpreted similarly as 'dry' localization events, in contrast to 'wet' localization, where water or melts play a role.

Mechanism transitions The processes leading to microstructural and/or deformation mechanism transitions in the Clarke Head granulites can be subtle. Inferring the absence of dislocations in small equant grains within high-strain zones to be an indicator of diffusion-controlled grain size sensitive flow, a transition in the dominant deformation regime occurs at about 5 um grain size (Fig. 8); that is, Zone II and Zone III ultramylonites are distinct. Additionally, grain size reduction by dislocation processes (dynamically recrystallization) is observed not to provide a path from grain sizes characteristic of Zone II to those of Zone III ultramylonite. Instead, the rarely observed and rarely preserved cases of fracturing, cataclasis and vein injection are responsible for strain localization and grain size reduction. Brittle deformation as an overprinted precursor to ductile deformation is commonly observed in experiments (e.g. Tullis & Yund 1987) and nature (Pennacchioni & Cesare 1997; Guermani & Pennacchioni 1998). The mechanism regime path then becomes:

34

J. C. WHITE

Fig. 8. A plot of the width of localized deformation zones (black bars) and the corresponding grain size of plagioclase within specific types of zones (shaded area). The localization zones are arrayed sequentially in terms of the degree to which they would exhibit a strain rate faster than the bulk rate; i.e. perfectly homogeneous deformation would give a ratio of 1.0 (see text). The figure illustrates that as the characteristic length scale of the localization decreases, so does the grain size; that is, the more discrete the localization, the higher the stress, and as interpreted here, the higher the strain rate during initiation of the localization. Based on microstructural observations, approximate boundaries are drawn between dislocation- and brittle-mediated localization, as well as the limit of dynamic grain size observed. It is suggested by these constraints that those ultramylonites observed in these rocks with characteristics of GSS flow required a brittle precursor. This is not to infer that this is an attribute for all examples of GSS flow.

dislocation-mediated creep —> fracture —> diffusion-mediated GSS creep. Deformation is sustained within the GSS creep field by the suppression of grain growth due by mutual pinning of the different. If there was some grain growth, it would be limited by the 'stable' recrystallized grain size appropriate for the dislocation creep regime (White 1982). These observations emphasize the possibility of transitional microstructures, for which little evidence is preserved, acting as necessary intermediary states between outwardly stable, but contrasting microstructures.

Deformation partitioning and localization Within the Clarke Head granulites, the localization of strain corresponds to progressive reductions in grain size. Development of discrete C-plane fabrics and shear bands reflects mechanical partitioning of the strain in response to inherent system hardening and/or a change in

boundary conditions. The contributory effects to the final microstructures of systemic hardening and localized softening are illustrated in Fig. 7. Strain hardening of an overall deforming system - a result of changes in kinematics, P-T fluid environment or fabric, where no one factor is independent of the other - will lead to increasing stress levels if imposed tectonic displacement rates continue. For continued accommodation of such displacements, there must eventually develop some mode of strain or strain-rate softening. In effect, system hardening leads to instability and localization at more extreme (e.g. higher stress) conditions than the pre-existing deformation state, while subsequent softening within the localized zones masks the processes of localization; softening can not be isolated from induced instability and the localization of deformation (as reviewed for experimental systems byHobbsetal. 1990). Microstructural steady state or stability during deformation must meet the criteria of

INSTABILITY AND DEFORMATION LOCALIZATION

Means (1981) that both weakening and softening be equally effective. The latter will be reflected as microstructures associated with a controlling micromechanism(s) (e.g. dislocation glide + climb) and which describe the deformation over a characteristic spatial and time scale. At any stage of a dislocation-mediated microstructural evolution, maintenance of an apparent steady-state microstructure is tied to grain size reduction (localization) events that enable sufficient softening to establish a degree of microstructural stability. Trigger events (or bifrucations) arising from system hardening provide the impetus for localization, and associated softening, i.e. in the absence of perfect spatial and temporal homogeneity, heterogeneous behaviour will be the rule. This is clearly demonstrated by the requirement for gradients in strain energy at the grain scale in order to induce dynamic recrystallization, even in nominally steady-state fabrics. The rarity with which these trigger events are preserved reflects the ubiquity of subsequent ductile overprints. The question of localization is a function of the length scale of interest and, in turn, a function of the prevailing defect structure. Central to this is the question of what parameters can be extracted from the rock record (e.g. palaeopiezometers), and to what volume of rock these are applicable. Within the Clarke Head granulites, each type of localization structure has its own characteristic length scale (thickness), plus a length scale (grain size) specific to its specific microstructure. Figure 8 relates the thickness of localized deformation zones to an estimate of the associated deviation of localized strain rate from the system or bulk strain rate. Ductile shear strain rate is partitioned as a function of zone thickness for a set ambient strain rate (White & Mawer 1992); for example, perfect homogeneity and steady-state at a bulk strain rate of 10~14 s"1 is taken as a ratio of 1.0; mylonites ~ 102; Type II ultramylonite « 104, etc. Typical frictional displacement rates are used to give a comparable strain rate estimate in the brittle regime. Each level of localization corresponds to a decrease in length scale (thickness of localized zone), an increase in the number of localized sites (e.g. foliation => foliation with multiple shear bands) and an increase in deformation intensity within that zone (smaller grain size). The variability of the thickness of the cherty ultramylonite distinguishes it from the other localized deformation zones, consistent with it forming by a unique process (pseudotachylyte). In cases of localized (heterogeneous) defor-

35

mation, there may not be a single answer to questions of strain rate and stress as determined from microstructures. During a phase of hardening, localization of more-or-less bulk mylonitization into discrete C-planes, suggests that the weak zones (C-planes) have the highest strength (smallest recrystallized grain size). If instead we couch questions in terms of how displacements are best accommodated by the rock mass, then the weak zones do carry evidence of a higher stress, but also higher strain rates distributed over smaller volumes of rock; that is, the timeaveraged rate of strain energy dissipation may stay more of less constant, whereas the path of displacement accommodation varies dramatically. Rutter (1999) has noted the need for sufficient strain rates in the localized zones to balance the bulk displacements. At a plate length scale, deformation rates of 10~14 s"1 could be appropriate, and the related microstructures could be non-descript. At the same time localization due to system hardening will record only stresses and strain rates greater than background. What are ultimately interpreted as the weakest zones may well be those with the initially fastest rate of strain hardening i.e. rock volumes within which stresses concentrated most quickly lead to a trigger that transforms microstructures to that characteristic of the new deformation regime.

Conclusions Sequences of mylonites and ultramylonites have granulite grade mineral assemblages and associated deformation microstructures for which thermobarometry indicates deformation at temperatures in excess of 700 °C and 750 MPa. In addition to high-temperature creep and grain size reduction by dynamic recrystallization, there is localization of deformation into discrete, but always narrower zones that are tied to introduction of a fluid phase (leucosome, veins, pseudotachylyte). Elevated fluid pressure transients (anatectic melts) that develop in concert with system hardening introduce material that is more amenable to deformation (strain softening) than the existing constitutive state. In extreme cases, frictional melting produces pseudotachylyte that is subsequently deformed ductily by grain size sensitive flow upon crystallization. The oxide-related localizations, as well as the more obvious pseudotachylyte events, may be coseismic events. The array of strain localization structures and concomitant variations in microstructures demonstrate the inherent heterogeneity of deformation, wherein characteristic length scales must be considered

36

J. C. WHITE

before assigning evidence from the rock record, such as palaeopiezometric measurements, to bulk behaviour of the lithosphere. The overprinting of microstructures associated with the initial localization appears to be common. Knowledge that such microstructural evidence has been lost may allow reconciliation of contradictions in the rock record, such as the production of ultra-fine-grained superplastic aggregates in what otherwise appears to be a dominantly coarser grained dislocation creep regime. Instrumentation and analytical facilities were supported through the UNB Electron Microscopy Unit. Financial support by the Natural Sciences and Engineering Research Council is gratefully acknowledged. Collaboration with Terry Gordon, Brendan Murphy and particularly Wes Gibbons, who provided the impetus for this work, has contributed significantly to these ideas, errors in which are solely the author's responsibility. Comments by R. Holdsworth and an anonymous reviewer contributed to clarification of the presentation.

References AGAR, S.M. & LLOYD, G.E. 1994. Rheology of shear zones in the lower ocean crust: the role of oxide deformation. EOS Transactions of the American Geophysical Union, 75/44, 650. BOULLIER, A.M. & GUEGUEN, Y. 1975. SP-mylonites: origin of some mylonites by superplastic flow. Contributions to Mineralogy and Petrology, 50, 93-104. CHEN, I.W. & XUE, L.A. 1990. Development of superplastic structural ceramics. Journal of the American Ceramic Society, 73, 2585-2609. DE BRESSER, J.H.P., TER HEEGE, J.H. & SPIERS, CJ. 2001. Grain size reduction by dynamic recrystallization: can it result in major rheological weakening? International Journal of Earth Sciences, 90, 28-45. EDINGTON, J.W., MELTON, K.N. & CUTLER, C.P. 1976. Superplasticity. Progress in Materials Science, 21, 63-170. GIBBONS, W. & MURPHY, J.B. 1995. Mylonitic mafic granulite in fault megabreccia at Clarke Head, Nova Scotia: a sample of Avalonian lower crust? Geological Magazine, 132, 81-90. GIBBONS, W., DOIG, R., GORDON, T., MURPHY, B., REYNOLDS, P. & WHITE, J.C. 1996. Mylonite to megabreccia: tracking fault events within a transcurrent terrane boundary in Nova Scotia, Canada. Geology, 24,411-414. GUERMANI, A. & PENNACCHIONI, G. 1998. Brittle precursors of plastic deformation in a granite; an example from the Mont Blanc Massif (Helvetic, Western Alps). In: RUTTER, E.H., BORIANI, A., BRODIE, K.H., BURLINI, L. & TREAGUS, S.H. (eds) Structures and properties of high strain zones in rocks. Journal of Structural Geology, 20,135-148.

HOBBS,B.E.,MULHAUS,H.B.& ORD,A. 1990. Instability, softening and localization of deformation. In: KNIPE, R.J. & RUTTER, E.H. (eds) Deformation mechanisms, rheology and tectonics. Geological Society, London, Special Publications, 54, 143-165. KEPPIE, ID. 1982. The Minas Geofracture. In: ST. JULIEN, P.& BELAND, J. (eds) Major structural zones and faults of the Northern Appalachians. Geological Association of Canada, Special Paper, 24, 263-280. KOCH, N. & MASCH, L. 1992. Formation of Alpine mylonites and pseudotachylytes at the base of the Silvretta nappe, Eastern Alps. Tectonophysics, 204, 289-306. KUSTER, M. & STOCKHERT, B. 1999. High differential stress and sublithostatic pore fluid pressure in the ductile regime - microstructural evidence for short-term post-seismic creep in the Sesia Zone, Western Alps. Tectonophysics, 303,263-277. LIN, A. 1999. Roundness of clasts in pseudotachylytes and cataclastic rocks as an indicator of frictional melting. Journal of Structural Geology, 21, 473-478. MAGLOUGHLIN, J.F. 1992. Microstructural and chemical changes associated with cataclasis and frictional melting at shallow crustal levels: the cataclasite B pseudotachylyte connection. Tectonophysics, 204,243-260. MEANS, W.D. 1981. The concept of steady-state foliation. Tectonophysics, 78,179-199. MEANS, W.D. 1995. Shear zones and rock history. Tectonophysics, 247,157-160. OLGAARD, D.L. 1990. The role of second phase in localizing deformation. In: KNIPE, R.J. & RUTTER, E.H. (eds) Deformation mechanisms, rheology and tectonics. Geological Society. London, Special Publications, 54,175-181. PANDA, PC, RAJ, R. & MORGAN, P.E.D. 1985. Superplastic deformation in fine-grained MgO.2Al2O3 spinel. Journal of the American Ceramic Society, 68, 522-529. PASSCHIER, C.W. 1982. Pseudotachylyte and the development of ultramylonite bands in the SaintBarthelemy massif, French Pyrenees. Journal of Structural Geology, 4, 69-79. PENNACCHIONI, G. & CESARE, B. 1997. Ductile-brittle transition in pre-Alpine amphibolite facies mylonites during evolution from water-present to water-deficient conditions (Mont Mary Nappe, Italian Western Alps). Journal of Metamorphic Geology, 15, 777-791. RUTTER, E.H. 1999. On the relationship between the formation of shear zones and the form of the flow law for rocks undergoing dynamic recrystallization. Tectonophysics, 303,147-158. SIBSON, R.H. 1980. Transient discontinuities in ductile shear zones. Journal of Structural Geology, 2, 165-171. SPRAY, J.G. 1992. A frictional basis for the frictional melting of some rock-forming minerals. Tectonophysics, 204, 205-221. TULLIS, J. & YUND, R.A. 1987. Transition from cataclastic flow to dislocation creep of feldspar:

INSTABILITY AND DEFORMATION LOCALIZATION mechanisms and microstructures. Geology, 15, 606-609. VISSERS, R.L.M, DRURY, M.R., NEWMAN. J. & FLIERVOET, T.F. 1997. Mylonitic deformation in upper mantle peridotites of the North Pyrenean Zone (France): implications for strength and strain localization in the lithosphere. Tectonophysics, 279, 303-325. WALDRON, J.W.F., PIPER, DJ.W. & PE-PIPER, G. 1989. Deformation of the Cape Chignecto pluton, Cobequid highlands, Nova Scotia: thrusting at the Meguma-Avalon boundary. Atlantic Geology, 25, 51-62. WEBSTER, T.L., MURPHY, IB. & BARR, S.M. 1998. Anatomy of a terrane boundary: an integrated structural, geographic information system, and remote sensing study of the late Paleozoic Avalon-Meguma boundary. Canadian Journal of Earth Sciences, 35, 787-801.

37

WHITE, J.C. 1982. Quartz deformation and the recognition of recrystallization regimes in the Flinton Group conglomerates, Ontario. Canadian Journal of Earth Sciences, 19, 81-93. WHITE, J.C. 1993. Melt instabilities and superplasticity during ductile shear of silicates. Proceedings of the Microscopical Society of Canada, 20,118-119. WHITE, J.C. 1996. Transient discontinuities revisited: pseudotachylyte, plastic instability and the influence of low pore fluid pressures on deformation mechanisms in the mid-crust. Journal of Structural Geology, 18,1471-1486. WHITE, J.C. & MAWER, C.K. 1992. Deep-crustal deformation textures along megathrusts from Newfoundland and Ontario: implications for microstructural preservation, strain rates and strength of the lithosphere. Canadian Journal of Earth Sciences, 29,129-141.

This page intentionally left blank

Microstructural evolution in a mylonitic quartz simple shear zone: the significant roles of dauphine twinning and misorientation GEOFFREY E. LLOYD School of Earth Sciences, The University, Leeds LS2 9JT, UK Abstract: SEM/EBSD-based orientation and misorientation analyses are described for a lower amphibolite facies simple shear zone (Torridon, NW Scotland). It is shown that as well as conventional crystal-slip processes (i.e. basal-a, prism-a, rhomb-a and negative second order rhomb-a slip), dauphine twinning also plays a role in both microstructural and petrofabric evolution. Twinning assists in the initial grain size comminution processes, including dynamic recrystallization, from originally coarse wall rock grains to a typical mylonitic microstructure in the centre of the shear zone. Subsequently, twinning helps to accommodate high shear strains in the mylonite whilst maintaining a stable microstructure and constant 'single crystal' petrofabric. The role of dauphine twinning appears to be to allow efficient switching between relatively 'soft' and relatively 'hard' slip directions that possibly exploit a distinction between negative and positive crystal forms. Misorientation analysis emphasizes the relationships between crystal-slip systems and grain boundary network, including dauphine twin planes, and suggests that the mylonitic microstructure contains preferred orientations of both tilt and twist boundaries that help to explain shear zone microstructural evolution and stability.

Shear zones occur in many deformed rocks on scales ranging from the microscopic to the macroscopic. They act to localize deformation and accommodate much of the ductile displacement that occurs during orogenesis. Consequently, the physical, mechanical and chemical processes involved in the formation and evolution of shear zones have attracted considerable and wide-ranging attention (e.g. Ramsay 1980; Rutter et al. 1998). In particular, many shear zones develop strong crystal lattice preferred orientation (LPO) during formation. However, the processes by which LPO develops during shear zone evolution remains controversial, but it must be possible to derive this knowledge from study of shear zone microstructure (e.g. Law 1987). Until recently, the only techniques widely available for studying microstructural and LPO evolution have been universal stage optical microscopy and X-ray texture goniometry. Although both techniques remain useful, they have their limitations (e.g. the former is restricted to minerals for which unique crystal directions can be identified optically, whereas the latter is restricted to a specific grain size range and derives only bulk data lacking microstructural control). Fortunately, recent developments in scanning electron microscopy (SEM), specifically electron-backscattered diffraction (EBSD), have provided solutions to these problems. SEM/EBSD (Venables & Harland 1973) permits the recognition of all crystal orientations for most minerals in one-

to-one relationship with microstructure, and automation provides statistically significant datasets. This contribution assesses the question of shear zone microstructural evolution and LPO development from the point of view of an almost perfectly simple shear zone developed within a quartz vein, using SEM/EBSD to derive the crucial observations and data.

Specimen and analytical details Sample details The sample used in this study (Fig. 1) was collected (Wheeler 1984) from a 30cm wide deformed quartz vein within Proterozoic gneisses at the head of Upper Loch Torridon, NW Scotland (UK GR NG 840530). The shear zone developed under lower amphibolite facies conditions typical of the mid-lower crust. Law et al. (1990) have previously described this sample in terms of its finite microstructure and LPO, as follows (see also Lloyd et al. 1992). The vein was deformed by crystal plastic processes to form a dextral shear zone with an intense mylonitic foliation and lineation (e.g. Ramsay & Graham 1970). In XZ section (i.e. where X>Y>Z), the vein microstructure is that of a classic Type II S-C mylonite (Lister & Snoke 1984) and consists of two planar domains (A and B) aligned parallel to the macroscopic mylonitic foliation (SA). A domains consist of

From: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 39-61. 0305-8719/$15.00 © The Geological Society of London 2004.

40

G. E. LLOYD

MICROSTRUCTURAL EVOLUTION

equant, less than 5 um quartz grains and subsidiary feldspar, whereas (the volumetrically more important) B-domains consist of dynamically recrystallized, elongate, less than 100 um quartz grains with long axis (SB) oblique to SA. The obliquity between SB and SA confirms a dextral shear sense. Shear zone microstructures (see Law et al. 1990, fig. 2) are indicative of constant volume, strongly non-coaxial, essentially plane strain deformation that closely approximates to simple shear. The development of a typically mylonitie microstructure is achieved via crystal slip systems that exploit the ease of slip parallel to (e.g. basal-a, rhomb-a and/or {±7i}<«>), in accordance with a simple shear kinematic framework that aligns <«> with lineation (X) and (c) normal to foliation (XY). The LPO (e.g. Fig. 2) therefore can be explained in terms of the simple shear kinematic framework indicated by shear zone geometry, in agreement with the fabric evolution model for bulk simple shear proposed by Etchecopar (1977). Furthermore, it appears that slip is favoured on negative rather than positive forms, which might suggest a lower resistance to slip on the former compared to the latter. Alternatively, the non-equivalence of positive and negative rhombs and the r point maximum in XZ at 30-50° to XY, subparallel to the NW-SE trending maximum principal stress direction inferred from the simple shear kinematic framework, is consistent with dauphine twinning (e.g. Tullis & Tullis 1972). The significance of dauphine twinning during quartz deformation has received relatively little attention (e.g. Tullis 1970; Baker & Wenk 1972; Baker & Riekels 1977; Olesen & Schmidt 1990; Lloyd et al. 1991; Barber & Wenk 1991; Mainprice et al. 1993; Neumann 2000; Lloyd 2000). The role of dauphine twinning during the microstructural and LPO evolution of the Torridon quartz simple shear zone is considered in detail in this contribution.

SEM/EBSD This contribution makes use of several complementary aspects of SEM crystallographic analysis, namely EBSD (e.g. Prior et al. 1999)

41

and orientation contrast imaging of specimen microstructure (e.g. Lloyd 1987; Adams et al. 1993; Trimby & Prior 1999). Although EBSD is used primarily to index crystal orientations, it can also provide several simulated images of samples (e.g. Jensen & Schmidt 1991; Adams et al. 1993; Field 1997; Lloyd 2000). Here, use is made of the grey-scale coded band contrast (BC) image (e.g. Fig. Ib), which reflects variations in the quality of the electron backscattered diffraction patterns (EBSP). EBSP are indexed via computer pattern recognition programs (e.g. Schmidt & Olesen 1989), which provide conventional petrofabric diagram representations and their misorientation equivalents. Several general or detailed automated and manual SEM/EBSD analyses were performed (see Fig. la for locations and Table 1 for details). In general, automated EBSD indexing success rates were poor and a dependence on both grid step size and sample grain size was observed. As the former increased and/or latter decreased, indexing success rate also decreased due to the increased probability of sampling 'artefacts' (i.e. grain boundaries, 'overlapping' grains, fractures, etc.). Particular care was taken to avoid the incorrect indexing of certain EBSP due to 'pseudosymmetry' effects (e.g. Prior et al. 1999; Lloyd 2000). Automated SEM/EBSD analysis of quartz is especially prone to this problem because the dauphine twinning operation, a 180° (crystallographically equivalent to 60°) rotation about the c-axis, results in very similar pattern configurations that are readily confused. This problem typically leads to an apparent increase in frequency of dauphine twinning in the SEM/EBSD LPO. However, as both Law et al. (1991) and Lloyd et al. (1992) recognized a real contribution from dauphine twinning in the mylonitie LPO in their studies using different techniques, it was important to minimise any mis-indexing due to pseudo-symmetry. The approach used here involved reducing the overall pattern contrast, such that the background diffraction bands that contain the significant differences in configuration are emphasized relative to the common main (brightest)

Fig. 1. SEM images of Torridon shear zone (XZ section plane), (a) Electron channelling orientation contrast image of the whole sample: SZWR: shear zone wall rock (analysis G 030600, see Table 1 for details); SZM: shear zone margin (G040800); MM: mature mylonite (G270700); and MD: mylonite detail (G050800). Also indicated are the positions of the detailed analyses (see Figs 5 & 6). (b) The same image based on EBSP band contrast. The grey scale reflects the quality of the EBSP at each analytical point (dark, poor; bright, good). This image clearly reveals significant detail of the microstructure (e.g. quartz grain size reduction and foliation development and deflection into the shear zone; break-up of feldspar; twinning and subgrains in wall rock feldspar; etc.).

Fig. 2. Equal area, upper hemisphere quartz pole figures for different parts of the Torridon shear zone constructed via auto-EBSD analysis (see Table 1) using the program Pf2k (D. Mainprice; see also Mainprice 1990, Mainprice & Humbert 1994). (a) Shear zone wall rock (SZWR). (b) Shear zone margin (SZM). (c) Mature mylonite (MM), (d) Mylonite detail (MD).

MICROSTRUCTURAL EVOLUTION

43

Table 1. Summary ofTorridon shear zone auto-EBSD experiments (see Fig. la for locations; BC: band contrast; BS: band slope; BN: band number; n/a, not available). In all but G030100, plagioclase was ignored. JOB (format: username-day-month-year)

DETAILS G030600 Specific comments Area analysed

* blown filament shear zone wall rock (Fig. la) Dimensions (mm) 10.0 X 13.0 No. data pts. 136640 Grid step (urn) 50 Time (hr:min:sec.) 41:47:25 Index rate (sec/ebsp) 1.1 Minerals indexed quartz & plagioclase (Anl6) % Low BC *46.6 % Low BS 0.0 % Low BN 0.1 % Not indexed 27.0 % Indexed 26.2 No. indexed 35800

G040800

G270700

G050800

G120702

user break shear zone margin (Fig. la) 17.0 X 6.0 120537 25 29:37:52 0.88 quartz

_ mature mylonite (Fig. la) 13.5 X 7.25 228825 20 41:18:56 0.65 quartz

_ mylonite detail (Fig. la) 1.5 x 1.0 150001 1 22:11:31 0.53 quartz

*blown filament linear traverse (Fig. 6) 0.1 X 4.0 334087 1 69:06:00 0.36 quartz

2.2 0.0 13.0 68.5 16.2 19527

5.1 0.0 0.2 87.2 7.5 17162

1.3 0.0 2.1 74.8 21.8 32700

n/a n/a n/a 27.49 72.51 290040

diffraction bands (Lloyd 2000). In addition, a relatively tight maximum 'mean angular deviation' (i.e. the difference between observed and predicted pattern configurations) of 0.75 was used (e.g. Krieger-Lassen 1996). Unfortunately, this meant that some 'good' patterns were rejected, further reducing successful indexing rates (see Table 1). In spite of these considerations, indexing rates were sufficient to simulate microstructural detail accurately (e.g. compare Fig. la, b) and to derive representative LPO (see Figs 2 & 6).

Crystallographic misorientation analysis The advent of SEM/EBSD analysis has focused attention on the crystallographic misorientation relationships that exist between adjacent regions (i.e. grains, subgrains, twins, etc.) and therefore upon the formation and crystallographic orientation of the separating boundaries (e.g. Lloyd et al 1997; Trimby et al 1998, 2000; Paul & Fitz Gerald 1999; Fliervoet et al. 1999; Prior 1999; Neumann 2000; Wheeler et al. 2001). The misorientation petrofabric or LPO of deformed rocks may prove significant as the (crystallographic) properties and impact of grain boundaries, particularly where they differ from the grain properties, are largely unknown (see Lloyd & Kendall, in press). Although misorientation data can be dis-

played simply as histograms of misorientation angle frequency, the misorientation between two adjacent crystal lattices is more accurately represented by a single axis about which one of the lattices needs to be rotated by a specific angle to bring it into complete coincidence with the other lattice. This leads to the concept of the misorientation axis/angle pair (e.g. Pospiech et al. 1986; Randle & Ralph 1986; Randle 1992, 1993; Mainprice et al. 1993; Lloyd et al. 1997; Kruse et al. 2001). Any misorientation can be defined by a number of symmetrically equivalent axis/angle pair combinations but current convention is to adopt the pair with the minimum misorientation angle (e.g. Mainprice et al. 1993; Lloyd et al. 1997). Misorientation axis/angle pair data can be represented by either pole figures or inverse pole figures. The former plots the data in 'sample space', whereas the latter plots the data in 'crystal space'. Both provide useful but different information on the nature and origin of (sub)grain boundary misorientations (Lloyd et al. 1997, Wheeler et al. 2001). Misorientation pole and inverse pole figure diagrams are typically presented as series in which each individual figure covers a narrow range of misorientation angles. However, due to problems in determining misorientation axes accurately as misorientation angles approach zero, plots of misorientations less than 5° are usually excluded (Prior 1999).

44

G. E. LLOYD

3b), referred to as the 'forbidden region' (Wheeler et al. 2001). The maximum misorientation angle possible is 104.5° about {2021}. Quartz exhibits also enantiomorphism and its left- and right-handed forms cannot be distinguished via electron diffraction techniques due to Friedel's law, which imposes an arbitrary centre of symmetry. Current convention therefore decrees that a right-handed pseudo-hexagonal definition is used for misorientation analyses of polycrystalline quartz studies (e.g. Neumann 2000). However, for individual quartz grains, EBSD does distinguish positive and negative forms (e.g. <±«>, {r} and {z}, [n] and {^'}, etc.), such that an unambiguous crystallographic indexing of quartz to its trigonal symmetry is possible. The possibility exists therefore that a trigonal construction may be more appropriate for quartz misorientation analysis (see below and also Mainprice et al. 1993).

Grain boundary (mis)orientation analysis

Fig. 3. Quartz crystallography, (a) Principal directions in quartz (after Linker et al. 1984) and comparison of the 60° hexagonal (darker shading) and 120° trigonal (lighter shading) inverse pole figure constructions used in misorientation analysis, (b) 60° misorientation inverse pole figure construction and the definition of 'forbidden regions' that develop progressively for misorientation axes above 60° (indicated by lower and upper bound angles). This construction fails to distinguish between <+a> or <-a>.

The range of possible misorientation angles is determined by the symmetry of the mineral concerned. Quartz has trigonal D3 - 32 symmetry but for convenience often has assumed to be pseudo-hexagonal (Fig. 3a). The former requires that the basal plane subtends a 120° angle with the c-axis, whereas the latter requires only a 60° unit triangle. The hexagonal definition imposes a constraint that misorientation angles above 60° lie progressively further from the c-axis (compare upper and lower bounds in Fig.

A full definition of grain boundary orientation involves the dimensional orientation of the boundary plane as well as the crystallographic misorientation across the boundary (e.g. McLaren 1986). It is difficult to measure accurately and rapidly the dimensional orientations of boundaries (e.g. Kruhl & Peternell 2002), particularly in the large quantities now being produced regularly for misorientation data via EBSD analysis. Fortunately, Randle (1992) has shown that the misorientation axis/angle pair definition alone often provides a meaningful approximation. In addition, the operation of crystal-slip (and twin) systems can be associated with the formation of different types of boundary in specific orientations (Lloyd et al. 1997; Neumann 2000; Kruse et al. 2001). The procedure for predicting boundary orientation from misorientation axis/angle pair data and slip/twin systems is as follows. Each crystal-slip system has a unique rotation axis (R) about which individual crystal directions migrate or disperse on small/great circles during slip (Lloyd & Freeman 1991, 1994). In contrast, different types of boundary have specific misorientation axes (i). The misorientation axis for tilt boundaries (-^TIB) is parallel both to the boundary plane (Fig. 4a) and also R (Fig. 4b). Such boundaries therefore form parallel to jR/^TiB and the slip plane normal (SPN) but normal to the slip direction (SD). The relationships between crystal-slip systems recognized for the Torridon shear zone (Law et al. 1990) and the predicted orientations of misorientation axes and tilt boundaries are given in Table 2.

45

MICROSTRUCTURAL EVOLUTION

requires at least two slip systems with different Burgers vectors (e.g. Hull & Bacon 1984), these ideal relationships cannot be sustained. In his study of recrystallization in quartzites, Neumann (2000, fig. 2b) included cross-slip systems (e.g. [m} and {z}) in his application of the edge dislocation model and predicted misorientation axes by combining the orientations appropriately (i.e. £{1122} and ^'{1132} respectively; see Fig. 3). His approach suggests a potential modification of the tilt boundary-edge dislocation to twist boundaries. Twist boundaries may form as a compromise to the combinations of slip systems responsible. Thus, -6TWB is most likely to form parallel to the Fig. 4. Relationships between crystal slip systems resultant of the slip plane normal directions (i.e. and boundary orientations: SD: slip direction; SPN: SPN), whereas the boundary plane is most likely slip plane normal; R: slip system rotation axis; £TlB to form_parallel to the resultants of the rotation and -^TwB? tilt and twist boundary misorientation axes, axes (R) and slip directions (SD) and hence (a) Tilt boundaries (TiB). (b) General relationship normal to SPN/£TwB. For example, if (c) and between TiB and crystal slip systems, (c) Twist (c)<m> slip contribute equally to twist boundary boundaries (TwB). (d) General relationship between formation, then SD = + <m> = {2130}, R = TwB and crystal slip systems. The actual situation is <m> + = {2130} and SPN/£TwE = (c) + (c) = more complex - see text for discussion. (0001). The relationships between crystal slip systems recognized in the Torridon shear zone (Law et al. 1990) and the orientations of twist The situation for twist boundaries is more boundaries predicted using this approach are complex because the relationship envisaged given in Table 3. It may therefore be possible to reconcile slip between slip systems, misorientation axes and boundary formation strictly applies only to slip systems and rotation axes with formation of due to edge dislocations. By definition, the mis- both tilt and twist boundaries via a combination orientation axis for twist boundaries (-^TWB) *s of EBSD-based LPO and misorientation analynormal to the boundary plane (Fig. 4c), which ses. This is the approach taken here, which also should be parallel to the slip plane and hence SD highlights the role of dauphine twinning in and R, whereas -€TWB *s parallel to SPN (Fig. 4d). boundary formation during shear zone However, as the formation of twist boundaries microstructural evolution.

Table 2. Relationships between specific quartz crystal slip systems (SPN: slip plane normal; SD: slip direction; R: unique rotation axis) recognized in the Torridon shear zone in both symbol and Miller-Bravais forms (after Linker et al. 1984), tilt boundary misorientation axes (-C-Tis) and potential crystallographic orientation of tilt boundaries. Orientations of boundaries that may be due to dauphine twinning are indicated by shading. Crystal-slip systems Symbol

Miller-Bravais {hkil} [UVTW]

SPN

SD

SPN

SD

(c) (c) {m} W [A M

<±a> {m} <±a> <±a> <±a> <±a>

(0001) (0001) {10-10} {1-102} {10-11} {01-11}

[11-20] [1-100] [1-210] [11-20] [1-210] [2-1-10]

Tilt boundary orientations R/^TiB

M

<±fl> W

M M M

SPN

SD

(c) (c)

<+a>

(m} M M M

{»}

<±a> <+«> <±«> <±«>

Table 3. Relationships between quartz, crystal-slip systems (R: unique rotation axis; SD: slip direction; SPN: slip plane normal) recognized in the Torridon shear zone (see Table 2) and twist boundary mis orientation axes (&TWB)> assuming slip systems operate equally in combinations (see also Neumann 2000). Also shown (first row) is fcj + (c){m} combination, although latter has not been recognized in the shear zone. The top line in each row gives individual orientations of R, SD and SPN/£TwB for each combination of slip systems read horizontally and vertically. Other lines in each row give resultant crystallographic orientations ofR, SD and SPN/£TwB. Slip system Slip system

(c) [±a]

R

W

SPN

R

SD

[m\ + <±a> [±a] + <m> {2130} (2130)

SPN

R

SD

-P-TwB

£-TwB

(c) [m]

{^} [±a]

(ml [±«]

M [±«]

SPN

SD

•^TwB

(c)

-

-

-

-

SPN -P-TwB

_

-

(m][±a]

{m} + (c) r{1011} z{0111}

<±a> [1120]

(c) +{m] r{1011) z{0111)

-

-

-

-

M[±«]

W +W {2021}

<±a> [1120]

(c) + M (0113)

(c) +(/•) ^(1012)

<±a> [1120]

{m} + {*} z{0111}

-

M[±«]

{m} + {*} z{0111)

<±a> [1120]

(c) +1/1 7r(1012[

(c) +(*} {1013}

<±a> [1120]

{m} + {r} {2021}

w{1123} +w

<±a> [1120]

{n'} + {r} {1123}

fz)[±a]

{jiijr{1011) +w

<±a> [1120]

7f|0112)

(c) +{^| (1013)

<±a> [1120]

{m} + {z} {0221}

W + {^1 {2023}

<±a> [1120]

{n'} + {z} {0123}

(c) +M

R

_

-

_

-

-

{n} + {n'} {1124}

<±a> [1120]

{r} + {z} ^{1122}

MICROSTRUCTURAL EVOLUTION

Results and interpretations Microstructure and LPO Optical (Law et al. 1990, figs 1 & 2), orientation contrast (Fig. la; see also Lloyd et al, 1992, figs 2-6) and EBSD-simulated microstructural images (Fig. Ib) show a rapid development of shear zone microstructure, including foliation development, grain size reduction and mylonitization, with fractured clasts of feldspar floating in a quartz mylonite. The shear zone strain gradient and dynamic recrystallization rates can be expected to have been concomitantly steep and/or rapid. The mylonite consists of highly flattened regions that appear to represent dynamically recrystallized original grains (Lloyd et al. 1992) and it is likely that these will exhibit similar crystal orientations or orientation relationships due to a parental grain orientation control during recrystallization (e.g. Lloyd & Freeman 1991,1994). Comparison of LPO from different regions within the shear zone (Fig. 2) indicates a relatively rapid migration of wall rock crystal pole directions towards those exhibited by the mature mylonite LPO, in agreement with microstructural observations. The mature mylonitic LPO exhibits an maximum parallel to X and a (c) maximum normal to the basal/XY foliation planes and approximates to that of a single crystal orientation. This observation, frequently observed in mylonites, poses the question as to how do these LPO and associated microstructures develop and persist. SEM/EBSD analysis reveals considerably more information on LPO and hence microstructural evolution than do conventional techniques. Such information is crucial to the understanding of this and similar shear zones. Close examination of the region where the shear zone margin transforms rapidly into the mature mylonite (e.g. Fig. 1) reveals a distinctive microstructure of narrow intragranular 'bands', comprising 'grains', that are being deflected progressively into the mylonitic foliation (see also Lloyd et al. 1992, figs 2-4). Manual EBSD analysis of 'grains' from one 'band' (e.g. Fig. 5a) identifies the crystal orientation relationships expected for dauphine twinning (Fig. 5b, d), such as six rather than three pairs of rhomb orientations. Furthermore, examination of the orientation relationships between adjacent bands (e.g. Fig. 5a, points 5/24, 6/23,10/21,15/19, etc.) also indicates dauphine twinning, although with a different Euler-3 angle relationship between 'parent' and 'twin' grains (Fig. 5d). Thus, the overall microstructure consists of not only alternating 'grains' within individual 'bands' but also

47

alternating 'bands', all of which have dauphine twin relationships to each other. The overall effect is a 60° 'corrugation' of the crystallographic microstructure in two effectively orthogonal directions to form long, narrow 'bands' and small, elongate grains (Fig. 5a). Although the mature mylonite LPO approximates to that of single crystal quartz, the rhomb poles display more than three orientation clusters (Fig. 2c, d), as would be expected for a single crystal orientation. Such a configuration is typical of dauphine twinning. This behaviour has been confirmed by a detailed SEM/EBSD linear traverse at 1 um step size (i.e. smaller than the grain sizes traversed) across part of the mylonite (see Fig. la for location). The pole figures for the traverse (Fig. 6a) are similar to those for the general mylonite petrofabric (Fig. 2c): an a-axis maximum subparallel to X; a great circle distribution of a-axes and m-poles subparallel to the XY foliation plane, normal to a strong c-axis maximum subparallel to Z; and superposition of the r and z poles. However, the dispersion patterns shown by the pole figures are stronger than for the general mylonite petrofabric and are compatible with slip on a prism system such as {m} (e.g. Lloyd & Freeman 1994), although this system was not recognized by Law et al. (1990) or Lloyd et al. (1992). The SEM/EBSD analyses show that dauphine twinning plays a role in both grain size reduction and mylonite development during shear zone evolution. Initial (i.e. relatively low shear strain) microstructural and petrofabric evolution is in part accommodated by dauphine twinning of original quartz grains, which helps to reduce grain size from several millimetres to a few 100 um. The mylonite LPO approximates a dauphine twinned quartz single crystal orientation onto which specific crystal slip systems have operated. However, a more complete appreciation of the role of dauphine twinning during shear zone microstructural and LPO evolution, and its relationship to crystal-slip systems that operate concurrently, can be achieved by assessing the misorientation relationships between adjacent (sub)grains, including the formation and orientation of (sub)grain boundaries.

Misorientation analysis The only EBSD data that can be used for 'correlated' (i.e. adjacent (sub)grains sharing a common boundary) misorientation analysis are from the manual analysis of the dauphine twin microstructure (Fig. 5a) and the linear traverse across the mylonite (Fig. la). The former

Fig. 5. SEM/EBSD analysis of localized dauphine twin microstructures (see Fig. la for location). Parts (c)-(f) were constructed using the program Channel5®. (a) Orientation contrast image, showing numbered points (1-18, along the band; 19-28 adjacent to the band on either side) from where EBSP were obtained, (b) Equal area, upper hemisphere pole figures illustrating basic crystallographic relationships for points 1-28 (dotted line indicates orientation of the analysed band), constructed via manual EBSD analysis and the program Pf2k (D. Mainprice; see also Mainprice 1990 and Mainprice & Humbert 1994). (c) Correlated (nearest neighbour) misorientation angle frequency histogram and expected random distribution curve for D3-32 symmetry class. Note distinctive peaks near 60° and at low angles (typically <10°). (d) Variation in Euler angles for points 1-28. The first two Euler angles are essentially constant but the third shows two distinct variations: points 1-18, Euler 3 has values of either 50-60° or 110-120°, indicative of a dauphine twin relationship; points 19-28, Euler 3 has values of 0-20°, a near dauphine twin relationship with points 1-18 with Euler-3 = 50-60°, which is assumed to be the parental value, (e) Correlated (nearest neighbour) misorientation axis/angle pair pole figure distributions in the range 5°(5°)1050. (f) Correlated (nearest neighbour) misorientation axis/angle pair inverse pole figure distributions in the range 5°(5°)105°.

Fig. 5. continued.

50

G. E. LLOYD

involved individual measurements from either side of manually recognized boundaries, whereas the latter involved a step size significantly smaller than the grain sizes encountered (i.e. 1 um compared to 5-20 urn). Mis orientation angle distributions. Figure 5c illustrates the misorientation angle distribution for the localized dauphine twin microstructure. The distribution is bimodal There is a distinct peak at the lowest misorientation angles (i.e. <15°), which rapidly diminishes to zero frequency at 25-35°. Misorientation frequencies then increase to a second peak at 60°. This peak is asymmetrical and misorientation frequencies drop almost instantaneously to zero beyond 60°. These misorientation data were derived from a single, local microstructure from which some of the elements present elsewhere in the shear zone microstructure were missing . Figure 6b illustrates the misorientation angle distribution derived from the traverse across the mylonite. In general, the distribution is similar to that observed for the localized dauphine twin microstructure (Fig. 5d) and is also bimodal with a low frequency saddle. Misorientation angles below 35° occur with greater frequency than expected for a random distribution, with frequencies increasing towards a distinct peak below 15°. Misorientation angles from 35-55° occur with the same frequency as the random distribution but then increase suddenly to the second peak at 60°. There is an equivalent sudden decrease in frequencies beyond 60° and misorientation angles from 65-105° occur less frequently than random. These observations suggest that distributions of misorientation angles are non-random and

also consist of different components (e.g. peaks at small and approximately 60° angles, low frequency regions, etc.). However, misorientation angle distributions alone do not provide a complete description of the misorientation characteristics (e.g. although misorientation angles of 60° are highly suggestive of dauphine twinning, they do not confirm its presence as the misorientation angles must also be allied with misorientation axes parallel to the c-axis). To obtain this information it is necessary to consider the misorientation axis/angle pair data. Misorientation axis/angle pairs. Misorientation axis/angle pair data for the linear traverse across the shear zone mylonite are shown in both pole and inverse pole figure forms (Fig. 6c, d). The following relationships can be recognized: (1) A very strong correlation between misorientation axes and the c-axis (and hence sample Z) at misorientation angles of 60°; (2) A tendency for misorientation axes to develop subparallel to the c-axis and Z for misorientation angles less than 60°, although the intensity does vary; (3) For misorientation angles above 60°, the correlation between misorientation axes and the c-axis immediately disappears due to symmetry constraints; (4) Concentrations of misorientation axes occur subparallel to the XZ plane for most misorientation angles, particularly subparallel to the {r/z} rhombs aligned 'NW-NNW and 'SW-SSW subparallel to the maximum principal stress direction (Law etal. 1990); (5) Minor concentrations parallel to the second

MICROSTRUCTURAL EVOLUTION

51

Fig. 6. Petrofabric and misorientation analysis of a linear traverse across the mylonite (see Fig. la for location). Petrofabric pole figures are non-polar equal area, upper hemisphere plots constructed via autoEBSD analysis (see Table 1) using the program Pf2k (D. Mainprice; see also Mainprice 1990 and Mainprice & Humbert 1994). Correlated (nearest neighbour) misorientation data are plotted in the range 5°(10°)105° using the program Channels®. Also indicated are positions of the main crystal poles and principal forms in the quartz 60° crystallographic unit triangle (see Fig. 3). All contours unit multiples of uniform distribution (m.u.d.) except where stated, (a) Petrofabrics: dashed line, 0.5 m.u.d.; c-axis pole figure, 2 x m.u.d; maximum and minimum values indicated lower left and right respectively for each pole figure. Also indicated is J-index (fabric intensity) for each pole figure, (b) Misorientation angle distribution: note the distinctive peaks near 60°. Also included is the expected random curve for D3 - 32 symmetry class, (c) Misorientation axis/angle pair pole figures, (d) Misorientation axis/angle pair inverse pole figures.

52

G. E. LLOYD

order rhombs \n/tf} for misorientation angles below 70°; (6) As misorientation angles approach the maximum allowed by symmetry, misorientation axes become parallel to either [m}_ (especially [m] closest to X) or {2021}/{0221}; (7) Minor concentrations of misorientation axes parallel _to non-specific^ crystal directions (e.g. {0113}, {1123}, {1122} and {0223}) for a range of different misorientation angles; (8) A notable absence of misorientation axes parallel to <«> for any misorientation angle. For completeness, a misorientation axis/angle pair analysis was performed also on the dauphine twin microstructure (Fig. 5a). Results from this specific microstructure are obviously dominated by 60° misorientation angles about the c-axis (Fig. 5e, f). However, a tendency for misorientation angles to develop about (c) for angles below 60° was observed. The most obvious feature of the misorientation axis/angle pair analyses is the association between misorientation axes and the c-axis pole figure for misorientation angles of 60°, which is consistent with a significant occurrence of dauphine twinning in the microstructure. This observation supports the suggestion made earlier that the mylonitic LPO approaches that of a dauphine-twinned quartz single crystal. However, there is clearly also a relationship between misorientation axes and c-axis for misorientation angles below 60°, which cannot be due to dauphine twinning. If misorientation axes are also rotation axes for crystal slip (see Fig. 4), a relationship between misorientation/rotation axis and c-axis is consistent with a prism slip system such as [m} (see Table 2). Prism-a slip was suggested by the general EBSD analyses described above (i.e. Fig. 2c, d), although it was not reported by either Law et al. (1990) and Lloyd et al. (1992) in their studies of this shear zone. These workers recognized only {-#}<«>, [r/z} and (c) slip systems, that have rotation axes parallel to {r}, {it/rf}, and {m} respectively (Table 2), for which concentrations of misorientation axes occur also for specific misorientation angles (Fig. 6c, d), indicating slip on these systems. Many misorientation inverse pole figure distributions are asymmetrical. For example, for misorientation angles of 85-90°, there are significantly more misorientation axes parallel to {2021} compared with the equivalent negative form {0221}, and the cluster parallel to {r} is

not matched by an equivalent cluster parallel to {z}.

Boundary formation Misorientation analysis can provide information on the formation and orientation of tilt (TiB) and twist (TwB) boundaries (see Fig. 4). The important parameters are the orientations of the boundary misorientation axes (i.e. -£TiB and -£TWB) and the unique crystal slip rotation axes (R), slip plane normals (SPN) and slip directions (SD) for the observed crystal slip systems relative to the LPO and crystal symmetry (Tables 2 and 3). This section considers the misorientation axis/angle pair data from the linear traverse (Fig. 6c, d) in terms of boundary formation and orientation. Results are summarized in Fig. 7a. For the purpose of this study, it is assumed that all boundaries developed due to either crystal slip or twinning processes, although in practice the development of many grain boundaries in such fine grain size mylonites may involve significant grain boundary sliding. The most obvious relationships between the misorientation axis and angle pair distributions (Fig. 6c, d) are clusters of misorientation axes about the c-axis for misorientation angles up to 60°. The analyses reported here and elsewhere (Law et al. 1990; Lloyd et al. 1992) suggest that the formation of the Torridon shear zone involved slip on {n'}, {r/z}<«>, (c)<«> and prism-fl systems. The rotation axes (R) for these systems are parallel to {r}, {n/n'\, [m] and (c) respectively. Tables 2 and 3 show that only prism [m} slip produces rotation and misorientation axes (i.e. R/€TiB) mutually parallel to the caxis. The only combination of common quartz slip systems that can form twist boundaries where SPN/£TwE are parallel to the c-axis is (c) + (c) and the latter system has not been recognized in the Torridon shear zone. Thus, boundaries associated with misorientation axes parallel to the c-axis can be interpreted as tilt boundaries parallel to prism planes due to prism [m} slip. However, as dauphine twinning has been shown to be a common feature of the shear zone, a proportion of the boundaries with misorientation axes parallel to the c-axis and misorientation angles of 60° must be dauphine twin planes. There is no contradiction here with the role of prism [m} slip because dauphine twin planes are parallel to prism planes by definition (Fig. 8a, b). Thus, as there appears to be no evidence that dauphine twinning forms boundaries normal to the twin or misorientation axis, such that SPM€TwB is parallel to the c-axis (e.g.

MICROSTRUCTURAL EVOLUTION

53

Fig. 7. Summary of grain boundary misorientation analyses results, (a) Relationship between misorientation axis/angle pair distributions, observed individual and combined slip systems, and the formation and orientation of tilt (TiB) and twist (TwB) boundaries for the linear traverse (Fig. 6d). (b) Relationship between grain boundary types and orientations and shear zone tectonic (XYZ) and crystal (, (c), {m}, {r} and jz}) framework, (c) Schematic representation and orientation of idealised mylonite grain ships based on their likely boundary configurations.

Fig. 8c), dauphine twin planes can be viewed as a special type of tilt boundaries that form parallel to prism planes by exploiting misorientations about the c-axis. Due to the dauphine twinned single crystal nature of the mylonite, this combination of prism-a slip and dauphine twinning has the potential to produce a strong dimensional preferred orientation distribution of grain boundary planes that might impact significantly on shear zone properties (see Lloyd & Kendall, in press), Another concentration of misorientation axes has developed parallel to {in} (Fig. 6c, d). Although no combination of observed slip systems can build twist boundaries with SPN/£TvfEparallel to {m}, Tables 2 and 3 show that basal-a slip builds tilt boundaries with

R/^TiB parallel to {m}. Thus, all boundaries with misorientation axes parallel to {m} are likely to be tilt boundaries formed via basal-a slip. Misorientation axes clusters parallel to {z} (Fig. 6d) cannot be produced by slip on any of the observed systems alone (Table 2). However, a combination of slip on either {m} + (c) or (n') + [m} systems results in SPM€TwB parallel to {z} (Table 3). The clusters are likely to be associated with twist boundaries due to these combined slip systems. The misorientation axes distributions (Fig. 6c, d) are characterized by a number of clusters parallel to non-specific crystal directions. As none of these clusters can be produced via slip on any of the observed slip systems alone (Table 2), they are unlikely to be associated with tilt

54

G. E. LLOYD

Fig. 8. Dauphine twinning and twin boundaries in quartz, (a) Quartz single crystal, (b) Conventional dauphine twin with twin plane defined by tilt boundary parallel to c-axis. (c) Postulated dauphine twin with twin plane defined by twist boundary parallel to basal plane (i.e. normal to c-axis).

boundaries. However, they can be explained by the observed slip systems acting in simple combinations, as follows (Table 3): (0113), (n') + (c)<0>; (0123), (z) + (n')\ (1123), (r) + «)<«>; (1122),_(r) + (z); (0221), {m} + (z); (2021), (r) +_{m}. For each of the combinations, SPM-€TwB is parallel to the non-specific crystal direction. Thus, these clusters are likely to be associated with twist boundaries due to combined crystal slip processes. Clusters of misorientation axes parallel to {^}, {n} and fr} (Fig. 6c, d) are more difficult to associate with either tilt or twist boundaries because both individual and combinations of slip systems produce misorientation axes parallel to either R/-£TiB or 57WA€TwB. Thus, the clusters could be related to tilt boundaries due to slip on [r/z} and {rf} systems respectively (Table 2), or they could be related to twist boundaries due to combinations of slip on {z}<«> + (c)<«>, [r} + (c) and [m} + (c) systems respectively (Table 3). Minor clustery of misorientation axes in the region of p'{1131} are difficult to explain. According to Neumann (2000, fig. 2b) this direction is the rotation axis for {z} slip and therefore is associated with tilt boundaries. However, this system has not been recognized previously in the Torridon shear zone. There is a notable absence in the misorientation axis distributions (Fig. 6d). Misorientation axes do not develop parallel to the a-axis. This non-occurrence can be explained easily by consideration of Tables 2 and 3. Only (c){m] slip is capable of producing -£TiB parallel to R and as this slip system has not been observed in this

particular shear zone, no associated tilt boundaries can develop. Furthermore, as no combination of observed slip systems produces -£TWB parallel to SPN, no twist boundaries can develop. In fact, because is the predominant slip direction during shear zone evolution, it can never be parallel to either R or -i and hence cannot be parallel or normal to any boundary that develops via crystal slip. Thus, the absence of misorientation clusters parallel to provides strong supporting evidence for the crucial role of crystal slip in quartz boundary formation during evolution of the Torridon shear zone.

Discussion EBSD-based LPO and misorientation analyses of the Torridon shear zone have shown that microstructural evolution involved dauphine twinning as well as conventional crystal slip processes. Initially dauphine twinning helps to reduce grain size at the shear zone margins but the twinning process persists within the mylonite and as such must cooperate with crystal slip systems in the overall strain accommodation process. In addition, the grain boundary microstructure develops also by a combination of crystal slip and dauphine twinning. The results of this study therefore raise several interesting and intriguing issues regarding microstructural and LPO evolution and stability in quartz shear zones.

Dauphine twinning and microstructural evolution Perhaps the most obvious feature of the Torridon shear zone is the rapid reduction in grain size from several millimetres to less than 20 um (Fig. 1; see also Lloyd et al. 1992, figs 2-6). Grain size reduction starts in original wall rock grains by the formation of intragranular 'bands' that curve into the shear zone. Dauphine twinning played a role in the formation and evolution of these intragranular bands and resulted in an orthogonally twinned or 'corrugated' microstructure and grain size reduction (Fig. 5a). Although many of the boundaries are obviously dauphine twin planes, they are also a special type of tilt boundary (Fig. 8b) and form part of a group of tilt boundaries with misorientation angles of up to 60° and -€TiB parallel to the c-axis that develop parallel to prism planes due mainly to prism-a slip (e.g. Table 2; Fig. 7a). Based on these observations, the following microstructural evolutionary model (Fig. 9) can be envisaged for the XZ section of the shear

MICROSTRUCTURAL EVOLUTION

55

and requires both further confirmation and refinement. However, if it proves characteristic for the development of quartz shear zones, it shows that dauphine twinning plays an important and hitherto unappreciated role in the process of mylonite formation, as well as contributing to LPO development and boundary formation. In addition, it suggests that the microstructure includes a strong component of grain boundary preferred orientation(s) that might impact significantly on physical properties (see Lloyd & Kendall, in press).

Dauphine twinning and (sub)grain boundary formation

Fig. 9. Schematic model for shear zone grain size reduction and mylonite formation due to dauphine twinning, (a) Initial large parental grain, (b) Intragranular 'bands' with alternating parental (P) and dauphine twin (T) relationships, (c) Refined microstructure of alternating parental (P) and dauphine twin (T) orientated 'grains'; (d) Protomylonitic microstructure of alternating parental (P) and dauphine twin (T) orientated 'grains' due to repeated twinning events.

zone. Extrapolation of this model to three dimensions will require similar observations from XY and YZ sections. An original grain (Fig. 9a) initially develops intragranular bands of alternating parent and dauphine twin orientations at a high angle to the shear zone foliation (Fig. 9b). A second generation of dauphine twins divides the bands into alternating parent and dauphine twin orientated 'grains' (Fig. 9c). This procedure is repeated cyclically to reduce grain size further (Fig. 9d). The operation of conventional crystal slip systems, and in particular prism-a slip, which shares a common misorientation axis with dauphine twinning parallel to (c), can contribute at any stage with consequent modification of the microstructure observed. This model, as presented, remains simplistic

Observations from the Torridon shear zone suggest that many grains are separated either by conventional 60° dauphine twin tilt boundaries parallel to prism planes (e.g. Fig. 8b) or tilt boundaries with misorientation angles up to 60° and -£TiB parallel to (c) due to prism-a slip (Fig. 7a). Because such boundaries are parallel to (c), they are constrained to form at high angles to the mylonitic (XY) foliation, which is subnormal to the c-axis LPO and subparallel to the quartz basal plane (Figs 2c, d and 6a). However, all of the slip systems recognized in the shear zone (i.e. [m}, fr/z}, {rt} and (c)) have a common slip direction parallel to and hence form tilt boundaries normal to this direction, parallel to prism planes. As R/^TIB f°r these slip systems (i.e. (c), [nVn], [r] and [m] respectively) lie progressively further from the c-axis, they can accommodate progressively larger misorientation angles. It is conceivable that as misorientation across prism-parallel tilt boundaries increases beyond 60°, there has to be a progressive switch to slip systems capable of accommodating this misorientation. The role of dauphine twinning in this process is to be able to produce 'instantaneously' prism-parallel tilt boundaries with 60° misorientations, thereby negating the need for the progressive accumulation of misorientation up to 60°. Boundaries that develop subparallel to the mylonitic foliation cannot be tilt boundaries unless -£TiB lies close to the foliation/basal plane (Table 2; Fig. 7a). Of the slip systems recognized for the shear zone, only basal-a has -€TIB parallel to [m] but it has also a slip direction parallel to and any tilt boundaries must develop normal to this direction and therefore at a high angle to the foliation/basal plane. Thus, boundaries subparallel to the mylonitic foliation are most likely to be twist boundaries that develop due to a combination of slip systems that result

56

G. E. LLOYD

in-£TwB at high angles to the basal/foliation plane (i.e. subparallel to (c)/Z). The most likely candidate is combined {7t?} + (c)<«> slip, which has a resultant -£TwB parallel to {01-13} (Table 3). Although the former is an unusual slip system, it was identified in earlier work on this shear zone by Law et al. (1990). The only other candidates exploit combinations of {rlz} + (c) and produce resultant -6TwB parallel to {n/tf}, but form twist boundaries somewhat oblique to the foliation/basal plane. A number of other tilt and twist boundary misorientation axes have been identified from the observed slip systems (Fig. 7a; Tables 2 and 3). These may be 'microstructurally necessary' boundaries that maintain microstructural continuity and integrity during shear zone evolution. The relationships between grain boundary types and orientations and shear zone tectonic (i.e. XYZ) and crystal frameworks is summarized in Fig. 7b. Many of the tilt boundaries due to dauphine twinning, prism-a or other slip systems with slip directions parallel to have developed parallel to the YZ plane. In contrast, twist boundaries due to combined slip on {n'} + (c) have developed (sub)parallel to the XY foliation plane. No tilt boundaries have developed parallel to XZ as this plane contains the ubiquitous slip direction <#>, which is always normal to any tilt boundary. Thus, boundaries that have developed parallel to XZ must be twist boundaries with misorientation axes located close to Y. The obvious explanation is a combination of {r/z}<#> + {m} slip systems, which have -£TWB parallel to {2021} and {0221} and accommodate the maximum possible misorientation angles (Fig. 3b and Table 3). The main boundary planes recognized form simple, idealized 'cubic', 'hexagonal' and 'octahedral' shapes (Fig. 7c), with the potential for modification of the apices by minor 'microstructurally necessary' boundary planes non-parallel to the tectonic framework, that fit together easily to form the mylonitic grain microstructure.

Dauphine twinning and microstructural stability A common feature of many shear zones is the preservation of both microstructures and LPO to high shear strains once a relatively fine (mylonitic) grain size has developed. Significant modification to the microstructural and LPO characteristics typically requires changes in environmental parameters (e.g. temperature, pressure, stress, strain-rate, etc.) affecting the

shear zone. The most obvious change has traditionally been termed post-deformation recovery and is interpreted in terms of attempts to achieve so-called microstructural equilibrium. However, the shear zone microstructure during deformation was clearly in equilibrium with the prevailing conditions. The processes by which microstructures remain stable whilst still managing to accommodate increasing strain have proven difficult, and sometimes contentious, to identify (e.g. Law 1987). It is apparent from the preceding comments on microstructural and boundary evolution that although the bulk microstructure and LPO can remain constant, individual grains grow and/or are consumed and boundaries (including dauphine twin planes; Neumann 2000) migrate via dynamic recrystallization. Thus, any perception of shear zone microstructural evolution is really one of scale, with statistical stability on the larger scale and change on the smaller scale. Although grain growth and dynamic recrystallization is inherently both stable and cyclical and maintains the overall mylonitic grain size distribution, it does not necessarily explain the constancy of the LPO. If the (generations) of neoblasts that develop are due to subgrain boundary rotation dynamic recrystallization, such that they always have a crystallographic orientation relationship to their parental orientation, there is little potential for deviation from a stable LPO. However, if grain boundary migration dynamic recrystallization processes are significant, the LPO may vary considerably. Furthermore, there remain serious questions concerning the tendency for many shear zones to exhibit almost single crystal LPO (e.g. Fig. 2c, d), particularly when the microstructure contains boundaries with a wide range of misorientations (e.g. Figs 6 & 7). The stored energy characteristics associated with these grains and boundaries may be sufficient to allow grain boundary mobility to absorb the cyclical growth and destruction of individual grains. The observations presented in this contribution offer a potential explanation for microstructural stability and constant (single crystal) LPO by recognizing the hitherto unappreciated role of dauphine twinning in shear zone microstructural and LPO evolution. In general, dauphine twinning exchanges the positive and negative forms in the quartz crystallography (i.e. {r} and {z}, <+«> and <-a>, \n] and {#'}, etc.). It has been suggested (e.g. Frondel 1962; Law et al. 1990; Lloyd et al 1992; Mainprice et al. 1993) that the critical resolved shear stress (CRSS) for crystal slip appears to be

MICROSTRUCTURAL EVOLUTION

easier on negative compared to positive forms. As many quartz crystal slip systems exploit for the slip direction (see Table 2), this means that slip in the <-a> direction would be easier than slip in the <+a> direction. In this respect, Neumann (2000) reported clusters of rotation axes and axial dispersion paths indicative of slip systems related to adjacent a-axes. Such a situation might imply different Burgers vectors parallel to <-a> and <+a>, although this has not been recognized. Alternatively, it may simply represent a difference in polarity (e.g. analogous to walking either up or down a dip slope). Whatever the cause, dauphine twinning can interchange relatively 'hard' <+a> orientations for relatively 'soft5 <-a> orientations for little appreciable effort. The net result is that the <±a> directions for any grain instantaneously change position in the LPO by 60° while the caxis positions remain constant. Thus, an apparently stable single crystal LPO (i.e. as measured by optical and/or X-ray methods) can persist, although in detail there is an oscillation of the <±a> directions, presumably about the macroscopic shear zone extension direction (i.e. X), whilst simultaneously accommodating large misorientations of individual components. This process has no effect on the observed (fine grain size) microstructure that develops, except perhaps in helping it to persist. Rather, grains continually 'switch' orientations between parent and twin configurations, thereby accommodating deformation and displacement. Furthermore, as progressive misorientation about the c-axis up to (a maximum) of 60° by prism-a slip effectively results in a prism-parallel dauphine twin tilt boundary, the boundary perhaps can be (immediately) 'untwinned' by the application of an opposite dauphine twin operation, thereby allowing misorientations to be continuously accrued across the boundary. The role of dauphine twinning is thus continually to 'refresh' (recover?) shear zone microstructure, allowing very large strains to be achieved under essentially steady-state conditions.

Misorientation angle distributions In an earlier section, misorientation angle distributions were described from a local dauphine twinning microstructure and a linear traverse (Figs 5d and 6b) and it was suggested that they consisted of several different components. Similar distributions have recently been reported by other workers (e.g. Neumann 2000, M. Drury pers. comm. 2002, P. Trimby pers. comm. 2002). However, without more specific information provided by misorientation

57

axis/angle pair data, it was not possible to interpret these distributions. Potential explanations are now considered. (1) High frequencies of misorientation angles below 15° correspond to simple subgrain boundaries that have developed mainly by subgrain rotation dynamic recrystallization. The distribution of misorientation axes responsible is clearly dispersed (Fig. 6d). This may be a consequence of the need to maintain compatibility between a large number of subgrains and their parental orientation as they nucleate and evolve, which probably involves the operation of a large number of very minor 'accommodating' slip systems (e.g. Neumann 2000). It may also be due to difficulties in accurately locating misorientation axes as misorientation angles approach zero (Prior 1999). (2) The occurrence of low frequency 'saddles' at about 30° may appear incompatible with the continuous achievement of increasing misorientation due to progressive subgrain rotation dynamic recrystallization. A logical explanation would be for a change in recrystallization mechanism (e.g. to grain boundary migration). Further detailed work is needed to verify this (see comments in Neumann 2000). However, the combined and/or successive operation of several crystal-slip systems does not necessarily result in an increase in misorientation angle. For example, a progressive increase in misorientation angle due to basal-a slip can be followed by an apparent decrease if the slip system changes to prism-a. Thus, low frequency misorientation saddles may be compatible with progressive misorientation due to subgrain rotation dynamic recrystallization. (3) The progressive increase in misorientation angle frequencies from about 30° to 60° can be explained by several processes, including: operation of specific slip systems, either individually or in additive combinations (e.g. Neumann 2000), grain boundary migration, grain boundary sliding and/or minor quartz twin laws (e.g. Frondel 1962; Kruhl and Peternell 2002). Careful analysis of individual boundaries (e.g. Trimby et al. 1998; Bestmann & Prior 2003) is needed to distinguish these potential contributions to microstructural and misorientation evolution. (4) The maximum frequency in misorientation angle distribution that occurs at about 60° is both exaggerated and asymmetrical. The exaggerated peak can be explained by the superposition of misorientation angles due to both crystal-slip systems and dauphine twinning. Misorientations due to the former can accrue progressively up to 60°, whereas the latter involves

58

G. E. LLOYD

a precise misorientation of 60°. The asymmetry can be explained by the fact that slip systems with a misorientation axis parallel to the c-axis are limited to a maximum misorientation angle of 60° by symmetry considerations. Thus, there is an instantaneous loss of contribution to misorientation angle frequency by these slip systems beyond 60°. (5) The decrease in misorientation angle frequencies beyond 60° can also be explained by limitations placed on the maximum misorientation angles possible due to symmetry considerations. Depending on the position of the misorientation axis, the maximum misorientation angle possible falls in the range 60-109°, with larger values having misorientation axes lying further from the c-axis (see Tables 2 and 3). Thus, for misorientations above 60° there is an apparent decrease in the number of slip systems available. However, the misorientation due to a particular slip system cannot simply stop at its prescribed maximum value without introducing mechanical instabilities into the microstructure. Misorientation may continue to be accumulated but be transferred to a symmetrically equivalent lower misorientation value. For example, a true misorientation of 61° could be recognized as either 59° or 1°, with the former contributing to the asymmetrical peak about 60° and the latter to the high frequency of misorientation angles less than 15°, although this implies that high angle grain boundaries can suddenly become subgrain boundaries! Two explanations for this apparent paradox have been provided already. First, it is possible that because some slip systems share a common slip direction (e.g. ) and hence tend to build mutually parallel boundaries (e.g. parallel to prism planes), misorientation can be transferred progressively to a system with a relatively higher maximum permitted misorientation value. Secondly, cyclical slip and dauphine twinning can potentially accrue misorientation continuously across a boundary without apparently exceeding the maximum value permitted. Nevertheless, the maximum misorientation angle that can be measured between two adjacent quartz 'grains' is 104.5° irrespective of the deformation mechanism or process responsible (i.e. twinning, crystal slip, grain boundary migration or sliding, fracturing, etc.). This constraint may cast doubt on the current development status of misorientation analysis in quartz.

Quartz misorientation analysis - a caveat The description and interpretation of quartz misorientation angles and axis/angle pairs may be influenced and biased by the 60° symmetry

assumed in the construction of the misorientation inverse pole figures (e.g. Fig. 3). Although it is beyond the scope of the present contribution to discuss this problem in detail, the following consequences should be considered in future quartz misorientation analyses. The 60°construction constrains the maximum possible misorientation angles about specific crystal orientations (e.g. the maximum misorientation angles about (c), and {m} are 60°, 90° and about 100° respectively). Thus, misorientation distributions about the two {m} positions in the inverse pole figure should be symmetrically equivalent. However, this is commonly not the case (e.g. see Fig. 6d). Mainprice et al, (1993) recognized a similar asymmetry, particularly involving <+a> and <-«>, in their study of misorientations within an individual quartz grains and accordingly adopted a 120° misorientation inverse pole figure construction. Although any misorientation can be defined by a number of symmetrically equivalent axis/angle pair combinations, current convention is to adopt the pair with the minimum misorientation angle (e.g. Mainprice et al. 1993; Lloyd et al. 1997). However, this convention is known to cause problems in specific situations. For example, misorientation analysis defines etwinning in calcite as a misorientation of 78° about an axis of <2021> rather than 180° about <4041> (Cooper 2002). Both are symmetrically equivalent, but the latter is the twin law. It is possible therefore that all symmetrically equivalent combinations of misorientation axis/angle pairs should be considered (see also Cross & Randle 2003). Finally, there may be a difference in the propensity for crystal slip between the <-a> and <+a> directions (e.g. Law et al. 1990; Lloyd et al. 1992; Mainprice et al. 1993). In other words, the a-axes should be considered as polar rather than non-polar lines. A 60° misorientation inverse pole figure construction cannot differentiate between these directions and may result in a significant loss of information.

Conclusions This contribution has used SEM/EBSD orientation and misorientation analyses to investigate the microstructural and petrofabric (LPO) evolution of an amphibolite facies quartz simple shear zone. The study has focused attention on the importance of the grain boundary network, which develops by a combination of dauphine twinning and crystal slip processes, as well as the conventional grain microstructure and petrofabric. In particular, it has highlighted the hitherto generally unappreciated role of dauphine

MICROSTRUCTURAL EVOLUTION twinning in microstructural and petrofabric evolution. Dauphine twinning participates initially in the grain size reduction process and subsequently in the development and preservation of a fine grain mylonitic microstructure. It combines with slip systems that exploit as the slip direction to produce a 'dauphine twinned single crystal' LPO and a microstructure dominated by misorientations about the caxis that form tilt boundaries parallel to prism (and often YZ tectonic) planes also. Misorientation angles greater than 60° appear to be achieved by transfer of slip from prism-n to systems that are capable of accommodating progressively higher misorientations up to the maximum of 104.5° allowed by symmetry (i.e. {r/z}<«>, {tf} and (c)<«> respectively). Although no other tilt boundaries are formed, the same slip systems operate in simple combinations to produce twist boundaries, particularly parallel to the XY foliation/basal planes and XZ tectonic plane. The combination of grain and grain boundary microstructures, misorientations and petrofabrics may provide explanations for a number of long-standing problems that remain extant in our understanding of shear zones, such as microstructural stability and the occurrence and preservation of steady-state LPO to high shear strains. In this respect, dauphine twinning may play a crucial role by exchanging <+«> and <-«> directions and thereby determining whether a grain has a 'hard' or 'soft' orientation for the operation of crystal slip systems that exploit as the slip direction. In a companion paper (Lloyd & Kendall, in press), the implications of the microstructural and petrofabric evolution described in this paper are considered in terms of the seismic properties exhibited by this shear zone. Thanks to Dave Mainprice for use of his crystal lattice pole figure plotting programs. Numerous discussions with Martin Casey, Jane Cooper, Martyn Drury, Andy Farmer, Andy Lind, Dave Mainprice, Berndt Neumann, Dave Prior, Pat Trimby and John Wheeler have helped to develop my thoughts on SEM/EBSD misorientation analysis. The detailed comments of Michel Bestmann and an anonymous referee helped to improve the original version of this manuscript. Thanks also to the editors of the special publication, Ian Alsop and Bob Holdsworth, for their assistance and understanding. A part of the automated SEM/EBSD facilities was funded by the UK NERC Small Grant GR9/3223.

References ADAMS, B.L., WRIGHT, S.I. & KUNZE, K. 1993. Orientation imaging: the emergence of a new microscopy. Metallurgical Transactions, 24A, 819-831.

59

BAKER, D.W. & RIEKELS, L.M. 1977. Dauphine twinning in quartz mylonite. Journal of Geology, 85, 15-26. BAKER, D.W. & WENK, H.-R. 1972. Preferred orientation in a low-symmetry quartz mylonite. Journal of Geology, 80, 81-105. BARBER, DJ. & WENK, H.-R. 1991. Dauphine twinning in deformed quartzites - implications of an in-situ TEM study of the alpha-beta phase transformation. Physics and Chemistry of Minerals, 17, 492-502. BESTMANN, M. & PRIOR, D.J. 2003. Intragranular dynamic recrystallisation in naturally deformed calcite marble: a case study by means of misorientation analysis. Journal of Structural Geology, 25, 1597-1613. CROSS, I. & RANDLE, V. 2003. Lowest angle solution versus low-index axis solution for misorientations. Scripta Materialia, 48,1587-1591. COOPER, J.L. 2002. Diffusional mass transfer and grain boundary processes in calcite. Unpublished Ph.D. thesis, University of Leeds, UK. ETCHECOPAR, A. 1977. A plane kinematic model of progressive deformation in a polycrystalline aggregate. Tectonophysics, 39,121-139. PAUL, U.H. & Frrz GERALD, ID. 1999. Grain misorientations in partially molten olivine aggregates: an electron backscatter diffraction study. Physics and Chemistry of Minerals, 26,187-197. FIELD, DP. 1997. Recent advances in the application of orientation imaging. Ultramicroscopy, 61,1-9. FLIERVOET, T.F, DRURY, M.R. & CHOPRA, PN. 1999. Crystallographic preferred orientations and misorientations in some olivine rocks deformed by diffusion or dislocation creep. Tectonophysics, 303,1-27. FRONDEL, C. 1962. Dana's system of mineralogy: III Silica minerals. John Wiley & Sons, 334 pp. HULL, D. & BACON, D.J. 1984. Introduction to dislocations. Pergamon, Oxford. JENSEN, DJ. & SCHMIDT, N.H. 1991. Local texture measurements by EBSP - new computer procedures. Textures and Microstructures, 14,97-102. KRIEGER-LASSEN, N.C. 1996. The relative precision of crystal orientations measured from electron backscattering patterns. Journal of Microscopy, 181,72-81. KRUHL, J.H. & PETERNELL, M. 2002. The equilibration of high-angle grain boundaries in dynamically recrystallized quartz: the effect of crystallography and temperature. Journal of Structural Geology, 24,1125-1137. KRUSE, R., STUNITZ, H. & KUNZE, K. 2001. Dynamic recrystallisation processes in plagioclase porphyroclasts. Journal of Structural Geology, 23, 1781-1802. LAW, R.D. 1987. Crystallographic fabrics and deformation histories. Journal of the Geological Society of London, 144, 675-676. LAW, R.D., SCHMID, S.M. & WHEELER, J. 1990. Simple shear deformation and quartz Crystallographic fabrics - a possible natural example from the Torridon area of NW Scotland. Journal of Structural Geology, 12, 29-45. LINKER, M.F, KIRBY, S.H., ORD, A. & CHRISTIE, J.M.

60

G. E. LLOYD

1984. Effects of compression direction the plasticity and rheology of hydrolytically weakened synthetic quartz crystals at atmospheric pressure. Journal of Geophysical Research, 89,4241-4255. LISTER, G.S. & SNOKE, A. 1984. S-C Mylonites. Journal of Structural Geology, 6, 617-638. LLOYD, G.E. 1987. Atomic number and erystallographic contrast images with the SEM: a review of backscattered electron techniques. Minemlogical Magazine, 51, 3-19. LLOYD, G.E. 2000. Grain boundary contact effects during faulting of quartzite: an SEM/EBSD analysis. Journal of Structural Geology, 22,1675-1693. LLOYD, G.E. & FREEMAN, B. 1991. SEM electron channelling analysis of dynamic recrystallisation in a quartz grain. Journal of Structural Geology, 13, 945-953. LLOYD, G.E. & FREEMAN, B. 1994. Dynamic recrystallisation of quartz and quartzites. Journal of Structural Geology, 16, 867-881. LLOYD, G.E. & KENDALL, J.-M. in press. Petrofabric and seismic property evolution in a mylonitic quartz simple shear zone. In: HARVEY, P.K. & BREWER, T. (eds) Petrophysical properties of crystalline rocks. Geological Society, London, Special Publications. LLOYD, G.E., SCHMIDT, N-H., MAINPRICE, D. & PRIOR, DJ. 1991. Crystallographic textures. Mineralogical Magazine, 55, 331-345. LLOYD, G.E., LAW, R.D., MAINPRICE, D. & WHEELER, J. 1992. Microstructural and crystal fabric evolution during shear zone formation. Journal of Structural Geology, 14,1079-1100. LLOYD, G.E., FARMER, A.B. & MAINPRICE, D. 1997. Misorientation analysis and the formation and orientation of subgrain and grain boundaries. Tectonophysics, 279, 55-78. MAINPRICE, D. http://www.isteem.univ-montp2.fr/ TECTONOPHY/petrophysics/software/ petrophysics_software.html MAINPRICE, D. 1990. An efficient FORTRAN program to calculate seismic anisotropy from the lattice preferred orientations of minerals. Computers and Geosciences, 16, 385-393. MAINPRICE, D. & HUMBERT, M. 1994. Methods of calculating petrophysical properties from lattice preferred orientation data. Survey Geophysics, 15, 572-592. MAINPRICE, D., LLOYD, G.E. & CASEY, M. 1993. Individual orientation measurements in quartz polycrystals - advantages and limitations for texture and petrophysical property determinations. Journal of Structural Geology, 15,1169-1187. MCLAREN, A.C. 1986. Some speculations on the nature of high-angle grain boundaries in quartz rocks. In: HOBBS, B.E. & HEARD, H.C. (eds) Mineral and rock deformation: laboratory studies, The Paterson Volume, Geophysical Monograph Series, vol. 36, pp. 233-247. American Geophysical Union, Washington, D.C. NEUMANN, B. 2000. Texture development of recrystallised quartz polycrystals unravelled by orientation and misorientation characteristics. Journal of Structural Geology, 22,1695-1711.

OLESEN, N.O. & SCHMIDT, N.H. 1990. The SEM/ECP technique applied on twinned quartz crystals. In: KNIPE, RJ. & RUTTER, E.H. (eds) Deformation mechanisms, rheology and tectonics. Geological Society, London, Special Publications, 54, 369-374. POSPIECH, I, SZTWIERTNIA, K. & HAESSENER, F. 1986.

The misorientation distribution function. Textures and Microstructures, 6, 201-215. PRIOR, DJ. 1999. Problems in determining the orientation of crystal misorientation axes for small angular misorientations, using electron backscatter diffraction in the SEM. Journal of Microscopy, 195,217-225. PRIOR, D. I, BOYLE, A.P, BRENKER, E, CHEADLE, M.C., DAY, A., LOPEZ, G, POTTS, G.J., REDDY. S., SPIESS, R., TIMMS, N, TRIMBY, P., WHEELER, J. & ZETTERSTROM, L. 1999. The application of electron backscatter diffraction and orientation contrast imaging in the SEM to textural problems in rocks. American Mineralogist, 84,1741-1759. RAMSAY, J.G. 1980. Shear zone geometry: a review. Journal of Structural Geology, 2, 83-100. RAMSAY, J.G. & GRAHAM, R.H. 1970. Strain variation in shear belts. Canadian Journal of Earth Sciences, 1, 786-813. RANDLE, V. 1992. Microtexture determination and its application. The Institute of Materials, London, 174 pp. RANDLE, V. 1993. The measurement of grain boundary geometry. Institute of Physics Publishing, Bristol, 169 pp. RANDLE, V. & RALPH, B. 1986. A practical approach to the determination of the crystallography of grain boundaries. Journal of Materials Science, 21, 3823-3828. RUTTER, E.H., BORIANI, A., BRODIE, K.H. & BURLINI, L. 1998. Special Issue: Structures and properties of high strain zones in rocks. Journal of Structural Geology, 20,200 pp. SCHMIDT, N.H. & OLESEN. N.0.1989. Computer-aided determination of crystal-lattice orientation from electron channelling patterns in the SEM. Canadian Mineralogist, 21,15-22. TRIMBY, P.W. & PRIOR, DJ. 1999. Microstructural imaging techniques: a comparison between light and scanning electron microscopy. Tectonophysics, 303, 71-81. TRIMBY, P.W., PRIOR, DJ. & WHEELER, J. 1998. Grain boundary hierarchy development in a quartz mylonite. Journal of Structural Geology, 20, 917-935. TRIMBY, P.W., DRURY, M.R. & SPIERS, CJ. 2000. Misorientations across etched boundaries in deformed rock salt: a study using electron backscatter diffraction. Journal of Structural Geology, 22,81-89. TULLIS, J.A. 1970. Quartz preferred orientation in rocks produced by Dauphine twinning. Science, 168,1342-1344. TULLIS, J.A. & TULLIS, T. 1972. Preferred orientation of quartz produced by mechanical twinning: thermodynamics and axial experiments. In: Flow and Fracture of Rocks - the Griggs Volume, American Geophysical Union, 16, 67-82.

MICROSTRUCTURAL EVOLUTION TULLIS, X, YUND, R. & FARVER, 11996. Deformationenhanced fluid distribution in feldspar aggregates and implications for ductile shear zones. Geology, 24,63-66. VENABLES, XA. & HARLAND, C.X 1973. Electron backscattering patterns - a new technique for obtaining crystallographic in formation in the scanning electron microscope. Philosophical Magazine, 27,1193-1200.

61

WHEELER, J. 1984. A new plot to display the strain of elliptical markers. Journal of Structural Geology, 6, 417-423. WHEELER, X, PRIOR, D.X, JIANG, Z., SPEISS, R. & TRIMBY, P.W. 2001. The petrological significance of misorientations between grains. Contributions to Mineralogy and Petrology, 141,109-124.

This page intentionally left blank

The application of GIS to unravel patterns of deformation in high grade terrains: a case study of indentor tectonics from west Greenland SANDRA PIAZOLO13, G. I. ALSOP2, B. M0LLER NIELSEN1, J. A. M. VAN GOOL1 1 Geological Survey of Denmark and Greenland, 0ster Voldgade 10,1350 Copenhagen K, Denmark 2 Crustal Geodynamics Group, School of Geography & Geosciences, University of St Andrews, St Andrews, Fife, Scotland KYI6 9AL, UK (e-mail: [email protected]) ^currently at: Department of Earth Sciences, University of Liverpool, 4 Brownlow Street, Liverpool, L69 3GP, UK (e-mail:[email protected]) Abstract: The ability to compare, integrate and knit together multidisciplinary datasets in terms of subject, space and scale is critical to the recognition of geological patterns. In this contribution, we show that the use of Geographic Information Systems (GIS) is extremely valuable in detecting patterns associated with broad zones of deformation in high grade terrains. The GIS methodology facilitates the geological interpretation and development of models as it permits an easy and quick investigation of several geoscientific datasets by subject, space and scale. The GIS-based integration of structural, metamorphic, fabric type and aeromagnetic datasets collected in west Greenland shows that patterns seen within one dataset coincide with patterns observed in other datasets. Consequently, two major domains are recognized that are separated by a broad boundary zone. The southern block is characterized by a distinct, irregular magnetic signal coupled with granulite facies metamorphism and dominant S-type fabrics. The map scale geometry of this block controls the patterns observed within the amphibolite facies domain further north. Foliation and lineation patterns form an arcuate swing in strike about the southern block. Fabric types vary both around the strike swing and across strike. An indentor model that incorporates a rigid, cooled granulite block in the south bounded to the north by a Theologically weaker amphibolite facies domain can explain these patterns. The preserved metamorphic grade governs the rheology of the different, but essentially authochthonous blocks with the amphibolite facies domain being plastered and 'moulded around' the rigid granulite indentor. As patterns of remote geophysical and geological data closely correspond with one another, greater confidence may be placed in the application of remote geophysics in areas which lack abundant ground-based data.

In order to unravel the geological history of a complex area, it is essential to examine data and recognize patterns in terms of subject, space and scale. Integration, combination and knitting together of diverse subjects such as directional structural data, fabric type, petrological and geophysical data enables a clearer understanding and interpretation of processes. The spatial distribution and relationships between these subjects allows pattern recognition and refinement of models. In addition, pattern recognition is strongly scale dependent. Patterns that appear vague and unstructured at a small scale e.g. outcrop maps, may be recognized as a much more significant, distinct and ordered at a regional scale e.g. regional maps. Thus, the ability to represent geological subjects spatially and at a

variety of scales is crucial to the identification of geological patterns and hence the understanding and interpretation of geological problems, It is often difficult to detect large scale deformation features such as broad high strain zones by traditional methods of geological mapping, This problem becomes particularly acute within high grade gneiss terrains. In such areas structural interpretations are often hindered by broad zones of deformation coupled with a lack of stratigraphy. Diffuse areas of high strain may develop (Tikoff et al. 2002; Van Gool & Piazolo 2002) as both the overall viscosity and rheological contrasts of rocks undergoing high temperature deformation are low (e.g. Rutter 1999; Kirby & Kronenberg 1987; Treagus 1988). Therefore the effective use of all available

From: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 63-78. 0305-8719/$15.00 © The Geological Society of London 2004.

64

S.PIAZOLOETAL.

multidisciplinary data in terms of subject, space and scale is essential for the geological interpretation of terrains where diffuse high strain zones are expected and which may or may not contain localized zones of discrete displacement. A powerful tool with which to represent and filter such subject, spatial and scalar data and thereby decipher patterns is Geographic Information Systems (GIS). GIS is an information system that is used to input, store, retrieve, manipulate, analyse and visualize geographically-referenced or geospatial data. The key components of GIS are a computer system (hardware and software) and geospatial data. GIS has already proven useful in several disciplines including mineral exploration on local to global scale (e.g. Bonham-Carter et al 1990; Goodwin et al 1996; Knox-Robinson & Wyborn 1997; Harris etal 2001), palaeontology (e.g. Carrasco & Barnosky 2000) and environmental assessment (e.g. Books 2000; Wilson et al 2000; True et al 1999). In this contribution, we show GIS based datasets from various geoscientific disciplines, which allow the recognition of large scale geological and geophysical patterns within a high grade terrain and thus enable interpretation and development of models. The case study that demonstrates the strength of combining different multidisciplinary datasets within GIS is taken from a large scale geological and geophysical mapping project (Van Gool et al 20026) which forms part of the major Central West Greenland programme of GEUS. This area is particularly well suited to such a GIS-based approach as it is well exposed and therefore significant datasets encompassing several disciplines, i.e. structural and metamorphic geology and geophysics, could be collected in a collaborative programme. The datasets are comparable in scale and resolution, which thus permits full integration into GIS. In addition, the exceptional level of exposure allows a strong field based program of data collection and thereby provides an excellent opportunity to test the accuracy and limits of remote geophysical data. In this case study, the correspondence of geophysical and geological data allows greater confidence to be placed in the application of remote geophysics in areas that lack abundant ground-based data.

The West Greenland case study The case study area covers over 3000 km2 and extends from 68 °N-68 °30' N and 52 °W-53 °15' W (Fig. 1). This region is well suited for such a study, as it is transected by major fjord systems which allow data collection

Fig. 1. Schematic map of the geology of the Nagssugtoqidian Orogen and adjacent forelands. SNO, Southern Nagssugtoqidian Orogen; CNO, Central Nagssugtoqidian Orogen; NNO, Northern Nagssugtoqidian Orogen; SNF, Southern Nagssugtoqidian Front. Modified after Van Gool et al (1996). Outlined box represents the location of the study area.

along exposed coastal sections. Previous reconnaissance work (Noe-Nygaard & Ramberg 1961; Henderson 1969; Marker et al 1995; Kalsbeek & Nutman 1996; Mengel et al 1998; Connelly et al 2000) suggests that within this area, distinct crustal scale shears are lacking which would overpower more delicate relationships. Therefore, this area is well suited to test the applicability of a GIS based approach in the detection of subtle geological relationships and patterns. The area represents the northern part of the approximately 300 km wide, generally easttrending Palaeoproterozoic Nagssugtoqidian orogen (Fig. 1). In the broadest sense, this tectonic belt resulted from a continent-continent collision between the Archaean North Atlantic Craton to the south and an Archaean continental mass to the north (e.g. Kalsbeek et al 1987; Connolly et al 2000; Van Gool et al 20020). The orogen is generally characterized by E-W trending kilometre-scale folds and ENE-trending linear belts which overprint an Archaean fabric. On the basis of the grade of metamorphic reworking, Ramberg (1949) and later Marker et al (1995) defined three subdivisions: the southern, central and northern Nagssugtoqidian Orogen (SNO, CNO, and NNO, Fig. 1). The southern foreland (south of SNO) consists of an Archaean complex of the North Atlantic Craton with granulite facies orthogneisses and minor units of supracrustals. In the SNO these

APPLICATION OF GIS TO HIGH GRADE ROCKS Archaean granulite fades gneisses are reworked at amphibolite facies during south-directed thrusting and folding (Marker et al 1995). A 5-10 km wide belt of straight gneisses associated with abundant supracrustal units marks the CNO-SNO boundary. The CNO comprises Archaean orthogneisses, sequences of Archaean and Proterozoic metasediments (Connelly & Mengel 2000) and two main bodies of Palaeoproterozoic calc-alkaline intrusive rocks, which are interpreted as remnants of magmatic arcs associated with subduction in the core of the orogen (Kalsbeek et al 1987). The CNO is at granulite facies, with the exception of the northeastern part, which is at amphibolite grade. Its northern boundary is formed by the Nordre Str0mfjord shear zone that may represent a sinistral, transpressional shear zone with minor displacement (Fig. 1; Marker etal 1995, Hanmer et al. 1997; Passchier et al 1998; Van Gool et al 20020). At the time of commencement of this study, the geological evolution of the NNO was the least known of the orogen, but previous studies have shown that the metamorphic grade generally decreases northwards and is predominantly amphibolite facies, with granulite facies rocks preserved in the southwestern corner (e.g. Marker et al 1995).

The GIS database Lithological, metamorphic, directional structural and fabric-type datasets were collected during a regional mapping programme conducted at a scale of 1:40 000, and high-resolution aeromagnetic data (Rasmussen & Van Gool 2000) were flown at a similar resolution with a data collection spacing of about 18 metres along flight lines orientated east-west for every 0.5 km and tie-lines for every 5 km in a north-south direction, respectively (Rasmussen & Van Gool 2000, and references therein). The nominal flight altitude was 300 m above the terrain. The aeromagnetic data needs to be transformed and filtered through various techniques (Appendix A) before being added to the GIS database. All geological and geophysical datasets have been incorporated into a commercial fullfeatured GIS database in which data can be analysed in terms of subject, space and scale. Within this GIS system thematic maps can be created that can then be used to display information stored in the respective datasets. Each map is based on a combination of different subjects such as lithology or structural measurements, which by using symbols with colour coded orientations show chosen attributes and various values at geographical locations. GIS

65

stores all information in a common GIS database. Thereby information from various subjects can be quickly accessed, queried, integrated, combined and evaluated. For this study we use the ArcView® version 3.2 which runs on Windows NT® and Windows 98® operating systems. In addition, an interactive GIS stereoplot program provided by Knox-Robinson & Gar doll (1998) was incorporated.

Lithological dataset Quartzo-feldspathic orthogneiss clearly dominates the map area; however, the overall map pattern is governed by discontinuous NE-SW trending belts of metasediments (Fig. 2). These supracrustals comprise two main units: (a) 2-3 km thick sequences of homogeneous medium-grained quartzo-feldspathic gneisses which are generally quartz-rich and biotitepoor, containing small (1-2 mm) garnets. This sequence is interpreted to represent psammitic rocks with subordinate mafic volcanic and pelitic intercalations, (b) Finely, layered mafic to intermediate sequences with calc-silicate pods make up a significant part in the supracrustal rocks of the area, which are interpreted to be predominantly volcanic in origin. This sequence is closely associated with pelitic to semipelitic schist containing biotite, quartz, plagioclase, with or without garnet and sillimanite. Only a few thin quartzo-feldspathic layers and occasional quartzites are present.

Metamorphic dataset The metamorphic grade on the map scale varies from granulite facies in the south to amphibolite grade further north. The contact between these two facies is transitional over a distance of 10 to 12 km and forms a gently curved boundary which is subparallel to the strike of the regional foliation (Fig. 2). In the SW of the area it is NE-SW trending, whereas towards the east it is more east-west trending. The granulite facies pyroxene bearing gneisses exhibit a brownish, soapy patina and are characterized by melt pockets and cross-cutting veins containing relict pyroxene. Discontinuous, slightly cross-cutting mafic sheets are frequent. Detailed petrological and thermo-barometric analyses (Piazolo 2002) point to peak temperatures of 800 ± 30 °C at medium pressures of 6-7.5 kbar. To the north of the granulite facies lighter coloured amphibolite grade gneisses contain fewer, biotite bearing melts and display assemblages suggesting P-T conditions of 650 ± 30 °C at 4-5 kbar. Within the supracrustal amphibolites growth of secondary,

66

S. PIAZOLO ETAL.

Fig. 2. Geological map of study area based on the 1:100 000 compilation of 1:40 000 field maps (modified after Van Gool et al 2QQ2a).

moderately aligned amphibole is commonly developed along foliation planes indicating a syntectonic fluid flux and associated metamorphism. Importantly, these rocks do not preserve any information of an early granulite facies metamorphism. Within the transitional boundary between the granulite and amphibolite facies, areas of metre-scale interfingering brown granulite to light coloured amphibolite facies rocks are observed. In addition, in both mafic and orthogneiss sequences granulite facies assemblages and mineral compositions are only partly retrogressed to amphibolite facies even in thin section scale (Piazolo 2002, unpublished data). Retrogression from granulite to amphibolite facies can only take place if sufficient fluid is present, both to catalyse the appropriate reactions and to provide the chemical components i.e. H2O which in turn allows the formation of the reaction product (e.g. amphibole). The described field observations and microtextural studies suggest that the described transitional boundary originates from incomplete retrogression of granulite to amphibolite facies. The boundary is irregular, subparallel to the foliation and is inclined at an angle of 20-30° towards the NNW. Due to the gentle-moderate inclination and irregularity of the boundary, the transition appears as a broad zone on map view (Fig. 2).

Directional structural dataset On the map scale the regional foliation defines a swing in strike from NE-SW trending in the west to more east-west strikes in the east (Fig. 3a). In the northwestern part of the area a very consistent NE-SW strike is observed. It is associated with variable dip directions to both the NW and SE. On the islands west of Tunorsuaq, the dominant direction of dip alternates across strike on a scale of approximately 5 km reflecting the wavelength of folding with such reversals in dip traceable along strike for over 40 km. The NE continuation of this belt on the eastern limb of the strike swing exhibits east-west strikes with variable dip directions but no obvious pattern of dip direction along strike. In the southern part of the study area, dip directions vary between NW and north around the strike swing with no map scale pattern of alternating dip directions. In general, dip values in metasedimentary belts are consistently greater than in the adjacent orthogneisses. Lineations within the area are defined by mineral alignment of amphibole and/or elongate quartz-feldspar aggregates. Outcrop-scale fold hinges are typically tight to isoclinal and colinear with the dominant mineral grain and aggregate lineation (Piazolo & Passchier 2002) reflecting

Fig. 3. Summary figures of structural dataset; (a) representation of foliation trend and dip direction of the study area; (b) representation of lineation trend and plunge direction of the study area; (c) Spatial distribution of fabric type (see Flinn 1978).

APPLICATION OF GIS TO HIGH GRADE ROCKS

67

68

S.PIAZOLOETAL.

high strains. As such the lineation trends on the map will also reflect the orientation of fold hinges. On the map scale, lineation trends define a similar pattern as described for the strike of foliation, with NE-SW trends in the west and becoming more east-west trending towards the east (Fig. 3a, b). In general, lineations are gently plunging although in mesoscopic fold hinges their plunge markedly increases. In the NW part of the area, the lineations plunge gently towards both the SW and NE to define a series of culminations and depressions. Reversals in the plunge of lineations may be traced across strike to define a series of alternating NW-SE trending culmination and depression surfaces. The sinuousity of this 'along strike porpoising' decreases from a wavelength of approximately 20 km in the NW to less than 10 km along the Tunorsuaq fjord and the islands to the west. In the SE of the area lineations plunge dominantly towards the SW and show no pronounced whalebackpatterns.

Fabric type dataset The regional correlation of polyphase fold and fabric sequences in high grade terrains is generally hindered due to continuous reworking, overprinting and transposition of local fabrics during progressive deformation. Therefore, in such terrains analyses of the finite shape fabric which is the result of the bulk deformation becomes the most useful method of recognizing significant patterns in regional deformation. This generalization of fabric types is of greatest value when applied to regional studies in areas dominated by broad scale deformation and limited heterogeneity. The finite planar (S) and linear (L) shape fabrics within a high strain rock may be qualitatively described (see Flinn 1978). The relative dominance of these respective components enables a distinction to be made between fabrics that are foliation dominated (S tectonite), lineation dominated (L tectonite) or a combination of foliation and lineation (S-L tectonite). The pattern shown by the spatial distribution of fabric types are easily represented in a GIS format and are most significant when supported by good data coverage. In the northwestern part of the study area there is a distinct, about 15 km wide, NE-SW trending belt of S-L dominated tectonites which runs along Tunorsuaq fjord and extends to the offshore island to the west (Fig. 3c). Within this belt, L dominated fabrics are also frequently observed whereas S tectonites are extremely rare. In the far NE a predominance of S fabrics with subordinate S-L and L tectonites is

recorded. Thus, the dominant fabric type varies around the regional strike swing from S-L dominated on the western side to S dominated on the eastern side (Fig. 3a, c). In the extreme NW of the study area fabric types are more variable with no pronounced dominance of S or L fabrics. Fabric type also shows some dependency on lithology, e.g. mafic supracrustals on the island of Qerqertarsuatsiaq (Fig. 2) are dominated by S tectonites whereas the surrounding orthogneiss predominantly displays S-L fabrics (Figs 2 & 3c). In addition, S dominated tectonites commonly exhibit lineations that are more east-west trending than surrounding S-L tectonites which contain NE-SW trending lineations (Fig. 3b, c). In the southern part of the study area the fabric type is clearly dominated by S tectonites with subordinate S-L tectonites and extremely rare L dominated fabrics (Fig. 3c). Here, the presence of different fabric types is independent of the trend of foliation and lineations (Fig. 3b, c) i.e. lineation trends and plunge directions do not deviate from the regional pattern in S-L dominated tectonites. The steepness of lineation plunge and foliation dip is also not reflected in fabric types. In addition, fabric types do not vary significantly with lithology (i.e. orthogneisses and metasediments). Overall, in the northern part of the study area fabric types clearly vary with trends in foliation and lineation and also to some extent rock type whereas in the southern part of the area fabric types show no relationship to either lithology or directional structural data. Another characteristic of shape fabric in rocks is the symmetry and asymmetry of the fabric host as shown by mantled porphyroclasts, micafish etc. (Passchier & Trouw 1996). In the study area, symmetrical shape fabrics dominate, whereas asymmetrical fabrics are largely absent and if present they are inconsistent in their sense of asymmetry.

Aeromagnetic dataset The most striking feature seen on the total magnetic field map of the study area is a marked change in the pattern of anomalies from south to north (Fig. 4a). In the south and especially the SE part, the total magnetic field pattern is very irregular with few linear features. It is generally governed by high amplitude anomalies. Towards the north, the magnetic signature exhibits lower amplitudes and pronounced elongate patterns. The elongate anomalies extend up to 40 km and show a general swing from NE-SW trending in the west to more east-west trending in the NE. In addition, distinct linear features are present, two that run in a north-south to NNE-SSW in

APPLICATION OF GIS TO HIGH GRADE ROCKS

the east and west part of the area and two other prominent, parallel features which trend NW-SE. These same linear features are even more pronounced in the vertical gradient map as the vertical gradient sharpens and resolves the edges of anomalies without any directional bias; whereas, the total magnetic field map emphasis the near-surface sources for magnetic anoma-

69

lies. The latter pronounced linear features coincide with a late dolerite dyke in the east, two major fjord systems which from field observations run along late brittle faults trending NNW-SSE and a conjugate fault to the west trending NNE-SSW (compare Fig. 2). The vertical gradient map shows a pronounced 'jump' in pattern from irregular, short wavelength, high

Fig. 4. Selected geophysical data; (a) total magnetic field data with shading from NNW and inclination of 45°; (b) vertical gradient of magnetic signal; physically, the vertical gradient is equal to measuring the magnetic field simultaneously at two points directly above each other, subtracting the two obtained values and then dividing by the vertical distance separating the points of measurements. The side set shows a part of the area; dashed lines highlight the trend of magnetic anomalies. Note the 'fanning' out and 'nucleation' of anomalies next to a jog in the magnetic anomaly to the south, (c) Analysis of the analytical signal of magnetics showing the depth and outline of sources. The analysis is of east-west orientated structures. See Appendix A for further detailed description of analysis techniques.

70

S. PIAZOLO ETAL.

Fig. 4. continued.

amplitude in the south to elongate, regular, banded, longer wavelength and calmer low amplitude pattern to the north. Elongate anomalies show a general swing from NE-SW trending in the west to more east-west trending in the east. More subtle changes in the trend of these elongate anomalies are also observed. In the area around 68°15' and 53° (next to Tunorsuaq fjord) the linear anomalies tighten at the geometrical continuation of a jog of an anomaly in the SE of the respective area. Away from the jog towards the SW and NE the anomalies 'fan out'. Also some 'new' anomalies are present close to the jog (see interpreted subset Fig. 4b). Figure 4c depicts an analysis of the analytical signal of magnetics in which all bodies with the same geometry have the same analytical signal. Here, the depth and outline of east-west orientated structures is shown. The most striking feature is that the swing in the trend of anomalies in the central part of the area (compare Fig. 4a,b)

is mimicked and highlighted by the analysis of the analytical signal. The pattern in the south and to some extent to the very NW is less pronounced, although some short, more elongate anomalies are recorded. The trends shown in Fig. 4c correspond closely to foliation and lineation trends (Fig. 3a & 3b). Overall, the break of patterns as seen and described in both the directional structural data and fabric type data coincide closely to those observed in aeromagnetics.

Combined and integrated datasets Combined directional structural data and shaded total magnetic field map Fold-structures and the general strike of both foliations and lineations are reflected in the shaded total magnetic field (sidesets Fig. 5a, b).

Fig. 5. Combination of directional structural data and total magnetic signature; (a) overview map showing the position of areas that were analysed in detail. Two enlargements on the side depict representative overlays of foliation trend and dip direction and total magnetic field showing a close correspondence in patterns; (b) equal area stereoplots from the shown areas in (a), lineations are shown as filled circles and poles to foliation planes as open squares; in addition, three representative figures are shown. These are generated by the autocorrelation function (ACF) using NIH Image and the FFT macro (Rasband 1996; Panozzo Heilbronner 1992) which emphasize the patterns of the total magnetic filed data of the selected areas, (c) Same analysis technique as used for the analysis of the magnetic signature as described in (b) showing three larger areas marked as a, p, y on (a) which represent and correspond to the three main domains seen in the study area as distinguished in figure (b).

APPLICATION OF GIS TO HIGH GRADE ROCKS

71

72

S. PIAZOLO ETAL.

Even subtle changes in the orientation of magnetic anomalies coincide extremely well with foliation trends (upper sideset Fig. 5a). Patterns depicted in stereoplots showing directional structural i.e. lineation and foliation data from selected areas show distinct patterns and can be separated into three groups which coincide with three spatially defined domains. Group A (areas 1, 2 & 3) is the north to NE part of the area and is characterized by elongate NE-SW trending magnetic anomalies (Fig. 5a), a trend that is mimicked by both lineation and foliation trends (Fig. 5b). Lineations are generally shallow plunging to the NE and SW. Poles to foliation planes form a girdle pattern along a great circle. In general, foliations form two, dispersed maxima that are symmetrical to the great circle and therefore indicate upright, relatively open folding. This feature is most pronounced in area 1. Lineations are co-linear with the fold hinges and as such suggest high strain deformation during which fold hinges become parallel to the lineation. The dispersal of foliations observed in area 3 points to some interference with another, possibly older structural grain. The strong NE-SW trend seen in both datasets is mirrored by the shaded total magnetic field map where the dominant pattern is clearly NE-SW trending and exhibits a strong elongation (Fig. 5b). Group B (areas 4, 5 & 8) shows a more dispersed arrangement both in trend and sharpness of magnetic anomalies, and patterns of directional structural data presented on stereonets. Lineations vary from NE-SW trending in the west and centre (areas 5 & 8) to more east-west trending in the north (area 4) and therefore mimic the general swing in trend observed in the magnetic data. Lineation plunge variation from shallow to steep reflects increased sinuosity or 'porpoising' (for further explanation see above; Fig. 5b). In area 4, poles to foliation form a great circle associated with a single maximum cluster, which is in contrast to the bimodal distribution of Group A. This increasingly clustered pattern can be attributed to upright folding (Group A) becoming increasingly tighter, overturned and SE vergent (Group B). In addition, the patterns may also be a result of the increased sinuosity of fold hinges and associated lineations. The more diffuse pattern is also represented by the autocorrelation image of the magnetic signature that shows a slightly elongate pattern which is significantly less pronounced than the Group A arrangement. Group C (areas 6, 7 & 9) is characterized by an even more diffuse pattern, lineations are generally rare and are moderate to steep plunging; shallow plunges are largely absent. Lin-

eation trends vary from SW-NE to west-east with no systematic correlation in the spatial position of analysed areas. Poles to foliations show one dispersed maximum, and in contrast to Group A & B do not form a great circle. Foliation dips are predominately shallow, with a marked absence of steeper dips as observed in Group A and B. The broad patterns seen in foliation and lineation trends and dips/plunges coincide with a point pattern in the total magnetic field. The autocorrelation image shows only a very vague SE-NW trend. The general character of the three different domains in terms of directional structural data coincides with distinct total magnetic field data patterns (Fig. 5c). Domain A is dominated by elongate magnetic belts which are also seen in the pronounced foliation and lineation trends and near horizontal plunges, domain B shows some linearity which is less pronounced, and which is mimicked by the more diffuse trends of foliation and lineation and lineation plunge steepness, and domain C exhibits a pointed pattern with no strong trend, which coincides with the dispersed foliation and lineation patterns.

Vertical gradient of total magnetic field, foliation trends and metamorphic data combined The pattern of the vertical gradient of the total magnetic field is closely mirrored by the change in metamorphic grade in the area. The area which has been metamorphosed to granulite grade and still exhibits partially preserved granulite facies coincides with a lower magnetic response and higher amplitude signal and more irregular pattern in the vertical gradient in the south (Fig. 6). The north is characterized by genuine amphibolite facies rocks with a generally higher magnetic response and calmer, lower amplitude pattern with elongate 'belts' (Fig. 6). Such a close correspondence of magnetization and metamorphic grade is commonly observed (e.g. Clark 1997). In addition, the elongation of magnetic anomalies is very well mimicked by the representative foliation trends.

Combination of all available datasets: directional structural data, fabric type, metamorphic data and total magnetic field map The close correspondence of patterns seen in different datasets is illustrated in Fig. 7. The

APPLICATION OF CIS TO HIGH GRADE ROCKS

73

geometry of the metamorphic and magnetic pattern 'jump' notably coincides with the swing in foliation trend and trend of elongate magnetic anomalies. In addition, lineations show very similar patterns. Although, the trends in foliation and lineation within the granulite block generally show similar strike swing geometries, they are more irregular and locally inconsistent. Towards the NW the magnetic amplitude increases, the elongation of magnetic anomalies is less pronounced, the orientation of lineations and foliations less accentuated and the fabric type more variable. Fig. 6. The vertical gradient of magnetics with overlay of metamorphic boundary and representative foliation trends.

granulite fades block in the south of the area is dominated by S-type tectonites and a high amplitude, irregular magnetic signature. The metamorphic boundary is closely mimicked by the dominance of S-L tectonites, although at the very NE, S dominated fabrics are present. The

Summary and discussion of observations The combination and integration of the available datasets show that patterns seen within one dataset coincide with patterns observed in the others. Consequently, two main domains are recognized which are separated by a broad boundary zone. The two domains are clearly distinguished by differences in metamorphic grade, magnetic signature, fabric type and structural patterns. The southern block is characterized by the dominance of orthogneisses with granulite f acies

Fig. 7. The total magnetic intensity with overlay of the metamorphic grade, representative spatial distribution of fabric type, foliation trends and lineation trend and direction of plunge.

74

S.PIAZOLOETAL.

Fig. 8. Proposed indentor model and the resulting structural features: increase in sinuosity of whalebacking/porpoising, tightening of folds and change from upright fold axial plane to tilted fold axial plane towards the rigid, granulite indentor; strike swing in foliation, lineation and magnetic anomaly trend mimicking indentor shape; dominance of S tectonite at sites of orthogonal deformation and SL tectonites in oblique deformation area. Note also the shallow plunge along the sides of the block which are due to a significant strike-slip component.

and/or partially retrogressed assemblages, a high amplitude, but irregular magnetic signature and predominately S to S>L tectonites. The foliation is generally moderately dipping and patterns suggest large-scale, sub-isoclinal folding with relatively weak lineations. The broad boundary zone between the southern granulite facies block and a northern amphibolite facies domain is characterized by granulite facies metamorphism that has been partially and locally retrogressed during a later amphibolite grade event. Within this zone, several linear magnetic signatures define a swing in strike from NE-SW trending in the west to more east-west trends further east. They extend over 40-50 km and coincide with discontinuous metasedimentary belts. These correspond to a pronounced preferred orientation of both lineations and foliations and display a well developed S-L fabric. In the central and western part of the area, lineations are generally shallowly plunging. Kilometre-scale reversals in the direction of lineation plunge about culminations and depressions suggest pronounced sinuous 'porpoising' i.e. kilometre-scale reversals of plunge of fold hinges and associated lineations which also follow the regional strike swing. The northern

domain is marked by amphibolite facies metamorphism together with a low amplitude and slightly linear magnetic pattern. This domain displays no clear dominance of a specific fabric type. Patterns of lineations and colinear fold hinges display 'whalebacking' i.e. tens of kilometres scale reversals of gently plunging lineation. Hinge-line vergence (Alsop & Holdsworth 1999) within this northern domain is also of a larger wavelength and shows less deflection around the regional strike swing. Folds are more open and upright. The juxtaposition of amphibolite and granulite facies domains may be achieved in two principal ways. In the first scenario the domains are allochthonous to one another and have been technically juxtaposed during regional orogenesis. The boundary between the domains would then be entirely tectonic. In the second model, the domains are essentially autochthonous and the boundary is largely non-tectonic although it may be reactivated and locally reworked. In this latter case we are therefore viewing an inclined, essentially autochthonous, transitional boundary between granulite and amphibolite facies rocks. Although tectonic interleaving may tend to generate narrow and possibly more discrete

APPLICATION OF GIS TO HIGH GRADE ROCKS

75

Fig. 9. Schematic indentor model with expected fabric type patterns.

contacts, the observed boundary zone between the two domains is relatively broad. In addition, the metamorphic grade should jump sharply across a tectonic boundary, whereas the observed change from granulite to amphibolite facies is gradual and broad in the study area. The boundary zone is parallel to the regional structural fabrics that also display a degree of continuity across the contact. The contact thus appears too gradational in many aspects to be considered purely tectonic and we therefore prefer the scenario of an inclined metamorphic boundary displaying local reworking along and parallel to a pre-existing rheological boundary. The northern boundary to the cooled, granulite facies block forms an arcuate trace which curves from NE-SW in the west to more east-west in the east. The systematic variation of fabric types around this boundary which marks the regional swing cannot be explained by a simple late refold in which early fabric types would remain uniform around the strike swing. In addition, there is no evidence for such refolds at outcrop. The coincidence of the swing in foliation strike/lineation trend and associated fabric type points to a common cause. In addition, the similarity in geometry of the block boundary and strike swing also suggests a generic link. The dominance of S tectonites adjacent to east-west trending segments of the southern, granulite block (Fig. 3c) in contrast to the S-L dominated fabric associated with NE-SW trending segments implies that the shape and orientation of the block boundaries governs variations in the fabric type. Therefore, we suggest that the irregular shaped, cooled, granulite block behaves in a relatively rigid manner compared to the adjacent amphibolite facies domain and thus acts as an indentor. Such an indentor model also readily explains the

observed decreasing fold wavelength and increasing hinge sinuosity towards the rigid, southern block. The shallow plunges of lineations would then point to a significant strike slip component at the 'sides' of the indentor. The 'fanning' of elongate magnetic anomalies next to a jog in the outline of the southern block (Fig. 4b, cf. Fig. 9) can also be attributed to such an indentation. A moderately north to NW dipping block boundary and relative movement of the 'rigid' southern block towards the north to NW explains the change from open, upright folding to tighter, overturned structures closer to the indentor (Fig. 8). According to such an indentor model, the main fabrics and structural patterns in the granulite 'block' were formed during granulite facies conditions before indentation took place whereas those in the northern, amphibolite domain, were created at amphibolite facies conditions during near horizontal indentation of the rigid, granulite block. Such a horizontal movement towards the north to NE is in accordance with the large scale tectonic regime in Palaeoproterozoic times (Van Gool et al 20020). Although, the overall pattern of deformation is controlled by the rheological contrast that is induced by variations in the metamorphic grade, the discontinuous metasedimentary belts partition and localize deformation on a more local scale. The observed lack of consistent shear sense may reflect the highly deformed but generally homogeneous nature of the orthogneisses which is not conducive for shear criteria. The diffuse and distributed style of deformation within these high grade rocks limits pronounced relative motions. Furthermore, a strong flattening component in the region i.e. abundance of S and S-L fabrics may hinder rotation and thereby a consistent sense of vorticity which is necessary

76

S. PIAZOLO ETAL.

for generation of reliable indicators of shear sense.

Conclusion In this case study, a 'block' has been identified in the southern part of the area that exhibits distinctly different geological and geophysical patterns to those developed to the north. The southern block is characterized by a distinct magnetic signal coupled with granulite facies metamorphism and dominant S-type fabrics. The map scale geometry of this block controls the patterns observed within the amphibolite facies domain further north. Foliation and lineation patterns define an arcuate swing in strike about the southern block. In addition, fabric types vary both around the strike swing and also across strike. The dip of foliation and wavelength of folds decrease towards the southern block and the sinuosity of associated lineations/hinges increases. The systematic variation of fabric types around the southern block indicates that the regional swing in strike is not a simple late refold. In contrast, observed patterns must be a result of deformation that controlled both the swing in foliation strike/lineation trend and associated fabric type. As such, an indentor model (Fig. 9) which incorporates a rigid, cooled granulite block in the south bounded to the north by a Theologically weaker amphibolite facies domain most readily explains the geological and geophysical patterns. The shape of the indentor is controlled by the geometry of the metamorphic boundary because at amphibolite facies conditions a cooled granulite is more rigid than amphibolite. Thus, the preserved metamorphic grade governs the rheology of the different, but essentially authochthonous blocks with the amphibolite facies domain being plastered and 'moulded around' a rigid granulite indentor. This study demonstrates that GIS enables the user to compare and/or test individual datasets quickly and easily. In this case study, we have shown that geophysical and geological data correspond closely. Therefore greater confidence can be placed in remote geophysics within areas which lack abundant ground-based data. A GISbased approach has the advantage of being able to combine and integrate multidisciplinary data and thereby recognize patterns in terms of subject, space and scale. It also allows the user to recognize and test the coincidence of patterns in different datasets and to investigate the physical reasons for such a coincidence. Thus the user can identify, focus and target future investigations into the critical key-relationships. Although in this study GIS proved to be of

great value in recognizing diffuse shear zones i.e. broad zones of high strain and thereby interpreting the geological history of a high grade terrain, it is also well suited to other problem areas, provided datasets are collected at comparable resolution and quality. As the degree of detected heterogeneity/homogeneity is a function of scale, pattern recognition is also scale dependent. As such, significant geological patterns are most recognizable at different scales in different geological environments. GIS allows easy access to database investigation and exploration of patterns at different scales and is therefore inherently valuable as an investigative tool in a wide range of geological settings. We thank the international participants and organizers of the geological and geophysical mapping project which was organized and financed by GEUS (Geological Survey of Denmark and Greenland) during the summer of 2001. Special thanks to J. W. Anderson, J. Lautrup and P. Bay for excellent logistic support and H. Myrup and M. Weyhe for their help in the execution of the expedition. Publication of this paper is authorized by the Geological Survey of Denmark and Greenland.

Appendix A: Details on aeromagnetic data and their transformation The total magnetization is the magnetic moment per unit volume. It represents the vector addition of remanent magnetic component that exists irrespective of any ambient external magnetic field and an induced magnetic component caused by the presence of the external magnetic field. The magnetic signature of rocks depend on the type, volume percentage, geometry, alignment and history of magnetic minerals present in the rock. The dominant ferromagnetic phases are members of the solid solution series of magnetite (Fe3O4) and ulvospinel (Fe2TiO4) present in the rock. (Clark & Emerson 1991; Clark 1997; Telford et al. 1998). On large scales, high magnetic areas are commonly related to large masses of igneous rocks and/or crystalline basement. Mafic to intermediate igneous rocks can be moderately to highly magnetic. In contrast, felsic igneous, metamorphic, altered rocks and especially sedimentary rocks have lower magnetic response. Meta-carbonates and graphitic meta-pelites are weakly magnetic, whereas metamorphosed greywackes and graphitic-free pelitic schists may have higher susceptibilities (Grant 1985; Clark & Emerson 1991; Clark 1997; Hildenbrand era/. 2001). Potential field data contains broadband information, where each reading includes the effect

APPLICATION OF GIS TO HIGH GRADE ROCKS of all the physical sources. To enhance and isolate specific sources contribution to the data, a wide range of linear and non-linear filtering and transformation techniques can be applied to the data. A first order technique which helps to emphasize the near-surface sources for anomalies is the shading of total magnetic field map (see

Fig. 4a).

Another direct approach is to calculate the vertical gradient that enhances shallow features, sharpens and resolves the edges of the anomalies, and attenuates the low frequencies relative to high frequencies (Fig. 4b). Physically, the vertical gradient is equal to measuring the magnetic field simultaneously at two points directly above each other, subtracting the two obtained values and then dividing by the vertical distance separating the points of measurements. The vertical derivatives have no directional bias. To obtain information on source parameters such as outline, depth and strike (Fig. 4c) (Nabighian 1972; Thurston & Smith 1997; Smith et al 1998) it is possible to filter the analytical signal mathematically. The analytical signal is a non-measurable parameter calculated from the magnetic derivatives, independent of the direction of the magnetization and the direction of the Earth's field. As a consequence all bodies with the same geometry have the same analytical signal. The position of the source is determined at the maximal analytical signal. The peak of the analytical signal is symmetrical and occurs directly over edges and centres of wide bodies and narrow bodies respectively. These characteristics can be used to highlight the shape and orientation of magnetic sources. Figure 4c shows the result of such an analysis where depth and outline of a source is found for east-west orientated structures.

77

a GIS-linked database to assess the effects of tectonic and climatic changes on mammalian evolution. Abstracts with Programs - Geological Society of America, 32,15. CLARK, D.A. 1997. Magnetic petrophysics and magnetic petrology: aids to geological interpretation of magnetic surveys. AGSO Journal of Australian Geology and Geophysics, 17, 83-103. CLARK, D.A. & EMERSON, D.W. 1991. Notes on rock magnetization characteristics in applied geophysical studies. Exploration Geophysics, 22, 547-555. CONNELLY, IN. & MENGEL EC. 2000. Evolution of the Archaean components of the Nagssutoqidian Orogen, West Greenland. Geological Society of America Bulletin, 112, 747-763. CONNELLY, J.N., VAN GOOL, J.A.M., & MENGEL EC. 2000. Temporal evolution of a deeply eroded orogen: The Nagssugtoqidian orogen, West Greenland. Canadian Journal of Earth Sciences, 37,1121-1142. FLINN,D. 1978. Construction and computation of threedimensional progressive deformations. Journal of the Geological Society, London, 135,291-305. GOODWIN, P.B., CHOINIERE, M.E., HARRIS, EW. & DEAN, B.P. 1996. Improving exploration with geographical information system (GIS) technology. American Association of Petroleum Geologists Bulletin, 80,1297. GRANT, ES. 1985. Aeromagnetics, geology, and ore enviroments, I. Magnetite in igneous, sedimentary and metamorphic rocks: an overview. Geoexploration, 23, 303-333. HANMER, S., MENGEL, E, CONNELLY, J. & VAN GOOL, J.A.M. 1997. Significance of crustal-scale shear zones and synkinematic mafic dykes in the Nagssugtoqidian orogen, SW Greenland: a reexamination. Journal of Structural Geology, 19, 59-75. HARRIS, J.R., WILKINSON, K., HEATHER, K., FUMERTON, S., BERNIER, M.A., AYER, J. & DAMN, R. 2001. Application of GIS processing tecniques for producing mineral prospectivity maps - A case study: Mesothermal Au in the Swayze Greenstone Belt, Ontario, Canada. Natural Resources Research, 10, 91-124. HENDERSON, G. 1969. The Precambrian rocks of the Egedesminde-Christianshab area, West Greenland. Rapport Gr0nlands Geologiske Unders0gelse, 23, 37 pp. HlLDENBRAND, T.G, BERGER, B., JACHENS, R.C. &

References ALSOP, G.I. & HOLDSWORTH, R.E. 1999. Vergence and facing patterns in large-scale sheath folds. Journal of Structural Geology, 21,1335-1349. BONHAM-CARTER, G.E, AGTERBERG, EP. & WRIGHT,

D.E 1990. Weights of evidence modelling: a new approach to mapping mineral potential. Geological Survey of Canada Paper, 89,171-183. BOOKS, CJ. 2000. Defining groundwater system recharge and vulnerability areas in regions of suburban expansion; overview of the northern Illinois example. Abstracts with Programs Geological Society of America, 33, 45. CARRASCO, M.A. & BARNOSKY, A.D. 2000. MIOMAP:

LUNDINGTON, S. 2001. Utility of magnetic and gravity data evaluation regional controls on mineralization: examples from the Western United States. Society of Economic Geologists Reviews, 14, 75-109. KALSBEEK, E & NUTMAN, A.P 1996. Anatomy of the Early Proterozoic Nagssugtoqidian orogen, West Greenland, explored by reconnaissance SHRIMP U-Pb dating. Geology, 24, 515-518. KALSBEEK, E, PIDGEON, R.T. & TAYLOR, P.N. 1987. Nagssugtoqidian mobile belt of West Greenland: cryptic 1850 Ma suture between two Archaean continents - chemical and isotopic evidence. Earth Planetary Science Letters, 85, 365-385. KIRBY, S.H. & KRONENBERG, A.K. 1987. Rheology of

78

S. PIAZOLO £TAL.

the lithosphere: selected topics. Reviews in Geophysics, 25,1219-1244. KNOX-ROBINSON, CM. & WEYBORN, L.A.I. 1997. Towards a holistic exploration strategy: using geographic information systems (GIS) as a tool to enhance exploration. Australian Journal of Earth Sciences, 44, 453-463. KNOX-ROBINSON, CM. & GARDOLL, S.J. 1998. Gisstereoplot: An interactive stereonet plotting module for Arcview 3.0 Geographic information system. Computers and Geosciences, 24, 243-250. MARKER, M., MENGEL, R, VAN GOOL, J.A.M. & FIELD PARTY 1995. Evolution of the Paleoproterozoic Nagssutoqidian Orogen: DLC investigations in West Greenland. Rapport Gr0nlands Geologiske Unders0gelse, 165,100-105. MENGEL, K, VAN GOOL, J.A.M. & KROGSTAD, E. AND THE 1997 FIELD CREW. 1998. Archaean and Palaeoproterozoic erogenic processes: Danish Lithosphere Centre studies of the Nagssugtoqidian orogen, West Greenland. Geology of Greenland Survey Bulletin, 180,100-110. NABIGHIAN, M.N. 1972. The analytical signal of twodimensional magnetic bodies with polygonal cross-section: its properties and use for automated interpretaion. Geophysics, 37, 507-517. NOE-NYGAARD, A. & RAMBERG, H. 1961: Geological reconnaissance map of the country between latitudes 69° N and 63°45 N, West Greenland. Meddelelser om Gr0nland, 123 pp. PANNOZZO HEILBRONNER, R. 1992. The autocorrelation function: an image processing tool for fabric analysis. Tectonophysics, 212, 351-370. PASSCHIER, C.P., DEN BROK, S.W.J., VAN GOOL, J.A.M., MARKER, M. & MANATSCHAL, G. 1998. A laterally constricted shear zone system - the Nordre Str0mfjord steep belt, Nagssugtoqidian orogen, W-Greenland. Terra Nova, 9,190-202. PASSCHIER, C.W. & TROUW, R.A.J. 1996. Microtectonics. Springer, Berlin. PIAZOLO, S. 2002. Overview of the metamorphic evolution of tonalitic gneisses and metasedimentary sequences from the Kangaatsiaq, Lersletten and Sydostbugten area - first comparison to adjacent areas. Danmarks og gr0nlands geologiske unders0gelse rapport 2002/9, Workshop on Nagssugtoqidian and Rinkian geology, West Greenland, 34-35. PIAZOLO, S. & PASSCHIER, C.W. 2002. Controls on lineation development in low to medium grade shear zones: A study from the Cap de Creus peninsula, NE Spain. Journal of Structural Geology, 24, 25-44. RAMBERG, H. 1949. On the petrogenesis of the gneiss complexes between Sukkertoppen and Christinashaab, West Greenland. Meddelelser Danske Geologiske Foreningen, 11, 312-327. RASBAND, W. 1996. NIH Image. National Institute of Health, Research Services Branch NIMH.

RASMUSSEN, T.M. & VAN GOOL, J.A.M. 2000. Aeromagnetic survey in southern West Greenland: project Aeromag 1999. Geological Survey of Denmark and Greenland Bulletin, 186, 73-77. RUTTER, E.H. 1999. On the relationship between the formation of shear zones and the form of the flow law for rocks undergoing dynaic recrystallization. Tectonophysics, 303,147-158. SMITH, R.S., THURSTON, IB., DAI T. & MACLEOD, I.N. 1998. iSPITM - the improved source parameter imaging method. Geophysical Prospecting, 46, 141-151. TELFORD, W.M., GELDART, L.P. & SHERIFF, R.E. 1998. Applied Geophysics, 2nd edn, 770 pp. Cambridge, Cambridge University Press. THURSTON, IB. & SMITH, R.S. 1997. Automatic conversion of magnetic data to depth, dip, and susceptibility contrast using the SPI (TM) method. Geophysics, 62, 807-813. TIKOFF, B., WATERS, C.L. & PRIOR, D.I 2002. Lithological and strain heterogeneity in granulite facies deformation, Arunta block, central Australia. Transport and flow processes in shear zones, The Geological Society, Burlington House, 2-3 September 2002, 85. TREAGUS, S.H. 1988. Strain refraction in layered systems. Journal of Structural Geology, 10, 517-526. TRUE, M.A., SLAWSKI, II & HIBNER, B.A. 1999. Assessment of environmental hazards in western Siberian oil fields using remotely sensed imagery from US and Russian national security systems. Proceedings of the thematic conference on geologic remote sensing, 13,13-20. VAN GOOL, J.A.M. & PIAZOLO, S. 2002. Preservation and recognition of deep crustal shear zones in high grade terranes: examples from the Nagssugtoqidian Orogen of West Greenland. Transport and flow processes in shear zones, The Geological Society, Burlington House, 2-3 September 2002, 86. VAN GOOL, J.A.M., CONNELLY, IN., MARKER, M. & MENGEL, EC 20020. The Nagssugtoqidian Orogen of West Greenland: tectonic evolution and regional correlations from a West Greenland perspective. Canadian Journal of Earth Sciences, 39, 665-686. VAN GOOL, J.A.M., ALSOP, G.I., ARTING, U, GARDE, A.A., KNUDSEN, C, KRAWIEC, A.W., MAZUR, S., NYGAARD, I, PIAZOLO, S., THOMAS, C.W. & THRANE K. 20026. Precambrian geology of the northern Nagssugtoqidian orogen: mapping in the Kangaatsiaq area. Geological Survey of Denmark and Greenland Bulletin, 191. WILSON, C.E., SCOTT, H.D., NORMAN, R.I, SLATON, N.A. & FRIZZELL, D.L. 2000. Spatial distribution of irrigation water quality parameters in Desha County, Arkansas. Abstracts with Programs, Geological Society of America, 32, 45.

Rheology of a two-phase material with applications to partially molten rocks, plastic deformation and saturated soils J. L. VIGNERESSE CREGU, UMR CNRS 7566 G2R, BP 23, F-54501 Vandoeuvre Cedex, France (e-mail: jean-louis. vigneresse@g2r. uhp-nancy.fr) Abstract: A global model is presented to account for the specific rheology of a two-phase material. Examples of observations are taken from a crystallizing magma and these are applied to a partially molten rock, plastic deformation and soil liquefaction. The general behaviour of the viscosity is drawn as a function of the strain rate and the amount of solid phase. It constitutes a 3D diagram developing a cubic surface. The cubic equation is justified by thermodynamic considerations. It results from the mixing of a Newtonian (n = 1) and a power law (n = 3) type of deformation. The diagram shows two types of rheological response. At high strain rate values, the viscosity contrast between the two phases is the lowest. The resulting en masse behaviour is observed during tectonic activity. It manifests itself by homogeneous transport of magma during emplacement and fabric development. An equivalent medium, with average viscosity is a good proxy. Conversely, at low strain rate values, the viscosity contrast between the two phases is the highest. The two end members behave according to their respective rheology. In between, a transitional state develops, in which instability occurs depending on the strain rate and stress conditions. In the 3D diagram it appears as a cusp shape. Rheology presents continuous jumps between the liquid-like and the solid-like rheology. They result in strain localization or phase segregation. The latter preferentially develops during magma crystallization. Deformation under a constant amount of each phase is also possible, resulting in pressure dissolutionlike processes. A bifurcation in the solution plane of the equation of viscous motion causes instability. It is comparable with strain softening. A similar situation should develop when mixing Newtonian and power law rheology, for example during diffusion and dislocation creep, or water-saturated sediment deformation. Owing to the continual jumps between the two types of rheology, hysteresis or memory effect may develop. Rapid cyclic deformation may drive strain to extreme straining. The effect of simple shear seems much more effective than pure shear (compaction) to separate the weak phase from its strong matrix. The development of instabilities and continuous jumps from one rheology to the other lead to discontinuous motion of the weak phase. In a molten region, it corresponds to discontinuous bursts of magma that are extracted.

Two-phase materials and rheology Rheology describes the response of a material to an applied force. Whatever the type of response (non-linear, time dependent), it pertains to continuum mechanics (Ranalli 1995). It assumes continuity, homogeneity and isotropy of the parameters. In a two-phase material, a distinct rheology characterizes each phase, representing a first departure to continuum mechanics. When the two properties are close enough, continuity can still be assumed. However homogeneity and isotropy must be taken into account except when the phases present a regular distribution, or when one phase is much more dominant compared to the other. Most geological objects are composed of several distinct phases. Rocks commonly include a limited number of minerals, each of them presenting a specific chemical, mechanical

or Theological property. A fluid phase, in variable quantity, can also be observed, which adds to the mineral phases, and may reach saturation in sedimentary rocks. Igneous rocks are also a good example of a two-phase material when they form. Partially molten rocks (PMR) represent good candidates for investigating the structural behaviour of such two-phase materials, as are water-saturated sediments or plastic deforming rocks. The rheology of a two-phase material is characterized by non-linear instabilities. The stress-strain curve for porous material is typically non-linear (Wang 2000). In saturated porous material, such as the Berea sandstone, the curve exhibits a shelf reflecting grain crushing and porosity collapse (Zhang et al. 1990). Compaction shear bands may form (Olsson 1999). Initially, they collect the fluid forming the second phase (Du Bernard et al. 2002;

From: ALSOP, G. L, HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 79-94. 0305-8719/$15.00 © The Geological Society of London 2004.

80

J. L. VIGNERESSE

Schmocker et al. 2003), but as the differential stress increases, they lose their porosity (Issen & Rudnicki 2000), expelling the fluid phase. Fluid extraction is discontinuous, like the matrix motion along those structures. The highly discontinuous response to stress appears to be characteristic of two-phase materials. It manifests through the fault valve mechanism (Sibson 1989), episodic fault motion (Heki et al. 1993), and cyclic magma extraction (Barraud et al. 2001; Vigneresse & Burg 2000). However, on a larger scale, deformation seems continuous and regular, as observed during large scale deformation of migmatite massifs (Brun & Martin 1978), homogeneous flow fabric in a sheared magma while flowing (Bouchez 1997) or during soil liquefaction (Ishihara 1993). Two-phase materials incorporate a varying proportion of each phase, each with a characteristic reaction to strain. These two phases are designated as the end-members, for example, in a PMR they would be the melt and its matrix. In a first step, the transition between the two endmembers has been assigned to a binary model in PMR (Arzi 1978). A sudden change in the viscosity marks the limit from a solid-like to a liquid-like behaviour. A coherent bulk behaviour is also observed during experimental deformation of two-phase systems (Burg & Wilson 1987; Ji & Zhao 1993; Handy 1994; Rosenberg & Handy 2001). The two mineral phases, or their analogue equivalent, present contrasting rheologies, but they have the same magnitude in terms of relaxation rate and length. They define a binary model of deformation related either to the load-bearing framework (LBF) structure or to the interconnected weak layer (IWL) structure. Nevertheless, experimental deformation on PMR is restricted to very low melt fractions of less than 12%, because of experimental difficulties (Kohlstedt et al. 2000). They also require fast strain rates, typically of the order of IQr6 to 10"4 s*1, even though experiments last several hours, or days. Composite materials also present two phases, but the rheological contrast between them must be low to guarantee cohesion (Hull & Clyne 2002). A bulk percentage of each phase can be estimated in a PMR. For instance, a 25% melt may be estimated in a migmatite. However, the proportion of the molten phase may vary from 100% in a leucosome to 0% in an unaffected zircon. The proportion varies from place to place, but also with time, owing to the large mobility of the weak phase. As a consequence, the definition of a bulk property, or transport property, is always subject to a scale problem in a two-phase material. Depending on place and time, the local properties cannot be estimated,

thus making local phenomena often unexplained. The rheology of the two phases has a high contrast in a PMR (Petford 2003). The viscosity contrast between the phases commonly ranges from 10 to 15 orders of magnitude (Burg & Vigneresse 2002). All experimental and numerical approaches report instabilities (Brown etal. 1995; Bercovici et al. 2001; Rabinowicz et al. 2001; Vigneresse & Burg 2001). These authors document that a single binary transition is not appropriate, and that a highly unstable intermediate state develops. Two thresholds have been defined that bracket the transitional state in a crystallizing PMR: 1) the rigid percolation threshold (RPT), which defines the onset of connection of the solid phase represented by the forming crystals; and 2) the particle locking threshold (PLT), corresponding to the close packing reached by the crystals (Vigneresse et al 1996). Thus, the rheology of PMR appears to be controlled by the respective amount of each phase. Other factors, such as the fluid content of the magma and melt polymerization (Petford 2003), seem to be restricted to second order effects. Two-phase materials are not restricted to Earth Sciences. They are commonly studied in liquid crystals (nematics), food engineering (pulp-bearing products), and various pastes and slurries (Okagawa et al. 1973; Forest et al. 1997; Fan et al. 1998). Recent developments in those fields have brought a series of new materials (polymers, composite materials, shape memory alloys) for which the reaction to stress is very important in industrial processes. They have resulted in many textbooks and/or journal articles (e.g. Adler et al. 1990; Liu & Masliyah 1996). They describe the various departures from a non-linear response of strain to stress, or thixotropic response (Barnes 1997; 1999). The second point that comes out from these studies is the effect of time. Time is present in the definition of strain rate, i.e. of viscosity, but, surprisingly it is absent in most models concerning the rheology of two-phase mixtures in Earth Sciences (e.g. Arzi 1978). The present paper focuses on the rheology of two-phase material during the transitional state. It incorporates several concepts that have been raised in a series of four papers on specific aspects of the rheology of PMR. Thresholds bracketing a specific behaviour of PMR in static conditions have been defined in Vigneresse et al. (1996). Subsequently, the role of the weak phase in partitioning strain and increasing vorticity, i.e. the non coaxial component of strain, was emphasized in Vigneresse & Tikoff (1999). It leads to the determination of the non-linear

RHEOLOGY OF TWO-PHASE MATERIAL

81

aspects of PMR rheology, with feedback loops that accelerate or damp the response to stress (Burg & Vigneresse 2002). Finally, catastrophic jumps from the rheology of one state to the other have been suggested (Burg & Vigneresse 2002). Here, a conceptual approach to the rheology of a two-phase material is presented. It takes examples from the different responses of PMR and focuses on the strain rate dependence of viscosity in a two-phase mixture. This is applied to water saturated sediments which undergo liquefaction during earthquakes (Ishihara 1993; Di Prisco & Imposimato 2002), and to plastic deformation.

Main principles of rheology The rheology of a rock characterizes its reaction, or strain (e), due to deformation when a stress (a) is applied (Ranalli 1995). This reaction may be elastic, and a linear equation relates the two parameters through an elastic modulus (K):

In the case of shear deformation, a notation relating shear strain (y) to tangential stress (T) is preferred, with a shear modulus (G):

Here, a for stress and £for strain are used indifferently for pure or simple shear. In the elastic domain, the strain returns to its initial conditions when the stress is removed, and strain is independent of time. In contrast, reaction to stress is time dependent for plastic bodies, in which case the strain rate (£) is related to stress by the viscosity (77):

Melts show a linear relationship between stress and strain. A constant viscosity characterizes those Newtonian bodies. Carbonatite magmas present a very low viscosity (5 10~3 to 10~2 Pa s), making them comparable to a liquid (Dobson et al. 1996). In contrast, felsic magmas are more viscous than basalts, with viscosity values in the range 104 to 106 Pa s (Clemens & Petford 1999). A specific linear relationship concerns plastic deformation activated by diffusion creep in crystal aggregates (Nicolas & Poirier 1976). It is a function of temperature (T), through the energy of activation (Q) and the gas perfect constant (R):

Fig. 1. Schematic map of plastic deformation. Tm is the melting temperature, whereas oVG is the ratio of stress to the shear modulus (about 5-50 GPa). The field of partially melted rock (PMR) is indicated by a grey shadow.

The relation may be related to dr2 or d~3, with d representing the grain size, depending on the way diffusion develops. It leads to Coble creep when diffusion develops along the grain boundaries (Coble 1963). In contrast, Nabarro-Herring creep develops through lattice diffusion within the grains (Nabarro 1950; Herring 1950). In both cases, the relation between strain rate and stress is linear. Coble creep develops preferentially at lower temperatures compared to Nabarro-Herring creep (Fig.l). In many other cases, non-linear descriptions are required, with an instantaneous viscosity. It is commonly taken as the local tangent to the curve relating stress to strain rate (Ranalli 1995). Here, it is computed simply as the ratio of stress to strain rate. Departures from the linear curve determine strain hardening when the response is less than the elastic response, which is commonly the case. The hardening coefficient (h doVde) remains positive, but decreases with increasing stress. Strain softening marks an increase in strain while stress decreases. Solid state physics (Green 1998) considers the mechanisms that rule plastic deformation (Fig. 1). In usual conditions, a power law describes the long term behaviour of crustal rocks (Kirby & Kronenberg 1987):

The different parameters, such as the stress exponent (n), the coefficient (^4), and the activation energy (Q), are experimentally determined for each particular rock type. At high temperatures, the stress exponent is close to 3

82

J. L. VIGNERESSE

for amphibolitic rocks (Kirby & Kronenberg 1987). It reflects a deformation accommodated by dislocation creep (Fig. 1). The mechanism develops at high temperatures, and higher stress values compared to diffusion creep. The adopted rheology for a PMR is typically that of melting or crystallization of a granitic magma, for which amphibolites offer a good proxy. The temperature is fixed at 800 °C, corresponding to the biotite breakdown reaction (Patino Douce & Beard 1995). The viscosity variations with temperature are not considered as they are a second order effect. Hence, the melt is Arrhenian, i.e. its viscosity exponentially decreases with temperature, according to the activation energy E:

A typical value for E is about 300 kJ/mole (Maal0e 1985), which is close to that of amphibolites (243 kJ/mole). As a consequence both phases, melt and solid, keep their viscosity contrast constant during the temperature interval of the phase transition. The viscosity of the granitic melt varies with high strain rate (Burg & Vigneresse 2002), but it also occurs at very high strain rate values (10"4-5 s-1), rarely observed in geological situations.

Fig. 2. Stress (a) versus strain rate (£) on a log-log diagram for a crystallizing magma. The Newtonian melt is taken at 106 Pa s. The solid phase follows a power law (n = 3). Stress values range from 0.01 MPa to 100 MPa, whereas strain rates range from 10~20 to 10° s-1. Viscosity (rj) values are indicated in grey. The zone corresponding to magma transport is indicated by a grey box. The zone where instabilities develop (see text) is hatched.

Rheology of a two-phase material The rheology of a two-phase material behaves according to the relative proportion of each phase. The case of a PMR was adopted for modelling purposes. The amount of the strong phase (F) is the intrinsic variable. The two endmembers have their own specific responses. They are progressively altered by the influence of the other phase. A transitional rheology develops which connects each of the two endmembers. This should be examined using a 3D diagram (rj-e-0). Viscosity (77) is used instead of stress (a) as it is better constrained by experimental values. The two end-members. Due to large variations in viscosity, the strain response to stress is plotted on a log-log diagram. For the melt, assumed to be a first order Newtonian equation, Eqn 3 translates into a line equation:

Here, the adopted viscosity value is 106 Pa s (Clemens & Petford 1999). The power law rheology (Eqn 5) of the matrix transforms into a linear plot, with slope n and a constant value (e0)

at the origin of co-ordinates in a log-log diagram:

Experimentally derived values for amphibolites (Kirby & Kronenberg 1987) have been adopted. Values for the parameters in Eqn 5 are n = 3, log A = -4.9 and Q = 243 kJ/mole and a temperature of 800 °C. Adopting these values, the pre-exponential and exponential terms combine, resulting in a constant value (%) of-32 and a slope of 3. In a log-log diagram the rheology of a PMR plots within the two lines that define the rheology of the two end-members (Fig. 2). Their intersection occurs at an unrealistic value of strain rate value (108 s"1). The two lines are separated by the usual range of stress and strain rate values. These range between 10~20 and 1 s"1, thus bracketing a common deformation rate estimated at 10-16 s-1 (Pfiffner & Ramsay 1982). In comparison, magma emplacement is estimated to take place at IQ-^s"1 (Koenders & Petford 2000). The fast deformation rate may be associated with seismic events. Stress values are

RHEOLOGY OF TWO-PHASE MATERIAL

83

bracketed between 0.01 and 100 MPa, thus resulting in logarithmic values of 4 to 8 in Figure 2. The transitional state. Determining the rheology of a PMR requires filling the area between the two end-members, leading to two possibilities. In the first case, extrapolation from the end-members provides a solution, as has been the case for previous experiments (e.g. Arzi 1978). However, the solution is not unique and remains poorly constrained. In the second case, thermodynamic considerations are necessary to provide additional constraints to the adopted solution, which is also controlled by experimental results.

Extrapolating the end-members For a suspension of solid particles in a viscous fluid of viscosity 7]0, the bulk relative viscosity increases with the amount of solid phase. Experiments have been carried out by Einstein (1906) and Roscoe (1952), resulting in a general equation:

The value of the maximum packing value ( resulting in a value of 1.8 (Krieger & Dougherty 1959). The viscosity increase due to the solid suspension is about 3 orders of magnitude when approaching the maximum packing of 0.75 (Vigneresse et al. 1996). This alters the viscosity value, which modifies the slope of the graph (Eqn7). The other end of the diagram corresponds to the solid phase with a low residual melt content. The completely solidified matrix follows a power law (Eqns 5 and 8) that is strain rate dependent. Its response to stress is prohibited until some melt phase is present, for example during initial melting (Vigneresse et al. 1996). Connection starts at Omax, but it extends up to loose packing. In both cases, the limit is a function, not defined at present, of the strain rate. From evidence, low

Fig. 3. Three-dimensional diagram (e-rj-O) showing the cusp shape of viscosity for low strain rate values.

strain rates provide more of a chance for the weak phase to be connected. The two surfaces overlap on a wide range of 0 (0.50-0.75). The connection between the two end-members takes the form of a cusp surface (Fig. 3). The viscosity adopts a cusp shape in a (£-/7-<2>) diagram, as suggested for other metastable situations dealing with the catastrophe theory (Thorn 1990; Petford 1995). However, the construction remains purely geometric, even if the two end members are experimentally controlled.

Thermodynamic considerations At present there is no exact theory that describes the transitional state physically and mathematically. There are partial solutions to determine the interactions between solid particles (Krieger & Dougherty 1959; Batchelor 1976), but they face close packing problems. Eulerian descriptions examine melt segregation under compaction and shear (Rabinowicz & Vigneresse 2004; Spiegelman 2004). They identify non-linear instabilities, the interactions of which concentrate the weakest phase into elongated pockets. Rabinowicz & Vigneresse (2004) and Spiegelman (2004) describe the consequences of the interactions, but these are not adapted to infer the difficulties. The total energy of a two-phase material should be considered. Assuming that pressure and volume are constant, and that temperature varies within a restricted range, the energy is the sum of the product of strain by stress, weighted for the respective amounts of weak and strong phase. The work needed to alter the system is

84

J. L. VIGNERESSE

the derivative of the energy with respect to time. It is the weighted sum of the respective products of strain rate by stress for both phases, and a coupling term:

The coupling term assumes a continuity of strain between the two phases. Equilibrium conditions implies no work variation (dW/£ - 0). Incorporating Eqns 4,5 and 9 into Eqn 10, deriving with respect to strain rate and rearranging for viscosity leads to the following equation for stability:

where p and q are a function of t and <2>. The cubic equation adopts a cusp shape when the term p becomes negative. At present, I am unable to quantify both terms p and especially q exactly. An approximation for the change of sign of p is when the viscosity of the strong phase is the cube of the weak one (7]matrix ~ ^meit3)Significance of the cusp shape. Describing the viscosity variation by a cubic equation during a phase transition is not fortuitous. Viscosity is the transport parameter that relates the two extrinsic variables, stress and strain rate. It also depends on the respective amount of a phase and the variable that describes the phase transition. It corresponds to an equation of state, with specific reference to rheology. The equation of state, describing pressure and volume as a function of temperature, during the liquid-gas transition, is in essence a cubic equation (Landau & Lifchitz 1967; Perez 2001). The cubic equation emphasizes the role of strain rate, i.e. time in rheology. Viscosity is in essence time dependent. This is shown by the strain rate: the derivative of strain with respect to time. Consequently, all mixtures of viscous fluid should incorporate a time dependence, which was not the case in previous models of rheology (e.g. Arzi 1978). The strain rate dependence and the cubic equation clearly delimit two types of behaviour. The first consists of a continuous evolution of the curved surface for high strain rates (Fig. 3). It corresponds to a homogeneous strain in response to stress. The second occurs beyond a critical point corresponding to the cusp. It is the domain in which local instabilities develop. A better illustration of the rheology of PMR is shown in Figure 4. This is simplified from Fig. 3, still refers to (?]-£-cP) and indicates the rheo-

Fig. 4. Three-dimensional diagram (£-r\-O) showing the cusp shape of viscosity for the high strain rate path (homogeneous deformation), and two paths at low strain rates under a common amount of phase and under a common strain rate value. They both show a cyclic path, determining hysteresis, or memory effect.

logical paths. In a (r]-<£) plane, equivalent to a section at constant t in the 3D diagram, the cusp region is shown by sudden changes from a weak phase to a strong phase, or stress hardening (Fig. 4). The region of overlapping branches is metastable. Three viscosity values can be assigned to a given point of fixed <2>. Two are metastable and correspond to the viscosity of the weak and strong phase. The third point, on the connecting branch, is unstable. The metastable region corresponds to a local maximum of the energy of the system, reflecting the fourth order equation of energy. The total energy, proportional to the product a.e, increases during strain hardening. As soon as strain softening develops, corresponding to a decrease of stress with strain, the energy decreases as well.

Rheological responses Owing to the cusp shape of the viscosity surfaces (Figs 3 and 4), a different response to stress is expected, depending on whether it develops at a high or low strain rate, i.e. whether instabilities occur or not. The transition from high to low strain rate values may be compared to a critical state that appears in the liquid-gas state equation (Perez 2001). It corresponds to the change of sign of coefficient p in Eqn 11. For a PMR, it occurs when the two values of viscosity are about in a cubic proportion. The critical point also corresponds to the coincidence between

RHEOLOGY OF TWO-PHASE MATERIAL

Fig. 5. Deformation at high strain rates, or high stress, as it occurs during tectonic activity. The deformation path depends on the average amount of solid, corresponding to a homogeneous, en masse, transport of the PMR.

loose and maximum packing (Fig. 3). Because maximum packing is partly strain dependent, an approximation could be 0.66. This means that the strain rate imposed on the melt prevents loose packing developing. At present, the shortage of data only permits estimates for the corresponding strain value. An approximate value, taken from Fig. 2, brackets the critical strain rate below 10~12 s"1, depending on the ambient stress value. Considering the strain rate during magma emplacement (10~12 s"1), it would place the critical transition below this, which fits with observations. This estimate is valid only for granitic magmas. High strain rate behaviour. High strain rates result from a high stress level applied to PMR. High stress may result from tectonic forces or the relatively fast motion of PMR. Magma emplacement is such a case, since the rate of emplacement, followed by crystallization, is about one order of magnitude faster than the tectonic strain rate (Harris et al. 2000; Ameglio & Vigneresse 2000). Experiments demonstrate that magma emplacement develops at any strain (Koenders & Petford 2000). Strain estimates, recorded by a magma pulse being emplaced to build a granitic pluton, provide the range of strain rate recorded by a flowing magma in comparison with the time of intrusion. A sequence of a crystallizing granitic magma, about 5 km in thickness, laterally displaced by about 15 km, results in a shear value of about 3. Both distances correspond to those observed in granitic plutons (Vigneresse 1995). If developing during the

85

intrusion time, that is about 102-104 years, this will result in a strain rate of 10~9 to 1Q-11 s-1. Faster strain rates are certainly recorded for smaller displacement of a much thinner unit of magma, which would be emplaced in a considerably shorter time before it crystallizes. Owing to the high strain rate recorded by a flowing magma, the viscosity contrast is the smallest between the melt and its matrix (Fig. 5). Because of this small contrast, each phase has a similar time to relax, leading to a bulk response of the PMR, the apparent viscosity of which is an average viscosity, depending on the respective amount of each phase. The strain recorded by a crystallizing magma while settling, appears homogeneous within a granitic pluton, at least at the scale of the emplacement and motion of one specific petrographic facies (Bouchez 1997). Likewise, when a migmatite body is subjected to external tectonic forces, the internal response, such as folds, are concordant with those of the surrounding rocks (Brun & Martin 1978). In such situations, average properties (Handy 1994; Rosenberg & Handy 2001) may be applied to estimate the bulk rheology for a two-phase material. Low strain rate response. At low strain rates, the response of a PMR varies according to the boundary conditions of the system. It depends on whether deformation develops under common stress, common stress rate or common percentage of each phase (Fig. 4). Three types of response with specific boundary conditions develop. Depending on the transition from solid to liquid or from liquid to solid, deformation may develop under common stress or common strain (see discussion). If the system is closed, the amount of each phase is fixed, avoiding large scale motion of one phase.

Response with a fixed amount of each phase This is the simplest case of stress-strain rate relationship and corresponds to a closed system. The volume of each phase is conserved, and strain is limited within the considered volume. A cyclic path develops in the (£-?]-<£) diagram along a fixed <2> value (Fig. 6). It first follows the liquid-like surface, jumping suddenly to the solid-like surface. This is shown by a decrease in the strain rate, i.e. a decrease in the viscosity of the solid-like phase. For instance, at 0.60 crystals, the decrease in strain rate goes to about 10~9 s"1, at which point the strain rate is equivalent to that of the matrix. A jump to solid-like

86

J. L. VIGNERESSE

behaviour allows a return to the initial condition. The proposed path (Fig. 6) points to two problems, which are the closed system and the decrease/increase in strain rate. A closed system is unlikely to develop at the onset of crystallization in a magma because of melt mobility. However, at the end of the crystallization, when the proportion of the solid phase is high, the possibility for the melt to move decreases drastically. This can lead to small scale closed systems (Fig. 6). The condition of varying the strain rate for the solid phase may be compared to pressure dissolution, which would lead to an enhanced stress gradient between touching crystals by dissolving one crystal to the detriment of the other (Grinfeld 1993). Such observations have been described in analogue crystal mushes (Means & Park 1994), leading to contact melting (Park & Means 1996). Natural examples of contact melting show the growth of one plagioclase crystal to the detriment of another in a gabbro (Nicolas & Ildefonse 1996). In both cases of contact melting, evidence exists that crystals were compacted with only a little residual liquid phase. In the case of the Oman gabbros, strain rate estimates are in the range 10~12 s"1 for crystals about 1 mm long, whereas the melt fraction is less than 20% (Nicolas & Ildefonse 1996).

Fig. 6. Deformation at constant amount of phase. It develops during the final stage of crystallization and acts as a closed system. Plagioclase crystals grow to the detriment of crystals that have already formed. Contact melting develops and crystals impinge altogether (inset redrawn from a photo of Oman gabbros, provided by B. Ildefonse (Nicolas & Ildefonse 1996).

Response under a common stress Under a common applied stress, the equivalent stress-strain rate diagram presents strain softening, which corresponds to the change from the solid-like curve to the liquid-like curve in a (cr-£) diagram. Both curves have a different viscosity, i.e. a different slope. The transition between them, or strain softening, is unstable (Bazant 1988; Goddard 2002) and localizes the strain (Fig. 7). Owing to the negative slope of the stress-strain rate curve, the weaker region localizes strain. Shear bands develop with specific orientations with regard to the major stress component orientation (Vermeer & De Boorst 1984). A complete theoretical analysis of shear band formation has been issued using a tensor notation (Rudnicki & Rice 1975), or using crystal dislocation mechanisms (Estrin & Kubin 1995). The constitutive tensor (L) is expressed as a function of the elastic (Young and Poisson coefficients) and plastic (friction and dilatancy angles) parameters. The change of the sign of the eigenvalues of the tensor nLn corresponds to shear bands developing along direction n. It is equivalent to compute the quantity (Rudnicki & Rice 1975; Bazant 1988). In a PMR, shear bands develop when an

Fig. 7. Deformation under common stress at low strain rates, as it prefers to develop during crystallization. Particle segregates, leading to banded features in the magma, shown in inset.

ambient stress is applied, leading to strain partitioning in the weak phase (Vigneresse & Tikoff 1999). In migmatites, shear bands are commonly observed that collect melt. In a crystallizing magma, shear zones develop which disrupt the original fabric (Guineberteau et al. 1987; Pons et al 1995; Smith 1998). They are local shear zones

RHEOLOGY OF TWO-PHASE MATERIAL

87

even melt rich segregation (Barriere 1981; Clarke & Clarke 1998; McCaffrey et al. 1999; Weinberg et al. 2001; Clarke et al. 2002). Schlieren are commonly interpreted as the final conversion of a mafic magma into a felsic magma. Segregation banding is associated with crystal sorting by size. Biotite and fine grained magma segregate on each rim, whereas coarsegrained plagioclase and alkali feldspars are more abundant at the centre of the flow. The occurrence of vertical banding rules out the action of gravity forces in the production of the observed sorting patterns. In migmatites, segregation of the residual biotite on each side of the flowing leucosome produces banding.

Fig. 8. Deformation under common low strain rates. Overlapping of the two curves for the viscosity of the weak and strong phase induces sudden jumps from one rheology to the other. Deformation develops through cycles, inducing hysteresis, i.e. memory effects. In inset, shear localization develops, leading to proto-faults.

developing in a magmatic stage and have been described as proto-faults (Brun et al. 1990). Proto-faults have been described, but no specific data on their respective orientation to the major stress component has been issued. Estimates of strain rate linked to such shear bands provide low values (10~14 s"4) for tectonic loading (Koenders & Petford 2000).

Response under a common strain rate Instability develops under a common strain rate more easily when the load is transferred from the weak to the solid phase (Fig. 8). It results from the lower slope of the weak phase compared to the strong one in a (o-£) diagram. To keep continuity between both phases, the strained material must segregate, leading to particle banding. Phase segregation has been described in flowing liquid crystals and in rigid rod-like particles in a shear flow (Olmsted & Golbart 1992). Commonly, particles become weakly aligned to the general flow motion. Two types of pattern are observed: onion-like and roll-like, which develop parallel to the shear gradient or to the vorticity, in the case of a Couette flow; a rotational shear. They reflect segregation under a common stress or a common strain rate, respectively (Olmsted 1999). Phase separation in a PMR corresponds to small scale features such as schlieren, banding or

Discontinuous motion of both phases. The preceding behaviours reflect the type of response of a two-phase material. The cusp shape also has implications in the way the material responds. Because of the alternating jumps from one rheology to the other, the two-phase material responds discontinuously in time to stress. It alternatively adopts the strong and the weak phase response, with strain partitioning alternating among the two phases. The situation is similar to stick-slip motion in dry (Marone 1998) or wet friction experiments (Geminard et al. 1999). Discontinuous extraction of the weakest phase, i.e. melt, in PMR has already been observed (Barraud et al. 2001; Vigneresse & Burg 2000). It can also be seen in mud volcanoes (Dimitrov 2002), in which the discontinuous release of fluids is controlled by the advancing front of the overlying pressure (Perez-Belzuz et al. 1997; Murton & Briggs 2003). An application for PMR relates to the proverbial 'chicken and egg problem' of the interaction between faults and plutons (Paterson & Schmidt 1999). Considering the two-phase rheology of a fault zone filled with magma, the continuous jumps from one rheology to the other predict alternate motions of the fault and the magma. Penetrative deformation in the magma and motion along the fault are observed in many places. Alternate strain partitioning in both phases rules out the question on whether the pluton or the fault controls the system, both being required.

Implications for other two-phase materials PMR represents a specific case of two-phase material. It manifests by a Newtonian melt and a power law matrix. It has been taken as an example because of the many available situations for which an explanation could be

88

J. L. VIGNERESSE

suggested by two-phase rheology. Many other materials mix two contrasting types of rheology. They could manifest similar types of response to stress. Diffusion/dislocation creep. Plastic deformation commonly develops through atoms moving or defects within crystal aggregates. Diffusion or dislocation creep (Nicolas & Poirier 1976) are such mechanisms. However, dislocation creep also involves other mechanisms, depending on the nature of the dislocations (single dislocation or dislocation loops), or on the nature of the motion: glide or climb controlled. They are described in solid state physics (Christian & Vitek 1970; Allnatt & Lidiard 1987; Green 1998). However, they are not specific to moderate stress commonly registered in rocks (Fig. 1). Strain rate is linear with respect to stress during diffusion creep, whereas it responds to the cube of stress during dislocation creep (Eqns 4 and 5). The energy involved when combining the two types of rheology writes like PMR. An equivalent viscosity can be computed. The condition for equilibrium (Eqn 11) thus also involves a cubic relationship, implying a cusp shape. Low strain rate effects should show different responses (and perhaps microstructures) depending on whether they develop at constant strain rate, constant stress or constant volume. Acoustic wave propagation. Acoustic wave propagation in poro-elastic media is a special case of alternate stress cycles. The general form of the equation of displacement (u) under shear writes, as a ID formulation for simplicity, with p as the density:

The second term incorporates the variation of stress with strain (h = do/de), the local tangent of the stress-strain curve. For strain hardening this term is positive and the equation is hyperbolic. Solutions write as complex exponential functions, i.e. usual trigonometric functions. They describe the propagation of an acoustic wave with velocity:

Loss of hyperbolicity occurs when one eigenvalue becomes null, or changes its sign. Such a situation develops with strain softening (h < 0). The system becomes elliptical, developing instabilities (Schaeffer 1992; Bazant 1988). Instability develops due to a bifurcation jump. The situ-

Fig. 9. Differences between melting and crystallization for the rheology of PMR. The transitional state from the respective end-members favours deformation under a common stress for melting since the system goes from the solid-like to the liquid-like rheology. In contrast, the opposite path that develops during magma crystallization favours deformation under a common strain rate.

ation mimics and exemplifies the formation of shear bands introduced above (Estrin & Kubin 1995; Besuelle 2001). Soil liquefaction. The rheology of a saturated sand pertains to that of two-phase materials, but under elastic conditions. Sand rheology is controlled by the shear modulus that relates stress to strain (Eqn 2). In usual conditions, G for sediments ranges from 0.1 to 0.4 GPa. However, because of the large porosity of the sediments, and particularly in sand, the relation is nonlinear and G is found to increase proportionally to shear stress, or T1/3. The power law relates to grain contacts and obeys the Hertz-Midlin relationship (Midlin 1949; Bachrach et al. 2001). As a consequence, granular soils cannot be considered as purely elastic when submitted to stress for a long period of time. This non-linear behaviour bears a similarity to viscosity that relates strain rate to stress. It corresponds to thixotropy (Barnes 1997) and is due to irreversible gain in soil strength with time at a constant amount of water (Bauer 1999). Liquefaction takes place in saturated soils submitted to earthquakes under undrained conditions (Seed & Lee 1966; Ishihara 1993; Vaid & Sivathayalan 2000). The material dilates or contracts with grain rearrangement under cyclic loading induced by the seismic waves. Under loading, packing normally increases. However, when compaction is prevented by the presence of a fluid within the interstitial pores, pore pressure also increases. The latter also decreases the effective mean pressure. Liquefaction occurs when the effective pore pressure becomes null. The collapse of pore volume and simultaneous increase in pore pressure can lead to vanishing

RHEOLOGY OF TWO-PHASE MATERIAL effective stress (Youd 1972). Nevertheless, irreversible compaction is not likely since saturated sands in regions of frequent earthquakes exhibit liquefaction at successive periods of time.

Importance of shear deformation in a twophase material Deformation occurs as pure shear, simple shear, or a combination of both. Pure shear essentially consists of a convergent (or divergent) component, whereas simple shear contains a rotational component (Jaeger 1969). A mixture of two viscous phases can accommodate both types of deformation without problems. The situation gets more complex when one phase has a very low viscosity, i.e. low compressibility. In that case, the mixture, for instance rocks and water, presents a contrasting reaction to deformation. Deformation of water-impregnated porous rocks produces alternate structures: dilatant bands where the fluid phase segregates and compaction bands in which the porosity is reduced (Antonellini & Ay din 1994). During sediment compaction, the fluid phase is expelled (McKenzie 1984), giving rise to alternate layers of pure compressional deformation, characterized by a reduced porosity (Olsson 1999; Issen & Rudnicki 2000). They have tabular structures that are perpendicular to the major stress component (Antonellini & Aydin 1994; Cashman & Cashman 2000). When a shear component is added to simple compaction, shear bands develop that are orientated ±25° to the major stress component. The dominant shear displacement gradient is accompanied by a porosity reduction and by the formation of a shear zone in which the fluid phase first accumulates and is later expelled. They form Riedel conjugate shear zones (Schmocker et al. 2003). Compacting and dilating shear bands have been predicted theoretically (Besuelle 2001) and experimentally reproduced (Du Bernard et al. 2002). Rudnicki & Rice (1975) and Estrin & Kubin (1995) extended the concept of localized deformation, perpendicular to the maximum compressive stress, as a bifurcation due to strain softening. Hence, the new approach takes into account the effect of the fluid phase. Sealing is a direct consequence of the development of alternate dilatant and compacted bands (Main et al. 2000). A second consequence of shear resides in the long term effect of fault development. During simple shear, associated Riedel fractures are initiated oblique to the plane of shear. They consequently weaken the material and constitute

89

preferential planes for future shear motion when shear is incremented. The situation is completely different during pure compaction. The lack of a rotational component in the strained material produces alternate bands that remain in the same position with regard to the stress pattern. Anisotropy develops within the twophase material, but cannot be reactivated later on. Thus in PMR shear deformation is more effective at segregating the melt than pure compaction. This was obvious after former studies of pure compaction were applied to magmas (McKenzie 1984). They resulted in inappropriate values of compaction length (about 100 m) that are not actually observed in the field. In contrast, when simple shear is added to compaction, the instabilities initiated by both types of deformation cooperate. Introducing values for magmas derived from the continental crust results in a compaction length (centimetre to millimetre) comparable to values that are commonly observed in migmatites (Rabinowicz & Vigneresse 2004). A similar reduction operates for mantle derived rocks (Holtzman et al. 2003). In both cases, the percentage of the fluid phase, through permeability, enters the equation of compaction length.

Discussion Experiments The model of metastable rheology presented in the diagram (?]-£-<£) is reflected in experimental studies. Few experiments document the rheological behaviour of PMR (Kohlstedt et al. 2000; Wirth & Franz 2000), or water impregnated quartz (Schmocker et al. 2003). In the case of olivine, the melt percentage always remains low (<12%) due to experimental difficulties. In the case of dry rocks, creep experiments report coefficients for a power law that are intermediate (l
90

J. L. VIGNERESSE

minerals may not be the best suited material, owing to their high viscosity, analogue materials should be tested. Simultaneously, theoretical and numerical models should address the development of instability. Non-linear effects. The non-linear effects in the rheology of two-phase materials have important consequences on the bulk description of PMR. Non-linear aspects of the melting and crystallization curves with temperature, varying viscosity during the phase transition, and a non-homogeneous spatial distribution of the two phases, have been considered in a preceding paper (Burg & Vigneresse 2002). They develop feedback loops that contribute to the acceleration or dampening of the rheological response during melting and crystallization, respectively. A second aspect of non-linearity is the development of cyclic curves, i.e. hysteretic, or a memory effect, during phase transitions. Hysteresis in stress-strain rate co-ordinates has long been known to develop in plastic material (Prandtl 1928; Brokate & Sprekels 1996). Viscosity variation according to the frequency of loading leads to several rules that are commonly used in soil studies. They are known as the Masing, or Cox-Mertz rules (Doraiswamy et al. 1991; Servais & Manson 1999). Renewed attention to hysteresis related to wave propagation in sediments impregnated with fluids provided a new insight into non-linear elasticity (Guyer et al. 1995; Kadish et al. 1996; Tutuncu et al. 1998). Hysteresis must be considered as a memory effect that delays the return to the initial condition (null strain) when the applied force (stress) is removed. However, hysteresis does not necessarily imply that a full cycle in melting/crystallization must occur to observe memory effects. It can also develop during partial cycling in (a-£) co-ordinates. This is the case when seismic waves pass through a PMR or water-saturated rock. Seismic waves consist of a succession of compressional/extensional small scale deformations with a characteristic frequency. Body waves are commonly in the range 1-50 Hz (Scholz 1990), which means that cycles of deformation alternate in the same range per second. When the material cannot relax, large strains add together, resulting in an amplification of the strain. This is similar to shifting the hysteresis cycle towards increasing strain by a spiral motion. Accumulating and adding strain may result in strain values that seriously overcome the possibility of the material to deform within a single cycle of loading (Fig. 10). Amplification of strain by spiral cycles is known in saturated sediments as cyclic mobility

Fig. 10. Hysteresis during cyclic loading and unloading of a PMR. Repeated loading may bring the hysteresis curve (thin line) towards unexpected high strain by spiralling cycles.

(Manzari & Dafalias 1997; Li & Ming 2000). Superplastic deformation could be re-examined with this in mind. Development of instabilities. The proposed cusp instability describing the rheology of two-phase material (Fig. 3) pertains to the six elementary catastrophes (Thorn 1990). Catastrophe theory has already been applied to Earth Sciences (Henley 1976; Petford 1995). The present approach attempts to quantify the cusp shape, determining a critical point that could separate the homogeneous response from a region where instabilities develop. The cusp corresponds to a third order surface, documented by a cubic equation (Eqn 10). It implies two internal variables (77 and
RHEOLOGY OF TWO-PHASE MATERIAL stiff material. When the plastic component is high, then the separation between the hyperbolic and elliptic solutions partly depends on the ratio of the elastic shear module to the instantaneous shear module (G/G'). Conclusions A global model is suggested which describes the rheology of a two-phase material according to strain rate. This is a first approach, still conceptual at present, which proposes a non-binary transition from a weak rheology to a strong one. It focuses on the interaction between two endmembers with a distinct and contrasting rheology. The model emphasizes the role of a critical strain rate value that separates a region of bulk homogeneous response from a region in which instabilities develop. Under low strain rates, local effects are seen that develop under constant stress, constant strain rate, or constant amount of each phase. Deformation while the amount of each phase is conserved leads to crystal melting. Deformation at a constant strain rate, which occurs during crystallization, leads to the banded segregation of magma and schlieren formation. Deformation at constant stress results in proto-faults and shear bands. This paper started from many discussions with people that resulted in series of papers on the rheology of PMR. Amongst them, P. Barbey, B. Tikoff and J. P. Burg provided valuable field arguments. D. B. Clarke kindly drew my attention to schlieren and banded granodiorites. He provided me with interesting discussions and helped in the editing of the paper. Nick Petford and Ken McCaffrey provided very constructive reviews. They convinced me of the importance of the topic, jointly with the lack of conclusive constitution equations for PMR. References ADLER, P., NADIM, A. & BRENNER, H. 1990. Rheological models of suspensions. Advances in Chemical Engineering, 15,1-67. ALLNATT, A.R. & LIDIARD, A.B. 1987. Statistical theories of atomic transport in crystalline solids. Report on Progress in Physics, 50, 373-472. AMEGLIO, L. & VIGNERESSE, XL. 2000. Onset of rigidity during cooling and crystallization of felsic magma intrusions. EUG10 Strasbourg. Journal of Conference Abstracts, 4, 615. ANTONELLINI, M.A. & AYDIN, A. 1994. Effect of faulting on fluid flow in porous sandstones: petrophysical properties. American Association of Petroleum Geologists Bulletin, 78, 355-377. ARZI, A.A. 1978. Critical phenomena in the rheology of partially melted rocks. Tectonophysics, 44, 173-184.

91

BACHRACH, R., NUR, A. & AGNON, A. 2001. Liquefaction and dynamic poroelasticity in soft sediments. Journal of Geophysical Research, B106, 13515-13526. BARNES, H.A. 1997. Thixotropy - A review. Journal of Non Newtonian Fluid Mechanics, 70,1-33. BARNES, H.A. 1999. The yield stress - a review or 'Ttocvia pel' - everything flows? Journal of NonNewtonian Fluid Mechanics, 81,133-178. BARRAUD, J., GARDIEN, V., ALLEMAND, P. & GRANDJEAN, P. 2001. Analog modelling of melt segregation and migration during deformation. Physics and Chemistry of the Earth, A26, 317-323. BARRIERE,M. 1981. On curved laminae, graded layers, convection currents and dynamic crystal sorting in the Ploumanac'h (Brittany) subalkaline granite. Contributions to Mineralogy and Petrology, 77, 217-224. BATCHELOR, G.K. 1976. Brownian diffusion of particles with hydrodynamic interaction. Journal of Fluid Mechanics, 74,1-25. BAUER, E. 1999. Analysis of shear band bifurcation with a hypoplastic model for a pressure and density sensitive granular material. Mechanics of Materials, 31, 597-609. BAZANT, Z.P. 1988. Softening instability: Part I Localization into a planar band. Journal of Applied Mechanics, 55, 517-522. BERCOVICI, D., RICARD, Y. & SCHUBERT, G. 2001. A two-phase model for compaction and damage. Part 3. Applications to shear localization and plate boundary formation. Journal of Geophysical Research, B106, 8925-8939. BESUELLE, P. 2001. Compacting and dilating shear bands in porous rock: Theoretical and experimental conditions. Journal of Geophysical Research, B107,13435-13442. BOUCHEZ, J.L. 1997. Granite is never isotropic: an introduction to AMS studies of granites. In: BOUCHEZ, J.L., HUTTON, D.H.W. & STEPHENS, W.E. (eds) Granite: from segregation of melt to emplacement fabrics. Kluwer Academic Publisher, Dordrecht, 95-112. BROKATE, M. & SPREKELS, J. 1996. Hysteresis and phase transitions. Applied Mathematics Sciences, 121, Springer Verlag, Heidelberg, 358 pp. BROWN, M., AVERKIN, Y.A., MCLELLAN, E.L. & SAWYER, E. 1995. Melt segregation in migmatites. Journal of Geophysical Research, 100, 15655-15679. BRUN, J.P. & MARTIN, H. 1978. Relations metamorphisme-deformation au cours de 1'evolution dynamique d'un dome migmatitique: le massif de Saint Malo (France). Bulletin de la Societe Geologique de France, 20, 91-101. BRUN, J.P, GAPAIS, D., COGNE, J.P, LEDRU, P. & VIGNERESSE, J.L. 1990. The Flamanville granite (NW France): An unequivocal example of an expanding pluton. Geological Journal, 25, 271-286. BURG, J.P & VIGNERESSE, J.L. 2002. Non-linear feedback loops in the rheology of cooling-crystallising felsic magma and heating-melting felsic rock. In: DRURY, M., DE MEER, S., BRESSER DE, J.H.,

92

J. L. VIGNERESSE

PENNOCK, G. (eds) Deformation mechanisms, rheology and tectonics. Geological Society, London, Special Publication, 200, 275-292. BURG, J.P & WILSON, CXL. 1987. Deformation of twophase systems with contrasting rheology. Tectonophysics, 135,199-205. CASHMAN, S. & CASHMAN, K. 2000. Cataclasis and deformation-band formation in unconsolidated marine terrace sand, Humboldt County, California. Geology, 28,111-114. CHRISTIAN, J.W. & VITEK, V. 1970. Dislocations and stacking faults. Report on Progress in Physics, 33, 307-411. CLARKE, D.B. & CLARKE, G.K.C. 1998. Layered granodiorites at Chebucto Head, South Mountain batholith, Nova Scotia. Journal of Structural Geology, 20,1305-1324. CLARKE, D.B., McCuiSH, K.L., VERNON, R.H., MAKASEV, V. & MILLER, B.V. 2002. The Port Mouton shear zone: intersection of a regional fault with a crystallising granitoid pluton. Lithos, 61,141-159. CLEMENS, ID. & PETFORD, N. 1999. Granitic melt viscosity and silicic magma dynamics in contrasting tectonic settings. Journal of the Geological Society of London, 156,1057-1060. COBLE, R.L. 1963. A model for boundary diffusion controlled creep in polycrystalline materials. Journal of Applied Physics, 34,1679-1682. DE KRUIF, C.G., VAN IERSEL, E.M.F., VRIJ, A. & RUSSEL, W.B. 1986. Hard sphere colloidal dispersions: viscosity as a function of shear rate and volume fraction. Journal of Chemical Physics, 83, 4717-4725. Di PRISCO, C. & IMPOSIMATO, S. 2002. Static liquefaction of a saturated loose sand stratum. International Journal of Solids and Structures, 39, 3523-3541. DIMITROV, L.I. 2002. Mud volcanoes - the most important pathway for degassing deeply buried sediments. Earth Science Reviews, 59, 49-76. DOBSON, D.P., JONES, A.P., RABE, R., SEKINE, T., KURITA, K., TANIGUCHI, T., KONDO, T., KATO, T., SHIMOMURA, O. & URAKAWA, S. 1996. In-situ measurement of viscosity and density of carbonate melts at high pressure. Earth and Planetary Science Letters, 143,207-215. DORAISWAMY, D., MUJUMDAR, A.N., TSAO, I., BERIS,

A.N., DANFORTH, S.C. & METZNER, A.B. 1991. The Cox-Mertz rule extended: a rheological model for concentrated suspensions and other materials with a yield stress. Journal of Rheology, 35, 647-685. Du BERNARD, X., EICHBUL, P. & AYDIN, A. 2002. Dilation bands: A new form of localized failure in granular media. Geophysical Research Letters, 29, 2176, doi: 10.1029/2002GL015666. EINSTEIN, A. 1906. Eine neue Bestimmung der Molekul-dimensionnen. Annales de Physique, 19, 289-306. ESTRIN, Y. & KUBIN, L.P. 1995. Spatial coupling and propagative plastic instabilities. In: MULHAUS, H.E (ed.), Continuum models for materials with micro-structure. Wiley & Sons, New York, 395-450.

FAN, X., PHAN-THIEN, N. & ZHENG, R. 1998. A direct simulation of fibre suspensions. Journal of NonNewtonian Fluid Mechanics, 74,113-135. FOREST, M.G., Qi, W. & BECHTEL, S.E. 1997. 1-D models for thin filaments of liquid-crystalline polymers: Coupling of orientation and flow in the stability of simple solutions. Physica, D99, 527-554. GEMINARD, J.C.,LOSERT,W. & GOLLUB, J.P. 1999. Frictional mechanics of wet granular material. Physical Review, E59, 5881-5890. GODDARD, ID. 2002. Material instability with stress localization. Journal of Non-Newtonian Fluid Mechanics, 102, 251-261. GREEN, D.J. 1998. Introduction to the mechanical properties of ceramics. Cambridge University Press, Cambridge, 348 pp. GRINFELD, M.A. 1993. The stress driven instability in elastic crystals: Mathematical models and physical manifestations. Journal of Nonlinear Science, 3, 35-83. GUINEBERTEAU, B., BOUCHEZ, XL. & VlGNERESSE, J.L.

1987. The Mortagne granite pluton (France) emplaced by pull apart along a shear zone: structural and gravimetric arguments and regional implication. Geological Society of America Bulletin, 99, 763-770. GUYER, R.A., McCALL, K.R. & BOITNOTT, G.N. 1995. Hysteresis, discrete memory, and nonlinear wave propagation in rock: A new paradigm. Physical Review Letters, 74, 3491-3494. HANDY, M.R. 1994. Flow laws for rocks containing two non-linear viscous phases: a phenomenological approach. Journal of Structural Geology, 16, 287-303. HARRIS, N., VANCE, D. & AYRES, M. 2000. From sediment to granites: timescales of anatexis in the upper crust. Chemical Geology, 162,155-167. HEKI, K., FOULGER, G.R., JULIAN, B.R. & JAHN, C.H. 1993. Plate dynamics near divergent boundaries: Geophysical implications of postrifting crustal deformation in NE Iceland. Journal of Geophysical Research, B98,14279-14297. HENLEY, S. 1976. Catastrophe theory models in geology. Mathematical Geology, 8, 649-655. HERRING, C. 1950. Diffusional viscosity of a polycrystalline solid. Journal of Applied Physics, 21, 437-445. HOLTZMAN, B.K., GROEBNER, N.J., ZIMMERMAN, M.E., GINSBERG, S.B. & KOHLSTEDT, D.L. 2003. Deformation-driven melt segregation in partially molten rocks. Geochemistry, Geophysics and Geosystems, 4, doi: 10.1029/2001GC000258. HULL, D. & CLYNE,T.W. 2002. An introduction to composite materials. Cambridge University Press, Cambridge, 342 pp. HUNT, G.W. 1986. Hidden (a)symmetries of elastic and plastic bifurcation. Applied Mechanical Review, 39,1165-1186. ISHIHARA, K. 1993. Liquefaction and flow failure during earthquakes. Geotechnique, 43, 351-415. ISSEN K.A. & RUDNICKI, J.W. 2000. Conditions for compaction bands in porous rocks. Journal of Geophysical Research, B105, 21529-21536.

RHEOLOGY OF TWO-PHASE MATERIAL JAEGER, J.C. 1969. Elasticity, fracture and flow with engineering and geological applications. Methuen, London, 268 pp. Ji, S. & ZHAO, P. 1993. Flow laws of multiphase rocks calculated from experimental data on constituent phases. Tectonophysics, 117,181-187. KADISH, A., JOHNSON, PA. & ZINSZNER, B. 1996. Evaluating hysteresis in earth materials under dynamic loading. Journal of Geophysical Research, B101, 29139-29147. KIRBY, S.H. & KRONENBERG, A.K. 1987. Rheology of the lithosphere: Selected topics. Reviews of Geophysics, 25,1219-1244. KOENDERS, M.A. & PETFORD, N. 2000. Quantitative analysis and scaling of sheared granitic magmas. Geophysical Research Letters, 27,1231-1234. KOHLSTEDT, D.L., BAi, Q, WAND, Z.C. & MEi, S. 2000. Rheology of partially molten rocks. In: BAGDASSAROV, N., LAPORTE, D. & THOMPSON, A.B. (eds) Physics and chemistry of partially molten rocks. Kluwer Academic Press, Dordrecht, 3-28. KRIEGER, I.M. & DOUGHERTY,T.J. 1959. A mechanism for non Newtonian flow in suspensions of rigid spheres. Transactions of the Society of Rheology, 3,137-152. LANDAU, L. & LIFCHITZ, E. 1967. Physique statistique. Mir Editions, Moscow, 584 pp. LEJEUNE,A.M. & RICHET,P 1995. Rheology of crystalbearing silicate melts: an experimental study of high viscosities. Journal of Geophysical Research, B100, 4215-4229. Li, X.S. & MING, H.V. 2000. Unified modeling of flow liquefaction and cyclic mobility. Soil Dynamics and Earthquake Engineering, 19, 363-369. Liu, S. & MASLIYAH, J.H. 1996. Rheology of suspensions. In: SCHRAMM, L.L. (ed.) Suspensions Fundamentals and Applications in Petroleum Industry. Advances in Chemical Series, 251, American Chemical Society, Washington, 107-176. MAAL0E, S. 1985. Principles of Igneous Petrology. Springer-Verlag, Heidelberg, 374 pp. MAIN, I.G., KWON, O., NGWENYA, B.T. & ELPHICK, S.C. 2000. Fault sealing during deformation-band growth in porous sandstone. Geology, 28, 1131-1134. MANZARI, M.T. & DAFALIAS, Y.F. 1997. A critical state two-surface plasticity model for sands. Geotechnique, 47, 255-272. MARONE, C. 1998. Laboratory-derived friction laws and their application to seismic faulting. Annual Reviews in Earth and Planetary Sciences, 26, 643-696. MCCAFFREY, K.J.W, MILLER, C.F., KARLSTROM, K.E. & SIMPSON, C. 1999. Synmagmatic deformation patterns in the Old Woman Mountains, SE California. Journal of Structural Geology, 21,335-349. McKENZiE, D. 1984. The generation and compaction of partially molten rock. Journal of Petrology, 25, 713-765. MEANS, W.D. & PARK, Y. 1994. New experimental approach to understanding igneous textures. Geology, 22, 323-326. MIDLIN, R.D. 1949. Compliance of elastic bodies in

93

contact. Journal of Applied Mechanics, 16, 259-268. MURTON, B.J. & BIGGS, J. 2003. Numerical modelling of mud volcanoes and their flows using constraints from the Gulf of Cadiz. Marine Geology, 195, 223-236. NABARRO, F.R.N. 1950. Steady state diffusional creep. Philosophical Magazine, 16, 231-237. NEEDLEMAN, A. 1979. Non-normality and bifurcation in plane strain tension and compression. Journal of Mechanics and Physics of Solids, 27, 231-254. NICOLAS, A. & ILDEFONSE, B. 1996. Flow mechanism and viscosity in basaltic magma chambers. Geophysical Research Letters, 23, 2013-2016. NICOLAS, A. & POIRIER, J.P 1976. Crystalline plasticity and solid state flow in metamorphic rocks. Wiley, London, 444 pp. OKAGAWA, A., Cox, R.G. & MASON, S.G. 1973. The kinetics of flowing dispersions. VI. Transient orientation and rheological phenomena of rods and discs in shear flow. Journal of Colloid and Interface Science, 45, 303-329. OLMSTED, P.D. & GOLBART, P.M. 1992. Isotropicnematic transition in shear flow. Strain selection, coexistence, phase transitions and critical behavior. Physical Review, A46, 4966-4993. OLMSTED, P.D. 1999. Two-state shear diagrams for complex fluids in shear flow. Europhysics Letters, 48, 339-345. OLSSON, W.A. 1999. Theoretical and experimental investigation of compaction bands in porous rocks. Journal of Geophysical Research, B104, 7219-7228. PARK, Y. & MEANS, W.D. 1996. Direct observation of deformation processes in crystal mushes. Journal of Structural Geology, 18, 847-858. PATERSON, M.S. 1995. A theory of granular flow accommodated by material transfer via an intergranular fluid. Tectonophysics, 245,135-151. PATERSON, S.R. & SCHMIDT, K.L. 1999. Is there a close spatial relationship between faults and plutons? Journal of Structural Geology, 21,1131-1142. PATINO DOUCE, A.E. & BEARD, J.S. 1995. Dehydrationmelting of biotite gneiss and quartz amphibolite from 3 to 15 kbar. Journal of Petrology, 36, 707-738. PEREZ, J.M. 2001. Materiaux non cristallins et science du desordre. Presses Polytechniques et Universitaires Romandes, Lausanne, 558 pp. PEREZ-BELZUZ, F, ALONSO, B. & ERCILLA, G. 1997. History of mud diapirism and trigger mechanisms in the Western Alboran Sea. Tectonophysics, 282, 399-422. PETFORD, N. 1995. Segregation of tonalitic-trondhjemitic melts in the continental crust: the mantle connection. Journal of Geophysical Research, B100,15735-15743. PETFORD, N. 2003. Rheology of granitic magmas during ascent and emplacement. Annual Review in Earth and Planetary Sciences, 31, 399-427. PFIFFNER, O.A., & RAMSAY, J.G. 1982. Constraints on geologic strain rates: arguments from finite strain states of naturally deformed rocks. Journal of Geophysical Research, 87, 311-321.

94

J. L. VIGNERESSE

PONS, J., BARBEY, P., DUPUIS, D. & LEGER, J.M. 1995. Mechanism of pluton emplacement and structural evolution of a 2.1 Ga juvenile continental crust: the Birimian of southwestern Niger. Precambrian Research, 70, 281-301. PRANDTL, L. 1928. Ein Gedankenmodell zur kinetischen Theorie der festen Korper. Zeitschrift fur Angewandte Mathematik und Mechanik, 8, 85-106. RABINOWICZ, M. & VIGNERESSE, J.L. 2004. Melt segregation under compaction and shear channelling: Application to granitic magma segregation in a continental crust. Journal of Geophysical Research, 109 (in press). RABINOWICZ, M., GENTHON, P., CEULENEER, G. & HILLAIRET, M. 2001. Compaction in a mantle mush with high melt concentrations and the generation of magma chambers. Earth and Planetary Science Letters, 188, 313-328. RANALLI, G. 1995. Rheology of the Earth. Chapman & Hall, London, 413 pp. ROGERS, C.D.E, DIJKSTRA,T.A. & SMALLEY,T.A. 1994. Particle packing from an Earthscience viewpoint. Earth-Science Reviews, 36, 59-82. ROSCOE, R. 1952. The viscosity of suspensions of rigid spheres. British Journal of Applied Physics, 3, 267-269. ROSENBERG, C.L. & HANDY, M.R. 2001. Mechanisms and orientation of melt segregation paths during pure shearing of a partially molten rock analogue (norcamphor-benzamide). Journal of Structural Geology, 23,1917-1932. RUDNICKI, J. & RICE, J. 1975. Conditions for the localization of deformation in pressure-sensitive solids. Journal of Mechanics and Physics of Solids, 23, 371-394. SCHAEFFER, D.G 1992. A mathematical model for localization in granular flow. Proceedings of the Royal Society of London, A436, 217-250. SCHMOCKER, M., BYSTRICKY, M., KUNZE, K., BURLINI, L., STUNITZ, H. & BURG, J.P. 2003. Granular flow and Riedel band formation in a water-rich quartz aggregates experimentally deformed in torsion. Journal of Geophysical Research, B108,2242, doi: 10.1029/2002JB001958. SCHOLZ, C.H. 1990. The mechanics of earthquakes and faulting. Cambridge University Press, New York, 439 pp. SEED,H.B. & LEE,K.L. 1966. Liquefaction of saturated sands during cyclic loading. Journal of Soil Mechanics Foundation Division, SM3, 698-708. SERVAIS, C. & MANSON, J.A.E. 1999. The relationship between steady state and oscillatory shear viscosity in planar randomly orientated concentrated fiber suspensions. Journal of Rheology, 43, 1019-1031. SIBSON, R.H. 1989. Earthquake faulting as a structural process. Journal of Structural Geology, 11,1-14.

SMITH, J.V. 1998. Interpretation of domainal groundmass textures in basalt lavas of the Southern Lamington volcanics, eastern Australia. Journal of Geophysical Research, B103, 27383-27391. SPIEGELMAN, M. 2004. Linear analysis of melt band formation by simple shear. Geochemistry, Geophysics and Geosystems (submitted). THOM, R. 1990. Apologie du logos. Hachette, Paris, 672 pp. TUTUNCU, A.N., PODIO, A.L. & SHARMA, M.M. 1998. Nonlinear viscoelastic behavior of sedimentary rocks, Part II: Hysteresis effects and influence of type of fluid on elastic moduli. Geophysics, 63, 195-203. VAID, Y.P. & SIVATHAYALAN, S. 2000. Fundamental factors affecting liquefaction susceptibility of sands. Canadian Geotechnical Journal, 37, 592-606. VERMEER, PA. & DE BOORST, R. 1984. Non-associated plasticity for soils, concrete and rocks. Heron, 29, 1-64. VIGNERESSE, J.L. & BURG, J.P. 2000. Continuous versus discontinuous melt segregation in migmatites: insights from a cellular automaton model. Terra Nova, 12,188-192. VIGNERESSE, J.L. & TIKOFF, B. 1999. Strain partitioning during partial melting and crystallizing felsic magmas. Tectonophysics, 312,117-132. VIGNERESSE, J.L. 1995. Control of granite emplacement by regional deformation. Tectonophysics, 249,173-186. VIGNERESSE, J.L., BARBEY, P. & CUNEY, M. 1996. Rheological transitions during partial melting and crystallization with application to felsic magma segregation and transfer. Journal of Petrology, 37, 1579-1600. WANG, H.E 2000. Theory of Linear Poroelasticity with Applications to Geomechanics and Hydrogeology. Princeton University Press, Princeton, 267 pp. WEINBERG, R.F., SIAL, A.N. & PESSOA, R.R. 2001. Magma flow within the Tavares pluton, northeast Brazil: Compositional and thermal convection. Geological Society of America Bulletin, 113, 508-520. WIRTH, R. & FRANZ, L. 2000. Thin amorphous intergranular layers at mineral interfaces in xenoliths: the early stage of melting. In: BAGDASSAROV, N., LAPORTE, D. & THOMPSON, A.B. (eds) Physics and chemistry of partially molten rocks. Kluwer Academic Press, Dordrecht, 229-268. YOUD, T.L. 1972. Compaction of sand by repeated straining. Journal of Soil Mechanics and Foundations Design, 98, 709-725. ZHANG, J., WONG, T.F. & DAVIS, D.M. 1990. Micromechanics of pressure-induced grain crushing in porous rocks. Journal of Geophysical Research, B95, 341-352.

A comparison of structural data and seismic images for low-angle normal faults in the Northern Apennines (Central Italy): constraints on activity C. COLLETTINI & M. R. BARCHI Dipartimento di Scienze della Terra, Universita di Perugia, Piazza dell'Universita 1, 06100, Perugia, Italia (e-mail: [email protected]) Abstract: During the past 18 Ma extensional tectonism has migrated from the Tyrrhenian sea eastward into the Northern Apennines of Italy. The extension is due in part to lowangle east-dipping normal faults, that are now exhumed in the Tyrrhenian islands and Tuscany, while additional extension is still occurring in the Apennine chain (Umbria region, c. 200 km eastward). This tectonic framework is an example where active extensional processes affecting the Umbria region can be studied in exhumed faults that are no longer active. Here a comparison between the Zuccale Fault (ZF), cropping out in the Isle of Elba, and the Altotiberina Fault (ATF), revealed by geophysical data, seismology and seismic profiles crossing the Umbria region, provide insights into the processes affecting low-angle normal fault development and evolution. Recorded microseismicity suggests that the ATF is presently active under a vertical
The presence of low-angle normal faults, LANFs, (dip < 30°) has been extensively documented in areas of continental extension. LANFs were firstly recognized in the Basin and Range province (Wernicke 1981; Lister & Davis 1989 for a comprehensive review) and then documented in other areas: in Greece (Lister et al 1984; Rigo et al 1996; Sorel 2000; Jolivet 2001); in the East African Rift System (Morley 1999) and the Northern Apennines of Italy, (Carmignani & Kligfield 1990; Storti 1995; Jolivet etal 1998; Pialli etal 1998). Although the LANFs are described in the literature, their origin, evolution and activity is still controversial. First, how do LANFs form? Is their low-angle attitude an original feature or the result of rotation? According to Anderson-Byerlee frictional fault mechanics, normal faults initiate at dips of about 60° then rotate in a domino fashion to frictional lockup angles of 40°-30° (Sibson 1985). Dips lower than the lockup angle would be achieved by domino-rotation produced by successive normal fault sets (Proffett 1977), or isostatic adjustments producing footwall flexure and uplift (Wernicke & Axen 1988). In contrast, some field observations (e.g. LANFs that truncate high angle normal faults, Scott & Lister 1992) and thermal constraints (John & Foster 1993) suggest initiation and movement along

LANFs at dips similar to their present attitude. Very low dips have also been explained as the result of dramatic departures from the Andersonian state of stress induced by severe topography (Abers et al, 1997), high shear stress at the base of the brittle crust (Westaway 1999) and the presence of fluid overpressures (Bradshaw & Zoback 1988). However, Wills & Buck (1997) critically examined the mechanisms of reorientating stress trajectories and they concluded that horizontal shear stress at depth is generally minor and insufficient to deviate <TI trajectories significantly from the vertical through most of the seismogenic crust. Second, assuming that significant displacement can be accommodated by LANFs, what is the slip behaviour (seismic, microseismic, aseismic)? Dip estimates for M > 5.5 continental normal-fault earthquakes, compiled from focal mechanisms with rupture planes unambiguously identified, extend over the range 65° > 5 > 30° with a marked peak at 8 ~ 45° (Jackson & White 1989; Collettini & Sibson 2001). There are no definitive examples of M > 5.5 normal slip earthquakes on faults dipping less than 30°. This result is consistent with active normal faults which follow the Anderson-Byerlee frictional fault mechanics. On the contrary, three possibly low-angle ruptures (10° < S< 30°; 6.0 < Mw < 6.8), though without positive discrimination of nodal

From: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. IW. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 95-112. 0305-8719/$15.00 © The Geological Society of London 2004.

Fig. 1. Crustal-scale cross-section from Elba to the Adriatic coast. The profile shows that part of the extension is accommodated by upper-crust, east-dipping LANFs. The extensional processes in the Tyrrhenian sea and Tuscany has been active enough to favour exhumation accompanied with high heat flow, positive Bouguer gravity anomaly, a shallow Mono, and deep release of CO2. Migration of extension from Corsica to Umbria, as calibrated by the time-space evolution of the syntectonic basins (data from seismic reflection profiles Barchi et al 1998; Pascucci et al 1999). In the active region crustal thinning is continuing. The ATF, located in the active region, is compared with the ZF that crops out in the Isle of Elba. Crustal structure is from the CROP03 seismic profile (Barchi et al. 1998), heat flow (Mongelli & Zito 1991), gravity anomaly (Marson et al 1998), and CO2 release (Chiodini et al. 2000). The position of the Moho is calibrated by deep seismic refraction profiles (Ponziani et al 1995). Brittle/ductile transition (Pauselli & Federico 2002).

ACTIVE LOW-ANGLE NORMAL FAULTS

plane activity, have been attributed to an active LANF in the Papua New Guinea region (Abers 1991; Wernicke 1995). The lack of moderate and large ruptures on LANFs in the contemporary seismic record has been adduced to their long recurrence intervals (Wernicke 1995). The presence of a possibly microseismically active LANF, dip 15°, has been suggested beneath the Gulf of Corinth in Greece (Rigo et al 1996; Rietbrock et al 1996), although Hatzfeld et al (2000) refute this interpretation and suggest that the shallow dipping structure may be the seismicaseismic transition (see also Westaway 2002). As a contribution to the LANFs debate, in this paper we present data for possible LANFs in the Northern Apennines where, during the past 18 Ma, the locus of extensional tectonics has moved from the Tyrrhenian Sea eastward into the Northern Apennines. As a result of this process, ancient low-angle east-dipping normal faults are now exhumed in the western sector (the Tyrrhenian islands and Tuscany), whilst active deformation continues in the eastern sector, i.e. the main Apennine chain. Geophysical data, seismic reflection profiles and seismology from the eastern active region are compared with structural data from the western exhumed region, and both are used to investigate movements on possible low-angle dip (<30°) faults.

Regional setting The Northern Apennines consist of a NE verging thrust-fold belt that formed as the result of the collision between the European continental margin (Sardinia-Corsica block) and the Adriatic microplate (e.g. Alvarez 1972; Reutter et al 1980). Northeastwards directed migration of compression is contemporaneous with hinterland extension. Geological and geophysical data locate the compressional front, presently active close to the Adriatic coast, and the extensional active front near the axial zone of the Northern Apennines (Fig. 1). The presence of adjacent sectors of compression and extension (Lavecchia et al 1994; Mariucci et al 1999), and the migration with time of the compressionextension pair from the Tyrrhenian Sea to the present position, may be explained by the rollback and retreat of a subducting slab (Reutter et al 1980; Doglioni 1991; Royden & Karner 1984; Keller et al 1994). Extensional tectonics in the Northern Apennines have been active enough (-20 Ma) to affect the geophysical properties of the lithosphere (Fig. 1). The extending sector is characterized by a widespread high heat flow, more than 90 mW/m2 (Mongelli & Zito 1991), long-wavelength positive Bouguer gravity

97

anomaly, around 30-40 mGals (Marson et al 1998), a widespread CO2flux, likely to be related to mantle degassing (Chiodini et al 2000), a shallow lithosphere/asthenosphere boundary, about 50 km deep (Suhadolc & Panza 1989), and a shallow Moho, 20-25 km deep (Ponziani et al 1995). The development of extensional deformation within the Northern Apennines is well known (Elter et al 1975), but the asymmetric character of extension has been revealed only recently. The CROP03 deep seismic reflection profile (Pialli et al 1998), which crosses the Northern Apennines from the Tyrrhenian to the Adriatic coast, shows that extension within the brittle upper crust is accommodated by a set of major ENE dipping LANFs (Barchi et al 1998; Decandia et al 1998), bounding syntectonic basins and associated antithetic high-angle normal faults (Fig. 1). Continuous migration of extension from west to east is well constrained by the age of the syntectonic basins (Fig. 1). The extensional process is accomplished by uplift, erosion and exhumation, allowing LANFs to crop out in the Tyrrhenian islands (e.g. Jolivet et al 1998). Within this tectonic setting, the exhumed faults detected in the internal sector may be ancient examples of presently active faults located in the active area corresponding to the axial zone of the Northern Apennines. Consequently, the easternmost LANF, the Altotiberina Fault (ATF), as defined by geophysical data in the active extending area (Umbria region), is compared with the structural attributes of the exhumed Zuccale Fault (ZF) cropping out in the Isle of Elba (Northern Tyrrhenian Sea). These two faults (Fig. 1) are regarded in this study as being different chapters of a common deformation history.

Active versus exhumed LANFs The Altotiberina fault in the Umbria region The Umbria region is an area of SSW-NNE orientated active extension (Montone et al 1999). The strongest instrumental, 5 < M < 6, and historical, intensity < XI, earthquakes of the area are related to a set of NNW-SSE trending, SW dipping normal faults (Barchi et al 2000), which border the intramountain Quaternary basins (Gubbio, Gualdo Tadino, Colfiorito, Castelluccio, Norcia) of the area (Fig. 2). The CROP03 deep seismic reflection profile (Pialli et al 1998), shows that the area is also affected by a crustal-scale east-dipping lowangle normal fault, named the Altotiberina

98

C. COLLETTINI & M. R. BARCHI

Fig. 2. Schematic structural map of the Umbria region. The map is derived from the compilation of available data and original geological surveys. Instrumental seismicity 5.5 < Mw < 6.0:1; September 19 1979, Norcia earthquake (eq.) Mw = 5.8; 2: April 24 1984, Gubbio eq. Mw = 5.6; 3: September 26 1997, 00.33, Colfiorito eq. Mw = 5.7; 4: September 26 1997, 09.40, Colfiorito eq. Mw = 6.0; 5: October 14 1997, Sellano eq. Mw = 5,6. Focal mechanism solutions 1, 3, 4 and 5 CMT; 2 Westaway et al. 1989. Historical seismicity (Boschi et al 1999) with epicentral intensity greater than IX on the Mercalli-Cancani-Sieberg scale (MCS): a: April 30 1279, Camerino eq.; b: December 1 1328, Norcia eq.; c: October 7 1639, Amatrice eq.; d: January 14 1703, Norcia eq.; e: July 27 1751, Gualdo Tadino eq.

ACTIVE LOW-ANGLE NORMAL FAULTS

fault, ATF (Barchi et al 1998; Boncio et al 1998; 2000; Collettini et al. 2000). The surface expression of the ATF is represented by a complex domino-like fault system mapped in the Perugia Mountains and interpreted as synthetic and antithetic splays detached on the ATF (Boncio et al. 1998; 2000). In this area two deep boreholes drilled the ATF close to its break-away zone (sensu Wernicke & Axen 1988): the ATF superposes the Miocene Turbidites (Marnoso Arenacea Fm) above the Triassic Evaporites (Burano Fm) at a depth of a hundred metres. The easternmost splay of the Perugia Mountains fault system borders the continental Tiber basin. In seismic reflection profiles the Upper PlioceneQuaternary fluvio-lacustrine sediments which infill the basin have an asymmetrical shape consistent with the activity of an east-dipping normal fault (Fig. 3). Good quality commercial seismic reflection data (Barchi et al. 1999), produced by oil exploration, are available for the region (see location in Fig. 2). These data have been integrated with the CROP03 profile in order to constrain the geometry of the ATF at depth and its relationship with the SW dipping normal faults. The seismic stratigraphy, calibrated on borehole data (Anelli et al 1994; Barchi et al 1998), is easily recognized within the study area. It consists of three main reflectors (Fig. 4) corresponding to: 1) the Marne a Fucoidi Fm (a marly interval within the Mesozoic-Cenozoic carbonates); 2) the top of the Triassic Evaporites; and 3) the top of the (Carboniferous?) phyllitic basement. The ATF is represented by the eastwarddipping alignment of the reflectors and by the strong interruption of the seismic markers in the fault hanging wall (Fig. 4). The Marne a Fucoidi reflector, well documented in seismic reflection profiles (Fig. 4), crops out in the Perugia Mountains area (Fig. 3). Considering this reflector, the displacement of the easternmost splay of the ATF, bounding the Tiber basin, can be evaluated as being about 2 km (Fig. 3). The Upper Pliocene-Quaternary age of the syntectonic sediments infilling the basin (Ambrosetti et al 1978) allow us to estimate a time average longterm slip rate of about 1 mm yr"1. The total displacement of the ATF is more difficult to address because of the low resolution of the seismic profiles in the fault footwall, and has been estimated in the range of 5-8 km (Boncio et al 1998; Collettini & Barchi 2002). In seismic profiles the ATF appears to be the master fault of the Umbria region, whereas the SW dipping fault system, hereafter termed the Umbria Fault System (UFS), to which the strongest seismicity is related, is antithetic to the ATF and does not

99

appear to cut the fault. The SW dipping Gubbio fault (Guf), one of the faults within the UFS, is also clearly defined on seismic profiles (Fig. 4): the displacement of the fault along the seismic profile of Fig. 4 is about 3 km, as evidenced by the Marne a Fucoidi reflector, which is located at 1.0 s (TWT) in the seismic reflection profile in the fault hanging wall, and which crops out in the fault footwall (see also Collettini et al 2003). The set of seismic reflection profiles constrains the geometry of the ATF over an area of 150km2. The fault has a constant NNW-SSE trend and an average dip of about 20° (Fig. 5d). During May-June 1987, a seismic survey (Fig. 5a) was conducted in the region between Perugia and the Apennine chain (Deschamps et al 1989) where nearly 400 microearthquakes with a local magnitude ranging from 0.6 to 3.0 were recorded. The distribution of the microseismicity plotted on a geological cross section, constructed on a depth-converted seismic reflection profile (Fig. 5b), shows that part of the seismicity fits the ATF trace: the hypocentres gently deepen towards the east from a depth of 6 km below the Tiber basin to about 14 km below the axial zone of the Apennine chain. The other recorded earthquakes nucleate in the ATF hanging wall, whereas only one of the events is located in the ATF footwall: the ATF, therefore, appears to separate an active hanging wall from an almost aseismic footwall. The inversion of the focal mechanisms (Fig. 5c) defines a main population of microseisms (N = 19 = 58%) homogeneously distributed in the area. These events are characterized by a stress tensor possessing a vertical GI, and a <73 orientated SSW-NNE (Boncio et al 2000). a3 is perpendicular to the fault trace and agrees with the mean direction of the regional active extension (Montone etal 1999). Most of these earthquakes have focal mechanisms with nodal planes trending parallel to the ATF (Fig. 5c). Recently, another microseismic network operating for nine months in the same area recorded more than 2000 earthquakes with magnitude 0.0 < M\ < 3.2. The distribution of this microseismicity is again consistent with the geometry of the ATF and defines a stress tensor possessing a vertical o\ and a 03 orientated SSW-NNE (Chiaraluce etal 2003). In our opinion, data available in the Umbria region allow us to make the following interpretations: 1) A presently active ENE dipping LANF (dip about 20°), the ATF, which is characterized by microseismicity; no strong instrumental nor historical earthquakes can be related to the ATF.

Fig. 3. Geological cross-section through the Perugia Mountains (see location in Fig. 2). Data sources: surface structural setting of the Perugia Mountains (Minelli 1992; Brozzetti 1995); stratigraphy of the San Donate borehole (Anelli et al. 1994); subsurface setting derived from the integration of seismic refraction (Ponziani et al 1995) and reflection surveys (see also Fig. 4). Seismic image of the Tiber basin.

Fig. 4. Line drawing of a commercial seismic reflection profile through the study area (see location in Fig. 2). Stratigraphy, not to scale, with main seismic reflectors.

102

C. COLLETTINI & M. R. BARCHI

Fig. 5. (a) Earthquake locations for the study area recorded by a detailed temporary microseismic survey May-June 1987; focal mechanisms constrained by more than eight reliable P-wave first motion polarities. Hypocentral errors less than 2 km in the XY plane and less than 4 km in the Z direction (Deschamps et al 1989; Boncio et al. 2000); (b) Plot of the microseismicity on a geological cross-section obtained from a depthconverted seismic reflection profile: the events plotted possess errors in focal depths less than 1 km and are located within a band with a half-width of 5.0 km. (c) Summary of stress analysis for the main population of the microearthquake focal mechanisms (n - 19) homogeneously distributed within the study area (Boncio et al 1996; 2000). Projected on the Wulff net are: seismic planes, observed (arrows) and theoretical (dots) slip vectors: note some gently eastward dipping planes. Orientation of the stress tensor, possessing vertical Oi, form inversion technique (after Boncio et al. 1996; 2000). Rose diagram constructed for the main population of the microearthquakes. (d) 3D reconstruction of the ATF constructed by using depth converted seismic reflection profiles. 2) An antithetic SW dipping system of normal faults (UFS dips 40°-60°) where the strongest earthquakes of the region occur (Figs. 2 & 4). 3) The ATF and the antithetic UFS represent a conjugate fault system, active under a regional stress field characterized by vertical trajectories of oi.

The Zuccale Fault in the Isle of Elba Since we cannot observe ATF fault rocks deeper than 3 km where the fault superposes the sedimentary cover above a phyllitic basement and where the microseismicity occurs, structural observations for LANFs in the Northern

ACTIVE LOW-ANGLE NORMAL FAULTS

Apennines have been carried out in the Isle of Elba, located in the innermost exhumed sector. The Isle of Elba (Figs 1 & 6) can be divided into three sectors (Keller 1990; Keller & Pialli 1990). The western sector consists of the M. Capanne granodioritic intrusion (about 7 Ma) (Serri et al 1993). The central sector is represented by a thrust stack of five imbricated complexes (from top to bottom: Late Cretaceous flysch, Ligurian ophiolites, Tuscan carbonate sequence, Tuscan metamorphic sequence and basement schists; Trevisan 1950) formed during the east-verging Upper Oligocene-Lower Miocene collisional phase. The stratigraphy of the five complexes is reported schematically in Fig. 6a. The same thrust stack is repeated in the eastern sector. The area dividing these two latter sectors has a low topography. This tectonic framework has been interpreted to be due to movements along an east-dipping LANF called the Zuccale Fault, ZF (Figs 6b & 7), that cuts down-section through the imbricate stack formed during the compressional phase (Keller et al. 1994). The present-day fault dip ranges from 5° to 15° and the eastward down-dip displacement is estimated in the order of at least 7 km from stratigraphic separation (Keller et al. 1994). The ZF brings the Upper Cretaceous Helminthoid flysch of complex V onto the Paleozoic basement schists (i.e. complex I), representing a normal-sense of metamorphic break (Fig. 7). The contact possesses a pervasively foliated fault zone with thicknesses ranging from 1 m to 6 m, characterized by abundant C and C' type shear bands. In this region extensional tectonics started at around 13 Ma (Punta Ala basin, Pascucci et al. 1999) and was active until 4 Ma; small eastdipping low-angle extensional faults cut N-S trending aplitic dykes of Late Miocene age which are related to the emplacement of the Porto Azzurro monzogranite intrusion (5.1 Ma by Rb/Sr and 5.9 ± 0.5 Ma; Saupe et al 1982). The exact degree of exhumation of the fault zone is difficult to determine. However, it can be inferred from two independent lines of evidence: stratigraphic separation and the metamorphic grade of displaced rocks. First, the fault carries the Upper Cretaceous flysch, which rested on the top of the compressional crustal stack, into contact with the basement schists, which were at the bottom of the imbricated crustal stack. Considering the stratigraphy of the five complexes, reported schematically in Fig. 6a, we can evaluate the thickness of the imbricated units originally separating the hanging wall and the footwall rocks, inferring a minimum depth of exhumation of about 3 km. Second, low-grade greenschist metamorphism

103

(2 kbar) experienced by complex II (Duranti et al. 1992) constrains the thickness of the crustal stack at the end of the compressional phase at about 6 km. This corresponds to the maximum depth of exhumation of the footwall rocks. The ZF, exhumed from a depth in the range 3-6 km superposing a sedimentary sequence on the phyllitic basement, is considered to be an analogue of the ATF at depths below 3 km. At depths greater than 3 km the ATF superposes a sedimentary sequence above a phyllitic basement and is characterized by microseismicity (Figs 4 & 5). A detailed microstructural analysis of the fault rocks would provide details of the deformation mechanisms during the fault activity and further constrain the degree of exhumation of the fault. In this study, we focus on the geometry and kinematics of the ZF. Structural data (fault planes, C and C' planes, slickenlines, foliations and extensional fractures) have been collected along the ZF zone (Fig. 8). We have studied the exposure of the fault at Punta di Zuccale, (see location in Fig. 6) where the fault zone crops out along the coast for about 200 m, showing good sections along the fault dip and strike. The fault zone lies within two distinct planar surfaces possessing a gentle dip of less than 15°. In the western portion of the outcrop the fault is westward dipping, whereas in the eastern portion it is eastward dipping (e.g. Fig. 8). The presence of a gently west-dipping portion of the ZF plane is interpreted as bending produced by footwall unloading during exhumation (e.g. Wernicke & Axen 1988). The original, eastdipping attitude of the ZF and the regional importance of the fault can be observed in a seismic profile located 10 km south of the Isle of Elba (Keller & Coward 1996; Fig. 12d). The mean trend of the ZF at the surface is N-S and the distribution of the slickenlines (striae and aligned calcite fibers) gives a N110°E direction of movement, showing kinematics close to pure dip-slip (stereoplots of Fig. 8). The fault zone is characterized by abundant C-type and C'-type shear bands (e.g. Berthe et al. 1979), bounding S foliated domains homogeneously distributed within the fault zone (Fig. 8). The attitudes of the C, C' and S surfaces are shown in Fig. 9: the C surfaces are subhorizontal; the C' surfaces, more abundant in areas of higher deformation, show gentle eastward dips; and the associated S foliation is westward dipping (15° < 8 < 40°). All of these kinematic indicators are consistent with a top-to-the-east sense of shear. Within the fault zone minor high angle normal faults, arranged in conjugate Andersonian systems are locally present, some of them

Fig. 6. (a) Schematic stratigraphy of the five complexes exposed in the Isle of Elba (after Trevisan 1950; Keller 1990). (b) Schematic geological and structural map of the island (after Keller et at. 1994). (c) Geological cross-section showing the geometry of the Zuccale fault (after Keller et al 1994).

ACTIVE LOW-ANGLE NORMAL FAULTS

105

Fig. 7. Outcrop photo of the Zuccale fault. The fault separates Upper Cretaceous flysch in the hanging wall from Palaeozoic basement in the footwall. The pervasively foliated fault zone, characterized by abundant C and C' type shear bands, indicates top-to-the-east movements.

rotated according to a top-to-the-east sense of shear (f &r in Fig. 8). The ZF zone is also characterized by closely spaced extension fractures infilled with fibrous calcite (Figs 8 & 10), showing multiple increments of opening and a median parting demonstrating that the growth occurred as the result of crack-and-seal mechanisms (Fig. 10 a, b). Two main vein systems can be recognized (Fig. lOc): an older system (Vx), which lies parallel to the foliation, cross cut by a younger system (V2). This later system consists of three different sets: two sets are subvertical, one parallel (i.e. E-W trending) and one perpendicular (i.e. N-S trending) to the slip-direction marked by the slickenlines; the third set is subhorizontal and parallel to the base of the fault-related shear zone. The N-S trending veins are locally deflected into west-dipping segments (Fig. 10 a). There is some evidence suggesting that the (V2) vein system was formed during the ZF activity: 1) the westdipping fractures are rotated in a top to-the-east sense of shear, whereas east-dipping veins are almost absent; 2) in a few examples the veins are cross-cut and displaced by C' surfaces (e in Fig. 8). Since some of the veins are observed to cross-

cut each side of the fault, they could be related to the final stage of the ZF activity. Structural data collected within the fault zone suggest movements on a shallow east-dipping normal fault and under vertical GI trajectories. The conjugate Andersonian system within the fault zone is formed under vertical GI and later rotated by further top-to-the-east movements along the ZF. The vertical vein system, which shows incremental growth, is considered to be due to hydraulic fracturing parallel to the a\ direction, hence implying a vertical a\ during the final stage of the fault activity. The (V2) faultboundaries, parallel veins and the (Vj) foliationparallel veins may have developed either as vertical systems rotated by the ZF activity or as the result of high rock permeability in the direction parallel to the fault zone (Faulkner & Rutter 2001).

Discussion According to the interpretation of regional seismic reflection profiles (Barchi et al 1998; Decandia et al 1998) the Neogene extension of the Northern Apennines is driven by

Fig. 8. Cartoon showing the architecture of the ZF and the minor structures distributed within the fault zone. Stereoplots (Schmidt equal area projection lower hemisphere) for the ZF plane and slickenlines.

ACTIVE LOW-ANGLE NORMAL FAULTS

107

Fig. 9. (a) Outcrop photo of the C-type shear bands within the fault zone, (b) Outcrop photo of the C'-type shear bands within the fault zone, (c) Stereoplots (Schmidt equal area projection lower hemisphere) for the C and C' shear bands and the S foliation.

east-dipping LANFs (dips <30°). In this paper an exhumed possible LANF (the ZF) and a possible active LANF (the ATF) have been described using structural and geophysical data respectively. Geophysical data in the active extensional region of the Northern Apennines suggest the presence of an east-dipping LANF (dip 20°), the ATF, that borders the Upper PlioceneQuaternary syntectonic Tiber basin. In seismic profiles the fault appears as a prominent reflector with the SW dipping active normal faults

antithetic to it (the UFS). The ATF separates a seismically active hanging-wall block from an almost aseismic footwall block (e.g. Boncio et al 2000; Collettini et al 2000). The hanging-wall block has been displaced towards the NE since the late Pliocene, and the clustering of the microseismicity along the ATF trace indicates that this process is still active. The stress tensor calculated by the inversion of the focal mechanisms (Fig. 5c) implies movements on the fault under vertical oi trajectories. An exhumed analogue of the ATF has been

108

C. COLLETTINI & M. R. BARCHI

Fig. 10. (a) Outcrop photo of the younger system (V2) of the extensional fractures: vertical system crosscutting each side of the fault e\ westward dipping rotated veins /; fault boundary parallel vein g. (b) Outcrop photo showing the incremental growth of the extensional fractures, (c) Stereoplots (Schmidt equal area projection lower hemisphere) for the Vj & V2 extensional fractures.

studied in the Isle of Elba by means of structural geology. The ZF is exhumed from a depth ranging between 3 and 6 km, comparable with a portion of the ATF that is presently microseismically active. The presence of vertical extensional fractures and Andersonian conjugate systems of normal faults, likely to be contemporaneous to the final stage of the ZF activity, again suggests slip on a low-angle plane and under vertical o\ associated with fluid involvement during faulting. Neither historical nor measured strong earthquakes can clearly be related to the ATF. This is consistent with the absence in recent catalogues of big ruptures on positively discriminated planes at dips less than 30° (Collettini & Sibson 2001), although the possibility of long recurrence intervals for LANFs has to be considered (Wernicke 1995). Instrumental data show that the region containing the interpreted ATF is presently characterized by microseismicity.

Since, according to our study, part of the extensional deformation within the Northern Apennines is accommodated by movements on faults dipping less than 30° in an Andersonian stress regime (vertical GI), these faults have to be weak. But what are the weakening mechanisms? Could they be low friction coefficients, high fluid pressure, or both? Low friction coefficients (^s < 0.2) have been inferred from the analysis of stress orientation data from earthquake focal mechanisms adjacent to the San Andreas fault (Townend & Zoback 2001) and observed in laboratory experiments after microstructural modifications induced by grain-size reduction and foliation development (Bos & Spiers 2001). Nevertheless, data collected worldwide show that Byerlee's (1978) friction coefficients (0.6 < ps < 0.85) are consistent with in situ stress measurements (Brudy et al. 1997), and with dip angles of active faults, which generate big earthquakes (M > 5.5),

ACTIVE LOW-ANGLE NORMAL FAULTS in both compressional (Sibson & Xie 1998) and extensional environments (Collettini & Sibson 2001). Fluid overpressure in the study area might be the cause of the fault weakening mechanism, as also suggested by the presence of the extensional fractures within the ZF. Bearing in mind the difficulty in invoking fluid overpressures (Pf > <73) in extending crust with hydrofractures and fluid loss, central Italy is characterized by extremely vigorous and widespread non volcanic CO2 derived from mantle degassing (Chiodini et al 2000). This degassing activity provides a continuous fluid supply at the base of the brittle crust (e.g. Fig. 1). During their ascent, the CO2-rich fluids are likely to be entrapped as they encounter structural seals (e.g. the ATF or ZF), producing localized fluid overpressure which favours small earthquakes (e.g. Collettini & Barchi 2002). In conclusion, the Northern Apennines provide the opportunity of studying a set of eastdipping LANFs including both old/exhumed and young/active faults. Comparing an exhumed with an active LANF, and using one fault as an analogy for the other, can provide more constraints for understanding the development and evolution of LANFs. We are aware that detailed analysis of the Zuccale fault zone (e.g. Imber et al. 2001 for the Outer Hebrides Fault Zone) will further define the microstructural history of the fault and discriminate the weakening mechanisms and slip behaviour.

109

AMBROSETTI, P., CARBONI, M.G, CONTI, M.A., COSTANTINI, A., ESU, D., GADIN, A., GlROTTI, O.,

LAZZAROTTI, A., MAZZANTI, R., NICOSIA, U., PARISI, G. & SANDRELLI, F. 1978. Evoluzione paleogeografica e tettonica dei bacini tosco-umbrolaziali nel Pliocene e nel Pleistocene inferiore. Memorie della Societd Geologica Italiana, 19, 573-580. ANELLI, L., GORZA, M., FIERI, M. & RIVA, M. 1994. Subsurface well data in the Northern Apennines (Italy). Memorie della Societd Geologica Italiana, 48, 461-471. BARCHI, R.M., MINELLI, G. & PIALLI, G. 1998. The crop 03 profile: a synthesis of results on deep structures of the Northern Apennines. Memorie della Societd Geologica Italiana, 52, 383^00. BARCHI, M.R., PAOLACCI, S., PAUSELLI, C, PIALLI, G. & MERLINI, S. 1999. Geometria delle deformazioni estensionali recenti nel bacino dell'Alta Val Tiberina fra S. Giustino Umbro e Perugia: evidenze geofisiche e considerazioni geologiche. Bollettino Societd Geologica Italiana, 118, 617-625. BARCHI, M.R., GALADINI, F, LAVECCHIA, G, MESSINA, P., MICHETTI, A.M., PERUZZA, L., PIZZI, A.,TONDI, E. & VITTORI, E. 2000. Sintesi sulle conoscenze delle faglie attive in Italia Centrale. Gruppo Nazionale per la Difesa dei Terremoti. 62 pp. BERTHE, D., CHOUKROUNE, P. & JEGOUZO, P. 1979. Orthogneiss, mylonite and non-coaxial deformation of granites: the example of South Armorican shear zone. Journal of Structural Geology, 1, 31-42. BONCIO, P., BROZZETTI, F. & LAVECCHIA, G 1996. State of stress in the Northern Umbria-Marche Apennines (Central Italy): inferences from microearthquake and fault kinematics analyses. Annales Tectonicae, 10, 80-97. BONCIO, P., BROZZETTI, F, PONZIANI, F, BARCHI, M.R., The authors are grateful to G. Pialli, who led the StrucLAVECCHIA, G. & PIALLI, G. 1998. Seismicity and tural Geology Group of Perugia until his death in 1999, extensional tectonics in the Northern Umbriashowed them the outcrop at Zuccale and greatly Marche Apennines. Memorie della Societd Geoencouraged their research on extensional tectonics in logica Italiana, 52, 539-555. the Northern Apennines. We thank D. Cowen, R. E. BONCIO, P., BROZZETTI, F. & LAVECCHIA, G. 2000. Holdsworth and J. V. A. Keller for their useful discusArchitecture and seismotectonics of a regional sions, and Hemin Koyi and Ken McCaffrey for their low-angle normal fault zone in central Italy. useful discussions and for reading an early version of Tectonics, 19,1038-1055. this manuscript. ENI-AGIP made available the Bos, B. & SPIERS, C.J. 2001. Experimental investigation seismic profile shown in Fig. 4. Enrico Tavarnelli and into the microstructural and mechanical evolGerald Roberts provided detailed and helpful reviews. ution of phyllosilicate-bearing fault rock under This work was supported by MIURO1 UR Perugia. conditions favouring pressure solution. Journal of Structural Geology, 23,1187-1202. BOSCHI, E., GUIDOBONI, E., FERRARI, G, VALENSISE, G. References & GASPERINI, P. 1999. Catalogue of Strong Italian ABERS, G.A. 1991. Possible seismogenic shallowEarthquakes from 461 BC to 1990. Annali di dipping normal faults in the Woodlark-D'EntreGeofisica, 42,868 p. http://www.ingrm.it/homita.htm casteaux extensional province, Papua New BRADSHAW, G.A. & ZOBACK, M.D. 1988. Listric Guinea. Geology, 1,1205-1208. normal faulting, stress refraction, and the state of ABERS, G.A., MUTTER C.Z., & FANG, G 1997. Shallow stress in the Gulf Coast basin. Geology, 16, dips of normal faults during rapid extension: 271-274. Earthquakes in the Woodlark-D'Entrecasteaux BROZZETTI, F. 1995. Stile deformativo della tettonica rift system, Papua New Guinea. Journal of Geodistensiva neH'Umbria Occidentale: 1'esempio dei physical Research, 102,15301-15317. Massicci Mesozoici Perugini. Studi Geologici ALVAREZ, W. 1972. Rotation of the Corsica-Sardinia Camerti, 1,105-119. microplate. Nature, 248, 309-314. BRUDY, M., ZOBACK, M.D., FUCHS, K., RUMMEL, F. &

110

C. COLLETTINI & M. R. BARCHI

BAUMGARTNER, J. 1997. Estimation of the complete stress tensor to 8 km depth in the KTB scientific drilling holes; implication for crustal strength. Journal of Geophysical Research, 102, 18453-18475. BYERLEE, J. 1978. Friction of Rocks. Pure and Applied Geophysics, 116, 615-626. CARMIGNANI, L. & KLIGFIELD, R. 1990. Crustal extension in the Northern Apennines: The transition from compression to extension in the Alpi Apuane core complex. Tectonics, 9,1275-1303. CHIARALUCE, L., PICCININI, D. & CHIARABBA, C. 2003. High resolution micro-seismicity data constraining the geometry and kinematic of an active very low-angle normal fault in the Northern Apennine (Central Italy). Geophysical Research Abstract, 5,11930, EGS. CHIODINI, G., FRONDINI, E, CARDELLINI, C, PARELLO, F. & PERUZZI, L. 2000. Rate of diffuse carbon dioxide Earth degassing estimated from carbon balance of regional aquifers: The case of central Apennine, Italy. Journal of Geophysical Research, 105, 8423-8434. COLLETTINI, C. & SIBSON, R.H. 2001. Normal Faults Normal Friction? Geology, 29, 927-930. COLLETTINI, C. & BARCHI, M.R. 2002. A low angle normal fault in the Umbria region (Central Italy): a mechanical model for the related microseismicity. Tectonophysics, 359, 97-115. COLLETTINI, C, BARCHI, M.R., PAUSELLI, C., FEDERICO, C. & PIALLI, G. 2000. Seismic expression of active extensional faults in northern Umbria (central Italy). In: CELLO, G. & TONDI, E. (eds) The Resolution of Geological Analysis and Models for Earthquake Faulting Studies. Journal of Geodynamics, 29, 309-321. COLLETTINI, C., BARCHI, M.R., CHIARALUCE, L., MIRABELLA, F. & Pucci, S. 2003. The Gubbio fault: can different methods give pictures of the same object? Journal of Geodynamics, 36, 51-66. DECANDIA, F. A., LAZZAROTTO, A., LIOTTA, D., CERNOBORI, L. & NICOLICH, R. 1998. The CROP 03 traverse: insights on post-collisional evolution of the Northern Apennines. Memorie della Societd Geologica Italiana, 52, 413-425. DESCHAMPS, A., SCARPA, R. & SELVAGGI, G. 1989. Analisi sismologica del settore settentrionale del1'Appennino umbro-marchigiano. GNGTS 8th Conference, 1, 9-15. DOGLIONI, C. 1991. A proposal of kinematic modelling for W-dipping subductions - Possible applications in the Tyrrhenian-Apennines system. Terra Nova, 3, 423-434. DURANTI, S., PALMERI, R., PERTUSATI, P. C. & RICCI, C. A. 1992. Geological evolution and metamorphic petrology of the basal sequences of eastern Elba (complex II). Acta Vulcanologica, 2, 213-229. ELTER, P., GIGLIA, G, TONGIORGI, M. & TREVISAN, L. 1975. Tensional and contractional areas in recent (Tortonian to Present) evolution of the Northern Apennines. Bollettino di Geofisica Teorica ed Applicata, 17,1975. FAULKNER, D.R. & RUTTER, E.H. 2001. Can the main-

tenance of overpressured fluids in large strike-slip fault zones explain their apparent weakness? Geology, 29, 503-506. HATZFELD, D., KARAKOSTAS, V., ZIAZIA, M., KASSARAS, L, PAPADIMITRIOU, E., MAKROPOULOS, K., VOULGARIS, N. & PAPAIOANNOU, C. 2000. Microseismicity and faulting geometry in the Gulf of Corinth (Greece). Geophysical Journal International, 141, 438-456. IMBER, I, HOLDSWORTH, R.E., BUTLER, C.A. & STRACHAN, R.A. 2001. A reappraisal of the SibsonScholz fault model: The nature of the frictional to viscous ("brittle-ductile") transition along a longlived, crustal-scale fault, Outer Hebrides, Scotland. Tectonics, 20, 601-624. JACKSON, J.A. & WHITE, N.J. 1989. Normal faulting in the upper continental crust: observation from regions of active extension. Journal of Structural Geology, 11,15-36. JOHN, B.E. & FOSTER, D. A. 1993. Structural and thermal constraints on the initiation angle of detachment faulting in the southern Basin and Range: the Chemehuevi Mountains study. Geological Society of America Bulletin, 105,1091-1108. JOLIVET, L. 2001. A comparison of geodetic and finite strain pattern in the Aegean, geodynamic implications. Earth and Planetary Science Letters, 187, 95-104. JOLIVET, L., FACCENNA, C, GOFFE, B., MATTEI, M., ROSSETTI, F, BRUNET, C, STORTI, F, FUNICIELLO, R., CADET, J.P., D'AGOSTINO, N. & PARRA,T. 1998. Midcrustal shear zones in postorogenic extension: Example from the northern Tyrrhenian Sea. Journal of Geophysical Research, 103, 12123-12160. KELLER, J.V.A. 1990. Apennine Compressional Deformation and Tyrrhenian Extension on the Island of Elba, Italy. PhD Thesis, Imperial College, London. KELLER, J.V.A. & COWARD, M.P. 1996. The structure and evolution of the Northern Tyrrhenian Sea. Geological Magazine, 133,1-16. KELLER, J.V.A. & PIALLI, G. 1990. Tectonics of the island of Elba: a reappraisal. Bollettino Societd Geologica Italiana, 109, 413-425. KELLER, J.V.A., MINELLI, G. & PIALLI, G. 1994. Anatomy of late erogenic extension: the Northern Apennines case. Tectonophysics, 238, 275-294. LAVECCHIA, G, BROZZETTI, F, BARCHI, M.R., KELLER, J. & MENICHETTI, M. 1994. Seismotectonic zoning in east-central Italy deduced from the analysis of the Neogene to present deformations and related stress fields. Geological Society of America Bulletin, 106,1107-1120. LISTER, GS. & DAVIS, G.A. 1989. The origin of metamorphic core complexes and detachment faults formed during Tertiary continental extension in the northern Colorado River region, USA. Journal of Structural Geology, 11, 65-93. LISTER, G.S., BANGA, G. & FEENSTRA, A. 1984. Metamporphic core complexes of Cordilleran type in the Cyclades, Aegean Sea, Greece. Geology, 12, 221-225.

111

ACTIVE LOW-ANGLE NORMAL FAULTS MARIUCCI, M.T., AMATO, A. & MONTONE, P. 1999. Recent tectonic evolution and present stress in the Northern Apennines. Tectonics, 18,108-118. MARSON, I., CERNOBORI, L., NICOLICH, R., STOKA, M., LIOTTA, D., PALMIERI, F. & VELICOGNA, I. 1998. CROP-03 profile: a geophysical analysis of data and results. Memorie delta Societd Geologica Italiana, 52,123-137. MINELLI, G. 1992. Tectonic Evolution of the Perugia Massif Area, Northern Apennines (Italy). PhD Thesis, Imperial College, London. MONGELLI, F. & ZITO, G. 1991. Flusso di calore nella regione Toscana. Studi Geologici Camerti, 1, 91-98. MONTONE, P., AMATO, A. & PONDRELLI, S. 1999. Active stress map of Italy. Journal of Geophysical Research, 104, 25595-25610. MORLEY, C.K. 1999. Marked along-strike variations in dip of normal faults-the Lokichar fault, N. Kenya rift: a possible cause for metamorphic core complexes. Journal of Structural Geology, 21, 479^92. PASCUCCI, V., MERLINI, S. & MARTINI, I. P. 1999. Seismic stratigraphy in the Miocene-Pleistocene sedimentary basins of the Northern Tyrrhenian Sea and western Tuscany (Italy). Basin Research, 11, 337-356. PAUSELLI, C. & FEDERICO, C. 2002. The brittle/ductile transition along the Crop03 seismic profile: relationship with the geological features. Bollettino Societd Geologica Italiana, 1, 25-35. PIALLI, G, BARCHI, M.R. & MINELLI, G. 1998. Results of the CROP03 deep seismic reflection profile. Memorie della Societd Geologica Italiana, 52, 657 pp. PONZIANI, F, DE FRANCO, R., MINELLI, G, BIELLA, G, FEDERICO, C. & PIALLI, G. 1995. Crustal shortening and duplication of the Moho in the Northern Apennines: a view from seismic refraction data. Tectonophysics, 252, 391-418. PROFFETT, J.M. 1977. Cenozoic geology of the Yerington district, Nevada, and implications for the nature of Basin and Range faulting. Geological Society of America Bulletin, 88, 247-266. REUTTER, K.J., GIESE, P. & CLOSS, H. 1980. Lithospheric split in the descending plate: observations from the Northern Apennines. Tectonophysics, 64, T1-T9. RlETBROCK, A., TlBERI, C, SCHERBAUM, F. & LYON-

CAEN, H. 1996. Seismic slip on a low-angle normal fault in the Gulf of Corinth: evidence from highresolution cluster analysis of microearthquakes. Geophysical Research Letters, 23,1817-1820. RIGO, A., LYON-CAEN, H., ARMIJO, R., DESCHAMPS, A., HATZFELD, D., MAKROPOULOS, K., PAPADIMITRIOU, P. & KASSARAS, I. 1996. A microseismic study in the western part of the Gulf of Corinth (Greece): implications for large-scale normal faulting mechanisms. Geophysical Journal International, 126, 663-688. ROYDEN, L. & KARNERR, GD. 1984. Flexure of the litosphere beneath Apennine and the Carpathian foredeep basins: evidence for an insufficient topo-

graphic load. American Association Petreleum Geologists Bulettin, 68, 704-712. SAUPE, F, MARIGNAC, C, MOINE, B., SONET, J. & ZIMMERMANN, L. 1982. Datation par les methodes K/Ar et Rb/Sr de quelques roches de la partie orientale de Pile d'Elba (province de Livourne, Italic). Bulletin de Mineralogie, 105, 236-245. SCOTT, R.J. & LISTER, GD. 1992. Detachment faults: Evidence for a low-angle origin. Geology, 20, 833-836. SERRI, G, INNOCENTI, F & MANETTI, P. 1993. Geochemical and petrological evidence of the subduction of delaminated Adriatic continental litosphere in the genesis of the Neogene-Quaternary magmatism of central Italy. Tectonophysics, 223,117-147. SIBSON, R.H. 1985. A note on fault reactivation. Journal of Structural Geology, 1, 751-754. SIBSON, R. H. & XIE, G 1998. Dip Range for Intracontinental Reverse Fault Ruptures: Truth Not Stranger than Friction? Bulletin Seismological Society of America, 88,1014-1022. SOREL, D. 2000. A Pleistocene and still-active detachment fault and the origin of the Corinth-Patras rift, Greece. Geology, 28, 83-86. STORTI, F. 1995. Tectonics of the Punta Bianca promontory: Insights for the evolution of the Northern Apennine-Tyrrhenian Sea basin. Tectonics, 14, 832-847. SUHADOLC, P. & PANZA, GF. 1989. Physical properties of the lithosphere-astenosphere system in Europe from geophysical data. In: BORIANI, A. BONAFEDE, M., PICCARDO, G.B. & VAX, G.B. (eds) The lithosphere in Italy. Advanced Earth Science Researches Accademia Nazionale Of Lincei, Rome, 1989, 80,15-40. TOWNEND J. & ZOBACK M. Z. 2001. Implications of earthquakes focal mechanisms for the frictional strength of the San Andreas fault system. In: HOLDSWORTH,

R.E.,

STRACHAN,

R.A.,

MAGLOUGHLIN, J.F. & KNIPE, R.J. (eds) The Nature and Tectonic Significance of Fault Zone Weakening. Geological Society, London, Special Publications, 186,157-170. TREVISAN, L. 1950. Memorie dellTstituto Geologico dell'Universitd di Padova, 16,1-39. WERNICKE, B.P. 1981. Low-angle normal faults in the Basin and Range Province: nappe tectonics in an extending orogene. Nature, 291, 645-648. WERNICKE, B. 1995. Low-angle normal faults and seismicity: A review. Journal of Geophysical Research, 100, 20159-20174. WERNICKE, B. & AXEN, GJ. 1988. On the role of isostasy in the evolution of normal fault systems. Geology, 16, 848-851. WESTAWAY, R. 1999. The mechanical feasibility of lowangle normal faulting. Tectonophysics, 308, 407-443. WESTAWAY, R. 2002. The Quaternary evolution of the Gulf of Corinth, central Greece: coupling between surface processes and flow in the lower continental crust. Tectonophysics, 348, 269-318.

112

C. COLLETTINI & M. R. BARCHI

WESTAWAY, R., GAWTHORPE, R. & Tozzi, M. 1989. Seismological and field-observations of the 1984 Lazio-Abruzzo earthquakes: implications for the active tectonics of Italy. Geophysical Journal international 98, 489-514.

WILLS, S. & BUCK, W. R. 1997. Stress-field rotation and rooted detachment faults: a Coulomb failure analisys. Journal of Geophysical Research, 102, 20503-20514.

Shear deformation of pelitic rocks in a large-scale natural fault EMILIO CASCIELLO1, M. CESARANO1 & J. W. COSGROVE2 l Dipartimento S. T.A. T., Universita degli Studi del Molise, Via Mazzini 8, hernia, 86170, Italy 2 Department of Earth Science and Engineering, Royal School of Mines, Imperial College, Prince Consort Road, London, SW72BP, UK Abstract: Experimental tests on simulated clay gouges and data from shear zones developed in pelitic media at convergent plate margins provide contrasting evidence regarding the hydraulic characteristics and, in consequence, the frictional properties of sheared clays. The natural shear zone analysed in this work indicates that shear strain can induce mineralogical changes in smectite-bearing sediments that imply loss of water from the smectite minerals and their replacement with anhydrous illite minerals. The extreme localization of the illitization process and its geometric characteristics allow us to argue that the reaction is initiated by stress concentration along the shear zone and, once discrete shears develop, it is accelerated by both cataclasis and the frictional dehydration of smectites. This process would generate fresh water from within the shear zone, leading to fluid overpressure build up, and can account for the observed hydraulic circulation and salinity anomalies in modern accretionary prisms.

Shear deformation in clay-bearing materials produces planar zones of deformation that can display properties quite different from those characterizing the surrounding undeformed rock mass (Hicher et al. 1994; Dewhurst et al. 1996; Takizawa & Ogawa 1999; Saffer et al 2001). Clay minerals undergo mineral reactions and compositional changes when subjected to lithostatic pressures and/or temperature increases within sedimentary basins (Pollastro 1985; Deng et al. 1996; Masuda et al. 2001). In addition, when clays are sheared there is a reduction in their volume and permeability and they develop a permeability anisotropy along the shear zone (Arch & Maltman 1990; Hicher et al. 1994; Zhang et al. 1999). These observations suggest that clay gouges possess characteristics that can greatly affect the dynamics of faulting, and whose understanding is paramount for a comprehensive evaluation of faulting processes in clay-bearing rocks of the upper crust. Our knowledge of the textural characteristics and the physical properties of clay gouge derives mainly from laboratory experiments using natural or synthetic clays deformed under controlled environmental conditions. By comparison relatively few studies (i.e. Rutter et al. 1986; Vrolijk & Van der Pluijm 1999) have investigated natural shear-zones developed in argillaceous media, the fabrics they possess and the mineral composition of the clay gouges. The study described in this paper does not aim to provide an account of the geometry and fabric elements that characterize shear zones. Rather it

is an attempt to investigate a natural phenomenon whose specific characteristics (i.e. microfabric development, mineral compositions, hydraulic properties) have been analysed independently by means of experimental tests, but whose interplay is not clear in the natural context. The shear zone analysed, part of what is known as the Scorciabuoi fault, is the result of left-lateral deformation of Upper PlioceneLower Pleistocene mudstones on a regional scale fault of kilometric displacement. The lack of significant deformation prior to the faulting process, the good exposure conditions and a relatively simple kinematic history make the Scorciabuoi fault an ideal natural laboratory for the study of shear deformation in pelitic media.

Geological framework of the Scorciabuoi fault The Scorciabuoi fault (SBF) is a major tectonic discontinuity that cuts obliquely across ENE verging thrust sheets and their sedimentary cover in the frontal portion of the Southern Apennine belt (Fig.l). This fault is part of system of deep, sub-vertical discontinuities, characterized by a WNW-ESE trend and left lateral displacement, that affected peninsular Italy during the Late Pliocene-Pleistocene (Knott & Turco 1990; Monaco et al. 1998; Cello & Mazzoli 1999). This was a period when active spreading of the Tyrrhenian Basin accompanied the strong arcuation that the Southern

From: ALSOP, G. L, HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224,113-125. 0305-8719/$15.00 © The Geological Society of London 2004.

114

E. CASCIELLO ETAL.

Fig. 1. Geological map of the eastern foothills of the Southern Apennines belt and location of the analysed Scorciabuoi fault (SBF).

Apennines took on (Lonergan & White 1997; Kastens et al. 1988), and in which the WNW-ESE faults had a potentially significant role (Fig. 2).

The Scorciabuoi discontinuity offsets Miocene deposits with a left transtensional dislocation and also affects the almost coeval Middle Pliocene-Middle Pleistocene marine sediments

SHEARING IN A LARGE-SCALE NATURAL FAULT

Fig. 2. Sketch illustrating the role of WNW-ESE trending faults, including the SBF, in generating the arcuate shape of the Southern Apennines belt.

of the S. Arcangelo basin (Figs 1-3). Coeval with the strike-slip activity on the Scorciabuoi fault was the uplift of the Valsinni ridge (Fig. 1), which separated the S. Arcangelo basin from the

115

Pleistocene foredeep and in which the fault seems to terminate. The link between the two structures, suggested by the timing and the geometry, is further indicated by the amount of displacement that characterizes the two structures. Early deposits of the Late Pliocene 'thrusttop' basin have recorded approximately 6 km horizontal displacement, estimated by calculating the dislocation of a reference horizon along the measured slip lines. This is in good agreement with the 6 km shortening calculated for the Valsinni ridge using section balancing techniques (Patacca & Scandone 1999). These data are compatible with the interpretation of the SBF as an oblique lateral ramp to the thrust that defines the Valsinni block. The fault geometry within the PliocenePleistocene sedimentary cover shows, a gradual increase and reversal of the fault dip and a concomitant change in the slip vector when traced along strike (Fig. 4). At the surface the fault appears as a 2-10 m wide band (shear zone) of dark, intensely strained claystones. These have been analysed using a variety of techniques including field based meso-structural analysis, SEM study, X-ray diffraction and textural analysis.

Fig. 3. Simplified geological map showing the distribution of the Pliocene-Pleistocene smectite-bearing sediments deformed by the SBF; location map in Fig. 1.

116

E. CASCIELLO ETAL.

Fig. 4. Fault dip and lineation pitch angles of the SBF and their variation along the fault trace. Refer to Fig. 3 for location of measure sites.

Compositional attributes of the sheared clays The most apparent characteristic of the SBF, which makes it readily recognizable in the field, is its darker colouring with respect to the undeformed light grey mudstones of the footwall and hanging wall (Fig. 5). Thermogravimetric analyses indicate that this feature, which is common to other shear zones in the same sedimentary basin, cannot be attributed to organic matter staining. However, detailed X-ray diffraction (XRD) analyses of the fault zone mudstones reveal that the gouge mineral content is different to that of the neighbouring undeformed wall rock. The clay fraction of the mudstone consists of smectite, illite, chlorite and kaolinite and the non-clay component of quartz, calcite and feldspars. The mudstones in both the footwall and the hanging wall show mixed-layer illite/smectite clay minerals whereas the dark sheared mudstones contain

only illite (Fig. 6). This finding confirms similar results obtained by Vrolijk & Van der Pluijm (1999) who analysed the clay gouge composition of several large faults in the Canadian Rockies and reported a dramatic increase of the illite percentage in illite/smectite assemblages within metres from the faults. What is surprising in the SBF shear zone, is that the illitization process is found to be extremely localized and its characteristics are such that it can be directly related to the deformation process, rather than to environmental conditions (i.e. temperature, fluid chemistry). In order to confirm the correlation between the darkening of the mudstones and the process of illitization, XRD analyses where carried out on samples extracted from saw-cut hand specimens showing domains of darker and lighter colour. The results confirm that even at the subcentimetre scale the darker clays correspond to domains of illitization and smectite disappearance. This observation is crucial in trying to understand the causes of the illite/smectite reaction and the possible effects of this mineral reaction on the processes of faulting.

The smectite/illite transformation Smectites are hydrated clay minerals that can contain up to three layers of water molecules between alluminosilicate layers (Fig. 7), and are well known for their swelling properties. The hydration state, measured as the basal spacing d of theo alluminosilicate layers (i.e. 12.5 A; 15.5 A; 18.5 A for the I, II and III layer of water, respectively), is a function of temperature, pressure and fluid chemistry (Eberl & Hower 1977; Howard & Roy 1985; Hall etal 1986; Colten-Bradley 1987).

Fig. 5. The Scorciabuoi shear zone is recognized by the markedly darker colouring of the strained mudstones with respect to the undeformed sediments.

SHEARING IN A LARGE-SCALE NATURAL FAULT

117

Fig. 6. X-ray diffraction curves of glycolated samples from the shear zone mudstones. a, b, c show mixed-layer illite/smectite clay minerals in (a) the footwall and (b, c) the undeformed slivers within the shear zone, d, e, f show the disappearance of smectite from the dark sheared clays, c and f were obtained from sub-centimetre spaced domains within the same hand specimen.

Fig. 7. Schematic representation of smectite and its hydration state in relation to the imposed effective stress.

Loss of water from the interlayer area makes smectite highly reactive and prone to transform to illite, with the uptake of K+ in the crystal structure. This is a reaction well documented in numerous sedimentary basins (Pollastro 1985; Deng et al 1996; Masuda et al. 2001) where smectite is replaced, at depth, by illite through a series of intermediate steps in which smectite and illite coexist. Positive effective stresses, i.e. stresses supported by the mineral framework, can facilitate the expulsion of water and a collapse to the 15.5 A state with an effective stress as low as 1.3 MPa (Fitts & Brown 1999). The loss of the II layer of water, at which state the illitization process is already in progress, occurs at temperatures below 67°-81°C (Colten-Bradley 1987) and, given the availability of K+, the reaction can take place at depths as little as 550 m below sea floor (Masuda et al. 2001). However, these conditions pertain to an approximately static setting such as that of a stratigraphic succession subject to a simple overburden stress. These conditions may have occurred in the PliocenePleistocene sediments which make up the wall rock of the SBF, and contain both illite and smectite according to XRD analyses (Fig. 6). In contrast, the concentration of illite in domains of high shear strain within the shear zone indicates that the dynamics of faulting can in some way facilitate and speed up the conversion of smectite to illite, under near surface conditions.

118

E. CASCIELLO ETAL.

Fig. 8. Grain size profiles of undeformed wall rock mudstones (empty circles) and the adjacent sheared mudstones (full circles). Both examples illustrate grain size reduction and a proportional increase of the clay sized fraction, in the sheared mudstones.

Grain size analysis Grain size reduction by cataclasis and the associated foliation development are considered to be the dominant processes leading to strain weakening in phyllosilicate-bearing gouges above the brittle/ductile transition (Bos & Spiers 2001). To investigate the grain-size characteristics of the sheared material, samples of the sheared, dark coloured mudstones have been compared with samples of the adjacent undeformed wall rock in two locations along the shear zone (site C and D in Fig. 3). Grain-size profiles (Fig. 8) have been obtained using a laser grain size counter (Sympatec granulometer) on samples that had been previously disaggregated using H2O2? distilled water and ultrasound. This treatment is bound to alter the organization of clay crystallites in the clay fraction. By using the same treatment on all samples, they should be affected in the same way, and variations observed in the samples frequency curves should reflect actual variations in grain size. The undeformed samples from both localities have very similar curves, with a bimodal distribution that

peaks in the very fine silt classes (from 5.6 /^m to 11.3 fjm) and in the fractions below 3.2 jinn. In both examples the profile of the sheared samples presents a fine-ward shift in the coarser silty component and a marked increase in the clay sized fraction. As grain size reduction mainly affects the coarser (clastic) classes, the observed grain size reduction may not be a consequence of the smectite to illite reaction in the clay fraction. Thus, if the observed shift in the deformed sediments is interpreted as the effect of mechanical grain refinement processes due to shear strain, the difference in the two reported examples is explained by them representing different stages in fabric evolution (Fig. 9). A larger amount of reduction, with a marked flattening of the peak in the silty component, is found for the sample obtained from a highly mature shear domain (Figs 8b & 9e). A lower degree of refinement, affecting exclusively the coarser fractions, reflects the less intense shear fabric of the sampled domain (Figs 8a & 9c-d). The increase in surface area that accompanies grain refinement, and the associated modification of grain surface texture have been invoked as factors that can increase reaction

SHEARING IN A LARGE-SCALE NATURAL FAULT

119

has been analysed using field based meso-scale analysis, peels obtained from orientated samples and scanning electron microscopy (SEM).

Mesoscale observations

Fig. 9. Evolution of shear fabric with increasing strain, as reported from laboratory experiments (Rutter et al 1986; Logan et al 1992).

rates (e.g. Vrolijk & Van der Pluijm 1999). The concentration of illite in the sheared samples could therefore be considered a consequence of faulting processes including cataclasis.

Fabric of the shear zone One of the central outcomes of laboratory studies of sheared materials is the recognition that the shear fabric develops in a stepwise fashion (Fig. 9); that is, under given conditions, different fabric elements arise in a sequence and with an abundance that reflects the shear-strain state of the sample (Logan et al 1992; Hicher et al 1994; Dewhurst et al 1996; Bos & Spiers 2001; Saffer et al 2001). Natural shear zones are more complex than gouges produced experimentally under controlled conditions (i.e. variable displacement, applied stresses, strain rate, hydraulic conditions), but provide the possibility of investigating the effect of parameters that have, perhaps, not been taken into account in the laboratory. The fabric of the SBF shear zone

Field recognition of the shear zone is facilitated by the contrast between light grey undeformed sediments and the darker colour of the sheared mudstones (Figs 5 & 10). The width of the shear zone varies from 2 to 10 m and it exhibits sharp and straight margins; however, one margin is generally more marked than the other one (Fig. 11). Slivers (lenses) of relatively undeformed (light) material are enclosed within the deformation zone and have their long axes moderately oblique to the shear zone. Along the shear zone and within the less strained domains, a pervasive foliation is present (stage a, Fig. 9) and is indicated by the preferred orientation of desiccation cracks and a discontinuous colour banding. The average angle that the foliation makes with the shear-zone boundary, against which it sometimes terminates abruptly, is 135°-160°. Differently orientated shears cut through the foliation forming a variety of patterns. Shears, that in hand specimens appear as polished surfaces, can be grouped into R-, Pand Y-shears, (stages b, c & d, Fig. 9), adopting the terminology of Logan et al (1992). The orientation of the main shears measured at locations B and D (Fig. 3) is shown in Fig. 12. However, the distribution and relative abundance of the fabric elements is not homogeneous either along or across the shear zone. Domains of intense strain localization are recognized by a higher density of shears, which tend to overprint the foliation, and reduce the dimensions of the relatively undeformed slivers enclosed within the deformation zone (Fig. 10). The extreme concentration of Y shears produces bands, several decimetres thick, of very cohesive gouge that can produce positive topographic effects. This generally occurs along one or other of the shear zone margins.

Fine scale analysis To examine the fine texture of the sheared clays, orientated samples were collected from different shear domains, representing slivers of relatively undeformed material or sheared mudstones. Samples up to 20 cm across have been cut along the plane perpendicular to the tectonic foliation and containing the shear direction for the preparation of peels and SEM observation. The peels (Fig. 13) provided a detailed visualization of the centimetre scale fabric and acted

120

E. CASCIELLO ET AL.

Fig. 10. The shear zone, marked by dark, intensely strained, illitic clays, encloses slivers (or shear lenses) of relatively undeformed smectite-bearing mudstones and displays sharp and straight margins. HW: hanging wall; FW: footwall.

Fig. 11. Schematic cross-sections through the SBF shear zone illustrating the inhomogeneous localization of shear strain along the deformation zone. The density in the pattern is proportional to the intensity of shearing; the shadowed pattern indicates locations containing the most advanced stages in fabric evolution (stage e, Fig. 9).

as samples for SEM study. This technique enables SEM observations on specific features of the peel, integrating grain scale observations into the wider picture of the sample's centimetre scale structural features. The sharp contrast between light and darker portions of the peels (Fig. 13) is the result of the

different quantities of material trapped by the resin of the peel and, as indicated by XRD analysis on different portions of the same sample, it matches the changes in clay mineral composition. Domains of smectite/illite composition adhere more easily to the resin of the peel, compared to the illite domains which leave only a thin

SHEARING IN A LARGE-SCALE NATURAL FAULT

121

Fig. 12. Equal angle stereographic projections showing the orientation of meso-scale shears characterizing the SBF at locations B and D of Fig. 3.

film of material. This difference reflects the finer grain size and higher cohesion that characterizes illitic clays compared to the smectite. Thus in contrast to direct observations of the samples, the peels show the illite rich layers as light and the mixed layers as dark. However, to remove this confusion the photographs of the peels shown in Fig. 13 have been printed in reverse to maintain a compatibility between the colours observed in hand specimen and on the peels. A feature characteristic of most samples is the distribution of illite clays along planes parallel to the orientation of the P foliation. Experimental work has shown that the P foliation is generated during the initial stages of deformation as particles are reorientated to track the finite strain ellipse (Rutter et aL 1986; Logan et al 1992; Bos & Spiers 2001) (Figs 9 & 13). The alignment of platy particles is well illustrated in the SEM images and appears to affect both illite/smectite (i-s) and the illite only (i) clays (Fig. 14). The millimetre and sub-millimetre spacing between alternating bands of i—s and i clays (Fig. 13a, b) indicates convincingly that temperature is not the key factor governing the mineral reaction. As expected from the meso-scale characteristics of the sampled domains, samples extracted from the slivers of relatively undeformed material (Fig. 13a) display a lower proportion of i clays than the more deformed domains. Within the slivers, illite is found along the plane of foliation which is sharply deflected by R and Y shears, and along millimetre-wide shears parallel to the P and Y orientations. In more deformed samples (Fig. 13b, c) the proportion of / clays is dominant over i-s clays. The latter are represented mainly by bands parallel to the plane of the foliation which are deflected, more

or less as passive markers, into sigmoidal shapes (Fig. 13c) by the cross-cutting shears. Scanning electron images of the illite domains show a very tight packing (Fig. 14c) and frequent microshears which together generate S—C fabrics. Large rounded voids are also common (Fig. 14d) and appear very similar to voids that in previous SEM analyses have been interpreted as being fluid-generated, in overpressured conditions (Takizawa & Ogawa 1999). The most advanced stage of deformation is represented by example d in Fig. 13. This shows a band of pervasive Y shears that almost completely overprints earlier fabrics. As noted earlier, this extreme localization of shears is generally associated with one of the margins of the shear zone (Fig. 11) and produces the most intense grain size reduction (Fig. 8b). However, SEM imagery indicates that, even in these extremely deformed samples, micro-scale domains exist that have preserved a flattening fabric orientated parallel to the P foliation.

Discussion and inferences A growing body of geochemical and geological evidence indicates that clay gouges can actively evolve during the faulting process (Brown et al. 1994; Vrolijk & Van der Pluijm 1999) and may be responsible for the anomalous weakness of some large faults, when compared to the frictional properties of the faulted geological materials (Mount & Suppe 1987; Saffer et al. 2001). Our study supports this hypothesis indicating that natural faults can induce, in clayey sediments containing smectite, a miner alogical conversion of the smectite to illite that would otherwise not proceed in the undeformed sediments. Analysis of the fabric of the shear zone

122

E. CASCIELLO ETAL.

SHEARING IN A LARGE-SCALE NATURAL FAULT

123

Fig. 14. Scanning electron micrographs obtained from, the peels shown in Fig. 13. The alignment of platy clay particles along the P foliation is evident in both the smectite (a) and illite (b) domains. Microshears giving rise to S-C fabrics (c) are frequently observed in the sheared illitic domains. Illite domains are also characterized by rounded voids, indicated by arrows in (d), which may be indicative of high fluid pressure and channelling.

suggests that gouge containing the newly formed minerals deformed in a more ductile-like style than portions that have not undergone the complete reaction (compare Fig. 13a and 13c). This observation may be explained either assuming a high water content in the transformed gouge or, as indicated by experimental studies on phyllosilicate gouges, by intense cataclasis (Bos & Spiers 2001). The present authors argue that both the presence of water and the process of cataclasis played an important role in the evolution of the analysed shear zone. The presence

of water is made likely by the disappearance of smectite, which can contain over 20% in weight of water (Fitts & Brown 1999), and its replacement by anhydrous illite. Experimental studies investigating the hydraulic properties of sheared gouges have shown that shearing generates a permeability anisotropy along the shear zone, with longitudinal permeability 0.5 to 2 orders of magnitude higher than transverse permeability (Arch & Maltman 1990; Zhang et al 1999). However, experimental results also indicate that both transversal and longitudinal permeability

Fig. 13. Peels obtained from orientated samples showing different stages of fabric maturity. Peel a is representative of a relatively undeformed sliver domain composed mainly by light coloured smectitic clays; illite is present along the planes of P foliation and in millimetre wide shears. A higher degree of fabric maturity is displayed in b and c. These peels are composed predominantly of dark illitic clays, while remains of smectitic clays form light streaks parallel to the P foliation. Peel d shows the most advanced stage of fabric evolution (Fig. 9e). Diffuse Y shears form a shear band, several centimetres wide, in which no remains of smectite clays are found. Grain size analyses show that this sample has the highest degree of grain size refinement (Fig. 8b). It can be compared with the lower degree of refinement observed for the sample of peel a (Fig. 8a).

124

E. CASCIELLO ET AL.

are lower than in the undeformed material. Sheared clays represent therefore aequitards which do not promote fluid flow (Brown & Moore 1993; Dewhurst et al 1996; Zhang et al 1999). This is in contrast with evidence from accretionary wedges which shows that fluid flow, dilatancy and salinity anomalies are concentrated along major shear zones in pelitic sediments (Fitts & Brown 1999; Takizawa & Ogawa 1999). In particular, smectite dehydration has been advocated as a principal source of fresh water, generating overpressure and hydrofracturing (Brown et al 1994) to account for the high permeability predicted from numerical models of the hydraulic circulation inside the wedge (Screaton et al 1990). In our view, the fault analysed in the Italian Apennines constitutes a good approximation of the shear zones produced during wedge accretion in pelitic sediments. The possibility of a detailed examination of the fine structures within the shear zone and the distribution of illite allows some deductions to be made regarding the smectite dehydration and the development of the observed shear fabric. The concentration of illite along planes corresponding to the P foliation, suggests that the same flattening process that produces particle reorientation within the shear zone also stimulated the dehydration of some of the smectite. This may possibly be initiated around more resistant grains, in a process very similar to the initiation of pressure solution seams in carbonate rocks (Salvini et al 1999). The foliation and the smectite-illite banding are clearly dissected by shears (Fig. 13a); these would intercept local pockets of fluids orientated parallel to the P foliation which are produced by smectite dehydration and illitization. This fluid, together with that generated along the shears would lubricate the shears and locally increase the fluid pressure along them. Wider shears, that have probably drained more water from 'affluent' shears, appear to possess a higher ductility (Fig. 13c) and are completely illitic. A ductile appearance on a mesoscopic scale can be achieved by the occurrence of diffuse microshears (Maltman 1987; Arch & Maltman 1990) and by cataclasis on a microscopic scale (Bos & Spiers 2001). The occurrence of cataclasis in the SBF shear zone is confirmed by the observed grain-size reduction (Fig. 8), and by SEM analyses of the dark sheared clays captured on the peels. The concomitant and complete illitization of the sheared clays may have been enhanced by two factors: the reduction in grain size and the associated alteration of the grain surface texture which

results from cataclasis, and the frictional dehydration of smectite. Previous studies on the dehydration of smectite have focused their attention on fluid composition, temperature and vertical lithostatic stresses (Howard & Roy 1985; Hall et al 1986; Colten-Bradley 1987; Deng et al 1996; Fitts & Brown 1999; Vrolijk & Van der Pluijm 1999). We argue that the addition of a shearing component to the vertical load can enhance the expulsion of water from volume-sensitive smectitic clays. Frictional dehydration (i.e. the de-bonding of water layers due to shear strain) would allow the shear zone to self lubricate by consuming, with progressive displacement, the slivers of relatively undeformed smectite sediments, generating a source of fluids inside the shear zone. Considering the very strong anisotropy of permeability that develops in the deformation zone, fluids would migrate along the shear zone rather than escaping from it. The local increase in fluid pressure along the shear zone caused by the frictional dehydration of smectite bearing sediments could also explain the discrepancy between laboratory evidence of reduced permeability in the sheared clays and data from accretionary wedges which show the shears to be zones of enhanced fluid flow. The authors would like to thank A. Langella, M. De Gennaro and S. Bravi at the University of Naples, and F. Molisso at the CNR Geomare Institute (Naples) for the support provided during this study. H. Shaw at Imperial College of London is thanked for the stimulating discussions on clay mineralogy. We are also grateful to C. J. Spears and R. Law for comments and improvements. This work was supported by the grant MPI 60% 2001 G. Pappone.

References ARCH, J. & MALTMAN, A.I 1990. Anisotropic permeability and tortuosity in deformed wet sediments. Journal of Geophysical Research, 95, 9035-9046. Bos, B. & SPIERS, CJ. 2001. Experimental investigation into the microstructural and mechanical evolution of phyllosilicate-bearing fault rock under conditions favouring pressure solution. Journal of Structural Geology, 23,1187-1202. BROWN, K.M., BEKINS, B., CLENNELL, B., DEWHURST, D. & WESTBROOK, G. 1994. Heterogeneous hydrofracture development and accretionary fault dynamics. Geology, 22, 259-262. BROWN, K.M. & MOORE, J.C. 1993. Comment on Anisotropic Permeability and Tortuosity in Deformed Wet Sediments' by J. Arch & A. Maltman. Journal of Geophysical Research, 98, 859-864. CELLO, G. & MAZZOLI, S. 1999. Apennine tectonics in southern Italy: a review. Journal ofGeodynamics, 27,191-211.

SHEARING IN A LARGE-SCALE NATURAL FAULT COLTEN-BRADLEY, V.A. 1987. Role of pressure in smectite dehydration - effects on geopressure and smectite-to-illite transformation. American Association of Petroleum Geologists Bulletin, 71, 1414-1427. DENG, X., SUN, Y., LEI, X. & Lu, Q. 1996. Illite/Smectite diagenesis in the NanXiang, Yitong, and North China permian-carboniferous basins: application to petroleum exploration in China. American Association of Petroleum Geologists Bulletin, 80, 157-173. DEWHURST, D.N., CLENNELL, M.B., BROWN, K.M. & WESTBROOK, G.K. 1996. Fabric and hydraulic conductivity of sheared clays. Geotechnique, 46, 761-768. EBERL, D. & HOWER, 11977. The hydrothermal transformation of sodium and potassium smectite into mixed-layer clay. Clays and Clay Minerals, 25, 215-227. FiTTS,T.G. & BROWN, K.M. 1999. Stress-induced smectite dehydration: ramifications for patterns of freshening and fluid expulsion in the N. Barbados accretionary wedge. Earth and Planetary Science Letters, 172,179-197. HALL, P.L., ASTILL, D.M. & MCCONNELL, D.C. 1986. Thermodynamic and structural aspects of the dehydration of smectite in sedimentary rocks. Clay Minerals, 21, 633-648. HIGHER, P.Y., WAHYUDI, H. & TESSIER D. 1994. Microstructural analysis of strain localisation in clay. Computers and Geotechnics, 16, 205-222. HOWARD, JJ. & ROY, D.R. 1985. Development of layer charge and kinetics of experimental smectite alteration. Clays and Clay Minerals, 33, 81-88. KASTENS, K., MASCLE, J. & ODP Leg 107 Scientific staff, 1988. ODP Leg 107 in the Tyrrenian sea: insights into passive margin and back-arc basin evolution. Geological Society of America Bulletin, 100,1140-1156. KNOTT, S.D. & TURCO, E. 1991. Late Cenozoic kinematics of the calabrian arc, southern Apennines. Tectonics, 10,1164-1172. LOGAN, J.M., DENGO, C.A., HIGGS, N.G. & WANG, Z.Z. 1992. Fabrics of experimental fault zones: their development and relationship to mechanical behavior. In: EVANS, B. & WONG, T. (eds) Fault mechanics and transport properties of rocks. Academic Press, London, 33-67. LONERGAN, L. & WHITE, N. 1997. The origin of the Betic-Rif mountain belt. Tectonics, 16, 504-522. MALTMAN, A.J. 1987 Shear zones in argillaceous sediments - an experimental study. In: JONES, M.E. & PRESTON, R.M.F. (eds) Deformation of sedi-

125

mentary rocks. Geological Society, London, Special Publications, 29, 77-87. MASUDA, H., PEACOR, D.R. & DONG, H. 2001. Transmission electron microscopy study of conversion of smectite to illite in mudstones of the Nankai trough: contrast with coeval bentonites. Clay and Clay Minerals, 49,109-118. MONACO, C, TORTORICI, L. & PALTRINIERI, W. 1998. Structural evolution of the Lucanian Apennines, Southern Italy. Journal of Structural Geology, 20, 617-638. MOUNT, V.S. & SUPPE, J. 1987. State of stress near the San Andreas fault: Implications for wrench tectonics. Geology, 15,1143-1146. PATACCA, E. & SCANDONE, P. 1999. Late thrust propagation and sedimentary response in the thrust belt-foredeep system of the Southern Apennines (Pliocene - Pleistocene). In: VAI, G.B. & MARTINI, I.P. (eds) Anatomy of a mountain-belt: the Apennines and adjacent Mediterranean basins. Chapman and Hall, 1999. POLLASTRO, R.M. 1985. Mineralogical and morphological evidence for the formation of illite at the expense of illite/smectite. Clays and Clay Minerals, 33, 265-274. RUTTER, E.H., MADDOCK, R.H., HALL, S.H. & WHITE, S.H. 1986. Comparative microstructures of natural and experimentally produced claybearing gouges. Pure and Applied Geophysics, 124, 3-29. SAFFER, M., FRYE, K.M., MARONE, C. & MAIR, K. 2001. Laboratory results indicating complex and potentially unstable frictional behavior of smectite clay. Geophysical Research Letters, 28, 2297-2300. SALVINI, E, BILLI, A. & WISE, D.U. 1999. Strike-slip fault-propagation cleavage in carbonate rocks: the Mattinata Fault Zone, Southern Apennines, Italy. Journal of Structural Geology, 21, 1731-1749. SCREATON, E.J., WUTHRICH, D.R. & DREISS, S.J. 1990.

Permeabilities, fluid pressures and flow rates in the Barbados Ridge Complex. Journal of Geophysical Research, 95, 8997-9008. TAKIZAWA, S. & OGAWA, Y. 1999. Dilatant clayey microstructure in the Barbados decollement zone. Journal of Structural Geology, 21, 111-122. VROLIJK, P. & VAN DER PLUIJM, B.A. 1999. Clay gouge. Journal of Structural Geology, 21,1039-1048. ZHANG, S.,TuLLis,T.E. & SCRUGGS, VJ. 1999. Permeability anisotropy and pressure dependency of permeability in experimentally sheared gouge materials. Journal of Structural Geology, 21, 795-806.

This page intentionally left blank

Insights from the Ocean Drilling Program on shear and fluid-flow at the mega-faults between actively converging plates ALEX MALTMAN1 & PAOLA VANNUCCHI2 Institute of Geography and Earth Sciences, University of Wales, Aberystwyth, SY23 3DB, UK (e-mail: [email protected]) 2 Dipartimento di Scienze delta Terra, Universitd di Modena e Reggio Emilia, Piazzale S. Eufemia, 19, Modena, Italy (e-mail:[email protected])

1

Abstract: The mega-faults between actively converging plates have recently been penetrated by the Ocean Drilling Program at three plate margins: Barbados, Costa Rica and Nankai. Cores, downhole instrumentation and detailed seismic imagery provide data which may be helpful in interpreting ancient examples of shear zones. The mega-faults, developed in poorly lithified sediments, separate major lithospheric plates yet are merely tens of metres in thickness. They respond to ongoing strain by intensifying inwards rather than propagating outward splays and can grow thinner because of continuing compaction. Surprisingly, lithological influence on the localization of fault propagation seems slight, but lithology determines the deformation style within the faults. The resulting structures show asymmetric distributions within the zones but, in these flat-lying structures, tend to show a downward increase in strain. Upper margins are typically gradational whereas lower boundaries can be strikingly abrupt. The fluid-transport behaviours are complex. In some situations the horizontal flux is very diffuse but centred around the fault. Some faults can efficiently channelizefluids- for distances of tens of kilometres - while at the same curbing flow across them. The fluid transport is clearly episodic and heterogeneous. Fingers of pressured fluid migrate within the fault zone, in patterns that constantly change through time.

Perhaps the most striking and most important examples of the shear zone -fluidflowinterplay come from modern convergent plate-margins. Here, typically, sediments accumulating on an oceanic plate are being accreted onto or subducted below a continental plate. Either way, although the incoming sediments are highly porous and water saturated they eventually end up as virtually dry rock. Enormous volumes of pore-water have to be dispersed. Crucial in this process is the role of the zone of shearing between the two colliding plates, a structure variously referred to as the plate-boundary fault or mega-thrust, mega-shear zone or decollement. The extent to which this fault can assist drainage by channelling the escaping fluids affects petrogenetic processes at depth, life at the ocean floor, and the very architecture and dynamics of the plate margin. In addition, whether the structure traps or drains fluid influences its frictional behaviour, and hence seismogenesis and all that implies for hazards to society. Much of our knowledge of these processes at active plate margins comes from the Ocean Drilling Program (ODP). In recent years ODP

has succeeded in penetrating the plate-boundary faults at three active convergent margins. Retrieved drill-cores have yielded a wealth of laboratory analyses, and downhole wireline logging and logging-while-drilling (LWD) have yielded in situ measurements. Ongoing downhole instrumentation is monitoring variations over time, and associated 3D seismic surveys have provided detailed imagery. The three margins (Fig. 1) are those at Barbados, where the North Atlantic Ocean is being underthrust beneath the Caribbean plate (ODP Legs 110, 156 and 171A), Costa Rica, where the Cocos Plate descends beneath the Caribbean Plate (ODP Leg 1701), and Nankai, at the junction of the Philippine Sea and Eurasian plates (ODP Legs 131, 190, and 196). The present review begins by outlining the tectonic setting of each of these three margins and then summarizes, in turn, their deformational characteristics and hydrogeology. Such an overview allows us to draw comparisons and contrasts between the three faults, and hence features that may be useful in interpreting ancient examples of megashear zones.

1 Preliminary results of the most recent drilling at the Costa Rica margin, ODP Leg 205, are available at http://www-odp.tamu.edu/publications/prelim/205_prel/205toc.html.

From: ALSOP, G. L, HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224,127-140. 0305-8719/$15.00 © The Geological Society of London 2004.

128

A. MALTMAN & P. VANNUCCHI

Tectonic setting of the three margins Barbados The North Atlantic Plate is converging with the Caribbean Plate at a rate of 25-40 km Ma"1, and subduction produces the Lesser Antilles volcanic arc (Fig. la). As the plate converges, sediments are progressively scraped off to build a structure variously known as the (North) Barbados Ridge, Lesser Antilles, or - as in this paper - the Barbados accretionary prism. The island of Barbados represents the prism emerging above sea-level. The sediments being scraped off are remote from the Orinoco and Amazon fans that are their main sources and this, together with the presence of bathymetric barriers, leads to the sediments being distal and fine-grained, dominantly clay-rich, silts (e.g. Moore et al. 1988). This lithology may be a fundamental control on the exceptional dimensions of the Barbados accretionary prism, which is over 150 km wide and unusually thin, with a prism taper angle of only 2°. The plate-boundary fault acts as the basal decollement of the accretionary prism and, in the toe region, maintains a constant level as it propagates oceanward into the incoming sedimentary section. Relatively good drilling conditions have led ODP to focus on this plate boundary: Leg 110 cored a transect across the margin (Mascle et al. 1988), Leg 156 focused on coring and LWD of the decollement zone itself (Shipley et al 1995), and Leg 171A was dedicated to LWD across the fault (Moore et al. 19980, b).

Costa Rica

Fig. 1. Schematic maps and cross-sections to show the location and general plate tectonic setting of the (a) Barbados, (b) Costa Rica and (c) Nankai convergent margins, and the ODP drill sites referred to in this paper.

Off Costa Rica, the Cocos Plate is being subducted beneath the Caribbean Plate, at a convergence rate of 85-90 km Ma"1 (Fig. Ib). The incoming sedimentary section is only about 400 m thick and includes a thick (225 m) calcareous section. Drilling during ODP Leg 170 revealed that no frontal accretion is taking place at this margin (Kimura et al. 1997). A sedimentary wedge exists, with a 7.5° taper angle of and a width of only 4 km, that is built entirely from material sourced from the landward sedimentary apron by debris flow and mass movements; it includes almost no trench or ocean-plate deposits. In other words, virtually the entire incoming sedimentary section is being underthrust. The drilled plate-boundary fault, penetrated during ODP Leg 170 at two sites (1040 and 1043; Fig. Ib) therefore separates sediments of quite different origins. Unlike the

FAULTS BETWEEN ACTIVE CONVERGING PLATES

129

other two margins discussed here, at Costa Rica subduction erosion appears to be important, and has been invoked to explain a number of features of the recent tectonics of the margin (Ranero & Von Huene 2000; Von Huene et al 2001; Vannucchi et al. 2001).

Nankai Inflection of the Philippine Sea Plate as it drives beneath the Eurasian Plate off SW Japan creates the trench known as the Nankai Trough. Hence the accretionary prism formed by scraping off of the incoming sediments is referred to as the Nankai Prism (Moore etal. 2001a; Fig. Ic). In the area probed by the ODP (Hill etal. 1991; Moore et al. 20016; Mikada et al. 2002) the prism is around 70 km wide with a taper angle of 3°-4°. The plate convergence rate, at around 40km Ma"1, is the slowest of the three margins described here. On the other hand, the Nankai margin in the drilling area shows a remarkably high heat flow, probably because a fossilized spreading ridge is currently being subducted hereabouts. The sediments in the accretionary prism are largely terrigenous. Although the lower parts, including the sections through which the plate boundary fault is propagating, are dominated by hemipelagic silty clays with volcanic ash intervals in places, erosion of the nearby Japanese highlands and resulting alongtrench turbidite transport leads to the upper part of the prism being dominated by coarse sediments.

Deformation features of the mega-shear zones Barbados The plate-boundary fault at the Barbados margin is developed in partially lithified clayey siltstones, and consequently the shearing is mainly expressed as zones of scaly fabric (Labaume et al. 1998). Such zones also occur in the prism, especially where associated with thrusts, but they are markedly more intense in a 30-40 m zone that corresponds to the decollement seen on seismic images (Maltman et al. 1998). The boundaries of the zone are somewhat gradational (Fig. 2), especially at the base, but around the topmost part are narrow zones in which the claystones are fragmented into narrow millimetre- to submillimetre-scale chips. These breccia zones are only centimetres across, but, coinciding with geochemical anomalies, may be

Fig. 2. Annotated schematic log based on cores from the Barbados decollement at Site 948 (that with most complete recovery). Density of grey tint approximately indicates the intensity of deformation seen in the recovered cores. AMS: anisotropy of magnetic susceptibility. At 498 m below the sea floor, zones of scaly fabric become abruptly thicker and more closely spaced: on this basis the decollement zone is 31 m thick. Including the overlying breccia zones and level of magnetic anisotropy change gives a thickness of 39 m, equivalent to the 40 m reported from Site 671 cores. Although poorly defined, the decollement thickness at the oceanward Site 949 exceeds this, and may be as great as 90 m.

hydrogeologically important. There is an abrupt change in the orientation of magnetic anisotropy in the upper part, reflecting a marked decoupling of the overlying prism material from the decollement zone (Housen et al. 1996). The thickness of the zone below the prism toe varies by only a few metres, within the range 30-40 m, though it appears much thicker oceanward. A feature of the Barbados plate-boundary fault is that it straddles an important lithostratigraphic boundary, which leads to a marked difference in style of deformation between the upper and lower parts of the fault zone. The upper sediments are low density, radiolaria-rich

130

A. MALTMAN & P. VANNUCCHI

derived slope debris. The shear zone is defined structurally by an increase in the intensity of anastomosing, finely interpenetrative fracture systems (Fig. 3). In the upper part of the zone, the sediment is broken into millimetre to centimetre scale lenticular to blocky fragments, typically with polished and striated surfaces (Vannucchi & Tobin 2000), in a disaggregated matrix. Presumably because the sediments contain less clay than at Barbados, scaly fabric is rare and not well developed. Minor veinlets are present. The lower part of the zone was poorly recovered by coring, but it seems that in general the style of deformation is more plastic than the brittle upper part. Although comprising the same lithology, the lower domain shows centimetre-scale folds and 'swirled' mixing of laminae, reminiscent of the stratal disruption at Barbados. A common difficulty with rotary cores of poorly lithified sediment is the recognition of structures due to drilling disturbance, and it may well be that these ductile structures are entirely artificial (Kimura et al. 1997. Nevertheless, the abrupt change in deformation style Fig. 3. Annotated schematic logs based on cores from the Costa Rica decollement at Sites 1040 and and the plastic consistency of the cores clearly 1043. Density of grey tint approximately indicates the reflect a different rheology in the lower part of intensity of deformation seen in the recovered cores. the fault - as at Barbados - and suggests that the Note the brittle style in the upper parts at both sites, lower part of the fault is ductile in situ. No veins and a plastic rheology in the lower interval, and the have been recorded from this lower domain. The total thickness variation in the zone. actual base of the shear zone is very sharply defined. A curious feature of this fault, at least in the sediments showing good development of scaly region penetrated by the ODP drilling, is the fabrics interspersed with intervals of fracture variation in its thickness. Although the fault networks (Labaume et al. 1998; Maltman et at. zone might be expected, as at Barbados, to 1998). The presence of the radiolaria tests evolve by inwards intensification of strain and be enables the sediment to support an open frame- subject to progressive compaction arcward, work, such that in the proto-decollement zone beneath the wedge toe the fault zone appears to the porosities are as high as 70-75%. In the thicken substantially. Specifically, at Site 1043, lower half of the fault zone, radiolaria are virtu- 400 m arcward from the trench, it appears to be ally absent, and the clayey sediments are dis- 9 m in thickness but at Site 1040,1.5 km arcward, tinctly plastic and sticky to the touch. Scaly it has attained a thickness of 40 m, roughly the fabrics are less well developed. The deformation same as the decollements at the other two style is markedly ductile, and characterized by margins. Both the upper, brittle parts and the stratal disruption in which laminations and lower, ductile parts of the zone vary roughly in centimetre-scale folds are thinned and dis- proportion: the ductile domain occupies only membered. 50 cm at 1043 and 14.4 m at 1040. The explanation may lie in the decollement zone being bounded by undulating rather than planar surCosta Rica faces. This is in line with preliminary obserThe plate-boundary shear zone at the Costa vations from very recent, Leg 205, drilling Rica margin is developed in low-density, weakly (Ocean Drilling Program 2003) which suggest lithified silty claystones and sandy lenses, which the zone may onvolve a system of narrow (10m) directly overlie diatomaceous ooze (Kimura et duplexes, probably developed preferentially in al. 1997). The lithology is relatively mixed the softer silts and clays rather than the harder because the structure is developing in trench sand lenses. sediments and the base of the frontal sediA further curiosity of this plate-boundary mentary prism, which is formed by landward- fault is that although there are marked

FAULTS BETWEEN ACTIVE CONVERGING PLATES

131

fractured siltstone, with shear-induced brecciation with some local phyllosilicate reorientation, and a narrow zone of intense scaly fabric at its abrupt base (Fig. 4). At Site 1174 the fault is in its early stages of development and comprises a 32 m zone of brecciation, the intensity of which increases irregularly downwards. It is thoroughly comminuted in its lower part (and may have scaly fabric at its base, as at Site 808) and shows an abrupt basal separation from little deformed, underthrusting sediments below. At Site 1173 no lithostratigraphic or physical-properties features have been observed that point to the controlling influences on the level of fault propagation. There are a number of physical influences on the architecture of an accreting plate margin (e.g. Martinez et al. 2002) and it may be that the depth location of the mega-fault may be reflecting controls other than sediment properties.

Fluid transport in the mega-faults Fig. 4. Annotated schematic logs based on cores from the Nankai decollement at Sites 808 and 1174. Density of grey tint approximately indicates the intensity of deformation seen in the recovered cores. The more evolved structure at Site 808 is thinner, but both examples are characterized by an irregular downwards increase in intensity of brecciation, culminating at Site 808 (and possibly Site 1174) with a narrow interval of intense scaly fabric that abruptly separates the decollement from the virtually undeformed sediments being underthrust.

lithological variations in the adjacent sedimentary section, it appears not to propagate along any particularly weak litho-mechanical horizon but crops out on the seafloor. Consequently, as plate convergence progresses the form of the fault reflects the topography of the incoming sea floor. Whereas nearby the fault shows a more complex form due to the subduction of seamounts and normal faults developed in the incoming plate, in the drilling area the seafloor is morphology-free and the fault approximately planar.

Nankai The plate-boundary fault at the Nankai margin has been penetrated where it is well developed (Site 808; Byrne et al. 1993), at an incipient stage (Site 1174), and at the level where expected to propagate (Site 1173; Fig. Ic; Moore et al. 20015). It shows arcward thinning. At Site 808 it is a roughly 20 m thick zone of intensely

Barbados Sediments oceanward of the plate margin show high porosities (70-75%), which become compacted as they are thrust below the prism toe. Because of burial and tectonic compaction, porosities only 4 km inboard of the deformation front are reduced to 50%. These early stages of consolidation entail the expulsion of large amounts of pore water from the sediment, prompting the question of what happens to the water. There are a number of possibilities. The pore water could temporarily accumulate nearby, in pockets of high porosity sediment (possibly overpressured and possibly with dilated fractures; Moore et al. 1995), it could exit the sediment system entirely, either diffusely or by being focused along particular channels, or there could be some combination of all these processes. At Barbados, it seems that the sediment dewatering is a combination of channelized drainage along faults, particularly the plateboundary fault itself, and intermittent, localized sealing such that pockets of overpressured sediment arise along the faults. In fact, the hydrogeological data from the Barbados decollement show two striking features: localized mineral veining and anomalous pore fluid geochemistry indicating distinct fluid-flow channelling along the zone, which probably varies through time, and evidence from 3D seismic processing that anomalous high porosity zones within the plate boundary fault zone are indicating spatial heterogeneity of the fluid transport.

132

A. MALTMAN & P. VANNUCCHI

Fig. 5. Depth profiles of selected chemical species in pore-fluids across the Barbados decollernent (shaded zone), from core analyses at Sites 948 and 949. Note the pronounced negative and positive spikes, respectively, of Cl and CH4 at the very top of the zone, and similar spikes in Br/Cl and I/C1 in the upper part of the zone at both sites. The fewer data points at Site 949 are due to poor core recovery. From Moore et at. (1997).

Pronounced anomalies in chemical species such as chloride and methane along the faults in the accretionary prism and, especially, the plate boundary fault (Fig, 5) denote fault-focused transport. The data show that the flux is particularly focused along the top of and immediately above the decollernent, where the narrow zones of breccia mentioned earlier may be providing the channelways. The low chlorinity fluid may have been generated by clay mineral dehydration within the sediments, particularly those arcward of the prism toe, from where they are channelled along the decollement before eventually reaching the sea floor. Lines of mud diapirs (Sumner & Westbrook 2001; Lance et al. 1998; Olu etal. 1997) attest to fluids in this region reaching the sea floor predominantly along fracture zones. The geochemical signatures of the fluids in the accreted and the underthrust sedimentary sections are quite different, indicating that the hydrogeological regime is distinct from the prism. That is, the plate boundary fault must be acting both as a conduit for lateral transport and, at the same time, as a seal to vertical flow. Thus the Barbados mega-shear zone provides a fine example of the hydrogeological anisotropy that active faults can induce. The structure also provides an excellent illustration of how the fluid-flow behaviour of major faults can vary through time, though the details are disputed. Experimental deformation has indicated the importance of episodic flow (e.g. Bolton et al. 1999) and theoretical modelling provides one approach to quantifying the processes (e.g. Saffer & Bekins 1999). Henry (2000)

modelled the dynamics of the Cl anomaly and suggested that the fluid flow had been concentrated along the mega-fault for the last 2000-20 000 years, with earlier pulses generating the mineral veins and certain other geochemical anomalies in the zone. Fitts & Brown (1999) however, argued on the basis of smectite dehydration kinetics, that the fluid transport episodes must have been restricted to shorter pulses, on a time scale of several hundreds of years. The apparent spatial heterogeneity of the transport regime in the fault zone is deduced from the polarity pattern of seismic wave amplitudes derived from a 3D seismic survey of the prisrn toe. Figure 6 shows the data, from Shipley et al. (1994) and particularly the distinct patches of normal and negative signals along the fault. The exact significance of these areas of negative seismic amplitude is unclear. Although Shipley et al. (1994) argued at the time that they represent pockets of overpressuring, calibration from borehole logs (Moore et al. 19980) and LWD (Moore et al. 1998&) suggests they may be better regarded as zones of low sediment density (high porosity). The idea of dilated sediments within the fault zone is supported by detailed microstructural observations (Takizawa & Ogawa 1999). Of course the dilation could be due to elevated pore pressures but Screaton et al. (2000) argued that the permeabilities reported for these sediments are too great to support such overpressures for any length of time. They therefore argued that the high porosities are due to a sustained undercompaction because of high sediment strength, perhaps together with a continuous fluid

FAULTS BETWEEN ACTIVE CONVERGING PLATES

133

Fig. 6. Lateral fluid content variations at the Barbados plate-boundary fault inferred from seismic amplitude. The dark areas indicate negative seismic polarity, thought to represent areas of high fluid content. Based on Shipley etal (1996).

recharge from an arcward source. The notion of such fluid ingress is appealing if the reverse polarity patches are viewed as being concentrated in a roughly NE-SW zone, at least 10 km long and 1-2 km wide, which could represent a broad 'finger' of pressured fluid moving within a shear zone to sites of lower fluid potential, here up-dip towards the seafloor (Screaton & Ge 1997,2000; Bangs et al 1999; Moore & Silver 2002). Preliminary hydrogeological data are available from long-term downhole instrumental observations, but so far they are difficult to interpret. For example Screaton et al. (1997, 2000) thought that the ambient fluid pressures were approaching lithostatic values (but because of the high permeability values, as mentioned above, called on transient influxes of pressurized fluids), whereas Henry (2000) argued that the data are inconsistent with overpressuring. LWD data from Leg 171 (Fig. 7) show a marked offset of sediment density to higher values just below the decollement, and that the high porosities within the plate boundary fault are progressively lost arcward. Moreover, the upper part of the mega-shear loses porosity more than the lower part, presumably because the ongoing shear has a greater effect on the open, radiolaria-supported fabric (Moore et al. 1997). Whatever the details, this mega-fault is fundamental in controlling the hydrogeology of this plate margin.

Costa Rica Porosity changes in the sediments at the Costa Rica margin are striking and well constrained. On the incoming Cocos Plate porosities differ considerably with lithology but fall within the range 65-80%, reduced variously to 50-60% as subduction begins. This corresponds to a thickness reduction of 67-80% in the sediments beneath the toe of the wedge. Within the plate boundary fault, high and low porosity zones alternate, partly due to fine lithological changes. All this implies that large amounts of pore water are being expelled from the sediments and, as at Barbados, the mega-fault is clearly playing a crucial role. The decollement lacks well developed veins, but isolated, aligned crystals of manganese oxide and rhodochrosite in the upper domain, together with very localized geochemical anomalies, attest to fracture-hosted fluid transport in the upper part of the decollement (Tobin et al. 2001). This brittly deformed, upper portion of the decollement probably has a high lateral permeability because zoneparallel fractures are being dilated by elevated fluid pressures whereas the aligned clays in the highly anisotropic fabric provide a seal to vertical flux (Maltman & Bolton, in press). In fact LWD data indicate the presence of metre-scale conduits, not apparent from cores or laboratory

134

A. MALTMAN & P. VANNUCCHI

Fig. 7. Variations with depth of sediment density across decollement and proto-decollement zones at the Barbados convergent margin, derived from LWD. The undeformed section at Site 1044, oceanward of the margin, provides a baseline reference. Note the general increase in density (loss of porosity) arcward, with progressive consolidation. In line with this, although not depicted here the thickness of the zone may become reduced (based on Moore et aL 1997).

determinations, that are greatly aiding the flux (Saito & Goldberg 2001). The fluid in the decollement is freshened (Silver etal., 2000) and so is most likely sourced from mineral dehydration at greater temperatures and depths (Kastner et al. 1991) and migrates tens of kilometres up dip. The paucity of mineralization is perhaps surprising in view of the fact that carbonate-rich material is being subducted; however, there is geochemical evidence that, as at Barbados, the fault is able to isolate the underlying sediments hydrogeologically while itself acting as a conduit, at least in places (Fig. 8). For example, the presence of thermogenic C3 hydrocarbon (propane) in the cold decollement but not beneath it (Fig. 8) is not only an indicator of long distance lateral transport within the fault but that the flux is isolated from the hydrogeological regime below. Laboratory tests by Saffer et al.

(2000) also suggested that the underthrust sediments are draining independently. It seems that the plastic clays in the lower, ductile domain of the decollement provide an effective barrier to vertical fluid communication across the zone, even though immediately above is a major conduit. Furthermore, the very sharp change in physical properties at the junction with underthrust section suggests that the hydrogeological isolation is accompanied by a marked decoupling of the deformation mechanics.

Nankai The hydrogeology of the Nankai margin shows two striking aspects: the marked, abrupt increase in porosity immediately below the plate boundary fault, first observed in Leg 131 cores (Hill et al. 1991) but recently confirmed by the Leg 196 LWD (Moore et al 1998ft), and the

FAULTS BETWEEN ACTIVE CONVERGING PLATES

135

Fig. 8. Depth profiles of porosity, chloride and propane (C^) in pore-fluids across the Costa Rica decollement (shaded zone). Note the spikes in the chemical species in the fault zone and the pronounced porosity increase immediately below.

absence of evidence for fluid flow concentrated along the fault. The preservation of anomalously high porosity in the underthrust sediments presumably indicates that they have been unable to drain - they are hydrogeologically isolated (Screaton et al. 2002). Yet most workers have inferred from the brecciated nature of the zone a high permeability (Byrne et al. 1993). Part of the problem is the practical difficulty of measuring permeabilities in such highly fractured materials. Laboratory determinations tend to focus on the coherent, intact blocks and not the breccia as a whole, which in turn is difficult to assess in situ. Almost certainly there are distinct mechanical and hydrogeological differences between the blocks and the slip zones between them (e.g. Morgan & Karig 1995). The apparent hydrogeological isolation of the high-porosity underthrust sediments may be partly due to the interval of scaly clay at the very base of the decollement. Although it is only some 15 cm thick, such highly aligned clays may be reducing the vertical permeability sufficiently to seal the fluids below (Vannucchi et al. 2003). The resulting overpressuring and weakening of the sediments then account for the mechanical decoupling of the virtually undeformed, underthrust section (e.g. Maltman etal. 1993) from the over-riding materials. The broad geochemical patterns at Nankai of species such as chloride ion contrast with the situation at Barbados and Costa Rica, where narrow, spiked anomalies coincide with faults. The distributions at Nankai are at best very

diffuse. Perhaps the clearest example is the chloride reduction, but this occupies a depth interval of about 600 m, albeit centred on the plateboundary fault (Fig. 9). Together with the lack of other evidence of concentrated fluid flow, such as mineral veining - which is completely absent from cores recovered from the mega-fault - it has been argued that the Nankai margin is primarily undergoing diffuse rather than channelized dewatering (e.g. Maltman et al. 1992). The broad Cl decrease has been attributed to dilution by fresher waters moving long distances laterally from depth, but Brown et al. (2001) have argued that it is chiefly due to in situ clay dewatering, perhaps supplemented by inflow from elsewhere. Modelling of this effect by Saffer & Bekins (1998) concluded that 71% of water being lost from the incoming sediments leaves by diffuse flow out of the seafloor, with only 0-5% being focused along the decollement (only 1% of the water was thought to be subducted). Although the drainage along the megafault is less focused than elsewhere, it may also be episodic. The computations by Saffer and Bekins (1998) suggested that the decollement allowed transient pulses of pressurized fluids from outside, for periods on the order of 160 ka. Comparisons and contrasts Having reviewed each situation separately, we now summarize collectively various structural and fluid-transport aspects of the plate boundary faults at these three active margins.

136

A. MALTMAN & P. VANNUCCHI

Fig. 9. Depth profiles of porosity and chloride concentration across the Nankai decollement (shaded zone) at Sites 808 and 1174, and at the projected equivalent level at Site 1173, an oceanward baseline reference site. Note the evolving porosity decrease at the base of the fault but lack of a chloride spike.

Lithological influence on propagation The mechanical behaviour of different lithologies might be expected to exert a dominant influence on the levels at which a plate-boundary faults propagates into the incoming ocean plate section, but the idea is not supported by the present evidence. The Barbados margin seems the best candidate for lithological control, as the fault coincides with a boundary that separates high-porosity (radiolaria-rich) sediments above from barren material below. However, it is not clear why the fault involves both the brittle upper part and the lower, plastic domain rather than using solely whichever material is the weaker. Perhaps, despite the difference in rheology, both materials happen to be similar in strength and resistance to fracture propagation. At Costa Rica there are porous, and presumably weaker, horizons in the incoming section but they seem to have no influence on guiding the fault. Instead the fault crops out on the sea floor (Fig. Ic), with the palaeobathymetry of the incoming ocean floor controlling the form of the fault. Even more puzzling is the situation at Nankai, where, as at Barbados, the location of the mega-fault is propagating at a remarkably consistent level, such that its likely depth in the reference Site 1173 seems readily predictable. However, a concerted effort with high quality cores, wireline logs and LWD data have failed to detect any changes in physical properties at this depth. It would seem, therefore, that there is some more fundamental

control on the location of these mega-shear zones, related to the geometry and dimensions of the converging lithospheric plates (e.g. Martinez et al. 2002 and references therein), rather than to detailed sediment mechanics. Fault thicknesses In the case of all three margins, the mega-fault despite being the active, high-strain junction between major lithospheric plates - is no more than a few tens of metres thick. The seismic images and thickness variations at both Barbados and Nankai are consistent with the faults having parallel-bounded margins, with the zone losing thickness by consolidation. The ongoing strain, at least in these prism toe areas, is accomplished by the sheared intervals intensifying inwards, to give a better defined, more pervasively sheared zone. However, at Costa Rica for unclear reasons there is a comparatively large thickness variation. The zone as encountered in the two boreholes appears to thicken landward, perhaps because it is bounded by non-planar surfaces and hence of uneven thickness. Deformation structures within the faults The structural style and deformation features within the mega-faults depend primarily on lithology. Thus the fault in the clay-rich sediments at Barbados is dominated by scaly clay, whereas breccia zones are the chief feature in

FAULTS BETWEEN ACTIVE CONVERGING PLATES

the more silty materials at Costa Rica and Nankai.

Margins of the fault-zones The nature of the boundaries of the fault zones at the three plate margins shows no consistent pattern. The borders of the Barbados fault are gradational both at top and bottom; at Costa Rica, the fault has a gradational upper boundary but very abrupt at base; at Nankai the base is similarly abrupt (and distinctly separates very highly deformed sediments within the zone from the virtually undeformed down-going section) but the top is gradational, albeit over a thinner interval than at Barbados. The reasons for these differences are unclear.

Internal geometry of the fault zones The nature and distribution of the deformation structures within the three mega-faults differs. At Barbados the structures are distributed roughly symmetrically about the central lithological change, although with brittle structures above and a plastic style below. Structures in the Costa Rica also show a brittle style in the upper parts and ductile below, but the distribution is asymmetrical, with a downwards increase in strain. The Nankai fault shows no rheological variation but an irregular increase in deformation downwards.

Hydrogeology The hydrogeological regimes in and around the three mega-faults differ markedly. Focusing of lateral fluid transport along the plate boundary faults is well documented at Barbados and Costa Rica. At Nankai, however, along fault flow is diffuse, in a zone centred on the fault but about fifteen times thicker than the decollement itself. The dewatering regime seems to be dominated by vertical transport, pervasive within the prism sediments. However, in all three faults, the underthrust sections are hydrogeologically isolated, showing anomalously high porosities and/or pore-fluids of different chemical signatures to the material above. The lower parts of decollements appear to seal the underlying fluids, because of plastically deforming, low permeability clays at Barbados and Costa Rica, and because of a narrow zone of scaly clay at Nankai. Also at all three margins, the fluid flow is clearly episodic and heterogeneous. This is certainly the case at Barbados and Costa Rica. At Nankai, the present frontal thrust of the prism seems to be more active tectonically and hydrogeologically

137

than the basal decollement (Hill et al 1991), but this must be temporary. New thrusts will be created as sediments are progressively accreted and the prism continues to grow. In due course, the plate-convergence strain and fluid flux must transfer elsewhere, to parts of some other thrust or the basal decollement of the prism. Implications for other mega-shear zones In summary, the three active plate-margins reviewed here show that very large shear strains can be accomplished in narrow fault zones: major lithospheric plates pass each other along zones no more than tens of metres thick. Such zones may be parallel sided but may have an undulating, anastamosing geometry (to help explain the anomalous thickness variations at Costa Rica). Flat-lying fault zones in weakly lithified sediments can thin as the zone evolves, due to progressive compaction of the material; accumulating strains lead to inward intensification of shear structures rather than the outward propagation of new splays. In some cases, lithology appears to have a role in localizing a fault zone, and certainly in its style of deformation. In other situations, the mechanical influences are much more subtle and the localization of the propagating fault may be due to some larger scale, indirect, geometric constraint. These plate boundary faults are typically asymmetrical internally, either with deformation style varying within zone if the lithology changes, or with a progressive increase downwards (in these flatlying examples) of the shear strain. The fault margins are not always well defined, especially at the top, but the base can be markedly abrupt. The base can be very highly sheared yet adjacent to virtually undeformed material. Fluids can be channelled along these megashear zones, though in some cases the horizontal flux may be much more diffuse around the fault. The fluids may be generated through local dehydration reactions, especially the smeetite-illite transition in the case of sediment, but other fluid sources can be very distant, with the fault transporting fluids tens of kilometres and more. At the same time, such horizontal conduits can efficiently curb across zone, vertical flow. It is therefore simplistic to regard a fault zone as a conduit or a seal: its effect may differ greatly with the flux direction. In the fault-parallel direction alone the behaviour may vary, even within a narrow zone of shear. A common situation for approximately flat lying mega-shears, judging by those reported here, is for the flux to be focused in the upper part of the zone or even immediately above it, whereas the lower parts,

138

A. MALTMAN & P. VANNUCCHI

preserving an earlier, underconsolidated high porosity state. The pockets attempt to move to sites of lower fluid potential, dispersing or coalescing to various degrees, to give fingers of fluid moving within the fault zone, in patterns that are constantly changing through time. The evidence available from ancient shear zones in rocks may not allow such detailed analysis of their shear-strain and hydrogeological behaviours. Even so, it would be misleading -judging by the evidence from modern, convergent plate-boundary faults in sediments - to interpret them as simple structures. Active mega-faults are intricate, heterogeneous structures that vary through time and which generate complex fluid flow behaviours. Only three such faults have been discussed here and it is difficult to make generalizations. Each one shows differences in its strain and fluid transport behaviour. Constructive comments from the two referees, A. Robertson and T. Needham, are gratefully acknowledged.

References

Fig. 10. Summary diagram of features at the three mega-shear zones described in this paper.

as well as presenting a seal to vertical flow, are much less permeable to horizontal fluid transport. Moreover, there is evidence, for instance from geochemical modelling and experimental deformation, that flow during active shearing involves episodic pulses and that the distribution of fluids within the shear zone is highly heterogeneous. There are pockets of higher fluid content, perhaps occupying dilated, overpressured pods and/or remnant lenses of sediments

BANGS, N.L., SHIPLEY, T.H., MOORE, J.C. & MOORE, G. 1999. Fluid accumulation and channelling along the Northern Barbados Ridge decollement thrust. Journal of Geophysical Research, 104, 20399-20414. BOLTON, A.J., CLENNELL, M.B. & MALTMAN, A.J. 1999. Nonlinear stress dependence of permeability: a mechanism for episodic fluid flow in accretionary wedges. Geology, 27, 239-242. BOLTON, A.J., VANNUCCHI, P., CLENNELL, M.B. & MALTMAN, A.J. 2001. Microstructural and geomechanical constraints on fluid flow at the Costa Rica convergent margin, Ocean Drilling Program Leg 170. In: SILVER, E.A., KIMURA, G., SHIPLEY, T.H. (eds) Proceedings of the Ocean Drilling Program, 170 (CD-ROM). Available from; Ocean Drilling Program, Texas A & M University, College Station TX 77845-9547, USA. BROWN, K.M., SAFFER, D.M. & BEKINS, B.A. 2001. Smectite diagenesis, pore-water freshening, and fluid flow at the toe of the Nankai wedge. Earth and Planetary Science Letters, 194, 97-109. BYRNE, T., MALTMAN, A.J., STEPHENSON, E., SOH, W. & KNIPE, R. 1993. Deformation structures and fluid flow in the toe region of the Nankai accretionary prism. Proceedings of the Ocean Drilling Program, Scientific Results, 131, 83-101. FiTTS,T.G. & BROWN, K.M. 1999. Stress-induced smectite dehydration: ramifications for patterns of freshening and fluid expulsion in the N Barbados accretionary wedge. Earth and Planetary Science Letters, 172,179-197. HENRY, P. 2000. Fluid flow at the toe of the Barbados accretionary wedge constrained by thermal, chemical, and hydrogeologic observations and

FAULTS BETWEEN ACTIVE CONVERGING PLATES models. Journal of Geophysical Research, 105, 25855-25872. HILL, I.E., TAIRA, A. AND THE SHIPBOARD SCIENTIFIC PARTY. 1991. Ocean Drilling Program Leg 131, Nankai Trough. Proceedings of the Ocean Drilling Program, Initial Reports, 131, 434 pp. HOUSEN, B.A.,TOBIN, H.J., LABAUME, P., LEITCH, E.C., MALTMAN, A.J., AND THE LEG 156 SCIENCE PARTY. 1996. Strain decoupling across the decollement of the Barbados accretionary prism. Geology, 24, 127-130. KASTNER, M., ELDERFIELD, H. & MARTIN, J.B. 1991. Fluids in convergent margins: what do we know about their composition, origin, role in diagenesis and importance for oceanic chemical fluxes? Philosophical Transactions of the Royal Society, London, A, 335, 243-259. KIMURA, G, SILVER, E., BLUM, P. AND THE SHIPBOARD SCIENTIFIC PARTY. 1997. Proceedings of the Ocean Drilling Program, Initial Reports, 170. LABAUME, P., MALTMAN, A.J., BOLTON, A., TESSIER, D., OGAWA, Y. & TAKIZAWA, S. 1998. Scaly fabrics in sheared clays from the decollement zone of the Barbados accretionary prism. Proceedings of the Ocean Drilling Program, Scientific Results, 156, 59-77. LANCE, S., HENRY, P., LE PICHON, X., LALLEMANT, S., CHAMLEY, H., ROSTEK, F, FAUGERES, J.C., GONTHIER, E. & OLU, K. 1998. Submersible study of mud volcanoes seaward of the Barbados accretionary wedge: sedimentology, structure and rheology. Marine Geology, 145, 255-292. MALTMAN, A.J. & BOLTON, A. in press. A look at the mechanics of sediment mobilization. In: VAN RENSBERGEN, P., HILLIS, R., MALTMAN, A.J. & MORLEY, C. (eds) Subsurface Sediment Mobilization, Geological Society, London, Special Publications, 216, 9-20. MALTMAN, A.J., BYRNE, T, KARIG, D.E. & LALLEMANT, S.J. 1992. Structural evidence from ODP Leg 131 regarding fluid flow in the Nankai Prism, Japan. Earth and Planetary Science Letters, 109, 463-468. MALTMAN, A.J., BYRNE,T, KARIG, D.E. & LALLEMANT, S.J. 1993. Deformation at the toe of an active accretionary prism: synopsis of results from ODP Leg 131, Nankai, SW Japan. Journal of Structural Geology, 15, 949-964. MALTMAN, A., LABAUME, P. & HOUSEN, B. 1998. Structural geology of the decollement at the toe of the Barbados accretionary prism. Proceedings of the Ocean Drilling Program, Scientific Results, 56, 279-292. MARTINEZ, A., MALAVIEILLE I, LALLEMAND S. & COLLOT, J.-Y. 2002. Partition de la deformation dans un prisme d'accretion sedimentaire: approche experimentale. Bulletin de la Societe Geologique de France, 173,17-24. MASCLE, A., MOORE, J.C. & THE SHIPBOARD SCIENTIFIC PARTY. 1988. Proceedings of the Ocean Drilling Program, Initial reports 110. MIKADA, H., BECKER, K., MOORE, J.C., KLAUS, A. ET AL. 2002. Proceedings of the ODP, Initial Reports 196 (CD-ROM). Available from; Ocean Drilling

139

Program, Texas A & M University, College Station TX 77845-9547, U.S.A. MOORE, G.F., TAIRA, A., KLAUS, A., AND THE SHIPBOARD SCIENTIFIC PARTY. 20010. New insights into deformation and fluid flow processes in the Nankai Trough accretionary prism: results of Ocean Drilling Program Leg 190. Geochemistry, Geophysics and Geosystems, 2,1-22. MOORE, GF, TAIRA, A., KLAUS, A. & THE SHIPBOARD SCIENTIFIC PARTY. 20016. Nankai Trough, Japan, Leg 190 of the cruises of the Drilling Vessel JOIDES Resolution, Sites 1173-1178, Initial Reports of the Ocean Drilling Program 190, College Station, TX (Ocean Drilling Program). MOORE, J.C. AND 27 OTHERS. 1995. Abnormal fluid pressures and fault-zone dilation in the Barbados accretionary prism - evidence from logging while drilling. Geology, 23, 605-608. MOORE, J.C., KLAUS, A. & THE LEG 171A SCIENTIFIC PARTY. 1997. Initiation and evolution of fault zones: insights from Barbados accretionary prism logging-while-drilling ODP Leg 171A. JOIDES Journal, 23, 4-7. MOORE, J.C, MASCLE, A., TAYLOR, E. & THE SHIPBOARD SCIENTIFIC PARTY. 1988. Tectonics and hydrogeology of the Northern Barbados Ridge Results from Ocean Drilling Program Leg 110. Bulletin of the Geological Society of America, 100, 1578-1593. MOORE, J.C. AND 19 OTHERS. 19980. Consolidation patterns during initiation and evolution of a plateboundary decollement zone: Northern Barbados accretionary prism. Geology, 26, 811-814. MOORE, J.C, KLAUS, A. & THE SHIPBOARD SCIENTIFIC PARTY. 1998&. Northern Barbados accretionary prism: logging while drilling, Sites 1044-1048. Proceedings of the Ocean Drilling Program, Initial Reports, 171A. MOORE, J.C. & SILVER, E. 2002. Fluid flow in accreting and eroding convergent margins. JOIDES Journal, 28, 91-96. MORGAN, J.K. & KARIG, D.E. 1995. Decollement processes at the Nankai accretionary margin, southeast Japan - propagation, deformation, and dewatering. Journal of Geophysical Research, 100,15221-15231. OCEAN DRILLING PROGRAM. 2003. Leg 205 Preliminary Report: Fluid Flow and Subduction Fluxes across the Costa Rica Convergent Margin: Implications for the Seismogenic Zone and Subduction Factory, http://www-odp.tamu.edu/publications/ prelim/205_prel/205toc.html OLU, K., LANCE, S. & SIBUET, M. ETAL. 1997. Cold seep communities as indicators of fluid expulsion patterns through mud volcanoes seaward of the Barbados accretionary prism. Deep Sea Research, 44, 811-821. RANERO, C.R. & VON HUENE, R. 2000. Subduction erosion along the Middle America convergent margin. Nature, 404, 748-752. SAFFER, D.M. & BEKINS, B.A. 1998. Episodic fluid flow in the Nankai accretionary complex: timescale, geochemistry, flow rates, and fluid budgets. Journal of Geophysical Research, 103, 30351-30370.

140

A. MALTMAN & P. VANNUCCHI

SAFFER, D.M. & BEKINS, B.A. 1999. Fluid budgets at convergent plate margins: implications for the extent and duration of fault-zone dilation. Geology, 27,1095-1098. SAFFER, D.M., SILVER, E. A., FISHER, A.T., TOBIN, H. & MORAN, K. 2000. Inferred pore pressures at the Costa Rica subduction zone: implications for dewatering processes. Earth and Planetary Science Letters, 177,193-207. SAITO, S. & GOLDBERG, D. 2001. Compaction and dewatering processes of the oceanic sediments in the Costa Rica and Barbados subduction zones: estimates from in situ physical property measurements. Earth and Planetary Science Letters, 191, 283-293. SCREATON, E. & GE, S.M. 1997. An assessment of along-strike fluid and heat transport within the Barbados Ridge accretionary complex: results of preliminary modelling. Geophysical Research Letters, 24, 3085-3088. SCREATON, E. & GE, S.M. 2000. Anomalously high porosities in the proto-decollement zone of the Barbados accretionary complex: do they indicate overpressures? Geophysical Research Letters, 27, 1993-1996. SCREATON, E.J., WUTHRICH, D.R. & DREISS, S.J. 1990. Permeabilities, fluid pressures, and flow rates in the Barbados Ridge complex. Journal of Geophysical Research, 95, 8997-9007. SCREATON, E., FISHER, A.T., CARSON, B. & BECKER, K. 1997. Barbados Ridge hydrogeologic tests: implications for fluid migration along an active decollement. Geology, 25, 239-242. SCREATON, E., CARSON, B. & DAVIS, E. 2000. Permeability of a decollement zone: results from a twowell experiment in the Barbados accretionary complex. Journal of Geophysical Research, 105, 21403-21410. SCREATON, E., SAFFER, D., HENRY, P. & HUNZE, S. 2002. Porosity loss within the underthrust sediments of the Nankai accretionary complex: Implications for overpressures. Geology, 30,19-22. SHIPLEY, T.H., MOORE, G.F., BANGS, N.L., MOORE, 1C. & STOFFA, PL. 1994. Seismically inferred dila-

tancy distribution, Northern Barbados Ridge decollement - implications forfluidmigration and fault strength. Geology, 22, 411-414. SHIPLEY, T.H., OGAWA, Y, BLUM, P. & THE SHIPBOARD SCIENTIFIC PARTY. 1995. Northern Barbados Ridge, Leg 156 of the cruises of the Drilling Vessel JOIDES Resolution, Sites 947-949. Initial Reports of the Ocean Drilling Program, 156. SILVER, E., KASTNER, M., FISHER, A., MORRIS, I, MclNTOSH, K. & SAFFER, D. 2000. Fluid flow paths in the Middle America Trench and Costa Rica margin. Geology, 28, 679-682. SUMNER, R.H. & WESTBROOK, G.K. 2001. Mud diapirisrn in front of the Barbados accretionary wedge: the influence of fracture zones and North America-South America plate motions. Marine and Petroleum Geology, 18, 591-613. TAKIZAWA, S. & OGAWA, Y. 1999. Dilatant clayey microstructure in the Barbados decollement zone. Journal of Structural Geology, 21,117-122. TOBIN, H., VANNUCCHI, P. & MESCHEDE, M. 2001. Structure, inferred mechanical properties, and implications for fluid transport in the decollement zone, Costa Rica convergent margin. Geology, 29, 907-910. VANNUCCHI, P. & TOBIN, H. 2000. Deformation structures and implications for fluid flow at the Costa Rica convergent margin, ODP Sites 1040 and 1043, Leg 170. Journal of Structural Geology, 22, 1087-1100. VANNUCCHI, P., SCHOLL, D.W., MESCHEDE, M. & McDouo ALL-REID, K. 2001. Tectonic erosion and consequent collapse of the Pacific margin of Costa Rica: combined implications from ODP Leg 170, seismic offshore data and regional geology of the Nicoya Peninsula. Tectonics, 20, 649-668. VANNUCCHI, P., BETTELI, G. & MALTMAN, A.J. 2003. On the nature of scaly fabric and scaly clay. Journal of Structural Geology, 25, 673-688. VON HUENE, R., RANERO C.R., WEINREBE W. & HINZ K. 2001. Quaternary convergent margin tectonics of Costa Rica, segmentation of the Cocos plate, and Central American volcanism. Tectonics, 19, 314-334.

Contrasting styles of fluid-rock interaction within the West Fissure Zone in northern Chile C. JANSSEN, V. LUDERS & A. HOFFMANN-ROTHE GeoforschungsZentrum Potsdam, Telegrafenberg, 14473 Potsdam, Germany (e-mail: jans@gfz-potsdam. de) Abstract: Geochemical data from three profiles crossing the West Fissure Zone in northern Chile were used to describe the chemical effects of fluids on faulting processes. The results document considerable differences of fluid-rock interactions between the profiles (A-C). Within the area of profile A and C fluid activities have neither led to intensive exchange reactions between fluids and rocks nor to notable changes in whole rock and mineral composition of fault rocks relative to their undeformed host rock. Investigations of stable isotopes (813C, 818O) and fluid inclusions indicate infiltration of predominantly descending (meteoric) fluids and only subordinate involvement of ascending hydrothermal fluids. In the case of profile B, fluid-enhanced weakening mechanisms are dominant. Dissolution transfer has led to the formation of an alteration zone up to 400 m wide. Along profile B we have no indications for warm fluids derived from a deep source. Rather, a number of lines of evidence (e.g. 818O and 8D) show that fault rock alteration took place due to the infiltration of low temperature meteoric water. For this profile, all geochemical data indicate an open fluid system. Moreover, the chemical variations across the West Fissure Zone agree well with the structural variability observed at this traverse. The application of a variety of geochemical analyses of fault rocks shows the heterogeneity of largescale continental fault zones.

The importance of fluids in a variety of faulting processes and the role that they play in the earthquake cycle is widely accepted (e.g. Sibson etal. 1975; Chester & Logan 1986; Blanpied etal. 1992; Schulz & Evans 1998). Especially, the hypothesis of the weakness of some major strike-slip faults has led to a long-standing discussion concerning the effects of fluids on fault mechanisms. One possible interpretation is the notion that fluids are active in reducing the strength of faults by elevating fluid pressure (e.g. Rice 1992; Byerlee 1993). However, the chemical effects of fluids on fault zone rheology are probably as important as their mechanical role (e.g. see Goddard & Evans 1995; Wintsch et al 1995; Kharaka et al. 1999; Imber et al 2001). For example, the fluid reacts with the rocks producing mineral dissolution and/or precipitation reactions (Sausse etal. 2001). In particular, solution, mass transfer and cementation lead to a repeated destruction and creation of fault zone permeability and implies fault zone 'weakening' and 'strengthening' processes (e.g. Sleep & Blanpied 1992; Gray etal 1999). The infiltration of meteoric water during brittle deformation and resulting hydrothermal alteration processes have been reported for several fault zones (e.g. O'Hara 1988; Goddard & Evans 1995; Janssen et al 1997). Substantial progress in solving fluid-related problems has

predominantly been made by geochemical investigations of the San Andreas fault in California (e.g. Chester et al 1993; Evans & Chester 1995; Schulz & Evans 1998). However, in spite of the advanced understanding of the mechanical and chemical effects of fluids on faulting processes uncertainties still exist regarding the origin and distribution of fluids, the role of fluids in the mechanics of faulting and the maintenance of high fluid pressure. Our study aims to examine the role of fluid-rock interactions in faulting deformation using geochemical investigations on fault-rock samples which represent different segments of the West Fissure Zone (WFZ) in northern Chile (Fig. 1). The WFZ is an exhumed trace of the Precordilleran Fault System and represents excellent general conditions for geochemical studies. Some of the largest porphyry copper deposits in the world, e.g. Chuquicamata, Radimiro Tomic and the mines of Colahuasi, are related to the West Fissure. Numerous geochemical investigations of mineralization and alteration processes provide an unique opportunity to obtain information about the fluid histories (e.g. Reynolds etal 1998; Ossandon et al 2001). Particularly, the evidence of fluid flow from different fluid sources (high temperature magmatic fluids (>300 °C), hydrothermal fluids (around 100 °C) with magmatic and meteoric components, and shallow

From: ALSOP, G. L, HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224,141-160. 0305-8719/$15.00 © The Geological Society of London 2004.

142

C.JANSSEN£rAL.

Fig. 1. Overview sketch of Northern Chile showing morphological units and the West Fissure Zone (WFZ), according to Reutter et al. (1996). The profile locations are indicated by rectangular areas.

water of meteoric origin), inspired a more detailed investigation about the role of these different fluids in faulting processes. In two previous studies (Hoffmann-Rothe 2002; Janssen et al. 2002), geological fieldwork and audiomagnetotelluric ( AMT) data from two profiles crossing the WFZ were used to describe the geometry and structure of the fracture damaged zone surrounding this strike-slip fault, as well as the subsurface electrical conductivity structure. We now report about geochemical studies, which have been carried out along the same profiles. Primary goal is to describe the source and composition of fluids and the variation of fluid-rock interaction along the selected segments. Based on these investigations, we compare our results with observations from the well studied San Andreas Fault.

Geological setting The Chilean Precordillera is structurally characterized as a belt of crustal shortening leading to

the upfolding of basement anticlines and common reverse faults on their flanks (Reutter et al. 1991). Oblique subduction of the Pacific Nazca plate below South America led to the formation of the Precordilleran Fault System (PFS), consisting of several orogen-parallel faults (Hollister 1978). The Precordillera is composed of Palaeozoic basement rocks of different ages (Ordovician schists and amphibolites, Late Carboniferous and Permian granites, Carboniferous and Devonian sediments). Basement rocks are overlain by Late Carboniferous to Triassic volcanics and sediments (Reutter et al. 1996). The northern part of the PFS between 23° S and 20° S is known as WFZ. The WFZ can be traced along 170 km almost continuously from Calama towards the north beyond Quebrada Guatacondo (Fig. 1; Dilles etal 1997). Fault-related deformation initiated during the Late Eocene/Early Oligocene arc-normal shortening, contemporaneous with magmatisms and vein mineralization (Incaic Phase, 38 Ma; Reutter et al. 1996). The Incaic Tectonic phase

STYLES OF FLUID-ROCK INTERACTION

143

Sampling and analytical methods Geochemical sampling was undertaken along three selected profiles shown in Fig. 2: (1) Profile A in the Guatacondo region at 20°54'S (Fig. 3a); (2) Profile B in the area north of Chuquicamata at22°05'S(Fig.3b);and (3) Profile C in the Limon Verde region at 22°40'S (Fig. 3c). Different exposure conditions along these profiles avert the application of the entire spectrum of geochemical investigations at all locations; Fig. 2 and Table 1 state the applied methods for each traverse. Samples of veins and surrounding host rocks were used for analysis of stable isotopes and rare-earth elements (REE; profile A). X-ray fluorescence analyses of major and minor elements, stable isotope analyses (613C, 818O and 5D) and REE analyses were performed on sediments and igneous rocks along profiles B andC. The X-ray analyses were conducted at the GeoforschungsZentrum Potsdam. The results have an accuracy of ±10%. REE data were obtained by Inductively Coupled Plasma Mass Spectrometry (ICPMS). Precision and accuracy of the ICPMS data are better than ±10%. Stable isotope analyses were performed at the AlfredWegner-Institut in Potsdam and the University of Tubingen, using standard techniques (McCrea 1950; Vennemann & O'Neil 1993). The 618O and 5D values refer to Vienna-standard mean ocean water (SMOW), the 813C values to Cretaceous Pee Dee Belemnite (PDB). The precision is ±0.1%o or better for carbon, ±0.2%o or better for oxygen and 2%0 or better for hydrogen. Fluid inclusions were studied in quartz and calcite samples from veins (profile A and C) using a USGS heating/freezing system with a transmitted light microscope. For calibration, synthetic fluid inclusion standards were used. Fig. 2. Landsat TM scene showing the WFZ with the position of the three profiles, the stratigraphy of rock samples and the geochemical methods used uTM coordinates.

led to dextral strike-slip displacement followed by sinistral shear (31-18 Ma). The youngest event is the reactivation of the fault in a phase of dextral slip. The varying sense of displacement is commonly related to modifications in convergence rate and angle of the Nazca plate with respect to the continental South American plate (Lindsay etal 1995).

Description of fault rocks We briefly summarize the mesoscopic and microscopic structures of fault rocks that are relevant to fluid-rock interactions in faulting deformation of the WFZ. A more detailed description can be found in Hoffmann-Rothe (2002) and Janssen et al (2002). Profile A Profile A is located along the E-W trending Quebrada (valley) Guatacondo within the

144

C.JANSSENETAL.

Fig. 3. Geological maps of the areas of detailed investigations, (a) Quebrada Guatacondo (slightly modified after Carrasco et al. 1999). (b) Gorila region (slightly modified after Hoffmann-Rothe 2002). (c) Limon Verde area (slightly modified after Boric et al 1990).

STYLES OF FLUID-ROCK INTERACTION

145

formation at temperatures between 150-300 °C. The limestone host contains pellets, fossils and some non-carbonate detrital grains. Fracturing and pressure solution are the dominant deformation mechanisms (Fig. 4b).

Profile B Profile B is located north of the CalamaChuquicamata region (Fig. 3b). On aerial photographs a single major fault structure is visible which is characterized by a zone of extensive chemical alteration of the fault rocks. The fault separates an Eocene quartziferous monzodiorite from Triassic andesites and dacites. Following the terminology of Rawling et al (2001), three architectural elements are distinguished:

Fig. 3. continued.

Precordillera (Fig. 3a). The fault separates Mesozoic sediments from Tertiary volcanic rocks (Carrasco et al 1999). The AMT imaging shows a narrow zone of low electrical resistivity (100 m width, 50 m deep) which matches the fault trace (Janssen et al 2002). The fault core is not exposed. The fault-related deformation has resulted in the formation of fractured rocks as well as calcite veins (hosted in Late CretaceousPaleocene limestones). The number of these veins progressively increases towards the fault trace, suggesting their fault-related origin. This is further emphasized by their sub-parallel orientation to the fault (NE-SW), indicating their formation in a NW-SE extensional regime imposed by dextral shearing during the Incaic tectonic phase (Lindsay et al 1995; Reutter et al 1996). Veins are filled with fibrous and euhedral blocky calcite. The vein filling is generally twinned. Twin lamellae are thick (>5 um) and straight (Fig. 4a). According to their appearance, they are classified as type II twins (Burkhard 1993), suggesting considerable de-

(1) The undeformed and unaltered protolith, which comprises monzodiorites and andesites. The monzodiorite composition using X-ray diffraction measurements is approximately plagioclase (50%), orthoclase (19%), quartz (16%), hornblende (8%), diopside (4%) and chlorite (<2%) with accessory ilmenite (<2%). (2) The damaged zone is about 1000 m wide and contains small faults; mineral-filled veins are rare (Fig. 4c). (3) The fault core is composed of intensely altered rocks (Fig. 4d). This unlithified granular material marks a 400 m wide zone that coincides with a zone of enhanced electrical conductivity (Hoffmann-Rothe 2002). The composition of the very fine-grained alteration zone matrix (from X-ray diffraction measurements) is made up of quartz (32%), plagioclase (28%), orthoclase (24%), muscovite (12%), diopside (2%) and chlorite-and-clay minerals (2%). Some less altered blocks (relics) of rock within the alteration zone show brittle deformation features with brecciated to cataclastic fabric, similar to the damaged zone fabric. The wide zone of increased alteration processes allows the application of a broad range of geochemical studies. Therefore, whole-rock X-ray fluorescence analysis, REE and stable isotope determinations were applied to characterize the composition of fault-related rock samples from a 1500m traverse across the WFZ (Table 1; Fig. 3b).

Profile C Profile C is situated near Calama where the fault splits into two branches (Limon Verde Fault

C.3ANSSENETAL.

146

Table 1. Sampling profiles Deformation

Sample No. Distance to fault (m)

Lithology

Profile A Al A2 A3 A4 A5 A6 A7 A8 A9 A10

L,h+v L,h+v L,h+v L,h+v L,h+v L,h+v L,h+v Q,v L,h+v L,h+v

fr fr uf uf uf fr fr fr fr

100 100 500 500 500 75 75 0 120 120

Profile B Bl B2 B3 B4 B5 B6 B7 B8 B9 BIO Bll B12 B13 B14 B15 B16 B17 B18 B19 B20 B21 B22 B23 B24 B25 B26 B27 B28 B29 B30 B31 B32

70 300 300 270 270 240 220 200 200 200 125 80 80 50 0 0 70 70 170 260 350 375 430 580 600 (minor fault) 600 (minor fault) 100 25 330 out of profile out of profile out of profile

Qz Da (dyke) Da (dyke) L L L L/Da Da Da Da Da Da Da Da MoD? Fe-C MoD? MoD? MoD MoD MoD MoD MoD MoD MoD MoD MoD H2O Ba Granite Granite Granite

fr fr fr uf fr uf fr-a fr fr fr/uf uf fr fr-a fr fr-a fr-a fr-a fr-a fr-a fr fr fr fr uf fr-a fr-a fr fr fr fr

Profile C Cl C2 C3 C4 C5 C6

Distance to fault (cm) 20 10 0 20 out of profile out of profile

Granite Granite Granite Granite Sandstone Sandstone

uf fr fr-a uf fr fr

XRF

Stable isotpes

REE

FE

X X

X X X X X

X

X X X X X X

X X

X X X X X X X X

X

X X X X X X X X

X X X X X X X X X X X X X X X X

X

X

X

X X X X X X X X X X X

X

X X X X

X X X X X X X X X X X X X X X X X

X X X

X X X X X X X

X X X X X X

Note: Samples from Profile B were not suitable for fluid inclusions studies. Abreviations: L, limestone; Q, quartz; Qz, quartzite; Da, dacite; Fe-C, ferrous carbonate; MoD, monzodiorite; H2O, water; Ba, barite; h, host rock; v, vein; fr, fractured rock; fr-a, fractured rock-altered; uf, unfaulted.

STYLES OF FLUID-ROCK INTERACTION

147

Fig. 4. Photographs of fault-related deformation structures from the Guatacondo region (a-b) and Gorila region (c-d). (a) Photomicrograph of intersecting deformation twins in calcite vein. A dark pressure solution seam marks the boundary between twinned vein cement and unchanged limestone matrix (scale bar 100 um). (b) Brecciated limestone showing vein cement with fractures and a stylolitic seam (scale bar 100 um). (c) Exposure close to the fault trace with several minor faults (left: E; right W). (d) Fault core with unconsolidated alteration zone material (hammer: 33 cm).

Zone and WFZ; Fig. 3c). The main fault branch of the WFZ cuts through the Limon Verde basement block consisting of mainly granite-diorite with abundant gabbro and metamorphic rocks (Lucassen et al. 1999). The fault core is marked by thin shear zones, ranging from 1 m to greater than 10 m in length. These minor shear zones are clearly distinguished from the unfaulted granite (protolith) by an overall increase in fracturing and by yellow-brown colour due to alteration processes. The contact between fractured host rock (fracture spacing > 20 cm) and the alteration zone is extremely sharp and distinct. The alteration zone, ranging in thickness between 10-20 cm, consists of rocks, which are friable and highly fractured at orientations parallel to the minor fault. The subrounded to subangular fragment sizes range from centimetres to millimetres. Alteration processes have caused a

reduction in fragment size. Under the microscope, fracturing and pressure solution are the dominant deformation mechanisms in the cataclastic zone. Indications for syntectonic recrystallization are absent.

Geochemistry of fault rocks Profile A Fluid inclusions. Fluid inclusions occur in wafers of vein calcites from Profile A, but are rarely present. When present, they often show rounded or rhombohedral shapes, but their sizes seldom exceed 5 um. Therefore, only a few combined data of salinity and temperature of homogenization were obtained, from two samples of calcites. Fluid inclusions of primary origin

148

C. JANSSEN ETAL.

clathrate or ice could not be detected in these inclusions. All measured carbonic inclusions decrepitated prior to reaching total homogenization upon heating. Therefore, these inclusions are not suitable for pressure estimations. The final ice melting temperatures of primary two-phase aqueous fluid inclusions in both samples (Fig. 5a) indicate that calcite precipitated from low-salinity fluids (0.2-4.5 wt % NaCl equivalent). The homogenization temperatures fall into a range between 163.7 and 180.2 °C (sample A10 in Fig. 5b), and 219.3 and 225.4 °C (sample A4 in Fig. 5b). A comparison between the final ice melting temperatures and homogenization temperatures of both samples indicates that sample A10 precipitated from less saline fluids than sample A4, at somewhat lower temperatures (-50 °C; Fig. 5a, b), suggesting a possible time gap (e.g. due to different faulting events?).

Fig. 5. Histograms showing (a) ice melting temperature (profiles A and C). (b) homogenization temperature for calcite (Profile A), (c) homogenization temperature for quartz (Profile C).

(Roedder 1984) were found in samples A4 and A10 (Table 1; Fig. 5). Primary two-phase aqueous fluid inclusions in milky calcite are arranged along growth zones or parallel to crystal planes. In one sample, (A4) a few H2O-CO2 inclusions were observed arranged parallel to a cleavage plane. They show melting of solid CO2 at temperatures of about -70 °C and homogenization of the carbonic phases with the liquid between 16.5 and 17 °C. The melting temperatures of solid CO2 lie well below the triple point of pure CO2 (-56.6 °C) indicating additional components such as CH4 and/or N2. Unfortunately, confirmation by Laser-Raman Spectroscopy was not possible due to the high fluorescence of the host mineral. Melting of

$3C vs. $8O relationships in calcite. Carbon and oxygen isotopic compositions were determined in host rocks and vein calcites (Table 2; Fig. 6). The carbon and oxygen isotopic compositions of host rocks and veins are highly variable (Fig. 6a). For host rock samples, the 518O values range between 17.5 and 22%0 when normalized to SMOW with a mean of 20%o whereas 813C values show a larger variation (0.5 to -13%o PDB) with a mean of -4% (Fig. 6a). The 613C and 818O values of vein calcite plot in a similar field to that of the host rocks but with more scatter. However, one group of vein calcites (samples Al, A2, A6, A9, A10) show only small variation in their -813C values (-1 to -2.3%o) but a significant variation in 618O (15.6-20.3%o). Carbon and oxygen isotopic compositions in calcite grown from the same fluid are dependent on temperature and the various carbon species in solution. If no organic compounds are present in the fluid, only pH and temperature can control the speciation of HCO3~ and H2CO3apparant (ap). Carbon isotopic fraetionation between calcite and HCO3~ decreases between 300 and 200 °C, and becomes nearly constant below 200 °C (Ohmoto & Rye 1979). On the other hand, at temperatures below approximately 300 °C, isotopic fractionation between calcite and H2CO3ap (including dissolved H2CO3 and CO2) increases with decreasing temperature. Oxygen isotopic fractionation increases with decreasing temperature in all relevant fluid-mineral systems (O'Neil et al. 1969; Clayton et al. 1972). Thus, if the dissolved carbon in the fluid is H2CO3, calcite which is precipitated from the fluid at temperatures below 300 °C would show a positive correlation between 613C and 518O values simply due to

STYLES OF FLUID-ROCK INTERACTION

149

Table 2. Isotope data Sample Profile A No. Al A2 A4 A5 A6 A7 A9 A10 Sample Profile B No. B2 B8 Bll B13 B15 B16 B17 B18 B19 B20 B21 B23 B24 B27

813C PDB

818O SMOW

host 0.55 -0.93 -12.72 -12.94 0.03 -1.74

host 21.59 17.96 21.52 21.52 20.94 22.10

-4.34

8D SMOW host 13.0 9.9 11.8 11.0 13.9 7.9 10.2 9.7 10.6 7.9 7.4 7.2 7.2 9.1

17.50

813C PDB

818O SMOW

vein -1.02 -2.33 -24.46 -15.36 -1.40 6.43 -1.95 -1.61

vein 15.57 16.96 22.91 21.68 18.87 23.30 20.29 18.94

18

0 SMOW host -92 -89 -84 -76 -100 -71 -109 -96 -100 -89 -95 -103 -93 -84

818O and 813C of host rock and veins from the limestone sequence (Profile A). 818O and 8D of host rock (monzodiorite) and fault-related rocks (monzodiorite; Profile B).

temperature effect as observed in many hydrothermal vein-type deposits (e.g. Ltiders & Moller 1992). The observed trend of small carbon isotopic fractionation and larger oxygen isotopic fractionation in most of the vein calcites therefore requires HCO3~, as the dominant dissolved carbon species in the fluid and indicates formation temperatures of less than 200 °C (Ohmotho & Rye 1979). This temperature estimation of vein calcite formation is also supported by fluid inclusion data derived from sample A10. Therefore, it seems plausible that calcite precipitated due to changes of P-T conditions from formation waters or fluids that interacted close to the exchange equilibrium with the host rocks prior to their ascent in higher structural levels. Conversely, vein calcites (A4 and A5) which show a larger fractionation in 813C values (-24.5 and -15.4%o) but small variation in 818O may have precipitated by mixing of a deep-seated crustal carbon with a limestone carbon. On the other hand, the variations in 613C and S18O can also be explained by the incorporation of

organic matter in the host rocks and vein calcites, respectively. This assumption is supported by the presence of carbonic fluid inclusions probably with considerable amounts of CH4 hosted in one of these vein calcite samples (A4). In contrast to other fault zones (e.g. see Kerrich & Kyser 1994) a depletion in 818O of the vein calcite with respect to the oxygen isotopic composition of the host rocks was not observed in the studied samples. However, the clear differences in 513C and 818O isotopic compositions between the vein calcites and their surrounding host rocks (Fig. 6b) excludes the possibility that the isotopic composition of the fluid is controlled by exchange with the surrounding host rocks (rock-buffered system; Gray etal. 1991). Trace elements. Trace elements including REE and yttrium (REY) were analysed with the same powders used for determination of isotopic ratios. The results are given in Table 3 and Fig. 7a. High contents of elements typically related

150

C. JANSSEN£TAL.

rocks (Fig. 7b) shows similar features for all couplets. The similarity between the REYSN distribution in the vein calcites and their respective limestone hosts suggests that Y and REEs have been derived from a source similar to the local detritus-rich marine limestone. However, the detritus-rich marine limestone is common in deeper lying Mesozoic sequences of the sedimentary pile. A lower Y, REE, Rb, Cs, and Th content in the vein calcites compared to the limestone (Table 3) indicates partial dissolution of the limestone. The high similarity of the individual REYSN patterns may infer that, in most cases, the vein calcite-forming fluids have closely achieved exchange equilibrium conditions with detritus-rich marine limestone. The vein calcites either precipitated from formation water or from REY-poor meteoric- or seawater that penetrated the marine detritus-rich limestone. In the latter case, long-term water-rock interaction under closed system conditions was necessary to equilibrate with the limestone. In any case, the absence of positive Eu anomalies in the REYSN patterns of the vein calcites indicates that mineral precipitation did not occur from fluids that have experienced temperatures above 250 °C (Bau & Moller 1992), and precludes fluids being derived from extraneous sources at great depth. This is consistent with the results of fluid inclusion studies as well as with the stable isotope (C, O) data.

Profile B Fig. 6. Diagrams of stable isotope data, (a) Crossplot of the 818O and 813C of the host rock and calcite veins of profile A. The arrows relate each calcite vein cement to its respective host rock, (b) Differences in S18O and 813C between the veins and related host rocks are plotted on the x- and y-axes respectively.

to aluminosilicates (Rb, Cs, and Th) were measured in all samples, with the exception of sample Al-V (V denotes vein, H stands for host). This indicates that considerable amounts of clastic detritus are present in these samples. In spite of the shale-normalized (SN) REE and Y (REYSN) patterns of detritus-rich limestones, negative CeSN anomalies, and positive anomalies for LaSN, and YSN, in samples A3-V, A4-V, Al-H, Al-V, A2-H, A2-V and A5-V are characteristic of marine limestones precipitated from ambient seawater (Bau et al. 1995). A comparison of REYSN distributions in the vein calcites with those of their corresponding limestone host

Major and trace elements. Geochemical analyses of the selected fault rocks for major, and trace elements are shown in Table 4. The chemical compositions of samples vary considerably between unfaulted host rock, damaged zone and fault core (alteration zone). In order to examine the spatial variability of the geochemical signatures across the fault, whole rock analyses along the eastern half-profile, the area entirely composed of monzodiorite, have been undertaken. The samples are plotted with respect to distance across the fault (Fig. 8). The traverse, trending at right-angles to the fault trace towards the undeformed protolith in the east, shows clear evidence for a chemically defined fault core. For example, the alkaline earth elements (Ca, Mg, Sr, and Ba) are clearly depleted within the fault core while SiO2 and Zr show some enrichment. A12O3, and Fe2O3, on the other hand, show broad bands with no distinct trend. A different way to present the chemical data is the isocon diagram of Grant (1986) using the GRESENS computer program (Gresens 1967;

Table 3. Concentration of selected trace elements in calcite veins (V) and their limstone host (profile A) Ba ppm

La ppm

Ce ppm

Pr ppm

Nd ppm

Sm ppm

Eu ppm

Gd ppm

Tb ppm

95.8 336 173 2682 140 228 81.1 235 3859 188 230 4249 80.5 2103

10.6 2.45 12.2 2.94 9.26 11.4 9.29 11.9 8.62 4.92 17.6 23.3 14 4.47

10.1 2.67 16.8 4.93 19.4 16.7 20.5 21.6 17.3 6.97 36.6 48.5 31.2 9.72

1.68 0.48 3.06 1.08 2.4 1.9 2.79 2.75 2.32 1.41 4.72 5.7 4.17 1.21

6.34 1.99 12.8 5.33 9.1 7.1 10.8 10.5 9.38 5.88 17.8 21.3 16.5 4.46

1.2 0.41 3.03 1.68 1.84 1.27 2.35 1.74 2.03 1.42 3.42 3.72 3.23 0.78

0.33 0.13 0.75 0.1 0.48 0.36 0.59 0.47 0.1 0.42 0.83 0.1 0.82 0.1

1.55 0.59 3.42 2.14 1.78 1.44 2.25 1.78 2.18 1.85 2.86 2.83 2.64 0.66

0.22 0.08 0.45 0.33 0.25 0.18 0.29 0.2 0.27 0.24 0.38 0.31 0.34 0.08

Yb ppm

Lu ppm

Hf ppm

Pb ppm

Th ppm

U ppm

Y/Ho ppm

Zr/Hf ppm

Th/U ppm

0.58 0.18 1.04 1 0.47 0.25 0.5 0.25 0.388 0.49 0.55 0.36 0.56 0.14

0.08 0.03 0.15 0.15 0.07 0.03 0.07 0.04 0.058 0.07 0.08 0.05 0.08 0.02

<0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5

24.7 2.79 80.6 24.8 183 48.9 12.5 3.22 8.44 7.5 1.48 7.25 3.33 2.11

0.47 0.03 2.34 0.66 1.59 0.18 1.89 0.11 1.65 0.47 1.65 0.82 1.41 0.27

1.13 0.25 0.48 0.21 0.16 0.05 0.17 0.08 0.235 0.11 0.17 0.27 0.21 0.14

48.5 58.7 39.3 47.2 32.8 39 30.9 37 36.5 42.2 28.1 28.6 27.4 30.1

Samples

Rb ppm

Sr ppm

Y ppm

Zr ppm

Cs ppm

Al-H Al-V A2-H A2-V A3-H A3-V A4-H A4-V A5-H A5-V A9-H A9-V A10-H A10-V

2.31 0.05 18.6 2.89 15 1.28 18.1 1.46 8.59 2.13 19.3 21 31.5 12.8

336 352 136 388 82.1 135 101 71.8 209 104 51.3 187 59 143

13.5 6.3 19.4 20.9 7.72 6.64 8 6.75 8.95 10.4 8.63 6.53 8.11 2.13

<1 <1 6 1.4 3 <1 6 <1 1.9 <1 3 1.3 5 <1

0.43 0.02 3.37 0.83 2.92 0.33 4.03 1.92 1.43 1.96 6.18 7.75 14.5 5.18

Samples

Dy ppm

Ho ppm

Er ppm

Tm ppm

Al-H Al-V A2-H A2-V A3-H A3-V A4-H A4-V A5-H A5-V A9-H A9-V A10-H A10-V

1.32 0.519 2.57 2.05 1.31 0.91 1.52 1.04 1.37 1.32 1.86 1.45 1.73 0.43

0.28 0.11 0.49 0.44 0.24 0.17 0.26 0.18 0.245 0.25 0.31 0.23 0.3 0.07

0.79 0.29 1.33 1.25 0.6 0.41 0.66 0.44 0.587 0.64 0.73 0.52 0.74 0.19

0.1 0.04 0.17 0.17 0.08 0.05 0.08 0.04 0.07 0.08 0.09 0.06 0.1 0.02

0.41 0.11 4.88 3.2 9.72 3.5 11.26 1.38 7.03 4.46 9.82 3.05 6.64 2.02

152

CJANSSEN^r^L.

Fig. 7. continued.

Potdevin 1993; Fig. 9). In this diagram element concentrations in the altered rock (fault core) are compared to concentrations in the original rock (protolith). Gresens (1967) reasons that distinct elements are likely to be immobile during the alteration process. Based on numerous geochemical studies (e.g. O'Hara 1988; Goddard & Evens 1995), TiO2 was assumed to be the immobile constituent within the faultzone rocks. The constructed immobile element isocon (solid line which runs through both the TiO2 data point and the origin of the isocon diagram in Fig. 9) defines the volume change between host rock and fault core. This volume change can be calculated from the following equation (O'Hara 1988): Fig. 7. Diagrams of trace element distribution, (a) Shale-normalized rare-earths and ytrium (REY) patterns of vein calcites (V) and their respective limestone hosts (H). (b) Host-normalized REY patterns of vein calcites (calculated as REYvein ealcites/REYlimestone host).

where Cg represents the concentration of the element in the fault core, C0 is the concentration in the original (host) rock, and V represents the volume change. From the relative depletion of the immobile TiO2 in the fault core versus the protolith (Fig. 9), we estimate a volume gain of about 68% for the fault core (see Evans &

STYLES OF FLUID-ROCK INTERACTION

153

Table 4. Whole rock analysis of fault rocks from Profile B (oxides and trace elements) oxides

SiO2

B2 B3 B4 B5 B6 B7 B8 B9 Bll B12 B13 B14 B15 B16 B17 B18 B19 B20 B21 B22 B23

TiO2 A12O3 Fe2O3 MnO

%

%

%

%

0.62 0.48 0.12 0.09 0.13 0.58 0.63 0.59 0.62 0.53 0.52 0.83 0.63 0.79 0.37 0.38 0.69 0.46 0.61 0.78 0.97 0.99

0.11

0.54

9.89 10.44 1.99 2.24 3.35 14.19 15.04 14.27 14.56 13.41 12.63 16.2 14.67 12.83 13.17 14.57 15.43 14.91 15.43 15.47 15.95 16.07 16.08 15.69 5.75 14.34

2.72 3.13 0.38 0.38 0.52

B25 B26 B27

69.19 58.05 18.48 13.67 13.82 42.05 65.01 64.73 63.35 64.96 65.06 54.4 61.64 56.94 69.29 71.6 60.84 68.25 65.25 64.07 60.92 59.85 59.58 58.65 22.3 66.9

4.53 5.56 5.02 5.22 5.76 5.47 5.05 21.27 6.01 2.08 10.72 3.28 4.51 4.92 6.38 6.46 6.45 6.27 2.03 4.93

trace element

Rb ppm

Sr ppm

Y ppm

B2 B3 B4 B5 B6 B7 B8 B9 Bll B12 B13 B14 B15 B16 B17 B18 B19 B20 B21 B22 B23 B24 B25 B26 B27

97.6 28.6 18.7

231 168 242 225 261 553 165 162 175 256 253 111 208 531 162 227 418 281 394 348 397 498 578 168

31.2 26.6 10.2

B24a B24b

%

4.7

33.1

157 170 175 207 120 110 237 160

77.1

181 219 215 172 144 189 186 163 141

50.6

183

1

0.92

0.3

1022

9.4

11.5 21.5 22.1 25.7 23.2 38.2 36.2 22.6 24.8

18

19.5 28.2

21

14.5 17.4 25.5 23.5 21.9 23.5 11.2 9.69

4.6

0.1

0.02 0.02 0.05 0.23 0.16 0.15 0.16 0.14 0.15 0.15 0.13 0.01 0.03 0.01 0.02 0.05 0.05 0.07 0.09

CaO %

Na2O

K2O %

P205

H2O %

CO2

1.87 1.17 0.33

7.04 14.4 43.69 46.6 44.57

0.59 0.23 0.12 <0.1 0.16 0.64 2.94 1.95 0.41 2.22 2.38 0.31 1.43 1.14 2.97 3.48 0.52

2.91 1.14

0.34 0.27 0.13 0.11

1.96 3.98 1.03 1.42 0.79 3.19 2.12 2.46 2.04 2.06 1.59 3.35 3.98

2.24

2.65 99.89 6.45 99.84 33.29 100.08 35.26 100.22 34.76 99.84 10.51 96.85 0.06 99.86 0.08 99.94 1.18 99.94 1.23 99.85 1.93 99.97 4.78 99.92 1.52 99.9 0.06 99.13 0.25 99.81 0.07 99.94 99.83 0.4 0.07 99.92 0.09 99.91 0.07 99.98 0.06 99.85 0.14 99.82 0.14 99.54 0.43 99.8 26.55 99.74 0.52 99.23

Eu ppm

Zn ppm

Cr ppm

1.3

22 20 <10 <10 <10 117 176 123 67 71 55 85 72 24 101 34 139 81 25 45 61 67 71 205 27

51 29 <10 <10 <10 59 31 17 13 10 41 97 20 54 <10 <10 <10 <10 <10 10 17 26 23 <10 <10

0.3

0.49 1.08 2.71 2.46 1.64 1.33 1.36 1.92 1.73

0.5 0.6

15

1.98 2.97 5.37

6.3 6.7

7.77

5.6

0.39 0.55 0.15 0.75 1.87 3.43 3.47

%

0

0.71 1.52 1.63 2.01 1.98 2.82 3.03 3.01 2.57 1.63 0.78

4.69 4.72 36.02 0.95

3.32 3.31 3.19 3.28 3.29 2.56 0.71 3.39

Zr ppm

Cs ppm

Ba ppm

La ppm

284 201

480 4.97 164 10.5 505 2.04 105 3.86 669 2.63 20.1 11179 5.21 823 844 3.69 6.35 2323 1290 10

31.5 24.8 12.2 11.6 15.2 28.1 25.5 27.1 21.2 31.1 33.9 21.3 30.5 1.73 23.1 37.5 31.2 27.4 26.5 31.7 31.7 32.6 32.4 17.8 31.6

66.2 61.9 39.3

118 110 106 118 215 216 155 128 151 209 208 154

91.9

126 119

71.8 58.2 49.2 47.4 100.1

0.1 0.1 0.1

0.08

4.6

29.7 16.2 12.3 4.87 6.06 19.4 9.07 6.28 10.9 11.7

12

9.47 1.48

5.4

Sum %

MgO %

858 811 710 207 488 540 352 507 727 645 712 550 676 234 689

4.4 4.7

3.1

0.5

0.15

1.1

4.59 4.51 4.58 5.46 2.29 1.72 4.52 3.33 1.56 4.66 5.37 3.91 4.31 3.76 4.24 3.62 3.21

%

0.1

0.18 0.16 0.15 0.16 0.17 0.16 0.23 0.18 0.15 0.14 0.11 0.29 0.15

0.2 0.2

4.4

0.25 0.26 0.26 0.28 0.13 0.23

Ce ppm

U ppm

Sm ppm

61 46

4.21

14

2.42

5.92 4.45 1.84 1.65 1.87 4.75 4.61 4.89 4.25

14.5 14.3

44.1 55.9 56.8 46.8 60.7 68.2 46.4 50.7 4.16 47.4 79.6 63.1 57.2 57.2 70.1 69.8 69.3 66.2 26.5 63.2

3.2

3.73 1.53

3.5

1.67

1.6

6.5

2.94 2.75 2.58 2.51 2.23 2.69 3.25 2.99 9.99

6.8

16.2 4.14 4.05 2.18 1.79 4.59 5.12

1.5

5.39

6.3

7.22 4.48 4.83 1.35 3.77 5.84 5.08 4.36 5.09 6.55 6.53 6.02 6.24 2.38 4.09

3.5

1.77 1.39 4.75 1.82 1.26

1.4

1.19 1.73 1.73

3.9 2.7

0.956 0.406 0.328 0.53 1.61

0.9

1.07 0.903 1.73 1.69 1.15 1.03 0.465 0.551 0.764 0.887 0.74 0.969 0.97 1.21

1.2

1.26 0.503 0.625

%

154

C. JANSSEN ET AL.

Fig. 8. Whole-rock analyses of selected oxides (%) and trace elements (ppm) from monzodiorite samples are plotted with respect to distance across the fault.

Chester 1995; Zulauf et al 1999). Elements plotting above the immobile line, e.g. SiO2, K2O, Na2O, A12O3, Fe2O3, Ba, Zr, U are supplied to the system whereas the elements plotting below the immobile line, e.g. CaO, MgO, MnO, Zn, Cr have been removed from the system. ff8O vs. 8D relationships in monzodiorite. Stable isotope data for 14 whole rock samples are presented in Table 2. The 818O values range from around 9 to 14%o SMOW and the 6D values vary between -71 and -109%o SMOW. The graphical representation of these data in Fig. 10 shows

several trends. The oxygen and hydrogen isotopic compositions of fresh monzodiorites plot into or close to the box of primary magmatic water (Hoefs 1987). Less altered monzodiorites (chlorite content < 2%) are enriched in hydrogen but not in oxygen compared with fresh samples, suggesting low water-rock interaction with high-latitude meteoric water (trend 1; Taylor 1978). In contrast, the D/H and 18O/16O ratios of extremely altered fault core samples are significantly different from those of unaltered monzodiorites. They are enriched in 5D and 818O and plot closely to the kaolinite

STYLES OF FLUID-ROCK INTERACTION

155

Fig. 9. Grant-type plot of average composition of protolith (monzodiorite) versus average composition of altered fault rock (monzodiorite). Concentrations are in wt% (oxides) and ppm (trace elements). The deviation of the data points from the isocon defines the concentration change for the corresponding elements (Grant 1986).

Fig. 10. Isotopic composition of host rock (monzodiorite), fault-related rocks (monzodiorite) from the WFZ (Profile B). Field for metamorphic and primary magmatic waters according to White et al. (1973).

(weathering) line (trend 2, Sheppard & Gilg 1995), indicating low-temperature alteration conditions at high water-rock ratios and crystallization of authigenic clay minerals (Lawrence & Taylor 1972).

Profile C Fluid inclusions. Two samples were suitable for fluid inclusions studies: a cataclastic sample (granite, C2) from a secondary fault on profile C and a quartz vein sample in Triassic sandstones (C5) at around 500 m distance from the major fault. The extension direction of the NNE-SSW trending vein is in agreement with a north-south directed dextral shear zone and indicates the fault-related nature of the vein (see Janssen et al 2002).

Primary two-phase fluid inclusions in small quartz crystals from veins hosted in sandstone are orientated in growth zones and show rounded and elongated forms and consistent liquid/vapour ratios. Their sizes do not exceed 10 um. Final ice melting temperatures were determined to be between -7.8 and -5.2 °C (Fig. 5a), corresponding to salinities between 8.1 and 11.5 wt % NaCl equivalent. The homogenization temperatures range between 239.4 and 300.9 °C (Fig. 5c) with a mean temperature of about 268.7 °C. Under a hydrostatic pressure regime of 250 bar (= 2500 m water column), a maximum pressure dependent temperature correction of about 315 °C is required to obtain the true trapping temperatures (Potter 1977). Secondary quartz-hosted fluid inclusions in extremely altered granite samples from the fault

CJANSSENEr^L.

156

mass-volume isocon, indicating a volume loss between 15-20%. On the basis of this isocon, a moderate enrichment in MgO, Sr, Eu, La, Nd, Hf, Pr and a remarkable loss in SiO2, Ba and Pb is suggested for the minor shear zone. Discussion and conclusion

Fluid sources and fluid composition The types and origins of the fluids that circulated through the WFZ and the resulting fluid-rock interactions were not homogeneous in time and space. Within the Guatacondo region (Profile A) veining was formed by fracturing, pressure solution, and cementation during a first stage of faultrelated deformation (<38 Ma; Eocene-Early zone are restricted to healed cracks. They occur Oligocene; Carrasco et al. 1999). The 513C and as trails of rounded or irregulary shaped two- 618O relationship in vein calcite suggests nearly phase inclusions and show different ranges of closed-system conditions during mineral deposalinity as well as homogenization temperatures sition. Calcite precipitated by cooling from low(Fig. 5a, c) when compared with fluid inclusions salinity fluids at temperatures below 200 °C. in quartz from veins hosted in sandstone in a dis- These fluids have interacted intensively with tance from the fault zone. Fluid inclusions in detritus-rich limestone as indicated by the simisecondary trails showing low-salinity always larities of the REYSN patterns of vein calcites and exhibit the highest homogenization tempera- their respective host-rocks. Probable fluid tures (>150 °C), whereas trails hosting high- sources are formation waters, or meteoric-/ seasalinity fluid inclusions were formed at the water that penetrated the limestone and closely lowest temperatures. approached exchange equilibrium with the rocks. However, the parental fluids have not Major and trace elements of granite and minor experienced temperatures greater than 250 °C as shear zone material. The chemical data from the indicated by the absence of positive Eu anomagranite and the associated minor shear zones are lies in the REYSN patterns of vein calcites (Bau plotted in the isocon diagram shown in Fig. 11 & Moller 1992). Hence, it is very unlikely that the (see also Table 5). Volume changes are inferred, REY-bearing fluids derived from a deep-seated as was done for Profile B, with the immobile crustal fluid source or from a magmatic source. TiO2 isocon, in this case having a slope of 1.2. In In the area of Profile B, indications for highcontrast to the fault rock alteration in Profile B, temperature magmatic or hydrothermal (hypothe immobile isocon plots above the constant gene) fluids, as they are evidenced in the Fig. 11. Grant-type plot of average composition of host rock granite versus average composition of altered fault rock (granite) for samples taken from Profile C. Concentration are in wt% (oxides) and ppm (trace elements).

Table 5. Whole-rock analysis of fault rocks from Profile C (oxides and trace elements) oxides

Si02 %

Ti02 A1203 Fe203 MnO % % % %

Cl C2 C3 C4

89.0 74.6 87.8 93.5

0.04 0.04 0.08 0.05

1.4 1.6 2.5 1.6

7.30 1.28 1.48 1.13

trace

Ba

Cr

Rb ppm

Sr ppm

elements ppm ppm

Cl C2 C3 C4

1793.0 423.0 61.4 25.4

19.24 11.58 17.29 <10 <10 21.37 19.72 14.91

MgO %

CaO %

Na20

%

K20 %

P205

%

H2O %

C02

%

Sum %

0.00 0.09 0.03 0.02

0.15 0.64 0.24 0.31

0.17 10.78 0.42 0.58

<0.1 <0.1 <0.1 <0.1

0.26 0.46 0.56 0.41

0.03 0.03 0.05 0.04

0.90 0.80 2.08 0.76

0.48 9.48 4.34 0.79

99.68 99.82 99.65 99.19

Y ppm

Zr ppm

Pr ppm

Pb ppm

Nd ppm

Hf ppm

Eu ppm

Yb ppm

U ppm

29.49 32.16 59.87 38.97

1 6.3 1

10.1

4

0.6 0.6 0.4

0.4 0.5 0.2

0.9 0.7 0.6

1.2 0.6 0.7

52.83 <10 69.05 26.28 72.95 13.11 19.99 <10

9.7

86.2

22.8

3.9

STYLES OF FLUID-ROCK INTERACTION

well-studied copper mines near the profile (Ossandon et al. 2001), are not preserved. Rather, our results indicate low-temperature alteration conditions with alteration temperatures up to 100 °C. For example, samples from the alteration zone show enrichment in 818O and 5D toward the WFZ trace, with values in the range of those typical for kaolinite weathering conditions (Shepard & Gilg 1995). Furthermore, the altered rocks are depleted in alkaline earth elements with an SiO2 content almost unchanged compared to the undamaged protolith. This is indicative for a low-temperature regime, as fault zones that developed under high-temperature conditions would exhibit decreasing SiO2 content (Gray et al. 1999). We interpret our geochemical data in the context of supergene alteration events, which took place in the Early to Middle Miocene until 14 Ma (Sillitoe & McKee 1966). Supergene alteration processes result from the exhumation of the fault system and comprise oxidation and leaching of the hydrothermally mineralized parts within the oscillating ground water table. These processes only affected the uppermost several hundred metres. For Profile C, only few fluid-inclusion data were available to provide information about fluid sources and fluid-rock interaction within the fault zone. At first glance, the extreme differences in salinity as well as the differences in homogenization temperatures would suggest mixing of two different fluids, i.e. a highly-saline brine with a hotter low-salinity fluid. The source of the brine is not easy to explain since evaporites are not known in the sequence. Therefore, it seems more likely that the differences in salinity were produced by variations in fluid-rock interactions (albitization, chloritization, etc.) or evaporation of seawater. However, both samples indicate the involvement of hotter fluids than in Profile B. This assumption is supported by the depletion in SiO2 (Profile C), because quartz solubility depends strongly on the temperature (Fournier & Potter 1982). The small volume changes between host rock (granite) and fault zone material indicate that the fluid flux through the faulted granite was significantly lower than in profile B.

Variations of fluid-rock interaction Within the areas of Profiles A and C, pervasive mineralization and alteration processes, as described for many other faults, are absent. This may indicate very reduced fluid-rock interactions in these parts of the fault zone. Porosity reduction due to cementation may be inferred

157

from the results of AMT investigations (Janssen et al. 2002), because measurements along the Profiles A and C reveal limited electrical resistivity zones that match the fault core. The conductive zones are very narrow (100 and 200 m) and extend to shallow depths only (50 m in Profile A, 200m in Profile C). According to Renard et al. (2000), we assume that fault zone sealing at shallow depth in Profile A is supported by the carbonate lithology. The kinetics of calcite allow dissolution/precipitation processes at low temperatures and hence, shallow depth. Along Profile B, fluid-driven activities are markedly different from the observed, limited fluid-rock interaction at Profiles A and C. Our work indicates significant amounts of fluid-rock interaction during fault zone evolution. Fluidenhanced mineralogical alteration, mainly due to infiltration of meteoric water, led to the development of a strongly altered fault core with an apparent increase in porosity and permeability as can be inferred from the isocon plot and the electromagnetic studies. The 400 m wide alteration zone corresponds with the width of the conductivity anomaly. The AMT results show that this alteration zone is imaged as a conductive structure down to about 1500 m (Hoffmann-Rothe 2002). Evans & Chester (1995), who demonstrate similar differences in fluid-rock interaction and weakening processes along the San Gabriel Fault in California, reason that such variability may be a manifestation of hydrological segmentation of the fault. Lithological variations, as they are also present along the WFZ, could cause an inherent permeability heterogeneity within the fault that, in turn, produces a complex distribution of sealed and unsealed zones.

Comparison with the San Andreas Fault, California A variety of fault zone evolution models are discussed in the literature, predominantly based on data from the well-investigated San Andreas Fault (SAF). The basic assumption of these cyclic fault evolution models is the interplay of high fluid pressure within the fault zone that may trigger earthquake rupture and fluid-enhanced strength recovery due to healing and cementation of fault material in the inter-seismic period (e.g. Rice 1992; Sleep & Blanpied 1992; Byerlee 1993). Differences between fault models concern firstly the origin of fluids, and secondly maintenance of high fluid pressure in the fault zone. For example, observations from the SAF are explained in the continuous flow model of

158

C.JANSSEN£r,4L.

Rice (1992). In this model, pressurized fluids in the fault zone have been supplied from the root zone of the fault and include a mantle component. Models for episodic flow of high pressure water (e.g. Byerlee 1993) propose that meteoric water from the adjacent country rock becomes trapped in the fault zone following an earthquake. Kharaka et al. (1999) investigated the geochemistry and the role of fluids in the dynamics of the San Andreas Fault system. Their results show that the fluids are predominantly of meteoric origin and circulate at shallow to moderate depths (up to 6 km/80-150 °C). However, the composition and isotope abundances of noble gases indicate a significant mantle component for volatiles (especially CO2) which support a deep continuous flow model. CO2 influx into the fault led to the generation of lithostatic fluid pressures, thus weakening the fault (Kharaka et al. 1999). O'Neil (1985) measured the amount of D, 18O, and H2O in fault gouge samples also collected from the San Andreas Fault. He found that the permeability of this gouge is high enough to allow meteoric waters to descend and mix with a 'brine' that is entering the system from below or laterally. Looking at the results from the three WFZ traverses, we can show that a mantle component seemed to have played a minor role, if at all, in the composition of formerly and/or presently circulating fluids. The lithified record indicates that fluids have not experienced temperatures greater 250 °C. Even todays hot spring waters of the Precordillera exhibit a pure radiogenic helium signal alone, indicating that these fluids do not contain a deep crustal or mantle component (e.g. Hilton et al. 1993; Hoke et al. 1994). For all three traverses geochemical data indicate that fluids are predominantly meteoric in origin. It is thus more likely that fluids are derived from lithologies/sequences adjacent to the fault. Different from the SAF, where the presence of broad gouge zones is held liable for the build up of high fluid pressure, the accumulated faulting activity has not led to the development of a broad clay-rich gouge. The authors wish to thank to P. Dulski and R. Naumann for geochemical analysis. Special thanks are addressed to G. Dresen and K. Hahne for their constructive criticisms. We thank A. Hendrich for help with drafting. A. Maltman and J. Imber provided constructive reviews that significantly improved the manuscript. This work was funded by the DFG (German Research Foundation) within the framework of the Collaborative Research project SFB 267 Deformation Process in the Andes and by the GeoForschungsZentrum.

References BAU, M. & MOLLER, P. 1992. Rare earth element fractionation in metamorphogenic hydrothermal calcite, magnesite and siderite. Mineralogy and Petrology, 45, 231-246. BAU, M., DULSKI, P. & MOLLER, P. 1995. Yttrium and holmium in South Pacific seawater: Vertical distribution and possible fractionation mechanisms. Chemie der Erde, 55,1-55. BLANPIED, ML., LOCKNER, D.A. & BYERLEE, J.D. 1992. An earthquake mechanism based on rapid sealing of faults. Nature, 358. BORIC, R., DIAZ, F. & MAKSAEV, V. 1990. Geologfa y Yasimientos Mataliferos de la Region de Antofagasta. Servicio Nacional de Geologia y Mineria Santiago, 40. BURKHARD, M. 1993. Calcite-twins, their geometry, apparance and significance as stress-strain markers and indicators of tectonic regime: A review. Journal of Structural Geology, 15, 351-368. BYERLEE, J.D. 1993. Model for episodic flow of highpressure water in fault zones before earthquakes. Geology, 21, 303-306. CARRASCO, P., WILKE, H. & SCHNEIDER, H. 1999. Post Eocene deformational events in the north segment of the Precordilleran fault system, Copaqiri (21°S). Fourth ISAG, Goettingen, Germany, 132-135. CHESTER, KM. & LOGAN, J.M. 1986. Implication for mechanical properties of brittle faults from observations of Punchbowl fault zone, California. Pure and Applied Geophysics, 124,79-106. CHESTER, F.M., EVANS, IP. & BIEGEL, R.L. 1993. Internal structure and weakening mechanisms of the San Andreas Fault. Journal of Geophysical Research, 98, 771-786. CLAYTON, R.N., O'NEIL, J.R. & MAYEDA, T.K. 1972. Oxygen isotope exchange between quartz and water. Journal of Geophysical Research, 77, 3057-3067. DlLLES, J.H.,

TOMLINSON, A.J.,

MARTIN, M.W.

&

BLANCO, N. 1997. El Abra and Fortuna Complexes: A porphyry copper batholith sinistrally displaced by the Falla Oeste. Congreso Geologico Chileno, 8th, 1997, Adas, 3. EVANS, J.P. & CHESTER, F.M. 1995. Fluid-rock interaction in faults of the San Andreas system: Inferences from San Gabriel fault rock geochemistry and microstructures. Journal of Geophysical Research, 100,13,007-13,020. FOURNIER, R.O. & POTTER, R.W. 1982. An equation correlating the solubility of quartz in water from 25 to 900°C at pressure up to 10,000 bars. Geochimica et Cosmochimica Acta, 46,1969-1973. GODDARD, J.V. & EVANS, J.P. 1995. Chemical changes and fluid-rock interaction in faults of crystalline thrust sheets, northwestern Wyoming, U.S.A. Journal of Structural Geology, 17, 533-547. GRANT, J. A. 1986. The isocon diagram - a simple solution to Gresens equation for metasomatic alteration. Economic Geology, 81,1976-1982. GRAY, D.R., GREGORY, R.T. & DURNEY, D.W. 1991. Rock-buffered fluid-rock interaction in deformed

STYLES OF FLUID-ROCK INTERACTION quartz-rich turbidite sequences, Eastern Australia. Journal of Geophysical Research, 96, 19681-19704. GRAY, D.R., JANSSEN, C. & WAPNIK, J. 1999. Deformation character and fluid flow across a wrench fault within a Paleozoic accretionary wedge: Waratah Fault Zone, southeastern Australia. Journal of Structural Geology, 21,194-214. GRESENS, R.L. 1967. Composition-volume relationships of metasomatism. Chemical Geology, 2, 47-65. HILTON, D.R., HAMMERSCHMIDT, K., TEUFEL, S. & FRIEDRICHSEN, H. 1993. Helium isotope characteristics af Andean geothermal fluids and lavas. Earth and Planetary Science Letters, 120,265-282. HOEFS, J. 1987. Stable Isotope Geochemistry. 3rd Edn, Springer-Verlag. HOFFMANN-ROTHE, A. 2002. Combined structural and magnetotelluric investigation across the West Fault Zone in northern Chile. PhD thesis, Universitat Potsdam, Scientific technical report (STR) 12/02. HOKE, L., HILTON, D.R., LAMB, S.H., HAMMERSCHMIDT, K. & FRIEDRICHSEN, H. 1994. The evidence for a wide zone of active mantle melting beneath the Central Andes. Earth and Planetary Science Letters, 128, 341-355. HOLLISTER, V.E. 1978. Geology of the porphyry copper deposits of the Western Hemisphere: AIME, 219 pp. IMBER, J., HOLDSWORTH, R.E., BUTLER, C.A. & STRACHAN, R.A. 2001. A reappraisal of the SibsonScholz fault zone model: The nature of frictional to viscous ('brittle-ductile') transition a longlived, crustal-scale fault, outer Hebrides, Scotland. Tectonics, 20, 601-624. JANSSEN, C, MICHEL, G.W., BAU, M., LUDERS, V. & MUHLE, K. 1997. The North Anatolian Fault Zone and the role of fluids in seismogenic deformation. Journal of Geology, 105, 387-403. JANSSEN, C., HOFFMANN-ROTHE, A., TAUBER, S. & WILKE, H. 2002. Internal structure of the Precordilleran fault system (Chile) - insights from structural and geophysical observations. Journal of Structural Geology, 24,123-143. KERRICH, R. & KYSER, T.K. 1994. The geochemistry and role of fluids in large continental structures: An overview. US Geological Survey, Proceedings of Workshop LXllI "The mechanical involvement of fluids in faulting, Open-file Report, 349-388. KHARAKA, Y.K., JAMES, J.T, EVANS, WC. & KENNEDY B.M. 1999. Geochemistry and hydromechanical interactions of fluids associated with the San Andreas Fault system, California. In: HANEBERG, W.C. (ed.) Faults and subsurface fluid flow in the shallow crust. Geophysical Monograph, 113, 129-148. LAWRENCE, J.R. & TAYLOR, H.P. 1972. Hydrogen and oxygen isotope systematics in weathering profiles. Geochimica et Cosmochimica Ada, 36, 1377-1393. LINDSAY, D.D., ZENTILLI, M. & DE LA RIVERA, J.R. 1995. Evolution of an active ductile to brittle

159

shear system controlling mineralization at the Chuquicamata porphyry copper deposit, Northern Chile. International Geological Review, 37, 945-958. LUCASSEN, F, FRANZ, G. & LABER, A. 1999. Permian high pressure rocks - the basement of the Sierra de Limon Verde in Northern Chile. Journal of South American Earth Sciences, 12,183-199. LUDERS, V. & MOLLER, P. 1992. Fluid evolution and ore deposition in the Harz Mountains (Germany). European Journal of Mineralogy, 4,1053-1068. McCREA, J.M. 1950. The isotopic chemistry of carbonates as a paleotemperature scale. Journal of Chemical Physics, 18, 849. O'HARA, K. 1988. Fluid flow and volume loss during mylonitization: an origin for phyllonite in an overthrust setting, North Carolina, U.S.A. Tectonophysics, 156, 21-36. O'NEIL, J.R. 1985. Water-Rock Interactions in Fault Gouge. PAGEOPH, 122, 440-446. O'NEIL, J.R., CLAYTON, R.N. & MAYEDA, T.K. 1969. Oxygen isotope fractionation in divalent metal carbonates. Journal of Chemical Physics, 51, 5547-5558. OHMOTHO, H. & RYE, R.0.1979. Isotopes of sulfur and carbon. In: BARNES, H.L., WILEY, J. & SONS (eds) Geochemistry of hydrothermal ore deposits, 2nd Edn., New York, 509-567. OSSANDON, G, FRERAUT, R., GUSTAFSON, L.B., LINDSAY, D.D. & ZENTILLI, M. 2001. Geology of the Chuquicamata mine: A progress report. Economical Geology, 96, 249-270. POTDEVIN, J. 1993. The Gresens 92 code. User manual (14/12/93). Universite des Sciences et Technologies de Lille, Villeneuve d'Asc. POTTER, R.W. 1977. Pressure corrections for fluid inclusion homogenization temperatures based on volumetric properties in the system NaCl-H2O. Journal of Research US Geological Survey, 5, 603-607. RAWLING, G.C, GOODWIN, L.B. & WILSON, J.L. 2001. Internal architecture, permeability structure, and hydrologic significance of contrasting fault zone types. Geology, 29, 43-46. RENARD, F, GRATIER, J.-P. & JAMTVEIT, B. 2000. Kinetics of crack-sealing, intergranular pressure solution, and compaction around active faults. Journal of Structural Geology, 22,1395-1407. REUTTER, K.J., SCHEUBER, E. & HELMCKE, D. 1991. Structural evidence of orogen-parallel strike slip displacements in the Precordillera of northern Chile. Geologische Rundschau, 80,135-153. REUTTER, K.J., SCHEUBER, E. & CHONG, G. 1996. The Precordilleran fault system of Chuquicamata, Northern Chile: evidence for reversals along arcparallel strike-slip faults. Tectonophysics, 259, 213-228. REYNOLDS, P., RAVENHURST, C., ZENTILLI, M. & LINDSAY, D. 1998. High-precision 40Ar/39Ar dating of two consecutive hydrothermal events in the Chuquicamata porphyr copper system, Chile. Chemical Geology, 148, 45-60. RICE, J.R. 1992. Fault stress states, pore pressure distributions, and the weakeness of the San Andreas

160

C. JANSSEN ETAL.

fault. In: EVANS, B. & WONG, T.F. (eds) Fault Mechanics and Transport Properties of Rocks. Academic Press New York, 475-503. ROEDDER, E. 1984. Fluid Inclusions. Reviews in Mineralogy, 12. SAUSSE, J., JACQUOT, E., FRITZ, B., LEROY, J. & LESPINASSE, M. 2001. Evolution of crack permeability during fluid-rock interaction. Example of the Brezouard granite (Vosges, France). Tectonophysics, 336,199-214. SCHULZ, S.E. & EVANS, IP. 1998. Spatial variability in microscopic deformation and composition of the Punchbowl fault, southern California: implications for mechanisms, fluid-rock interaction, and fault morphology. Tectonophysics, 295, 223-244. SHEPPARD, S.M.F. & GILG, H.A. 1995. Stable isotope geochemistry of clay minerals. Clay Mineralogy, 31,1-24. SIBSON, R.H., MOORE, J.M. & RANKIN, A.H. 1975. Seismic pumping - A hydrothermal fluid transport mechanism. Journal Geological Society, 131, 653-659. SILLITOE, R.H. & McKEE, E.H. 1996. Age of supergene oxidation and enrichment in the Chilean

porphyry copper province. Economical Geology, 91,164-179. SLEEP, N.H. & BLANPIED, M.L. 1992. Creep, compaction and the weak rheology of major faults. Nature, 359. TAYLOR, H.P. 1978. Oxygen and hydrogen isotope studies of plutonic rocks. Earth and Planetary Science Letters, 38,177-210. VENNEMANN,T.W. & O'NEIL, J.R. 1993. A simple and inexpensive method of hydrogen isotope and water analyses of minerals and rocks based on zinc reagent. Chemical Geology, 103, 227-234. WHITE, D.E., BARNES, I. & O'NEIL, J.R. 1973. Thermal and mineral water of nonmeteoric origin. California Coast Ranges. Geological Society of America Bulletin, 84, 547-560. WINTSCH, R.P., CHRISTOPHERSON, R. & KRONBERG, A.K. 1995. Fluid-rock reaction weakening of fault zones. Journal of Geophysical Research, 100, 13021-13032. ZULAUF, G., PALM, S., PETSCHICK, R. & SPIES, 0.1999. Element mobility and volumetric strain in brittle and brittle-viscous shear zones of the superdeep well KTB (Germany). Chemical Geology, 156, 135-149.

Ductile shearing, hydrous fluid channelling and high-pressure metamorphism along the basement-cover contact on Sikinos, Cyclades, Greece S. GUPTA1 & M. J. BICKLE2 1 Department of Geology & Geophysics, I.I.T., Kharagpur 721 302, India 2 Deptartment of Earth Sciences, Downing St, Cambridge CB2 3EQ, UK (e-mail: [email protected]) Abstract: On the island of Sikinos in the Cyclades a schistose carapace separates a marble-blueschist cover sequence from underlying basement rocks. The basement 'core', comprising metapelitic gneisses, biotite-bearing granodiorites and aplites, becomes increasingly strained towards the carapace with progressive obliteration of earlier structures and intensification of a mylonitic foliation that becomes pervasive within the carapace. Granodiorites in the 'core' can be traced into microcline schists within the carapace, whereas metapelitic gneisses are converted to garnet-mica schists. The carapace is therefore a simple shear zone comprising basement rocks mylonitized during overthrusting of the cover. Biotite clusters in granodiorites of the basement 'core' are partially altered to phengite, whereas plagioclase shows incipient sericitization. In more strained rocks these hydration reactions show enhanced progress, until biotite and plagioclase are finally eliminated in the carapace. Rare glaucophane and chloritoid inclusions within garnets of metapelitic gneisses adjacent to the carapace are also attributed to hydration reactions. The association of higher strain with increased hydration in the basement suggests localization of fluids in the strained carapace zone, with limited percolation into underlying rocks. The restricted availability of water outside the carapace may be responsible for preservation of pre-Alpine assemblages in large parts of the Cycladic basement.

The association of quartzofeldspathic rocks of continental crustal affinity with blueschists and eclogites has been reported from many orogenic belts, e.g. the western Alps (Chopin 1984), Corsica (Gibbons & Horak 1984) and the Dabie Shan province of China (Okay et al 1989). This crustal material may have assisted the uplift and exhumation of high-pressure rocks subsequent to decoupling from the subducting slab, on account of their inherent buoyancy (Davies & von Blanckenberg 1995). Such buoyancy-driven exhumation mechanisms can be invoked only by demonstrating that continental material was subducted along with the associated highpressure rocks. However, the signature of highpressure, low-temperature metamorphism in continental rocks may not be easily recognizable (Hermann 2002). In rare cases, crustal rocks preserve independent evidence of high-pressure, low-temperature metamorphism, in the form of jadeite-, coesite- and/or diamond-bearing assemblages, which irrefutably prove subduction or burial to mantle depths. In such cases, the buoyant crustal components may have contributed substantially to the uplift and exhumation of the associated blueschists and eclogites.

The Attic-Cycladic massif (Fig. 1) provides an ideal setting for investigating the phenomenon of continental crustal subduction, and its involvement in the subsequent uplift of blueschists and eclogites. The massif consists of a lower pre-Alpine continental basement unit of Hercynian age (Henjes-Kunst & Kreuzer 1982; Andriessen et al. 1987), underlying an upper Alpine cover sequence of marbles and blueschists (Dlirr et al. 1978; Van der Maar et al. 1981; Papanikolaou 1987). Much of the sequence lies below the Aegean Sea. Seismic studies (Makris 1978; Mascle & Martin 1990) indicate that the Aegean Sea is floored by thin continental crust. Although there are clear indications of a greenschist (M2) imprint of Alpine age on the basement (Van der Maar et al. 1981; Franz et al. 1993), only zones proximal to the cover preserve evidence (rare) of an early Alpine high-pressure, low-temperature (M^) metamorphic event. The reasons for the restricted occurrence of the high-pressure imprint in the basement are unclear, and form the subject matter of this paper. The basement is exposed on the islands of Naxos, los and Sikinos (Van der Maar 1980,1981;

From. ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224,161-175. 0305-8719/$15.00 © The Geological Society of London 2004.

162

S. GUPTA & M. J. BICKLE

Fig. 1. Geological map of the Attic-Cycladic massif, modified after Van der Maar and Jansen (1983).

Franz et al. 1993). On Naxos, the situation is complicated by a late high temperature (amphibolite facies) overprint in the basement that has resulted in migmatization (Jansen & Schuiling 1976; Buick & Holland 1991), and has consequently erased all traces of any earlier high pressure mineralogy. In contrast, the basementcover contact is well-exposed on los, but its nature remains controversial. The uppermost structural levels of the los basement are intensely mylonitized along the 'South Cycladic Shear Zone' (Lister et al. 1984; Vandenberg & Lister 1996), which juxtaposes the basement against the overlying marble-blueschist cover sequence. Van der Maar (1980) and Van der Maar and Jansen (1983) interpreted the contact as a pre- to syn-blueschist-facies thrust plane localized along an earlier erosional surface. Grlitter (1993) also characterized the contact as a thrust, but reassigned the contact to a location higher in the structural sequence. In contrast, Vandenberg & Lister (1996) suggested that the present contact is a north-dipping normal fault that cut through the SCSZ. Subsequently, Forster & Lister (1999) considered the SCSZ to be part of a single anastomosing normal fault system (the los detachment system) related to Oligo-Miocene extensional tectonics in the Aegean. On the island of Sikinos, which neighbours los, the basement-cover contact is also exposed (Van der Maar et al. 1981; Franz et al. 1993), but has not been the subject of comparable study. Most workers have assumed that the basement has undergone the Eocene high-pressure metamorphic event (Van der Maar 1981; Van der Maar & Jansen 1983; Franz et al. 1993;

Vandenberg & Lister 1996), though the inference is based on limited mineralogical criteria, such as high silicon content in phengites. The reason for the paucity of diagnostic highpressure assemblages in the basement has never been investigated, in spite of its obvious tectonic implications. The high-pressure imprint may be lacking simply because the basement was never subjected to blueschist facies metamorphism, and was only juxtaposed against the cover sequence during exhumation. Vandenberg & Lister (1996) and Forster & Lister (1999) have suggested that the basement-cover sequence contact is related to exhumation. Alternatively, the basement may have been subducted, but did not develop the characteristic mineralogy. The reason for such a phenomenon must therefore be investigated. This study focuses on the Sikinos basement, with a view to reassessing the nature of the basement-cover contact, to examine the evidence for a high-pressure metamorphic signature in the basement, and to suggest why diagnostic blueschist facies assemblages are largely absent in the Cycladic basement.

Geology of the cover sequence on Sikinos Much of Sikinos is composed of a marble-schist cover sequence, with the pre-Alpine basement exposed in the eastern and southeastern parts (Van der Maar et al 1981; Franz et al. 1993; Fig. 2). The sequential evolution of mineralogical assemblages with structural fabrics in the cover sequence was described in detail by Gupta (1995), and is outlined below. The cover sequence comprises a variety of rock types that

BASEMENT-COVER CONTACT ON SIKINOS, GREECE

163

Fig. 2. Geological map of Sikinos (after Van der Maar et al. 1981 and Franz et al. 1993).

are all intensely deformed. Of the five identified deformation events, the first three were fabricforming. D! involved isoclinal folding resulting in the development of a pervasive axial planar foliation Sj on original litho-contacts. Si is defined by sodic amphibole in metabasites, phengite in marly metasediments and paragonite in metapelites. Tight to isoclinal folding of Sj along shallow north-south trending axes during a second deformation event D2 gave rise to a very prominent lineation L2, generally defined by crenulations on Si in all lithologies. D2 terminated at the time lawsonite (now pseudomorphed) in metabasites and calcareous metasediments broke down by reaction with albite, to form clinozoisite/epidote, paragonite and quartz (Heinrich & Althaus 1980); this reaction took place at pressures of around 8.5 kbar and temperatures around 350 °C (Gupta 1995). In metabasites, epidotes formed at this stage are large and randomly orientated. Synchronously, randomly orientated sodic amphibole overprinting D! and D2 fabrics formed in associated phengite-rich, marly metasediments. Similar sodic amphiboles also formed in metamorphosed cherts and acidic rocks. The absence of preferred orientation in minerals across a range of bulk compositions, and especially in some apparently incompetent lithologies suggests a period of low strain, rather than selective strain partitioning around competent rocks. Low strain conditions apparently persisted till the peak of metamorphism in the epidote-blueschist facies (as defined by Evans 1990). Peak metamorphic conditions of 11-14 kbar, 480 ± 20 °C were inferred (Gupta 1995) from metapelitic assemblages containing glaucophane-chloritoid-garnet, a characteristic

high pressure association (Guiraud et al. 1990; Okay & Kelley 1994). Deformation resumed during exhumation with an extensional deformation (D3) characterized by the development of S-C fabrics. Extensional deformation was accompanied by retrogression through the greenschist facies, and continued into the brittle zone with the development of late normal faults. Decompression was near-isothermal. Since diaspore remained stable in metabauxites through the P-T history, the temperature did not exceed 500 °C, the temperature at which diaspore dehydrates to corundum. D4 was a phase of broad open folding giving rise to local warps and broad crenulations, and D5 is repesented by post-metamorphic high angle normal faults cutting across all earlier structures. D! and D2 in the cover sequence preceded the peak of high-pressure metamorphism, and are therefore prograde. Thereafter, pressure and temperature continued to increase, in the absence of deformation, till the metamorphic peak. D3, D4 and D5 can all be interpreted in the light of proposed models for the exhumation of metamorphic core complexes by extensional tectonics (e.g. Davis 1980; Wernicke & Burchfiel 1982; Lister et al 1984). The timing of fabric-forming events Db D2 and D3 with respect to the P-T path for the cover sequence is shown in Fig. 9.

Basement lithology Basement outcrops on Sikinos are not as extensive as on los, but here too, a schistose carapace with a single penetrative fabric mantles a relatively less deformed core (Van der Maar 1980). The southeastern basement exposures abut marbles of the cover sequence, and are

164

S. GUPTA & M. J. BICKLE

Fig. 3. Geological map of the eastern basement outcrop on Sikinos. Note that much of the basement-cover contact is a normal fault, although a ductile shear zone (thrust) can be identified in the west where the fault cuts upsection into the cover.

consequently less useful for fabric correlation. On the other hand, the smaller eastern outcrop is juxtaposed against a sequence comprising mica schists and marbles; this region has been mapped in detail to compare structural features of the basement and cover (Fig. 3). The main part (the 'core') of the eastern basement exposures includes metapelitic gneisses, granodiorites, aplite dykes and rare metabasite pockets. Variably orientated aplite dykes cut across these rock types. Garnet-mica schists and microcline schists are confined to the carapace. The contact between the metapelitic gneisses in the core and the adjacent garnet-mica schists in the carapace is not exposed. However, the biotite-bearing granodiorite can be traced into the microcline schists of the carapace. The change can be correlated with an increase in strain, suggesting that the carapace schists represent recrystallized, mylonitized equivalents of basement rocks in the core. Relevant mineral analytical data in the ensuing description and subsequent geothermobarometric calculations is provided in a Supplementary Publication which has been deposited with the Society Library and British Library at Boston Spa, W Yorkshire, No. SUP 18198 (7 pages). It is also available online at http://www.geolsoc.org.uk/SUP18198. These were obtained on a wavelength dispersive CAMECA SX-50 machine operating at 20 kV and 10 nA, at the Department of Earth Sciences, University of Cambridge.

Pre-Alpine deformation The metapelitic gneisses in the core preserve the earliest evidence of deformation, represented by the gneissic layering, SH1, comprising alternating mica (phengite)-rich and quartzo-feldspathic bands. Importantly, there is no alignment of micas in the micaceous layers. SH1 was isoclinally folded during a second phase of deformation DHI (Fig- 4a), with no identifiable associated fabric. The contact between the basement gneisses and biotite-bearing granodiorites is faulted, but xenoliths of gneisses within the intrusions confirm the sequence of events. SH1 and FH2 folds within the xenoliths are truncated by several generations of aplite dykes (Fig. 4b), which also cut across the relatively undeformed granodiorites in the core. DH1 and DH2 are therefore considered to be pre-Alpine in age. Minor faults, with strikes varying from north-south to east-west, displace all lithologies and lithocontacts within the basement core; these faults cannot be traced into the schistose carapace.

Alpine deformation The basement 'core' Alpine deformation in the basement resulted in a progressive increase in strain from the basement 'core' to the schistose carapace. The strain is manifested by the development of shear bands

BASEMENT-COVER CONTACT ON SIKINOS, GREECE

165

Fig. 4. Field relations in basement rocks, (a) FH2 isoclinal folding of the SH1 gneissic layering in metapelitic gneiss xenolith truncated by intruding granodiorite. (b) SH1 and FH2 folds truncated by aplite dyke. Note minor dragging of fabrics along dyke. Biotite cluster present alongside the hinge of the fold, (c) Adjacent to the carapace, SA1 shear bands cut across the metapelitic gneiss and intruding aplite. (d) Quartz vein in carapace, initially parallel to SA1, is isoclinally folded. Note development of new axial planar SA2 foliation.

(C-surfaces) that cut across the metapelitic gneisses, granodiorites and intruding aplite dykes (Fig. 4c). These bands are thin (on a millimetre scale) and widely spaced in the basement 'core'. Towards the contact with the cover sequence, the width and frequency of the bands increase progressively till the rocks are completely schistose within the carapace. In metapelitic gneisses, a new, widely-spaced foliation has developed axial planar to broad, asymmetrical folds on SHi. Unlike the pre-Alpine structures in the gneisses, these folds are not penetrative in nature and are not apparent on the mesoscopic scale. This foliation is overprinted by an Alpine mineralogy, and itself curves tangentially into the fabric within the shear bands, SA], which represent the C-surfaces of the non-coaxial event. S-C fabrics are especially distinct in the granodiorites of the basement 'core'. As in the metapelitic gneisses, shear bands (C-surfaces) are present but are very widely spaced (separ-

ation of about 50 cm). The angle between the S- and C-planes is about 40°. The density of Cplanes increases progressively towards the contact with the cover sequence, whereas the angle between the S- and C-surfaces decreases systematically as the coarse-grained rock with no preferred orientation is converted to a strongly foliated microcline schist. The sense of movement is top-to-the-south on north-dipping C-planes, consistent with thrust-sense movement. Finally, at around 75 m from the contact, the C-plane predominates and the S-foliation cannot be distinguished at all. This marks the beginning of the carapace.

The carapace The schistose carapace includes garnet-mica schists grading from the metapelitic basement gneisses of the 'core', and microcline schists which are mylonitized equivalents of the granodiorites. Aplite dykes, which cut across all

166

S. GUPTA & M. J. BICKLE

Fig. 5. Equal area foliation pole and lineation plots for the basement and cover rocks in the Eastern Basement outcrops. Note parallelism of all structural elements.

rocks in the basement 'core' at varying orientations, are also present in the carapace, but are almost completely rotated parallel to the dominant foliation. Exclusively within the carapace, the mylonitic foliation SA1 is isoclinally refolded (FA2) about north-south axes, with local development of an axial planar foliation SA2 (Fig. 4d). The resulting strong crenulation lineation LA2 is parallel to the L2 lineation in the overlying cover sequence. The SA2 foliation wraps around garnets in the mica schists. A few, late shears in which late chlorites are aligned parallel to the foliation are observed at low angles to the SA2 foliation; these are probably extensional, developed during uplift of the amalgamated basement-cover sequence. Refolding of the SAI foliation is also observed in microcline schists of the carapace; these folds are isoclinal and reclined, with axes parallel to L2 in the over-

lying cover. Furthermore, the SA2 foliation is characterized by the development of an intense stretching lineation parallel to the LA2 lineation defined by minor fold axes. The DA2 deformation was followed by late faults that cut across all rock types. Some are minor, with lateral displacements of up to a metre, whereas others are more significant, separating rock groups such as the garnet-mica schists and microcline schists (Fig. 3). The SA1/A2 foliation in the basement carapace is everywhere parallel to the S1/2 foliation in the overlying cover sequence. The attitude of the foliation varies from northerly in the north, to north-NNW in the west, to west-NW in the south. Equal area plots of the poles to the foliation (Fig. 5) spread about an axis trending northerly, similar to that observed in the cover, and parallel to the LA2 lineation in the

BASEMENT-COVER CONTACT ON SIKINOS, GREECE

167

Fig. 6. Schematic representation of basement-cover relations on Sikinos. Metapelitic gneisses and granodiorites in the basement 'core' are mylonitized within the carapace. Aplite dykes are rotated parallel to the pervasive carapace foliation, which is parallel to and cofolded with, the overlying cover sequence.

basement, and the L2 lineation in the cover. The field relations are schematically depicted in Fig. 6. Evidence for non-coaxial strain in the basement rocks towards the contact with the cover sequence, the parallel planar and linear fabric elements in the basement and cover, and the sense of movement inferred from the S-C fabrics, where distinguishable, indicate that the cover sequence was initially thrust over the basement. Thereafter, the two units had a shared history, though much of the subsequent deformation in the basement remained restricted to the ductile carapace zone of schistose rocks.

Mineralogical transformations in the basement - core to carapace Biotite-bearing granodiorite to microcline-quartz schist Granodiorites in the basement 'core' have a predominantly igneous texture and contain quartz, plagioclase, and microcline. Characteristically, these rocks also contain clusters of kinked and cracked biotite flakes associated with rare garnet and epidote that are dispersed through the main granodiorite body. The clusters are interpreted to be fragments (xenoliths / xenocrysts) of metapelitic gneisses within the intrusions. The transformation of the coarse-

grained, unfoliated intrusions in the 'core' to the strongly deformed microcline schists of the 'carapace' is accompanied by increasing evidence of ductile strain in the rock, accompanied by the replacement of biotite and plagioclase by phengitic muscovite, which ultimately defines the foliation in the carapace schists. The sequence of transformation has been documented through a series of samples collected systematically during a traverse approaching the contact from well within the 'core'. Quartz within unfoliated rocks of the basement core (Fig. 7a) shows some strain in the form of weak undulose extinction. Incipient replacement of large plagioclase grains by sericite along cleavages is common, while microcline is largely unaffected. Increasing strain results in the development of spectacular chess-board deformation lamellae in quartz, and flame perthites (Pryer & Robin 1995; Vernon 1999) in Kfeldspar, with increasing sericitization of plagioclase (Fig. 7b). At this stage, there is some mobilization of the phengites into a crude SA1 foliation that wraps around deformed quartz and microcline porphyroclasts. Within the carapace, plagioclase is entirely absent and the remobilized sericites (phengites) define a strong SAI fabric. Quartz forms elongate ribbons which are internally strained, and in places marginally recrystallized to aggregates of strain-free subgrains (Fig. 7c). Adjacent to the contact, recovery of strained quartz domains is widespread, and the rock is a fine-grained, recrystallized microcline schist (Fig. 7d).

168

S. GUPTA & M. J. BICKLE

Fig. 7. Photomicrographs (a) to (d) depict the transformation from granodiorite in the 'core' to microcline schist in the carapace, (a) Igneous textures preserved in the 'core'. Quartz shows weak undulose extinction; plagioclase laths are partially sericitized [Sample 12(i)]. (b) Closer to the carapace, sericites derived from plagioclase are mobilized into the SA1 foliation which wraps around quartz porphyroclasts showing spectacular chess-board deformation lamellae [Sample 104(ii)]. (c) Within the carapace SA1 is prominent, and quartz forms ribbons with deformation bands. Note the incipient marginal recrystallization in quartz [Sample 112(i)]. (d) The SAI foliation is pervasive adjacent to the contact. Quartz is recrystallized to aggregates of strain-free subgrains, while porphyroclasts of microcline survive [Sample 112(ii)]. (e) Transformation of biotite to phengite in the basement 'core'. The highest Si contents in phengites are preserved in cores of biotites, while phengite Si contents are lower at biotite rims, which have reequilibrated at lower pressures [Sample 36(ii)]. (f) Replacement perthites in basement granodiorites. Note the SA1 foliation wrapping the microcline host [Sample 104(i)].

BASEMENT-COVER CONTACT ON SIKINOS, GREECE

Elimination of biotite and plagioclase in the carapace can be attributed to separate hydration reactions. Transformation of biotite to phengite takes place along rims, cleavages and cracks by the reaction (Fig. 7e):

This reaction forms the basis of the phengite barometer proposed by Massone & Schreyer (1987). Phengites replacing biotite cores have the highest Si content (~ 3.4), while the rims are characterized by phengites with lower Si (~ 3.07). Towards the contact with the cover sequence, reaction (1) proceeded to completion and the original, kinked biotite clusters were replaced by flakes of unstrained phengite. Within microcline schists of the carapace, biotites are completely absent, suggesting complete replacement of all original biotite. Sericitization of plagioclase was accompanied by replacement of K-feldspar by albite, generating replacement perthites (Pryer & Robin 1995). Sodium is released from plagioclase during muscovite-forming reactions, and Kfeldspar is replaced by albite along the Murchison plane in a set of cyclic reactions (Vernon 1999) that can be summarized as follows:

This reaction leads to an increase in the albite content of the host plagioclase as well as within the flames. Relicts of original plagioclase grains seen in the less altered 'core' have an almost albitic composition. This may reflect original sodic plagioclase, or may be a result of the replacement process. The amount of zoisite formed in the altered plagioclase during reaction (2) is limited.

169

Idiomorphic garnets with dusty cores (~Alm60Gr27) and clear rims (~Alm70Gr15) containing rare glaucophane with and without chloritoid inclusions truncate (post-date) the SAI foliation defined by remobilized phengites (Fig. 8b). Unlike chloritoid, glaucophane also occurs in the matrix and like the phengites, shows no preferred orientation. Large chlorite flakes overgrow SA1, and in places, rim garnet. In basement schists, the SA2 foliation, which is axial planar to isoclinal folds on SA1, wraps garnet porphyroblasts with distinct pressure shadows (Fig. 8c). These garnet porphyroblasts have core and rim compositions indistinguishable from those in the gneisses, and the rims contain inclusions of chloritoid and sodic amphibole. Chlorites overgrow SA2, but are aligned within extensional shear planes that occur locally within the carapace, at low angles to the SA2 foliation (SUP 18198). Reactions leading to glaucophane and chloritoid formation in the basement gneisses and schists are difficult to model, since the early ferromagnesian phase contributing components for amphibole formation has been entirely consumed. In the metapelites, this phase is likely to have been either prograde chlorite or biotite. Although there is no evidence that prograde chlorite was indeed present, biotite clusters occur as xenocrysts and within xenoliths of gneiss enclosed in the intruding granodiorite body (e.g. Fig. 4a). Reports from the basement of los (Van der Maar 1980) also confirm that biotite is 'ubiquitous' in garnet-mica schists. However, since chlorite is also present in these rocks (albeit retrograde), the possibility of it being a precursor phase in the metapelites cannot be entirely overruled. Accordingly, reactions involving chlorite, biotite and muscovite have all been modelled to produce glaucophane, chloritoid and the celadonite component in phengites.

Basement gneiss to basement schist Unlike the granodiorite to microcline schist transformation, exposures documenting the change from basement gneisses in the 'core' to basement schists in the 'carapace' are not continuous. Quartzo-feldspathic layers of the SH1 foliation in the gneisses comprise quartz and albite, with subsidiary phengite, garnet and chlorite. The micaceous layers are composed predominantly of randomly orientated phengite flakes, garnet, albite and some chlorite, with rare sodic amphibole (with a high glaucophane component), chloritoid, rutile and quartz (Fig. 8a).

Reactions (3) and (4) lead to the formation of glaucophane-bearing assemblages at the expense of chlorite in metapelites. Reaction (3)

170

S. GUPTA & M. J. BICKLE

has often been postulated to represent the blueschist to greenschist transition reaction (e.g. Schliestedt & Matthews 1987; Brocker 1990), and reaction (4) introduces the glaucophanechloritoid assemblage into metapelites (Guiraud et al. 1990; El-Shazly & Liou 1991; Okay & Kelley 1994). Both reactions produce water, and stabilize paragonite-glaucophane or chloriteglaucophane associations on completion. Paragonite is minor in the basement metasediments, whereas chlorite is clearly late and post-dates glaucophane-bearing assemblages. Reactions (3) and (4) are therefore unlikely to have been the main glaucophane-producing reactions during Alpine high pressure metamorphism. Reactions (5) and (6), in conjunction, can lead to the formation of glaucophane-chloritoid assemblages from biotite. These reactions have the additional merit of accounting for the observed increase in the celadonite content of white micas in the basement rocks. However, these reactions consume, rather than produce, water; if these were the main glaucophaneforming reactions, they would explain why even the restricted occurrences of sodic amphibole, as seen on Sikinos, are not observed in the extensive basement outcrops on los. Unlike sodic amphibole and chloritoid, garnet is ubiquitous in basement metapelites, irrespective of their location with respect to the contact. This may be explained through reactions such

as:

which do not require water, and consequently stabilize garnet in basement rocks irrespective of their spatial proximity to the cover. Some garnet may also be produced by the reaction:

Fig. 8. Transformation sequence from metapelitic gneiss in the basement 'core' to garnet-mica schist in the carapace, (a) SH1 gneissosity in metapelitic gneiss, defined by quartzofeldspathic and micaceous layers. Phengites in the micaceous layers are randomly orientated. Note the abundance of euhedral garnets. A crude SAI foliation cuts across SH1 [Sample 2c(i)]. (b) Remobilized phengites defining the SA1 foliation are truncated by garnets containing inclusions of glaucophane and chloritoid [Sample 116]. (c) The SA2 foliation in garnet-mica schists of the carapace wraps around garnets with distinct pressure shadows [Sample 126(ii)].

This reaction may account for the minor paragonite observed near glaucophanechloritoid-garnet assemblages.

Pressure-temperature conditions of metamorphism Pressure-temperature conditions during Alpine high-pressure metamorphism of the Sikinos basement were estimated using the program THERMOCALC (v 2.5), with the updated, internally consistent dataset of Holland &

BASEMENT-COVER CONTACT ON SIKINOS, GREECE

171

Fig. 9. Pressure-temperature estimates for high-pressure metamorphism in basement rocks calculated from the assemblage glaucophane-chloritoid-garnet-paragonite/phengite-quartz, from four samples of metapelitic gneiss and one garnet-mica schist [Sample 126(v)]. Average temperatures with their errors calculated using THERMOCALC plot as bands in P-T space. A central overlap zone is obtained from the four gneiss samples. Si content of phengite isopleths (after Massone and Schreyer 1987) are drawn for the limiting K-feldsparphlogopite-quartz observed in the granodiorite. The highest Si content (-3.4) isopleth intersects the average temperature overlap zone at 11 kbar, 475 °C. This estimate lies below the albite breakdown curve. The P-T path for the cover sequence, with the approximate timing of Db D2 and D3 deformation events is marked for comparison. The lawsonite and diaspore breakdown curves are shown for reference.

Powell (1998). The assemblage used for the thermobarometric study was glaucophanechloritoid-garnet-quartz ± paragonite ± phengite in the metapelitic gneisses and schists. Since the transformation reactions stabilizing the high pressure mineralogy in the carapace schists and adjoining gneisses are considered to be hydration reactions, finite water activities need to be incorporated into the calculations. In the cover sequence, high-pressure metamorphism was accompanied by a hydrous fluid phase, with unit water activity (Gupta 1995). In the basement, the equilibrating assemblage cannot precisely constrain water activity. Unit water activity was assumed in all calculations; lower values yield lower temperature intersections, with correspondingly lower pressure estimates. Average temperatures were calculated by THERMOCALC using two or three independent sets of reactions, with the end-members glaucophane, ferroglaucophane, magnesiochloritoid, ferrochloritoid, almandine, quartz, H2O, and either paragonite (where present) or muscovite + celadonite. Mineral compositions used were from contiguous phases, and calculated activities were based on models incorporated in the computer program AX, designed for use in the program THERMOCALC (http://www.earth-

sci.unimelb.edu.au/tpg/thermocalc/pdff/axnotes .pdf). Activities calculated are listed in Supplementary Publication No SUP 18198. The complete assemblage was detected in four samples of basement gneiss and a single sample of basement schist. Calculated temperatures in the pressure range 1-20 kbar vary from 400-480 °C for the basement gneiss samples, and plot as bands in P-T space when the errors in temperature estimates are taken into consideration (Fig. 9). Correspondence between all samples is near-perfect, with an overlap zone of calculated temperatures. Pressures were estimated using the phengite geobarometer of Massone & Schreyer (1987). The granodiorites contain the limiting assemblage of K-feldspar, phlogopite and quartz, and consequently, the silicon content isopleths give direct estimates of pressure. Using the highest observed silicon content (3.4), and the average overlap temperature recorded in the gneisses, peak metamorphic conditions of around 11 kbar, 475 °C can be inferred. The basement schist sample 126(v), however, which contains foliation parallel chlorite and is inferred to lie within an extensional shear plane, appears to have equilibrated to lower temperatures of around 435 °C at c. 10 kbar.

172

S. GUPTA & M. J. BICKLE

Discussion

The basement-cover contact on Sikinos Much of the basement-cover contact (particularly marbles of the cover sequence) on Sikinos is intensely brecciated, with minor, local discordance between carapace schists and overlying cover rocks suggesting the presence of a fault. This fault dips at shallow angles and is clearly not equivalent to the simple, high angle normal faults in the cover (D5). However, in a 150 m stretch along the western boundary of the basement (Fig. 3) the fault cuts up sequence into the cover, and an older, unbrecciated basementcover contact is visible. Here, garnet-mica schists of the basement are directly overlain by a thin (1 m), discontinuous layer of actinolite-bearing metabasites (not on mappable scale), and then by marbles of the cover sequence. Attitudes of SAi/SA2 foliations in the basement carapace are parallel to the S1/2 fabric in the overlying cover sequence. Dt and D2 deformation in the cover took place on the prograde part of the P-T loop (Fig. 9); fabrics correctable with these events in the basement therefore suggest burial, with initial juxtaposition of the two units by thrusting. Shear sense indicators are also consistent with the emplacement of the cover sequence on basement rocks. The brecciation along much of the contact is therefore a late feature, probably related to exhumation. An association with exhumation is also suggested by the presence of late chlorite-bearing extensional shear planes in the carapace which are sub-parallel to all foliations as well as the contact itself. These shears are attributed to D3 extension in the cover. In view of the prevailing confusion over the nature of the basement-cover contact on los, these observations on Sikinos assume extreme importance. Vandenberg & Lister (1996), and Forster & Lister (1999) used similar criteria (i.e. a brecciated hanging wall and ductilely deformed footwall) to characterize the contact either as a single, low angle detachment fault, or part of an anastomosing detachment system. They suggested that the contrasting rheological behaviour of the hanging wall and footwall implied derivation from different crustal levels, with significant relative motion on the low angle faults that ultimately juxtaposed the basement and cover. This study, in turn, argues that the contact was primarily a thrust, which was subsequently reactivated during D3 extension and uplift, but with little significant displacement. The point of argument, therefore, is the amount of later normal-sense movement accommodated on the original thrust plane during uplift.

Low angle detachment faults result in major excision of crustal sections, and invariably juxtapose deeper footwall rocks against shallower crustal levels. On Sikinos, such a model is not substantiated by the metamorphic data. Pressures and temperatures retrieved from the first documented complete, high-pressure metamorphic assemblage in the basement (glaucophanechloritoid-garnet-paragonite/phengite) indicate compatibility with estimates of peak pressures and temperatures (11-14 kbar, 480 ± 20 °C) from the cover sequence of Sikinos (Gupta 1995) and los (Griitter 1993). If the water activity in basement samples is considered to be lower than unity, the pressure estimates are lower. The basement therefore, represents either similar or shallower crustal levels. The contrasting rheological response of the cover and basement can be ascribed to small amounts of differential movement between hydrated schists of the carapace and marbles of the cover in the late stages of extensional uplift. However, the anhydrous basement 'core' remained rigid and brittle throughout. Evidence of differential movement along lithocontacts is also widespread in the cover sequence. For instance, nearly all marbles (even those immediately adjacent to the eastern basement outcrops) are brecciated along contacts with schists of the cover sequence. Since it is hardly likely that significant movement occurred along all lithocontacts, these features are best explained by relative movement between lithologies of contrasting competence in response to extensional deformation associated with uplift (Gupta 1995). It is thus unnecessary to assume that a large part of the basement crust has been excised during uplift.

High pressure metamorphic imprint in the Cycladic basement Irrespective of the difference (or similarity) between exposed structural levels of basement and cover rocks, the basement underlies the cover sequence and was subjected to the Eocene high-pressure event M!. What is more difficult to explain is why mineralogical evidence of M! in the basement is so restricted. In particular, the apparent scarcity of sodic pyroxene (jadeite) and sodic amphibole (glaucophane) (Van der Maar 1981) must be accounted for, given the extensive outcrops of granitoid rocks and metapelites on los and Sikinos. P-T estimates for high-pressure metamorphism of the Sikinos basement (see Fig. 9) indicate that peak pressures did not exceed the pure albite breakdown curve. Indeed, plagioclase

BASEMENT-COVER CONTACT ON SIKINOS, GREECE

laths associated with igneous textures, and perthites containing pure albite show no evidence of breakdown to jadeite and quartz; this is in consonance with the pressures inferred from application of the phengite geobarometer to the limiting assemblage of K-feldspar-phlogopitequartz preserved in the granodiorites. These pressures are compatible with those in the cover sequence (Gupta 1995), where, too, rocks were not subjected to metamorphic conditions necessary for albite breakdown. Consequently, the absence of jadeite from basement rocks is easily explained. However, sodic amphibole can be stabilized under these conditions and accounting for its rarity is more complicated. Appropriate bulk compositions for developing sodic amphibole obviously did exist, since some gneisses and garnet-mica schists do contain some (albeit rare) glaucophane. These metapelites are also not substantially different from those in the cover, which contain abundant glaucophane. Since reactions leading to sodic amphibole formation in the basement gneisses and schists are essentially hydration reactions, they can only take place provided adequate quantities of hydrous fluids are available. The transition from granodiorites in the basement 'core' to microcline-quartz schists in the 'carapace' is accompanied by the addition of water. The availability of hydrous fluid increases notably in the vicinity of the contact. Ductile deformation of the carapace occurs concomitantly with hydration reactions, and the rheology appears to be controlled by transformation and flow of feldspars rather than quartz (Janecke & Evans 1988). Away from the carapace, the limited effects of ductile strain, well-preserved pre-Alpine structures, and occurrence of brittle faults not seen in the carapace suggest that the basement 'core' remained dry and brittle through the Alpine event. Sodic amphibole was restricted to those metapelitic gneisses in which Alpine fabrics developed, allowing percolation of water. The most intensely strained basement rocks, which comprise the carapace, contain the highest proportion of hydrous minerals. Correspondingly, rocks with lower strain (in the 'core') preserve limited evidence of hydrous alteration, with only incomplete replacement of plagioclase and biotite by phengite. These replacement reactions are characteristic of high pressure metamorphism, and are therefore attributed to the Alpine event. The limited progress of hydration reactions in the weakly deformed basement 'core' compared to the carapace suggest higher flux of water into zones which are more strained. This suggests channelling of fluids into the shear zone, thereby con-

173

fining the high-pressure imprint to zones within and immediately adjacent to the 'carapace'. This rheological behaviour of the basement persisted into the M2 greenschist facies event, since late extensional shears related to uplift are also restricted to the carapace, leaving little or no impression on the 'core'. Indeed, there is evidence that some glaucophane-bearing assemblages in the carapace may have been obliterated by later fluids accompanying M2 metamorphism. The source of the hydrous fluids, however, remains enigmatic. Since high pressure metamorphism in the cover also occurred in the presence of a hydrous fluid phase (Gupta 1995), the cover sequence itself may have been a possible source. An important implication of these observations is that, inspite of being subjected to high pressure, low-temperature conditions, the basement did not generally develop the characteristic blueschist facies mineralogy and therefore maintained its low density throughout the Eocene and Miocene events. Thus, the buoyancy of the basement may have assisted in the subsequent uplift of the associated blueschists and eclogites. Conclusions Field relations and structures along the basement-cover contact on Sikinos confirm that the blueschist facies cover sequence was thrust onto older crystalline basement during the Eocene high-pressure event. The basement rocks are extensively mylonitized at the contact, generating a 'carapace' zone of deformed rocks which grades into the weakly deformed basment 'core'. A series of hydration reactions accompanied the transformation of the basement granodiorites to microcline schists of the carapace. Similarly, the basement metapelitic gneisses are converted to garnet-mica schists in the carapace. Rareglaucophane-chloritoid-garnet assemblages in the gneisses and schists are also attributed to hydration reactions, and confirm that the Cycladic basement was indeed subducted during the Eocene metamorphic event. Much of the present basement-cover contact on Sikinos is a late fault that reactivated the older thrust plane, possibly during the Oligo-Miocene extensional event. However, this cannot be a major detachment surface involving excision of a large portion of the basement, since pressure-temperature estimates from basement rocks are identical to those retrieved from the cover. The reason for the lack of diagnostic high pressure assemblages in the Cycladic basement is

174

S. GUPTA & M. J. BICKLE

therefore two-fold: metamorphic P-7 conditions did not cross the pure albite breakdown curve to form jadeite and quartz, and hydrous fluid percolation was restricted to the contact shear zone, resulting in the bulk of the basement remaining relatively anhydrous and rigid, with sodic amphibole-forming hydration reactions restricted to the vicinity of the contact. Earlier workers (e.g. Austrheim 1987; Boundy et al 1992; Kullerud et al. 2001) have suggested 'eclogitization' of parts of the crust as a consequence of fluid infiltration along shear zones. In the Cyclades, since the bulk of the basement was not 'eclogitized', its density remained low through the entire event, and may have contributed to the uplift of the overlying Cycladic high pressure rocks after being detached from the subducting slab. This work forms part of the Ph.D. thesis of S. Gupta at the University of Cambridge, UK. S.G. thanks the Nehru Trust for Cambridge University for the Nehru Scholarship during the tenure of the work. Financial assistance from the Cambridge Philosophical Society and the Lundgren Research Body is also gratefully acknowledged. I.G.M.E., Athens, is thanked for permitting the conduct of the study. J. Baker is especially thanked for advice during fieldwork, and enthusiastic discussions at various stages of the study. H. Grutter, R. Poulton and T. Holland are also thanked for their assistance during the course of the work. The manuscript has benefited from the critical reviews and suggestions made by Bruce Yardley and Rebecca Jamieson.

References ANDRIESSEN, P.A.M., BANGA, G. & HEBEDA, E. 1987. Isotopic age study of pre-Alpine rocks in the basal units on Naxos, Sikinos and los, Greek Cyclades. Geologic en Mijnbouw, 66, 3-14. AUSTRHEIM, H. 1987. Eclogitization of lower crustal granulites by fluid migration through shear zones. Earth and Planetary Science Letters, 81, 221-232. BOUNDY,T.M., FOUNTAIN, D.M. & AUSTRHEIM, H. 1992. Structural development and petrofabrics of eclogite facies shear zones, Bergen arcs, western Norway: implications for deep crustal deformational processes. Journal of Metamorphic Geology, 10,127-146. BROCKER, M. 1990. Blueschist-to-greenschist transition in metabasites from Tinos Island, Cyclades, Greece: Compositional control or fluid infiltration? Lithos, 25, 25-39. BUICK, I.S. & HOLLAND, TJ.B. 1991. The nature and distribution of fluids during amphibolite facies metamorphism, Naxos (Greece). Journal of Metamorphic Geology, 9, 301-314. CHOPIN, C. 1984. Coesite and pure pyrope in highgrade pelitic blueschists in the Western Alps: a first record and some consequences. Contributions to Mineralogy and Petrology, 86,107-118. DAVIES, J.H. & VON BLANCKENBERG, F. 1995. Slab

breakoff: a model of lithosphere detachment and its test in the magmatism and deformation of collisional orogens. Earth and Planetary Science Letters, 129, 85-102. DAVIS, G.H. 1980. Structural characteristics of metamorphic core complexes, southern Arizona. Geological Society of America Memoirs, 153, 35-77. DURR, S., ALTHERR, R., KELLER, J., OKRUSCH, M. & SEIDEL, E. 1978. The Median Aegean Crystalline Belt: stratigraphy, structure, metamorphism, magmatism. In: ROEDER, H.C. & SCHMIDT, K. (eds) Alps, Apennines, Hellenides. Stuttgart, 455-477. EL-SHAZLY, A.K. & Liou, J.G. 1991. Glaucophane chloritoid-bearing assemblages from NE Oman: petrologic significance and a petrogenetic grid for high P metapelites. Contributions to Mineralogy and Petrology, 107,180-201. EVANS, B.W. 1990. Phase relations of epidoteblueschists. Lithos, 25, 3-23. FORSTER, M.A. & LISTER, G.S. 1999. Detachment faults in the Aegean core complex of los, Cyclades, Greece. In: RING, U, BRANDON, M.T., LISTER, G.S. & WILLETT, S.D. (eds) Exhumation processes: normal faulting, ductile flow and erosion. Geological Society, London, Special Publications, 154, 305-323. FRANZ, L., OKRUSCH, M. & BROCKER, M. 1993. Polymetamorphic evolution of the pre-Alpidic basement rocks on the island of Sikinos (Cyclades, Greece). Neues Jahrbuch fiir Mineralogie Monatschefning, H.4,145-162. GIBBONS, W. & HORAK, J. 1984. Alpine metamorphism of Hercynian hornblende granodiorite beneath the blueschist facies schistes lustres nappe of NE Corsica. Journal of Metamorphic Geology, 2, 95-113. GRUTTER, H.S. 1993. Structural and metamorphic studies on los, Cyclades, Greece. Unpublished Ph.D. thesis, University of Cambridge, UK. GUIRAUD, M., HOLLAND,T. & POWELL, R. 1990. Calculated mineral equilibria in the greenschistblueschist-eclogite facies in Na2O-FeO-MgOAl2O3-SiO2-H2O. Methods, results and geological applications. Contributions to Mineralogy and Petrology, 104, 85-98. GUPTA, S. 1995. Structure and metamorphism of Sikinos, Cyclades, Greece. Unpublished Ph.D. thesis, University of Cambridge, 248 pp. HEINRICH, W. & ALTHAUS, E. 1980. Die obere stabilitatsgrense von lawsonit plus albit bzw jadeit. Fortschritte der Mineralogie, 58, 49-50. HENJES-KUNST, F. & KREUZER, H. 1982. Isotopic dating of the pre-Alpidic rocks from the island of los (Cyclades, Greece). Contributions to Mineralogy and Petrology, 80, 245-253. HERMANN, J. 2002. Experimental constraints on phase relations in subducted continental crust. Contributions to Mineralogy and Petrology, 143, 219-235. HOLLAND, TJ.B. & BLUNDY, J.D. 1994. Non-ideal interactions in calcic amphiboles and their bearing on amphibole-plagioclase thermometry. Contributions to Mineralogy and Petrology, 116, 433-447.

BASEMENT-COVER CONTACT ON SIKINOS, GREECE HOLLAND,TJ.B. & POWELL, R. 1990. An enalarged and updated internally consistent thermodynamic dataset with uncertainties and correlations: the system K2O-Na2O-CaO-MgO-MnO-FeOFe2O3-Al2O3-TiO2-SiO2-C-H2-O2. Journal of Metamorphic Geology, 8, 89-124. HOLLAND, TJ.B. & POWELL, R. 1998. An internally consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology, 16, 309-343. JANECKE, S.U & EVANS, J.P. 1988. Feldspar-influenced rock rheologies. Geology, 16,1064-1067. JANSEN, J.B.H. & SCHUILING, R.D. 1976. Metamorphism on Naxos: petrology and geothermal gradients. American Journal of Science, 276,1225-1253. KULLERUD, K., FLAAT, K. & DAVIDSEN, B. 2001. Highpressure fluid-rock reactions involving Clbearing fluids in lower-crustal ductile shear zones of the Flakstad0y Basic Complex, Lofoten, Norway. Journal of Petrology, 42,1349-1372. LISTER, G.S., BANGA, G. & FEENSTRA, A. 1984. Metamorphic core complexes of the Cordilleran type in the Cyclades, Aegean Sea, Greece. Geology, 12, 221-225. MAKRIS, J. 1978. The crust and upper mantle of the Aegean region from deep seismic soundings. Tectonophysics, 46, 269-284. MASCLE, J. & MARTIN, L. 1990. Shallow structure and recent evolution of the Aegean Sea: a synthesis based on continuous refection profiles. Marine Geology, 94, 271-299. MASSONE, H.J. & SCHREYER, W. 1987. Phengite geobarometry based on the limiting assemblage with K-feldspar, phlogopite and quartz. Contributions to Mineralogy and Petrology, 96, 212-224. OKAY, A.I., Xu, S. T. & SENGOR, A.M.C. 1989. Coesite from the Dabie Shan eclogites, central China. European Journal of Mineralogy, 1, 595-598. OKAY, A.I. & KELLEY, S.P. 1994. Tectonic setting, petrology and geochronology of jadeite + glaucophane and chloritoid + glaucophane schists from north-west Turkey. Journal of Metamorphic Geology, 12, 455-466.

175

PAPANIKOLAOU, D.J. 1987. Tectonic evolution of the Cycladic blueschist belt (Aegean Sea, Greece). In: HELGESON, H.C. (ed.) Chemical transport in metasomatic processes. Reidel, 429-450. PRYER, L.L. & ROBIN, P.Y.F. 1995. Retrograde metamorphic reactions in deforming granites and the origin of flame perthite. Journal of Metamorphic Geology, 14, 645-658. SCHLIESTEDT, M. & MATTHEWS, A. 1987. Transformation of blueschist to greenschist facies rocks as a consequence of fluid infiltration, Sifnos (Cyclades), Greece. Contributions to Mineralogy and Petrology, 97, 237-250. VANDENBERG, L.C. & LISTER, G.S. 1996. Structural analysis of basement tectonites from the Aegean metamorphic core complex of los, Cyclades, Greece. Journal of Structural Geology, 18, 1437-1454. VAN DER MAAR, PA. 1980. The geology and petrology of los, Cyclades, Greece. Annales Geologiques des Pays Helleniques, 30, Athens, 206-224. VAN DER MAAR, P.A. 1981. Metamorphism on los and the geological history of the southern Cyclades, Greece. Geologica Ultrajectina, 28,1-142. VAN DER MAAR PA. & JANSEN, J.B.H. 1983. The geology of the polymetamorphic complex of los, Cyclades, Greece, and its significance for the Cycladic Massic. Geologisches Rundschau, 72, 283-299. VAN DER MAAR, PA., FEENSTRA, A., MANDERS, B. & JANSEN, J.B.H. 1981. The petrology of the island of Sikinos, Cyclades, Greece, in comparison with that of the adjacent island of los. Neues Jahrbuch fur Mineralogie Monatschefning, 10, 459-469. VERNON, R.H. 1999. Flame perthite in metapelitic gneisses at Cooma, SE Australia. American Mineralogist, 84,1760-1765. WERNICKE, B. & BURCHFIEL, B.C. 1982. Modes of extensional tectonics. Journal of Structural Geology, 4,105-115.

This page intentionally left blank

Shear zone folds: records of flow perturbation or structural inheritance? G. I. ALSOP1 & R. E. HOLDSWORTH2 l Crustal Geodynamics Group, School of Geography & Geosciences, University of St Andrews, St Andrews, Fife, Scotland KYI6 9AL, UK (e-mail: gia@st-andrews. ac. uk) ^Reactivation Research Group, Department of Earth Sciences, University of Durham, Durham, DH1 3LE, UK Abstract: Deformation within shear zones can be both temporally and spatially variable, resulting in multiple generations of folds which display a range of scales and overprinting relationships in mylonitic rocks associated with high strain zones. Despite such complexities, two main fold associations are broadly recognized in many shear zone settings: early tight to isoclinal sheath folds, often with mylonitic limbs that are post-dated by one or more local generations of synshearing folds which are preserved within, or root downwards into mylonitic high strain zones. These latter structures locally fold the mylonitic foliation and lineation whilst displaying geometric characteristics that are kinematically compatible with the movement regime of the major shear zone. Using examples related to ductile thrusting in Moine metasediments of north Scotland, we show that both types of fold display predictable geometric patterns on fabric topology plots. Fold axes and axial surfaces display consistent changes in asymmetry and sense of obliquity relative to local, transport-parallel mineral lineations that can be used to map out a series of culminations and depression zones. The sheath folds preserve more acute, but almost identical geometric patterns compared to the later synshearing folds, with culmination and depression zones often coinciding in location and scale. Detailed analysis also demonstrates that the distribution of finite strain is systematically linked to the architecture of all folds and that clear and predictable relationships exist between the fabric topologies of both the sheath folds and synshearing folds. These consistent topological relationships could be explained in terms of a fold evolution model, where sheath folds represent a more highly deformed and evolved variety of synshearing folds originally generated during perturbations in ductile flow. However, an alternative fold inheritance model predicts that the gross structural architecture generated during sheath folding may subsequently control the geometry and govern the orientation of the synshearing folds. Both models may be widely applicable in a broad range of shear zone environments.

Within orogenic systems, there is a close association between folding and localized zones of enhanced ductile deformation forming shear zones. This relationship has been studied by generations of geologists working on a variety of scales and tectono-metamorphic environments, Although a broad link between both contractional and extensional shear zones and folding is well established (e.g. see Harris et al. 2002 for a review), precise geometrical relationships between planar and linear shear zone fabrics with fold morphologies is the subject of much past and current research (e.g. Ramsay 1967; Hansen 1971; Coney 1974; Ramsay 1979; Fossen & Rykkelid 1990; Dillon et al 1990; Carreras 1997; Bolhar & Ring 2001). In particular, the role of folding with respect to the timing of shear zone development and progressive deformation within mylonites, together with the kinematic

interpretation of folding are the foci of continuing research (e.g. Fletcher & Bartley 1994; Ez 2000; Ramsay & Lisle 2000). This study concentrates on the evolution of shear zone-related folding and in particular the generation of synshearing flow folds, together with those modified by progressive shear culminating in the development of sheath folds (Fig. 1). Two broad fold associations are typically recognized within shear zones. First, an early phase of tight to isoclinal, highly curvilinear folds with mylonitic limbs and 'low-strain' hinges which may be cross-cut at low angles by broader zones of high strain such as ductile thrusts (e.g. Carreras et al. 1977; Quinquis et al 1978; Minigh 1979; Cobbold & Quinquis 1980; Berthe & Brun 1980; Henderson 1981; Lacassin & Mattauer 1985; Holdsworth 1989). Models of sheath folding are classically applied to such

From: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224,177-199. 0305-8719/$15.00 © The Geological Society of London 2004.

178

G. I. ALSOP & R. E. HOLDSWORTH

Fig. 1. Summary schematic diagrams illustrating increasing deformation and evolution of fold hinge-lines associated with sheath folds (a,b) and synshearing flow folds (c,d). Diagrams are plan view and show clockwise (Cw) and anticlockwise (A-Cw) rotations of fold hinge-lines which are marked by reversals in fold facing directions about the transport-parallel culmination and depression surfaces. Synshearing flow folds are generated by surging flow (large transport (Ln) arrow) and slackening flow (small transport (Ln) arrow) separated by differential sinistral and dextral shear. See text for further details.

folds (see Alsop & Holdsworth 1999 for a review) (Fig. la,b). Secondly, one or more generations of later, synshearing folds occur, which are temporally, spatially and kinematically related to the shear zones (e.g. Carreras et al. 1977; Bell 1978; Cobbold & Quinquis 1980; Evans & White 1984; Ridley 1986; Holdsworth 1990). Models of flow perturbation folding may be applied to these folds (see Alsop & Holdsworth 2002 for a review) (Fig. lc,d). The analysis of consistent and ordered scale-independent relationships between the transport direction and fold hinges, fold asymmetries and axial surfaces can provide a record of shear zone evolution within a broad range of coherent kinematic systems. Following a brief summary of both sheath and synshearing flow folding, details of a case study from north Scotland are presented which examines the sequential fabric relationships that exist between such evolving fold systems.

Sheath folds Sheath folds are defined here as folds in which high strain has been superimposed on the original fold geometry so that the hinges now display a curvature in excess of 90° (see also Ramsay & Huber 1987 p. 311) (Fig. la,b). Sheath folds may be further subdivided into tongue folds with a hinge line curvature between 90°-160° and tubular folds with a hairpin-like curvature above 160° (Skjernaa 1989). The subparallelism of large segments of curvilinear hinge-lines with the mineral elongation lineation (X) in high strain settings may be achieved via a variety of mechanisms. Accepted traditional explanations for co-linearity of hinges and mineral lineations assume bulk simple shear or plane strain non-coaxial flow in which fold hinges rotate (typically with associated interlimb tightening) towards the transport-parallel mineral lineation (X) during intense ductile

SHEAR ZONE FOLDS shearing (Bryant & Read 1969; Sanderson 1973; Escher & Watterson 1974; Rhodes & Gayer 1977; Bell 1978; Williams 1978; Mies 1991) (Fig. la,b). In such flow regimes, folds that are generated oblique to transport will display unidirectional rotation with all hinges rotating in the same (clockwise or anticlockwise) sense which is determined by the initial sense of hinge-transport obliquity. This results in asymmetric, typically tight-isoclinal cylindrical folds which, due to incomplete rotation, frequently preserve a small angle of obliquity to the transport lineation (see Alsop 1992). Fold hinges which initiate broadly orthogonal to shear may display reversals in the sense of hinge/lineation obliquity along their length, reflecting the initial growth and lateral propagation of the buckle fold hinge (Fig. la). Such mildly curvilinear hinges may subsequently undergo opposing senses of relative rotation resulting in extremely curvilinear sheath fold forms. Opposing senses of (clockwise and anticlockwise) rotation result in reversals in the polarity of minor fold facing which is directed outwards away from upwardsclosing antiformal culminations and inwards towards downwards-closing synformal depressions (e.g. Holdsworth 1988; Alsop & Holdsworth 1999) (Fig. Ib). Originally consistent buckle fold geometries which undergo opposing rotations at opposite ends will display either a S or Z down-plunge asymmetry depending on the sense of rotation, thus resulting in double-vergence geometries along the length of the individual hinge (e.g. Hobbs et al. 1976 p. 168; Holdsworth & Roberts 1984; Alsop & Holdsworth 1999). Reversals in vergence and the polarity of minor fold facing therefore mark transport-parallel and foliation-normal culmination/depression (medial) surfaces which bisect the resulting curvilinear hinge closures (Fig. Ib; Alsop & Holdsworth 1999). Thus, progressive ductile shearing is considered to result in the modification and rotation of originally highangle, gently curved fold hinges towards the transport direction. An alternative mechanism in which synshearing fold geometries initiate at variable angles to transport as a consequence of perturbations in ductile flow within shear zones will now be considered.

Synshearing flow folds Curvilinear fold patterns may be generated adjacent to shear zones via variations in the relative rate of ductile flow. Temporal and spatial variations in the shear strain rate can result in localized perturbations in flow which

179

may involve acceleration (surging flow) or deceleration (slackening flow) with respect to the adjacent flow (see Coward & Potts 1983; Platt 1983; Ridley 1986; Holdsworth 1990; Alsop & Holdsworth 1993, 2002; Alsop et al. 1996). Surging flow results in overall contraction at the front of the deformation cell resulting in folding and the development of transport-parallel and foliation-normal culmination surfaces which bisect the cell (Fig. Ic; Alsop & Holdsworth 2002). Conversely, slackening flow leads to a mirror image configuration with contraction and thickening at the rear of the cell associated with depression surfaces (Fig. Ic). On the lateral flanks of flow perturbation cells, S folds (viewed down-plunge) are developed anticlockwise of the mineral lineation during differential dextral shear, whereas Z folds are produced by differential sinistral shear (Fig. Id; Coward & Potts 1983; Alsop & Holdsworth 1993). Although analysis of synshearing folds may provide evidence of the scale and nature of heterogeneous flow and deformation cells in shear zones (e.g. Alsop et al. 1996), many uncertainties still exist. Although flow perturbations represent a plausible model to explain the generation of variably orientated synshearing folds, questions remain as to what extent the attitude of the pre-existing structures within the mylonite may influence the geometry and orientation of such folds. Clearly, variably orientated layering which is oblique to the strain ellipsoid will inherently create a range of possible fold orientations (e.g. Flinn 1962; Treagus & Treagus 1981: Carreras 1997). In order to address these issues, a detailed case study of curvilinear sheath folds and variably orientated synshearing folds from the Caledonides of north Scotland is described (Fig. 2).

Regional setting; Caledonian Moine Nappe, Sutherland The regional geology of NW Scotland is dominated by structures generated during the Lower Palaeozoic Caledonian orogeny (c. 460-420 Ma), the most important of which are the ESEdipping Moine Thrust and overlying Sgurr Beag/Naver Thrust (Fig. 2). In Sutherland, these thrusts, which underlie the Moine and Naver Nappes respectively, formed as a result of major west to NW-directed thrusting. The Moine Thrust separates imbricated, essentially unmetamorphosed sedimentary sequences overlying Precambrian metamorphic basement of the Lewisian complex (western foreland)

180

G. I. ALSOP & R. E. HOLDSWORTH

Fig. 2. Simplified geological map of the Moine and Naver Nappes in the Kyle of Tongue area highlighting the location of the study area (see also British Geological Survey 1997, 2002). Major NW-toWNW-directed Caledonian ductile thrusts which carry reworked Lewisian in the hanging wall are shown with solid barbs. The reference grid relates to the UK National Grid with the map area falling within the NC prefix quadrangle. The inset shows the location of the map in relation to Scotland with the Moine and Naver Nappes being carried on Caledonian thrusts.

from metamorphosed Neoproterozoic psammites and pelites forming the Moine Nappe. Further east the Naver Thrust juxtaposes migmatized (in the hanging wall) and nonmigmatized Moine psammites. Within both the Moine and Naver Nappes, psammites and subordinate pelites of the Moine Supergroup are tectonically interleaved with high grade, Lewisian-like basement upon which the Moine is thought to have been unconformably

deposited (Holdsworth 1989). Tectonic repetition is generated by Caledonian-age ductile thrusting and folding that forms the main (D2) fabrics observed throughout the region. These Caledonian structures, which are the focus of the present study, consistently overprint earlier D! fabrics and associated metamorphic assemblages considered to be of Neoproterozoic age (Holdsworth 1989). Regional D2 deformation has resulted in a gently east-dipping, approximately beddingparallel foliation (Sn) which intensifies into broad zones of platy mylonite that define Caledonian ductile thrusts formed at mid-crustal depths (greenschist-amphibolite facies) (see Holdsworth et al. 2001 and references therein). Within the plane of the foliation, a pronounced mineral lineation (Ln) defined by aligned micas and elongate quartz - feldspar aggregates plunges gently towards the east directly down the dip of the foliation. This is considered by most authors to represent the trend of tectonic transport within a regime of non-coaxial flow associated with bulk plane strain (Holdsworth & Grant 1990). D2 deformation has resulted in two generations of folding (termed F2 and F3) which show distinct and consistent relationships to the regional foliations and lineations. F2 folds of bedding display an axial planar foliation and are marked by highly curvilinear hinge-line patterns on all scales that have been interpreted previously in terms of sheath fold models (e.g. Holdsworth 1987, 1989, 1990; Alsop & Holdsworth 1999). F3 folds of the regional foliation and lineation are spatially associated with the D2 ductile thrust zones, and display axial traces that vary between transport parallel and transport normal. Such folds are developed in both the Moine and Naver Nappes (e.g. Holdsworth 1990; Alsop & Holdsworth 1993; Alsop et al. 1996) and have been previously interpreted in terms of flow perturbation models (e.g. Alsop & Holdsworth 2002). Detailed geometric analyses of the F2 sheath folding and F3 flow perturbation folding have been documented previously from the region in separate publications (Alsop & Holdsworth 1993, 1999, 2002). This study presents the first quantitative analysis of the relationships and associations that exist between F2 and F3 structures in a well-exposed region of the Moine Nappe (Figs 2 & 3). Similarities and differences between these folds will be described prior to a discussion on the origins of folding in terms of fold evolution models associated with long-lived flow perturbations in ductile mylonites and fold inheritance models governed by inherited structural architecture.

SHEAR ZONE FOLDS

181

Fig. 3. (a) Simplified structural map of the study area highlighting the location of ductile thrusts and a major WNW-trending culmination trace which divides the region in to northern and southern subareas. Summary facing domains (based on younging evidence) are highlighted in close stipple. When viewed down-plunge, culmination and depression surfaces are offset in a consistent sinistral (top-to-the-north) sense by E>2 ductile thrusts located along pelitic units, suggesting strain localization and continued progressive deformation within these weaker horizons. Detailed structural analysis within Lewisian and Moine rocks of the study area illustrate the vergence and facing relationships of minor folds, together with the relative obliquity between fold hinges and the Ln mineral lineation. The inset key shows the relative structural position of schematic (F3) flow folds (larger symbols on upper row) and (F2) sheath folds together with stereonet symbols. Equal area lower hemisphere stereographic projections of all fold hinges (b) subdivided into northern (c) and southern (d) domains clearly show clockwise rotation of north-facing hinges (stipple) and anticlockwise rotation of southfacing folds. Poles to all associated axial planes (e) subdivided into northern (f) and southern (g) domains clearly show anticlockwise rotation of S axial planes and clockwise rotation of Z axial planes towards the mean strike of the regional (Sn) foliation. Note that the intersection of the great circles representing the mean Z and S axial planes (starred) is typically parallel to the transport lineation (Ln). Solid and open symbols represent north-facing and south-facing structures respectively. Refer to Fig. 2 for location and text for further details.

The Melness folds case study The Melness area to the west of the Kyle of Tongue is an ideal place to compare and contrast sheath and synshearing folds as it contains diverse lithologies interlayered on all scales which encourages the development of folds. In addition, it is well exposed and readily accessible. Structurally, the Melness area is dominated by a Caledonian ductile thrust stack associated

with top-to-the-WNW shear criteria including the Achininver and Ben Hope thrusts which collectively form the Talmine Imbricate Zone (Holdsworth et al. 2001) (Figs 2 & 3a). The ductile thrusts have been openly folded by kilometre-scale folds which are predominantly parallel to the WNW-trend of tectonic transport and separate north and south facing domains. These upright folds define culmination/ depression surfaces which root downwards onto

182

G. I. ALSOP & R. E. HOLDSWORTH

underlying thrusts and have previously been attributed to perturbations in ductile flow within the thrust-related mylonitic shear zones (Holdsworth 1990; Alsop & Holdsworth 1993, 2002; Alsop et al 1996) (Figs 2 & 3a). Major culmination surfaces are orthogonal to the regional foliation surface and parallel to the transport direction marked by the mineral lineation. The largest WNW-ESE trending culmination surface in the study area intersects the gently east-dipping form of the detachment directly beneath the Moine Thrust Zone mylonites (Figs 2 & 3a). The planar geometry of this structure in the hinge of the culmination suggests that it has not been affected by the synshearing flow folding and has acted as the local basal decollement to the deformation (Alsop et al. 1996; Holdsworth et al. 2001 Alsop & Holdsworth 2002). Thus, the Melness study area is marked by a major culmination, which together with the presence of earlier F2 antiformal sheath folds cored by Lewisianoid inliers results in a dome-dominated structural setting (see Alsop & Holdsworth 1999, 2002 for further details). Detailed descriptions of the (F2) sheath folds and (F3) synshearing flow folds generated during the regional (D2) progressive ductile thrusting are presented below.

Synshearing (F3) flow folds F3 folds of the Melness area may be broadly categorized into major and minor structures (e.g. Holdsworth 1990). Major structures include the large-scale transport-parallel antiformal culmination and synformal depressions noted above (Fig. 3a). These open, upright folds lack overturned limbs and can be traced across strike for several kilometres where they may die out or root downwards into underlying ductile thrust mylonite zones (Fig. 3a). Minor F3 folds in the Melness area display complex field relationships with both Z and S folds often observed in single exposures. In some cases (e.g. NC 5188 6518), open Z folds are seen to refold earlier S folds (or vice versa), whereas in other situations changes in vergence arise due to the development of curvilinear hinge geometries (e.g. NC 5153 6509). In contrast, many exposures preserve both Z and S folds where no refolding or hinge curvature is apparent and, in these cases, the folds seem to form a conjugate set of asymmetric structures (e.g. NC 5127 6590) (Fig. 4a). In mylonites relatively unaffected by minor folding, centimetrespaced extensional shear bands are developed, consistently indicating top-to-the-west senses of shear (e.g. NC 5145 6675). Minor F3 folds are

best developed within and adjacent to mylonite zones and define extremely variable orientations ranging from transport-normal to transport-parallel. The dominant type of minor F3 folds are open-close, gently east-plunging cylindrical structures with hinges trending obliquely or subparallel to the adjacent mineral lineation (Ln) (Fig. 4b). Some, close to isoclinal curvilinear F3 fold hinges exhibit 90° of curvature and display overall west-directed vergence. Although F3 folds clearly refold both the mylonitic fabric (Sn) and the east-plunging transport lineation (Ln), they are considered to form part of a kinematicaly linked system of progressive deformation related to continuing west-directed ductile thrusting (Alsop & Holdsworth 2002). There appears to be a change in the dominant sense of minor fold vergence either side of the major culmination trace with Z folds predominating to the south whereas S folds are more common in the region to the north (Fig. 3a). This evidence has in the past been used to suggest that flow perturbations exist on a larger scale, forming structures analogous to surge zones in the more external parts of thrust belts (e.g. Alsop & Holdsworth 1993). Within the Melness area, S fold hinges are north facing and on average lie about 26° anticlockwise of the mineral lineation (Ln) defining the transport direction. Z folds are south facing and on average lie about 26° clockwise of Ln (Figs 3a,b & 5a,b). Minor F3 fold hinges measured on the overturned limbs of major F2 folds typically lie closer to the mineral lineation with mean F3 Z fold hinges 10° clockwise and mean F3 S fold hinges 8° anticlockwise of Ln. The strike of the associated fold axial surfaces also displays similar systematic relationships, with mean Z fold axial surfaces around 50° clockwise and mean S fold axial surfaces about 57° anticlockwise with respect to Ln (Figs 3a,e,f,g & 5a,b). The strike of minor F3 axial planes measured on the overturned limbs of major F2 folds is typically developed closer to the transport lineation. Thus, both S and Z fold hinges and axial planes display bimodal peaks on distribution graphs (Figs 5a,b & 6a,b) reflecting differing degrees of hinge and axial planar rotation towards ductile thrusts with higher strains concentrated on the overturned limbs of F2 folds (see Alsop & Holdsworth 2002). Both S and Z F3 fold hinges and axial planes display broadly similar orientations on either side of the major culmination trace, indicating a lack of subsequent folding or tilting around the major culminations and depressions (Figs 3a-g & 6a,b) (see also discussion of transection and topological relationships between F2 and F3 folds). On

SHEAR ZONE FOLDS

183

Fig. 4. (a) Asymmetric F3 folds showing a reversal in fold vergence across a minor culmination in mylonites derived from mixed acid-intermediate Lewisianoid gneiss. Photograph looking towards the east down the plunge of F3 folds on the eastern flank of Cnoc nan Gobhar, Melness. (b) Typical minor F3 asymmetric folds in mylonites derived from mixed acid-intermediate Lewisianoid gneiss. Note traces of refolded F2 isoclines and folding of mylonitic foliation. Photograph looking towards the east down the plunge of F3 folds on the eastern flank of Cnoc nan Gobhar, Melness. (c) F2 fold in Moine psammites with more highly attenuated lower (inverted) limb. The hinge of this fold plunges ESE subparallel to the regional mineral lineation (Ln). Photograph looking towards the east on the west flank of Creag Mhor. South of Port Vasgo, Melness. (d) Minor F2 sheath folds in interbanded Moine psammites and semi-pelites showing typical eye-structures and along-strike double vergence patterns. Photograph looking down plunge towards the east on the summit region of Creag Mhor, South of Port Vasgo, Melness.

stereographic projections, mean F3 S and Z axial planes intersect precisely parallel to the mineral lineation (Ln) (Fig. 3e-g). In summary, these minor structures appear to be typical of progressive deformation patterns developed during ductile flow in many mylonites (e.g. Evans & White 1984; Ridley 1986). Minor (F3) synshearing flow folds are mostly gentle to close fold pairs which typically plunge gently east, with associated shallowly east-dipping axial planes. Few curvilinear folds are preserved, with F3 structures dominated by cylindrical asymmetric S fold hinges and axial planes displaying anticlockwise obliquity to Ln, whereas Z fold hinges and axial planes are clockwise of Ln. All folds are associated with abundant top-to-the-WNW S-C fabrics and shear criteria confirming a progressive system of deformation associated with (02) ductile thrusting. The geometry of earlier (F2) sheath folds, which were generated during

the same progressive (D2) deformation, is now described.

Sheath fold (F2) structures Within the Melness area, large-scale transportparallel antiformal culminations and synformal depressions associated with major F2 folds almost exactly coincide with the F3 medial surfaces described above (see Holdsworth 1990; Alsop & Holdsworth 1993, 2002; Alsop et al. 1996; Fig. 3a). Minor (F2) sheath fold hinges are variably distributed over arcs approaching 90° from Ln, with north facing fold hinges consistently developed in a anticlockwise sense to Ln, whereas south facing hinges are clockwise (Fig. 3a,b). The relative structural position of minor folds about the major sheath culmination demonstrates that mean minor folds on the major overturned lower limb (i.e. north-facing Z

184

G. I. ALSOP & R. E. HOLDSWORTH

Fig. 5. Detailed structural analysis of folding within the study area illustrating the vergence and obliquity relationships of minor north- and south-facing S and Z folds respectively on either side of the major culmination (refer to Figure 2 for location). The relative obliquity between S and Z fold hinges and associated axial planes with the Ln mineral lineation is shown on rose diagrams for transecting (F3) flow folds (a,b), folds forming the upper limbs of (F2) sheath folds (c,d) and folds on the overturned lower limbs of sheath folds (e,f). Fold hinge and axial plane obliquity is measured relative to the adjacent mineral lineation (Ln) and is described as clockwise (Cw) or anticlockwise (A-Cw). F3 S fold hinges are north facing and on average lie 26° anticlockwise of the mineral lineation (Ln) defining the transport direction. Z folds are south facing and on average lie 26° clockwise of Ln. A consistent overall relationship exists between fold facing and sense of obliquity of minor fold hinges with north-facing hinges being (92%) anticlockwise and south-facing hinges (91%) clockwise of Ln. Associated S axial planes strike (82%) anticlockwise whilst Z axial planes are (84%) clockwise of Ln.

folds and south-facing S folds) plot closest to the mineral lineation (Ln) and have undergone the greatest rotation and deformation (Fig. 5c,d,e,f). A similar pattern is observed on both sides of the major culmination trace demonstrating a lack of pronounced reorientation by the later major synshearing flow folds (Fig. 3a,c,d). Rotations in the strike and dip of minor fold axial planes are governed by the sense of fold asymmetry (S/Z) rather than facing directions. Thus, with increasing deformation on the lower limbs of major sheath folds, mean S fold axial planar strike rotates anticlockwise and mean Z fold axial planes clockwise towards the regional foliation

(Sn) marking the shear plane (Figs 3a,e,f,g, 5c,d,e,f & 6c,d,e). F2 sheath folds are associated with abundant top-to-the WNW shear criteria generated during (D2) ductile thrusting. These folds form a series of transport-parallel antiformal culminations and synformal depressions with traces orthogonal to the mylonitic high strain zones. Major F2 sheath folds with overturned limbs are common, necessitating the use of fold facing directions to discriminate S and Z minor folds on normal or inverted fold limbs (see e.g. Alsop & Holdsworth 1999). Minor (F2) sheaths are tight to isoclinal folds with mylonitic limbs which on

SHEAR ZONE FOLDS

185

Fig. 6. Frequency distribution histograms of fold hinges and axial planes from the study area orientated relative to the trend of the adjacent Ln transport lineation. The graphs illustrate the vergence and obliquity relationships of minor S folds (solid ornament) and Z folds (stipple ornament) on either side of the major culmination and at varying structural positions around minor medial and axial surfaces (refer to central key for structural position and Fig. 2 for location). The relative obliquity between S and Z fold hinges and associated axial planes with the Ln mineral lineation is shown for transecting (F3) flow folds (a,b), folds forming the upper limbs of (F2) sheath folds (c,d) and folds on the overturned lower limbs of sheath folds (e,f). Note the increasing concentration of fold hinges towards Ln and axial planar strikes towards the transport normal from the upper to the lower overturned limbs of sheath folds.

average plunge gently towards the east, with associated shallowly east-dipping axial planes (Fig. 4c). Minor (F2) curvilinear sheath folds are ubiquitous with north-facing fold hinges anticlockwise of Ln and south-facing fold hinges clockwise of Ln (Fig. 4d). Irrespective of the facing direction, S fold axial planes are typically anticlockwise of Ln and Z fold axial planes clockwise of Ln. Thus, large-scale (F2) sheath folds are defined by medial surfaces which coincide in scale, orientation and location with the previously described (F3) culminations and depressions (Fig. 3a). This correspondence, taken together with the increasingly rotated nature of minor F3 fold hinges developed on the overturned limbs of F2 folds, suggests that major F2 and F3 structures are closely related in terms of structural mechanisms and controls. Significantly, both minor F2 and F3 folds pre-date secondary recrys-

tallization indicating that there are no metamorphic grounds to separate F2 and F3 structures (Holdsworth 1987 p. 124). Thus, there is compelling evidence indicating that F2 and F3 folds are linked to the same kinematic regime of (D2) deformation. This paper will now explore the detailed relationships between these fold patterns in order to investigate further the kinematic and geometric evolution of shear zone structures during progressive deformation.

Transection relationships between F2 sheath folds and F3 synshearing folds Geometric investigation of (F2) sheath folds and adjacent (F3) synshearing flow folds reveals consistent patterns which display marked structural trends. F3 flow folds, which obliquely overprint the adjacent F2 sheath folds may be usefully

186

G. I. ALSOP & R. E. HOLDSWORTH

Fig. 7. Frequency distribution histograms of transecting fold hinge angles from the study area orientated either clockwise (Cw) or anticlockwise (A-Cw) relative to the trend of the adjacent F2 fold hinge. The angle of transection of (F3) flow folds on sheath folds for all of the study area is given in (a), north-facing folds are shown in (b), whilst south-facing hinges are displayed in (c). North-facing hinges are typically (86%) transected by flow folds in an anticlockwise sense, whereas south-facing sheath folds normally (82%) display clockwise transection.

described in terms of transecting (clockwise or anticlockwise) relationships (Fig. 7). An analysis of fold hinge/lineation angles between (F2) sheath folds and adjacent (F3) flow folds reveals distinct patterns, with north facing F2 hinges consistently transected in a anticlockwise (negative) sense by adjacent F3 hinges, whereas south facing F2 hinges are transected with a clockwise (positive) obliquity (Fig. 7a-c). Importantly, the sense of transection is independent of fold asymmetry, which is governed by the sense of hinge rotation and hence fold facing. The sense of transection of F3 synshearing fold hinges and axial surfaces on adjacent minor F2 sheath folds allows the location of the minor F2 folds on the major sheath framework to be deduced. Combinations of culmination and depression medial surfaces on antiformal and synformal axial surfaces generate dome and basin (end-member) scenarios in which the eight possible domains are defined uniquely by the sense and combination of fold hinge and axial planar transection (Table 1). Although the sense of fold hinge transection is governed by the direction of F2 fold rotation (and thereby fold facing), the sense of axial planar transection is controlled by the geometry of the overprinting (Z or S) axial plane (Table 1). This transection grid (Table 1) enables the senses of hinge rotation to be deduced even if no evidence of younging or facing information is preserved. Major basinal sheath fold geometries generate hinge and axial planar transec-

tions of the same (synthetic rotations) sense, whilst domal sheaths produce transections of an opposing (antithetic) sense (Table 1).

Topological relationships between F2 sheath folds and F3 synshearing folds Topology may be defined as 'a branch of geometry concerned with those properties of a figure which remain unchanged even when the figure is bent or stretched' (Chambers Dictionary 1993). Fabric topology plots (FTPs) illustrate these invariant or irreversible geometric properties of a body under deformation, and have been used recently to emphasize and monitor planar and linear fabric inter-relationships during progressive shearing (e.g. Alsop & Holdsworth 2002). Importantly, FTPs clearly demonstrate that although amounts of angular obliquity between planar and linear fabric elements will vary systematically during deformation, reversals in the sense of obliquity will not occur reflecting the inability of planar and linear fabric elements to rotate through the shear plane and shear direction respectively (Ramsay 1980; see Passchier 1997). Consistent and ordered scale-independent relationships between the mineral lineation and fold hinges, fold asymmetries and axial surfaces illustrated on FTPs thus provide a potential record of the generation of folds within mylonites and their subsequent evolution during

187

SHEAR ZONE FOLDS Table 1. Summary transection grid associated with a dome (antiform on culmination) geometry (above) and basin (synform on depression) configuration (below). Dome geometry

Upper limb Antithetic transection Antiformal axial plane Lower limb Antithetic transection

Culmination surface Anticlockwise S hinge obliquity

Clockwise Z hinge obliquity

S folds Anticlockwise transecting hinge Clockwise transecting axial plane

Z folds Clockwise transecting hinge Anticlockwise transecting axial plane

Z folds Anticlockwise transecting hinge Clockwise transecting axial plane

S folds Clockwise transecting hinge Anticlockwise transecting axial plane

Basin geometry

Upper limb Synthetic transection Synformal axial plane Lower limb Synthetic transection

Depression surface Clockwise S hinge obliquity

Anticlockwise Z hinge obliquity

S folds Clockwise transecting hinge Clockwise transecting axial plane

Z folds Anticlockwise transecting hinge Anticlockwise transecting axial plane

Z folds Clockwise transecting hinge Clockwise transecting axial plane

S folds Anticlockwise transecting hinge Anticlockwise transecting axial plane

Clockwise rotation (relative to Ln trend) will result in the generation of anticlockwise transecting folds and visa versa. The rotation of an axial plane is measured through variations in strike and is governed by the fold vergence, whereas the sense of rotation of fold hinges is dependent on the facing direction and position relative to medial surfaces. Fold hinges may thus be transected with increasing strain in either the same (synthetic) or opposing (antithetic) sense to the associated axial plane depending on structural setting. Each of the eight scenarios around dome and basin geometries are uniquely denned by the combination and sense of fold hinge and axial planar transection of S and Z folds.

continued shearing. The topological relationships between minor (F3) synshearing folds, and parasitic (F2) folds developed on the upper and lower (overturned) limbs of major sheath folds are summarized on Fig. 8. The parameters illustrated on this figure are shown for a dip-slip system of deformation but equivalent relationships exist in other kinematic regimes. Fabric topology plots are shown in Fig. 9, and summary FTPs that highlight the relationships between mean data sets are shown on Fig. 10 (n = 750) and Fig. 11 (n = 670). The angle between the trend of the fold hinge and lineation displays a systematic reduction from transecting F3 flow folds to the overturned lower limbs of F2 sheath folds (Fig. 8a). This reflects the increasing rotation of fold hinges towards the shear direction with the greatest deformation on overturned F2 limbs. Although F3 hinges are developed closer to Ln on overturned F2 limbs, the angle they make with minor F2 hinges is actually larger as the latter are almost colinear with Ln, i.e. minor F2 hinges undergo a more significant rotation (than F3) from the upper to lower limbs of major F2

sheaths. Consequently, the mean angle of F3 transection increases from the upper to lower limbs of sheath folds (Figs 9a,b & lOa-t; Table 2). The systematic variation in the strike of axial planes about culmination/depression surfaces results in S fold axes (and the associated mineral lineation) typically pitching in a clockwise (positive) sense on NE-trending S axial planes, whilst Z fold hinges (and Ln) pitch in an anticlockwise (negative) sense on SE-striking Z axial planes. Compared to F2 sheath folds, F3 hinges display smaller angles of fold hinge pitch on associated axial planes (Figs 8b, 10c,f,m & lla-t). The angle of hinge pitch transection increases from the upper to lower limbs of sheath folds reflecting greater F2 hinge rotation on lower limbs (Table 2). As the angle (and sense) of lineation pitch is also measured from the strike of the (Z or S) axial plane, it therefore increases from F3 flow folds to the lower limbs of sheath folds (Figs 8c, 10d,n & lld,n). The angle of lineation pitch transection similarly increases from the upper to lower limbs of sheath folds (Table 2). Increasing rotation of fold hinges towards the shear direction with greater

188

G. I. ALSOP & R. E. HOLDSWORTH

Fig. 8. Schematic sketches illustrating the geometric consequences of fold and fabric rotations associated with increasing deformation from transecting (F3) flow folds (upper row) to the lower limbs of (F2) sheath folds (lower row). Lower strain situations for each parameter (a-g) are thus given in the upper row, whilst the higher strain situation for each relationship are shown below. Angular relationships between parameters may increase or decrease with increasing deformation, and are frequently related to the orientation of foliation (Sn) and the transport-parallel lineation (Ln).

deformation results in a systematic reduction in the angle between the pitch of a fold hinge and adjacent lineation from F3 folds to the overturned lower limbs of sheath folds (Fig. 8d). The mean angle of F3 transection thus increases from the upper to lower limbs of sheath folds (Figs 10e,g,q & lle,g,q; Table 2). The angle between the strike of the axial plane and the trend of the adjacent lineation displays a systematic increase from F3 folds to the overturned lower limbs of sheath folds (Fig. 8e). This reflects the increasing rotation of axial planes towards the shear plane with greater deformation. The mean angle of F3 axial planar transection therefore increases from the upper to lower limbs of sheath folds (Figs 10h,r & llh,r; Table 2). Increasing rotation of axial planes towards the foliation (shear) plane results in a systematic reduction in the angle between axial planar strike and the trend of the adjacent foliation from F3 flow folds to the lower limbs of sheath folds (Fig. 8f). Consequently, the mean angle of F3 axial planar transection increases from the upper to lower limbs of

sheath folds (Figs 10i,s, lla-j, s; Table 2). Similarly, the acute angle between the axial plane and the adjacent foliation displays a systematic reduction from F3 folds to the overturned lower limbs of sheath folds, whereas the mean angle of F3 axial planar transection shows a concomitant increase (Figs 8g, 10j,t & llj,t; Table 2). In summary, both F2 and F3 folds are associated with top-to-the-WNW thrusting, although their age relative to the local (Sn) fabric differs with F2 folds displaying mylonitic limbs and F3 structures refold this foliation. Minor fold hinges display sequential rotations towards the transport direction with the angle of obliquity systematically reducing from (F3) synshearing flow folds to the upper limbs of (F2) sheath folds to the higher strained and overturned lower limbs. Minor fold axial planes display concomitant rotations towards the foliation plane. Thus, fold hinge and axial plane transection angles increase from the upper to lower limbs of sheath folds reflecting the greater deformation and fabric rotation on the lower sheath limbs

SHEAR ZONE FOLDS

189

Figure 9. Fabric topology plots of sheath fold and flow fold data from the study area. South-facing data is shown by open symbols and north-facing data is shown by solid symbols. In each case, structures associated with Z (circles), or S (squares) (F2) sheath folds are shown with positive clockwise (Cw) or negative anticlockwise (A-Cw) trends relative to adjacent transecting (F3) flow folds. Fold trend values are given between 0° and 180°, but values of axial planar strike are displayed for 90° either side of 0° strike, to show the continuation of geometric trends across the north-south direction. The upper row of plots shows how the angle of (F2) sheath fold hinge transection by (F3) flow folds varies in relation to (a) trend of sheath fold hinges, (b) angle between sheath fold hinge and adjacent mineral lineation (Ln), (c) angle between adjacent transecting axial planes. The lower row of plots shows how the angle of axial planar transection between sheath folds and flow folds varies in relation to (d) strike of sheath fold axial planes, (e) strike of flow fold axial planes. A comparison of sheath and flow fold axial planar strikes and resultant angles of clockwise (Cw) and anticlockwise (A-Cw) transection is given in (f).

(Table 2). These geometric relationships are now discussed in terms of alternative models for overall fold and fabric evolution within a system of progressive (D2) deformation.

Discussion of curvilinear fold patterns and evolution General observations and inferences of progressive deformation forming successive generations of curvilinear folds have been made at a variety of scales from outcrop (e.g. Ghosh & Sengupta 1984, 1987; Ghosh et al 1999) to regional (e.g. Goscombe 1991; Alsop 1994). The detailed analyses of synshearing and sheath folds presented in this study is based on fabric development across both axial surfaces and medial (culmination/depression) surfaces reflecting the 3D nature of such deformation systems. Such an examination also enables the progressive struc-

tural development to be assessed. A variety of alternative models for the development of synshearing folds and their relationships to earlier sheath folds are now discussed.

Fold evolution model The fold evolution model suggests that the orientation and asymmetry of synshearing (F3) folds is directly controlled by variable displacement along underlying detachments. The orientation and asymmetry of the folding is a direct consequence and product of perturbations in ductile flow which are both transient and scale independent (Holdsworth 1990; Alsop & Holdsworth 1993, 2002). In this model, differential dextral and sinistral shear will generate S and Z folds with consistent anticlockwise and clockwise obliquities to Ln respectively (Fig. Ic). Such folds may then subsequently undergo a degree of hinge rotation towards the transport

Fig. 10. Summary fabric topology plots showing the mean orientation (n - 750) of Z (circles) and S (squares) fold hinge trends relative to the trend of the adjacent lineation (Ln) (a-j). In each case, north-facing folds are shown by the solid symbols; south-facing structures are given by the open symbols. Note how north-facing fold hinges are anticlockwise of Ln and south-facing folds are clockwise. A guide to symbols and structural position is given in the central transport-normal section/ key with transecting (F3) flow folds shown on the upper row (larger symbols). The relative location of the culmination surface and the antiformal axial plane results in the range of minor fold geometries depicted. The lower set of FTPs (k-t) show the angle of transection of mean (F2) sheath fold hinge trends by adjacent (F3) flow fold hinges. Note that north-facing fold hinges are consistently transected in a (negative) anticlockwise sense and south-facing hinges in a (positive) clockwise sense (see central plan for obliquities and symbols). The various structural parameters display systematic and sequential variations from F3 flow folds, to the upper limbs of F2 sheath folds to lower limbs of sheath folds.

Fig. 11. Summary fabric topology plots showing the mean orientation (n = 670) of Z (circles) and S (squares) fold axial planar strikes relative to the trend of the adjacent foliation (Sn) (a-j) and lineation (Ln) (e,f). In each case, north-facing folds are shown by the solid symbols; south-facing structures are given by the open symbols. Note how Z and S fold axial planes are consistently anticlockwise (negative) and clockwise of the mean Sn trend. A guide to symbols and structural position is given in the central transport-normal section/ key with transecting flow folds shown on the upper row (larger symbols). The relative location of the culmination surface and the antiformal axial plane results in the range of minor fold geometries depicted. The lower set of FTP's (k-t) show the angle of transection of (F2) sheath fold axial planar strike by adjacent (F3) flow fold axial planes. Note that S axial planes are consistently transected in a positive clockwise sense and Z axial planes in a (negative) anticlockwise sense (see central plan for obliquities and symbols). The various structural parameters display systematic variations from flow folds, to the upper and lower limbs of sheath folds.

G. I. ALSOP & R. E. HOLDSWORTH

194

Table 2. Summary table of angular obliquities associated with a-g structural parameters in the study area. North facing

Culmination surface

South facing

Structural parameter

Upper limb trans.

Difference in trans, across axial surface

Lower limb trans.

Lower limb trans.

Difference in trans, across axial surface

Upper limb trans.

a) b) c) d) e) f) g)

-14° +18° +9° -9° —3° +20° 19°

-9° +38° +20° -19° +17° -52° +3°

-23° +56° +29° -28° +14° -32° 22°

+22° -64° -39° +26° -9° +29° 22°

+18° -50° -27° +23° -9° +53° +4°

+3° -14° -12° +3° 0° -24° 18°

The table summarizes mean F3 transection (Trans.) data across the axial surface of major F2 folds (upper and lower limbs), on both the north-facing (left) and south-facing (right) margins of the major culmination. In each case, data refers to the sense and amount of mean angular transection and indicates that lower (F2) fold limbs have undergone greater rotation and hence larger angles of transection. Structural parameters: a) Angle between trend of fold hinge and lineation; b) Angle of fold hinge pitch on axial plane; c) Angle of lineation pitch on axial plane; d) Angle between pitch of fold hinge and lineation; e) Angle between axial planar strike and lineation trend; f) Angle between axial planar strike and trend of foliation; g) Acute angle between axial plane and long-limb foliation. Refer to Fig. 8 and text for further details.

direction raising the question as to whether sheath folds could initiate by a similar process (Fig. Id). In such a model F2 sheath folds simply represent early flow perturbation folds subsequently modified by high-strain shearing. Although the common evolutionary patterns on many fabric topology plots from F3 synshearing folds to F2 sheath folds may be taken to support such an interpretation, a number of important differences exist. The general lack of F3 sheath folds might suggest that a different process is responsible for creating the earlier (F2) sheath folds. Importantly, F2 sheath folds describe a complete arc of hinge orientations from transport-normal to transport-parallel, with some hinges still preserved at high angles to transport (Fig. Ib; Holdsworth 1988; Alsop & Holdsworth 1999). F3 folds do not describe such a pronounced arc and typically initiate at smaller (oblique) angles to transport than sheath folds (Fig. la,c). They also lack pronounced overturned limbs (see Alsop & Holdsworth 2002). If sheaths were initiated by flow perturbations, this suggests that the geometry of the flow cell was different from those that created the F3 folds. Thus, although F2 and F3 folds clearly belong to a progressive (D2) deformation, the angles of F2 and F3 fold initiation appear to be quite different. This may require an alternative model for generating geometrically distinct but related fold phases to be considered.

Fold inheritance model The fold inheritance model suggests that the orientation and asymmetry of synshearing (F3) folds is directly controlled by the pre-existing attitude of layering. The orientation of the mean regional foliation (Sn) and lineation (Ln) displays limited but recognizable variation to the north and south of the major transport-parallel culmination trace which bisects the Melness study area (Figs 3b-d & 12). This variation in fabric attitudes is simply a consequence of reverse asymmetry vergence where the long limbs of minor folds (with opposing vergence) remain subparallel to one another across the medial surface (Fig. 13; see also fig. 9 Alsop & Holdsworth 2002). Clearly, this differs from normal asymmetry vergence patterns where the long limbs of minor folds are rotated normally resulting in markedly different attitudes on either side of a major fold and the classical stereographic girdle pattern. The variations in the mean statistical orientation of the layering (Sn) about the major (F2) culminations may however be significant in relation to F3 fold development as layering which is oblique to the axes of the bulk strain ellipsoid will generate fold hinges that initiate perpendicular to the direction of shortening within that layer (Flinn 1962; Treagus & Treagus 1981). The orientation and geometry of such fold hinges, which consistently form parallel to the long axis of the sectional ellipse within the

SHEAR ZONE FOLDS

Fig. 12. Schematic stereographic plot illustrating how layering developed oblique to the X, Y, Z axes of the finite strain ellipsoid will develop folds of differing orientations (see Treagus & Treagus 1981). In this example, layering which strikes anticlockwise of the X—Yplane will initiate folding anticlockwise of transport (X), whilst layering which strikes clockwise of the X—Y plane will initiate folding clockwise of transport. The sense of subsequent fold hinge rotation during progressive deformation is also shown whereas the table shows the calculated mean data for either side of the major culmination in the Melness case study area. Different mean orientations of the folded surface (Sn) to the north and south of the culmination will encourage folds to initiate with an opposing sense of obliquity to Ln (X).

layer, will thus be governed by and vary according to, the relative obliquities between layering and the strain ellipse, together with subsequent rotations associated with angular migration of the fold axis during progressive deformation (Treagus & Treagus 1981; Carreras 1997) (Fig. 12). A reversal in the sense of obliquity of layering to the strain ellipsoid will generate a reversal in apparent differential shear as the layer rotates towards the X-Y plane. Thus, systematic variations in the orientation of layering relative to the bulk strain ellipsoid will result in folds initiating with reversed senses of clockwise or anticlockwise obliquity to the ^direction. These will subsequently undergo opposing senses of hinge rotation during progressive shear. To the north

195

of the major culmination, mean Sn layering intersects the mean S2 fabric representing the X-Y plane of the strain ellipsoid in a clockwise sense (relative to the intermediate Y axis), whereas to the south of the culmination the equivalent intersection is anticlockwise of Y (Fig. 12). In summary, antiformal culminations and synformal depressions associated with F2 sheath folding may impart a consistent obliquity between the bulk strain ellipse and layering, which will reverse on opposing flanks of culminations and depressions. This inherited architecture will also govern the orientation of subsequent F3 fold hinges and thereby control the sense of F2-F3 transection and direction of F3 hinge rotations. Thus, geometric patterns developed during F2 sheath folding could control subsequent F3 medial surfaces resulting in the observed coincidence of F2 and F3 culmination and depression surfaces (Fig. 13). Within the geometric model outlined above, the orientation and asymmetry of F3 synshearing folds is entirely controlled by the pre-existing attitude of the layering following F2 folding. A significant issue raised by this model is our inability to define the attitude of this post-F2 layering precisely. These post-F2 orientations are critical to the integrity of the fold inheritance model, but remain largely unresolved owing to the overprinting F3 deformation. In addition, layering may change orientation from the upper to lower overturned limbs of F2 folds, thereby also generating a reversal in subsequent F3 vergence and orientation. However, the ability to trace consistent F3 vergence across F2 axial surfaces suggests that such a switch in asymmetry does not generally develop. This may arguably reflect increasing D2 strain on overturned (F2) limbs which negates any significant obliquity in layering orientation (Fig. 13). Alternatively, this relationship may suggest that the overall coincidence of F2 and F3 patterns is not entirely explained by superposition of F3 on F2 structures in the fold inheritance fold model.

Hybrid fold model The hybrid fold model suggests that the orientation and asymmetry of synshearing (F3) folds is directly controlled by variable components of both the fold evolution and inheritance models. Regions lacking pronounced F2 sheath folding do not develop transport parallel F3 culminations and depressions (Alsop & Holdsworth 1993). In contrast, the Melness area is marked by a major F2 sheath culmination which coincides with transport-parallel F3 folding. Such

196

G. I. ALSOP & R. E. HOLDSWORTH

Fig. 13. Schematic 3D cartoon illustrating the geometry and orientation of synshearing flow folds together with sheath folds associated with transport-parallel culmination and depression surfaces. Increasing deformation towards the underlying detachment surface (marked by barbs) results in pronounced attenuation of the lower fold limb and reduction in the apical angle of the sheath fold from tongue folds on the upper limb to tubular folds on the lower limb. Culmination and depression surfaces are associated with transport-parallel neutral verging (M) open folds with steeply dipping axial planes. They separate regions of synshearing flow folds with S folds trending anticlockwise of transport from Z folds developed clockwise of transport.

variations have in the past been attributed to the different geometries of flow perturbation cells (e.g. Alsop & Holdsworth 1993; Alsop etal. 1996). However, they may also be interpreted in terms of F2 folds exerting a clear control on the geometry of subsequent F3 folds at least on a larger kilometric scale. Within such a scenario, pre-existing F2 sheaths may exert a geometric influence on F3 folds (fold inheritance model) whilst at the same time constraining the location and scale of subsequent flow perturbation cells (fold evolution model). Hybrid fold models incorporating flow perturbations and a component of inherited geometric constraint (perhaps concentrated in the lower-strain regions of earlier F2 folds) may thus also be applicable.

Conclusions These natural fold patterns systematically record and reflect the kinematics and evolution of flow within shear zones. However the inevitable overprinting of structures that occurs

during progressive deformation means that the interpretation of the preserved relationships is somewhat ambiguous. Fabric topology plots (FTPs) are of value in such situations as they clearly and effectively monitor fold and fabric relationships during progressive deformation. These plots demonstrate that synshearing F3 flow folds and F2 sheath folds share many geometric properties and display a common evolutionary sequence from larger through to smaller obliquities respectively (Fig. 13). FTPs can be used to highlight geometric patterns of structural evolution in which fabric relationships may undergo modification (but not reversal) during kinematically coherent flow. Synshearing F3 folds display a consistent sense of either clockwise (Z fold) or anticlockwise (S fold) hinge transection on the earlier F2 sheath folds. The sense of transection (which is governed by the direction of hinge rotation) switches across the medial (culmination/depression) surface and thus acts as a reliable indicator of clockwise or anticlockwise rotation. The sense of transection will therefore predict the

SHEAR ZONE FOLDS

polarity of fold facing even in areas lacking stratigraphic control and/or abundant evidence of younging. The amount of transection increases from the upper to lower limbs of sheath folds, reflecting larger fabric rotations associated with greater deformation on the overturned (lower) limb setting (Fig. 13). Similar, predictable patterns may occur in all situations where folds are associated with shear zones. The sense of axial planar transection is governed by the asymmetry of (Z or S) folding and therefore reverses across both medial surfaces and axial surfaces. F2 Z axial planar strike is transected (by F3) in an anticlockwise sense (reflecting a clockwise F2 rotation) whereas S fold axial planes are transected (by F3) in a clockwise sense. Thus, the sense of axial planar transection is governed by fold (Z or S) asymmetry, and the sense of hinge transection is controlled by the direction of hinge rotation and thereby the polarity of fold facing. Combining hinge and axial plane transection relationships on a transection grid allows the position of minor F2 folds to be uniquely located on major F2 sheath folds. Foliation-normal and transport-parallel culmination and depression surfaces generated during F2 and F3 folding correspond with one another in terms of scale, space and orientation reflecting common origins and generation during progressive (D2) shearing (Fig. 13). The concurrence of culminations and depressions defined by the (F2) sheath folds and synshearing (F3) folds is particularly important with regard to an overall understanding of the progressive (D2) deformation. This coincidental relationship is interpreted in the fold evolution model to suggest that the sheath folds are genetically and kinematically related to flow perturbation cells, similar to those proposed previously to explain the distribution and geometry of the F3 folds in Melness and other areas. However, F3 'eye-shaped' closures and obvious curvilinear folds are scarce, with F3 folds notably exhibiting different angles of hinge initiation compared to F2 sheath folds. Alternatively, the fold inheritance model suggests that development of the earlier (F2) sheath folds may influence and pre-determine the geometry and distribution of later synshearing (F3) flow folds. Although it is difficult to define the orientation of post-F2 layering, it may be expected to display differing attitudes across the axial surfaces of major F2 folds, and hence generate a reversal in F3 fold vergence. However, our detailed observations suggest the contrary with F3 folds maintaining constant asymmetry across the (F2) axial surfaces. A combination of these two models, as suggested in the hybrid fold model, is also entirely feasible and perhaps most applicable in

197

regions where major early fold systems arch mylonitic fabrics and foliations in which the synshearing folds will subsequently develop. Clearly such systems may become extremely complex and to some extent self sustaining as the geometric influence of successive fold sets systematically influences the nature and kinematics of subsequent deformation. Fieldwork for this paper was funded under the NERCBGS-Academic mapping programme awarded to the University of Durham (Grant F60/G2/36). Subsequent additional funding was provided from the Edinburgh Geological Society and the Welch bequest of the University of St Andrews. Jordi Carreras and Graham Potts provided careful and constructive reviews.

References ALSOP, G.I. 1992. Progressive deformation and the rotation of contemporary fold axes in the Ballybofey Nappe, northwest Ireland. Geological Journal, 27, 271-283. ALSOP, G.I. 1994. Relationships between distributed and localized shear in the tectonic evolution of a Caledonian fold and thrust zone, northwest Ireland. Geological Magazine, 131,123-136. ALSOP, G.I. & HOLDSWORTH, R.E. 1993. The distribution, geometry and kinematic significance of Caledonian buckle folds in the western Moine Nappe, northwestern Scotland. Geological Magazine, 130,353-362. ALSOP, G.I. & HOLDSWORTH, R.E. 1999. Vergence and facing patterns in large-scale sheath folds. Journal of Structural Geology, 21,1335-1349. ALSOP, G.I. & HOLDSWORTH, R.E. 2002. The geometry and kinematics of flow perturbation folds. Tectonophysics, 350, 99-125. ALSOP, G.I., HOLDSWORTH, R.E. & STRACHAN, R.A. 1996. Transport-parallel cross folds within a midcrustal Caledonian thrust stack, northern Scotland. Journal of Structural Geology, 18,783-790. BELL,T.H. 1978. Progressive deformation and reorientation of fold axes in a ductile mylonite zone: the Woodroffe thrust. Tectonophysics, 44, 285-321. BERTHE, D. & BRUN, J.P. 1980. Evolution of folds during progressive shear in the South Amorican Shear Zone, France. Journal of Structural Geology, 2,127-133. BOLHAR, R. & RING, U. 2001. Deformation history of the Yola Bolly terrane at Leech Lake Mountain, Eastern belt, Franciscan subduction complex, California Coast ranges. Geological Society of America Bulletin, 113,181-195. BRITISH GEOLOGICAL SURVEY 1997. Tongue, Scotland 114E. Solid Geology. 1:50,000. BRITISH GEOLOGICAL SURVEY 2002. Loch Eriboll, Scotland 114W. Solid Geology. 1:50,000. BRYANT, B. & REED, J.C 1969. Significance of lineation and minor folds near major thrust faults in the southern Appalachians and the British and Norwegian Caledonides. Geological Magazine, 106, 412-429.

198

G. I. ALSOP & R. E. HOLDSWORTH

CARRERAS, J. 1997. Shear zones in foliated rocks: geometry and kinematics. In: SENGUPTA, S. (ed.) Evolution of geological structures in micro- to macro-scales, pp. 185-201. Chapman & Hall, London. CARRERAS, J., ESTRADA, A. & WHITE, S. 1977. The effect of folding on the c-axis fabrics of a quartz mylonite. Tectonophysics, 39, 3-24. CHAMBERS DICTIONARY 1993. Chambers Harrap Publishers Ltd. Edinburgh. 2062 pp. COBBOLD, PR. & QUINQUIS, H. 1980. Development of sheath folds in shear regimes. Journal of Structural Geology, 2,119-126. CONEY, PJ. 1974. Structural analysis of the Snake Range 'Decollement', East-Central Nevada. Geological Society of America Bulletin, 85, 973-978. COWARD, M.P & POTTS, GJ. 1983. Complex strain patterns developed at the frontal and lateral tips to shear zones and thrust zones. Journal of Structural Geology, 5, 383-399. DILLON, J.T., HAXEL, G.B. & TOSDAL, R.M. 1990. Structural evidence for northeastward movement on the Chocolate Mountains Thrust, southeasternmost California. Journal of Geophysical Research, 95,19953-19971. ESCHER, A. & WATTERSON, J. 1974. Stretching fabrics, folds and crustal shortening. Tectonophysics, 22, 223-231. EVANS, DJ. & WHITE, S.H. 1984. Microstructural and fabric studies from the rocks of the Moine Nappe, Eriboll, NW Scotland. Journal of Structural Geology, 6, 369-389. Ez, V. 2000. When shearing is a cause of folding. Earth Science Reviews, 51,155-172. FLETCHER, J.M. & BARTLEY, J.M. 1994. Constrictional strain in a non-coaxial shear zone: implications for fold and rock fabric development, central Mojave metamorphic core complex, California. Journal of Structural Geology, 16, 555-570. FLINN, D. 1962. On folding during three-dimensional progressive deformation. Quarterly Journal of the Geological Society of London, 118, 385-433. FOSSEN, H. & RYKKELID, E. 1990. Shear zone structures in the 0ygarden area, West Norway. Tectonophysics, 174, 385-397. GHOSH, S.K. & SENGUPTA, S. 1984. Successive development of plane noncylindrical folds in progressive deformation. Journal of Structural Geology, 6, 703-709. GHOSH, S.K. & SENGUPTA, S. 1987. Progressive development of structures in a ductile shear zone. Journal of Structural Geology, 9, 277-287. GHOSH, S.K., HAZRA, S. & SENGUPTA, S. 1999. Planar, non-planar and refolded sheath folds in the Phulad Shear Zone, Rajasthan, India. Journal of Structural Geology, 21,1715-1729. GOSCOMBE, B. 1991. Intense non-coaxial shear and the development of mega-scale sheath folds in the Arunta Block, Central Australia. Journal of Structural Geology, 13, 299-318. HANSEN, E. 1971. Strain Fades. New York, SpringerVerlag. HENDERSON, J.R. 1981. Structural analysis of sheath

folds with horizontal X-axes, northeast Canada. Journal of Structural Geology, 3, 203-210. HARRIS, L.B., KOYI, H.A. & FOSSEN, H. 2002. Mechanisms of folding of high-grade rocks in extensional tectonic settings. Earth Science Reviews, 59, 163-210. HOBBS, B.E., MEANS, WD. & WILLIAMS, P.F. 1976. An outline of structural geology. John Wiley & Sons, New York. HOLDSWORTH, R.E. 1987. Basement/cover relationships, reworking and Caledonian ductile thrust tectonics of the Northern Moine, NW Scotland. Ph.D. thesis, University of Leeds. HOLDSWORTH, R.E. 1988. The stereographic analysis of facing. Journal of Structural Geology, 10,219-223. HOLDSWORTH, R.E. 1989. The geology and structural evolution of a Caledonian fold and ductile thrust zone, Kyle of Tongue region, Sutherland, N. Scotland. Journal of the Geological Society of London, 146, 809-823. HOLDSWORTH, R.E. 1990. Progressive deformation structures associated with ductile thrusts in the Moine Nappe, Sutherland, N. Scotland. Journal of Structural Geology, 12, 443-452. HOLDSWORTH, R.E. & GRANT, CJ. 1990. Convergencerelated 'dynamic spreading' in a mid-crustal ductile thrust zone: a possible orogenic wedge model. In: KNIPE, RJ. & RUTTER, E.H. (ed.) Deformation mechanisms, rheology and tectonics Geological Society, London, Special Publications, 54, 491-500. HOLDSWORTH, R.E. & ROBERTS, A.M. 1984. A study of early curvilinear fold structures and strain in the Moine of the Glen Garry region, Inverness-shire. Journal of the Geological Society of London, 141, 327-338. HOLDSWORTH, R.E., STRACHAN, R.A. & ALSOP, G.I. 2001. Geology of the Tongue District. Memoirs of the British Geological Survey. Sheet 114E (Scotland), 76 pp. LACASSIN, R. & MATTAUER, M. 1985. Kilometre-scale sheath fold at Mattmark and implications for transport directions in the Alps. Nature, 315, 739-742. MIES, J.W 1991. Planar dispersion of folds in ductile shear zones and kinematic interpretation of fold hinge girdles. Journal of Structural Geology, 13, 281-297. MINNIGH, L.D. 1979. Structural analysis of sheathfolds in a meta-chert from the Western Italian Alps. Journal of Structural Geology, 1, 275-282. PASSCHIER, C.W. 1997. The fabric attractor. Journal of Structural Geology, 19,113-127. PLATT, J.P. 1983. Progressive refolding in ductile shear zones. Journal of Structural Geology, 5, 619-622. QUINQUIS, H., AUDREN, C, BRUN, J.P. & COBBOLD, PR. 1978. Intense progressive shear in Ille de Groix blueschists and compatibility with subduction or obduction. Nature, 273, 43-45. RAMSAY, D.M. 1979. Analysis of rotation of folds during progressive deformation. Geological Society of America Bulletin, 90, 732-738. RAMSAY, J.G. 1967. Folding and fracturing of rocks. McGraw Hill.

SHEAR ZONE FOLDS RAMSAY, J.G. 1980. Shear zone geometry: a review. Journal of Structural Geology, 2, 83-99. RAMSAY, J.G. & HUBER, M. 1987. The Techniques of Modern Structural Geology. Volume 2: Folds and Fractures. Academic Press, London. RAMSAY, J.G. & LISLE, R. 2000. The techniques of modern structural geology. Volume 3: Applications of continuum mechanics in structural geology. Academic Press, London. RHODES, S. & GAYER, R.A. 1977. Non-cylindrical folds, linear structures in the X-direction and mylonite developed during translation of the Caledonian Kalak Nappe Complex of Finnmark. Geological Magazine, 114, 329-341. RIDLEY, J. 1986. Parallel stretching lineations and fold axes oblique to displacement direction-a model

199

and observations. Journal of Structural Geology, 8, 647-654. SANDERSON, DJ. 1973. The development of fold axes oblique to the regional trend. Tectonophysics, 16, 55-70. SKJERNAA, L. 1989. Tubular folds and sheath folds: definitions and conceptual models for their development with examples from the Grapesvare area, northern Sweden. Journal of Structural Geology, 11, 689-703. TREAGUS, J.E. & TREAGUS, S.H. 1981. Folds and the strain ellipsoid: a general model. Journal of Structural Geology, 3,1-17. WILLIAMS, G.D. 1978. Rotation of contemporary folds into the X direction during overthrust processes in Laksefjord, Finnmark. Tectonophysics, 48, 29-40.

This page intentionally left blank

Geometric and kinematic analysis of a transpression terrane boundary: Minas fault system. Nova Scotia, Canada E. A. MACINNES & J. C. WHITE Department of Geology, University of New Brunswick, Fredericton, NB Canada, E3B 5A3 (e-mail: ellie_macinnes@hotmail. com) Abstract: The Minas fault system is an ENE-WSW trending transpressional boundary between the Appalachian Meguma and Avalon tectono-stratigraphic terranes of Nova Scotia, Canada. Along this boundary there is large-scale partitioning of deformation into distinct external (contractional) and internal (shear) zones. With the increase in strain from external to internal zones there is progressive localization of deformation, culminating in the discrete shear band domain. Deformation has produced materially, temporally and spatially distinct folds and faults throughout the fault system history. Ductile structures are generally composite features derived from multiple transposition of pre-existing layers. The partitioning of deformation found amongst fault rock units can in turn be associated with contrasting deformation micromechanisms. The distinctive variation in mechanical response and microstructures provides an insight into the role of localization, partitioning and distribution of deformation. Kinematic analysis has demonstrated that the Minas fault system segment examined here is a thinning deformation zone, in which strain is accommodated within progressively narrower volumes of rock. Deformation can be summarized as a broad, initially diffuse zone of triclinic transpression that has evolved, with the accumulation of finite strain, into zones of distinct structural style and variation in finite strain. It is not possible to demarcate 'deformed shear zone' and 'undeformed host rocks'. Instead, the Minas fault system is described in terms of discontinuous transitions in finite strain and deformation style within a large scale movement picture.

The Minas fault system (MFS) is an ENE-WSW trending transpressional boundary between the Appalachian Meguma and Avalon tectonostratigraphic terranes (Donohoe & Wallace 1982; Keppie 1982; Waldron et al. 1989) of Nova Scotia, Canada (Fig. 1). Current interpretations have the Avalon terrane migrating from its original position along the Gondwanan margin during the Neoproterozoic (Murphy & Nance 1989; Murphy et al 2001) with accretion to Laurentia by the Late Ordovician (Cawood et al. 1994). Accretion of Meguma and Avalon terranes involved extensive displacements along what is now defined as the MFS during the Middle to Late Devonian, Acadian orogeny (415-370 Ma) (Keppie 1982) and Carboniferous, Alleghanian-Variscan orogeny. This convergence between Laurentia and Gondwana culminated in the amalgamation of Pangea (Dalziel et al 1994; Miller et al 1995; Williams et al 1995; Gibbons etal. 1996; Murphy etal 1999). The specific study area lies between latitudes 45°20'59" N and 45°24'22" N and longitudes 64°52'03" W and 64°32'44" W along 25 km of coastline on the north shore of the Minas Basin (Fig. la, b). This segment of the crustal scale deformation zone (Webster et al 1998) comprises rocks of varied provenance exhibiting

very low- to low-grade metamorphism caught up in a long-lived zone of intense deformation. Acadian plutons (eg. Cape Chignecto pluton) that were introduced into Avalon terrane basement at the Devonian-Carboniferous boundary exhibit northwest-vergent thrusts and ductile fabrics dated at 329±11 Ma (Namurian) (Waldron et al 1989). Volcanism (Fountain Lake Group) occurred during the earliest Carboniferous along the southern margin of the Maritimes basin (Piper et al 1999). The Devonian-Early Carboniferous (late Famennian-Tournaisian) Horton Group is an overstep sequence deposited during and following the Acadian orogeny. In the Horton Group, the Greville River Formation is the predominant unit deformed within the MFS. It is a distal facies of a single alluvial fan-fluviatile-lacustrine unit (Piper 1994) developed from Avalon and Meguma (Murphy 2000) and comprises grey and tan quartz wacke and arenite and red and green siltstone. The mainly marine strata of the Visean Windsor Group, which represents the only known marine incursion into the Maritimes Basin prior to Late Carboniferous, succeeded the Horton Group. The Early-Middle Carboniferous Mabou Group followed the final marine retreat when terrestrial strata were deposited.

From-. ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 201-214. 0305-8719/$15.00 © The Geological Society of London 2004.

202

E. A. MACINNES & J. C. WHITE

Fig. 1. Location map of the study area; (a) Northern shore Minas basin, Nova Scotia. MFS, Minas fault system; (b) Map showing the external (low strain) and internal (high strain) zones. The internal zone is subdivided into marginal, core and shear band (SBD) domains.

The MFS and associated discrete faults in northern Nova Scotia exerted a strong control on subsequent Upper Carboniferous sedimentation that is characterized by pull-apart basins exhibiting rapid facies changes and coarsening of elastics associated with active faulting (Yeo & Ruixiang 1987). Mesozoic opening of the Atlantic produced graben systems along the eastern margin of North America with contemporaneous volcanism, sedimentation (fluvial, aeolian, lacustrine) and faulting (Stevens 1980, 1987). During this fragmentation of Pangea the MFS was a sinistral transtensional boundary, although at least one episode of post-Alleghanian (Variscan) dextral displacement has produced localized brittle deformation and folding in Triassic sandstones. In order to establish the specifics of deformation in a crustal-scale transpression zone, a large data set comprising foliation type, orien-

tation, geometric history and timing relationships was collected. This will in principle allow identification of and testing for prescribed flow in such zones, and extension of our knowledge of rock behaviour in transpression zones beyond simple recognition of regional geometry.

Minas fault system; framework Detailed structural analysis of the MFS (see Structural elements and geometric relationships) shows it to be divisible into a high-strain internal zone, in which transcurrent motion is concentrated, and a lower strain external zone that is dominated by a related sequence of folds. The internal (high strain) zone can in turn be subdivided into three domains that reflect increases in finite strain from margin to core of the internal zone: (1) marginal domain, (2) core domain and (3) shear band domain (Fig. Ib).

KINEMATICS OF THE MINAS FAULT SYSTEM

The demarcation line between contrasting styles and intensity of deformation of the external and internal zones is sharp and effectively defined by the current Cobequid fault scarp that has late normal displacement (Donohoe & Wallace 1982). The external zone is characterized by lowstrain deformation features typically associated with contraction: bedding plane thrusts, ramp and flat structures, folding and duplex structures. Sedimentary units from the external zone (Fig. 2a, b) give an indication of the protoliths for the highly deformed fault rocks within the internal zone of the fault. The internal zone (Fig. 2c, d) is exposed over a width of about 180 m. Regional displacements have juxtaposed highly deformed foliated fault slices of basement, and Carboniferous units. Intense deformation has accommodated large displacements and produced a wide range of fault rocks and tectonic fabrics including transposed compositional layering, sheath folds, cleavages, cyclically introduced veins, S-C fabrics, shear bands, fault gouge and discrete fractures. The development of first-order bimodal architecture comprising internal high strain and external contractional zones is notably similar to models for transpressional zones (Sanderson & Marchini 1984; Jones & Tanner 1995).

Fault rocks Whereas protoliths can be recognized easily in the external zone, the bulk of internal zone lithologies are described in terms of three primary fault rock units derived from the Greville River Formation: an arkosic L-S tectonite, a pelitic black phyllite and a carbonatequartz-mica schist. The units are often juxtaposed and have experienced intense deformation. Their presumed differing initial sedimentary compositions (e.g. clay- versus quartz-rich) in concert with material flux during deformation (veining) result in contrasting deformation features. The distinctive variation in mechanical response and microstructures provides an insight into the role of localization, partitioning and distribution of deformation.

Structural elements and geometric relationships Folds Folding is common throughout the MFS, with the style and significance of folds dependent on

203

their location within the overall fault zone architecture. Four generations of folding are observed in the external zone. First generation folds are tight to isoclinal, second generation folds are open and recumbent, third generation folds are upright and NW-verging and fourth generation folds are large-scale warps with dextral vergence which have axial plane oblique to the fault trace (Fig. 4d). In general folding within the internal zone is much more intense, with few hinges of largescale folds observed. In the marginal domain of the internal zone recumbent folds of black and green phyllite plunge moderately to the NW with a strong axial planar cleavage that dips moderately to the NW. These folds are refolded into upright folds with axial planes parallel to the strike of the fault. The sense of rotation on the earlier folds is dextral. Within the core domain a fold sequence plunges steeply to the NE and SW with steeply dipping axial planes to the SE and NW. Typically, at any one location, isolated fold hinges lie within the plane of the compositional foliation and plunge both to the SW and NE (Fig. 4e). Transposition is seen at all scales and is fundamental to development of the dominant compositional foliation. Repeated transposition is observed, consistent with a cyclical process whereby progressive folding of original layering produces a composite transposed layering. Relict quartz fold hinges and intrafolial folded compositional layering highlight the multiple folding that characterizes this process. Sheath folds are common within and restricted to the very fine-grained carbonatequartz-mica schist. Cyclic transposition precludes determination of timing relationships other than those of local generation. The youngest folds that refold the transposed layering are most evident in the marginal domain and in the quartz-mica L-S tectonite. They plunge moderately to the south and SE with steep axial planes that dip to the west and east. The orientation of these folds and dextral offsets on the limbs suggest they are related to a late stage of dextral displacement. Individual fold limb orientations, as well as the distribution of foliation data from the marginal domain, suggest the existence of folds plunging to the south that correlate with these late stage folds (Fig. 4a). Foliations Foliations are most extensively developed within the internal zone where they act as markers that highlight cyclical folding, relative displacement and mechanism of strain

Fig. 2. Contrasting styles of deformation in the Minas fault; (a) Contractional duplex structure in the external zone; (b) In the external zone transposed foliation is folded by upright NW vergent third generation fold (F3) which overprint recumbent second generation folds (F2) and tight isoclinal first generation folds; (c) In the core domain, internal zone, deformation is intense. Regional transposition has juxtaposed highly deformed foliated fault slices of basement and Carboniferous units; (d) Intense foliation produced by transposition of compositional layering in the core domain, internal zone.

KINEMATICS OF THE MINAS FAULT SYSTEM

accommodation. Specifically, the ubiquitous compositional foliation is used as the primary marker to highlight fault zone history. The generation of distinct fault rock types within the internal zone is also directly related to the processes by which foliations are produced. The dominant foliation in the internal zone is a highly transposed compositional layering. It is oblique to the fault scarp, trends approximately 075°, dips steeply to the SSE and exhibits normal displacement. The compositional layering is a product of high strain and is defined by layers of mylonite, ultramylonite, vein quartz, and modal variations of quartz and micas. In both field exposure and thin section, up to four generations of folds are recognized with each generation contributing to transposition of any pre-existing foliation. The repeated folding of pre-existing folds highlights the cyclical nature of deformation. Orientation data for compositional layering, from the marginal and core domain respectively, are displayed in Fig. 4a and b. A similar pattern of foliation orientations exists throughout the internal zone, dominated by a concentration of foliations dipping SE at 30°-60°. The latter is consistent with regional transposition of the initial primary compositional layering (bedding?) into parallelism with the MFS trend; this foliation effectively defines the fault internal zone and trend of the MFS. The marginal domain has more dispersed orientations of compositional layering (Fig. 4a) whereas compositional layering in the core domain has been more highly transposed towards parallelism the MFS trend (Fig. 4b). The tight cluster of foliations gives a typical foliation orientation striking 057° and dipping 63°SE (typical values assigned using eigenvalue calculations, Spheristat© 1990-1998 Pangaea Scientific). Variations in the foliation orientations reflect the progressive and cyclical nature of transposition in the high strain zone whereby material planes are continually being re-orientated and re-deformed within the fault zone. Separation of compositional foliation data from within the L-S tectonite and the black phyllite shows L-S tectonite foliations to be more tightly clustered, consistent with the very regular nature of the compositional foliation observed in outcrop. The wider spread of foliation orientation in the black phyllite can be attributed to the reorientation of this foliation during development of S-C fabrics and shear bands.

Lineations The intensity of deformation and fine-grained nature of most of the fault zone rocks make the

205

identification and classification of lineations very difficult. Mineral stretching lineations (Fig. 4a, b) typically have shallow to moderate south to SW plunges (e.g. 28° toward 218°) and are best developed in the L-S tectonite. The latter unit has a mylonitic layering containing a strongly developed S-foliation in thin section, whereas the lineation on the transposed foliation is pervasive and defined by the alignment of elongated quartz and mica (Fig. 3a). The mineral lineation and associated quartz rods, boudins and fold hinges are all effectively parallel. The black phyllite lacks a definitive stretching lineation but has well developed crenulation lineations. In the marginal domain more than one orientation of crenulation lineation is observed on the surface of the transposed foliation and are at high angles to the stretching lineations (Fig. 4c). In the core domain a single strong crenulation lineation plunging shallowly to the ENE is observed. Rotation of crenulation lineations into parallelism with the main stretching lineation is common and sometimes makes differentiation between these two linear features difficult.

Veins Quartz veins are cyclically introduced into the fault zone throughout its history and act as both important mechanical elements and deformation markers. The appearance and orientation of the quartz veins are influenced strongly by the type of fault rock into which they are injected and their relative position within the fault zone architecture. Veins in the marginal domain are relatively undeformed, cut the transposed compositional foliation and are generally 0.5-2 cm wide. Individual quartz grains are typically 0.2-1 cm. They are generally formed in two orientations at a high angle to each other. The veins are predominantly quartz but have some calcite. In the core domain the level of repeat veining is so intense that the composition of fault rocks is significantly altered; the style of subsequent deformation is largely controlled by the presence of these veins. Repeated introduction of veins into the quartz-mica L-S tectonite has produced a fault rock that has a large component of deformed vein quartz. The quartz veins are folded with attenuated and boudinaged limbs. They have undergone dynamic recrystallization giving a preferred orientation of the new grains that is oblique to the compositional layering (S-foliation). Recrystallized and subsequent deformed grains have aspect ratios with an upper value of 10:1 and more

Fig. 3. Fault rock units in the internal zone of the Minas fault; (a) Arkosic L-S tectonite with well developed stretching lineation (top right to bottom left) observed on transposed foliation surface; (b) Mylonitic foliation in arkosic L-S tectonite. Compositional layering comprises quartz-mica, quartz and mica-quartz domains; (c) Quartz veins are folded with attenuated and boudinaged limbs in pelitic black phyllite. Dextral offset of the compositional layering by shear bands; (d) Crenulation of pre-existing fabric and folded quartz veins in pelitic black phyllite; (e) Fine-grained carbonate-quartz-mica schist with shallowly plunging sheath folds that are subsequently shortened normal to the transposed foliation; (f) Extensive flow folding produced sheath folds that reflect the dominance of grain-size-sensitive flow mechanisms in the fine-grained (<10 um) carbonate-quartz-mica schist.

Fig. 4. Equal area lower hemisphere stereographic projections; (a) Marginal domain, internal zone transposed foliation (black circles) and stretching lineation (open triangles). VNS, vorticity normal section; S, average transposed foliation; C, shear zone boundary; X, calculated displacement direction; (b) Core domain, internal zone transposed foliation (black circles) and stretching lineation (open triangles). VNS, vorticity normal section; S, average transposed foliation; C, shear zone boundary; X, calculated displacement direction; (c) Crenulation lineations lie on the S-plane in the core and shear band domains in the internal zone; (d) External zone, great circle in foliation (black circles), fold plunges (open triangles) and average axial plane (ap); (e) Plunge of intrafolial folds from core domain, internal zone.

208

E. A. MACINNES & J. C. WHITE

Fig. 5. Equal area lower hemisphere stereographic projections of foliations from the shear band domain, internal zone; (a) C, planes; (b) S, foliations; (c) Shear bands; (d) Planes to average S- and C-planes and shear bands (SB). VNS, vorticity normal section; X, calculated displacement direction.

commonly 3:1 with an individual grain size of up to 2 mm. The most intensely deformed veins are fully incorporated into the fault rock as components of the composite compositional layering. Quartz veins act as important markers within the black phyllite and are integral to the style of deformation by which this fault rock is characterized (Fig. 3c,d)« The quartz veins are transposed and the limbs attenuated with dismembered quartz veins forming sigmoidal pods that are aligned obliquely to the compositional layering. The length of the pods is between 0.5-5 cm. Calcite and chlorite form in the necks of boudinaged quartz veins and tensional fractures.

S-C fabrics S-C fabrics (Berthe et al 1979; Lister & Snoke 1984) are developed in both the L-S tectonite and the black phyllite of the core domain. Cplanes are sub-parallel to the compositional foliation that serves as the marker from which both S- and C-fabrics are developed. Within the L-S tectonite, S-foliation is defined by the preferred orientation of quartz, plus aligned mica and chlorite. C-planes are defined by aligned finegrained white mica and chlorite and very finegrained quartz. The black phyllite exhibits the strongest S-C

fabric. On a gross scale C-planes exhibit shear displacement of compositional layering; in thinsection individual C-planes are defined by pressure solution cleavage, gouge zones, ultramylonite, or fine-grained micas. S-planes are formed from segments of dismembered quartz vein and axial planes of crenulated micas. With increasing strain, S-surfaces rotate into parallelism with C-surfaces and produce a composite S-C fabric. C-planes from the black phyllite in the core domain typically strike 069° and dip 57° SE (Fig. 5a). The largest C-planes form as through going surfaces that bound zones of particularly intense shear band formation. S-foliations in the black phyllite are easily measured in outcrop (Figs 5b & 8a). This contrasts with the S-foliation in the L-S tectonite that can only be routinely established in thin section (Fig. 3b). S-planes formed in the black phyllite are oblique to the shear zone boundary and typically dip 57°SE with a strike of 044° (Figs 5b,d & 7). There is an overlap of the S-foliation and the transposed foliation orientations that reflects the fact that S-foliations are forming at all scales during heterogeneous deformation.

Crenulation cleavage Crenulation of pre-existing fabrics is particularly extensively developed in the very finegrained black phyllite but is common wherever

KINEMATICS OF THE MINAS FAULT SYSTEM

209

Fig. 6. Equal area lower hemisphere stereographic projections, (a) Internal zone dextral faults; (b) Internal zone sinistral faults.

alignment of layer silicates has occurred during deformation and transposition of the host rocks. As a result crenulations can be both pervasive and localized in response to progressive folding and transposition. The hinges of the crenulations are typically composed of chlorite, feldspar and white mica and the limbs have a concentration of pressure solution seams, ilmenite, titanium oxides, opaques and chlorite (Fig. 3d).

Shear bands Shear bands (Platt & Vissers 1980; Williams & Price 1990) are well developed (Figs 3c, d, 5c, 7 & 8) throughout the core domain. These occur as extensional features with discrete, penetrative surfaces which form at angles of less than 35° to the shear zone margin, offset the compositional layering and S-C fabrics in a dextral sense and are sigmoidal, curving asymptotically into C-planes. Shear bands are associated with S-C fabrics, and likewise are best developed in the black phyllite where they overprint and offset the compositional layering, S-foliation and C-planes. The typical shear band dips 57°S and strikes 095° (Figs 5c,d & 7). At outcrop scale, individual shear bands range in length from 1-15 cm. In the quartz-mica L-S tectonite, shear bands offset the compositional layering and are typically less than 1 cm in length. In thin section, they are associated with thin (2-7 mm) layers of crenulated and transposed mica-quartz domains. In thin section multiple generations of shear bands are identified. During progressive deformation, shear bands are rotated into near parallelism with the shear plane and are subsequently overprinted by younger sets of shear bands, as reported elsewhere by Alsop (1993). Faults Brittle faults are most common within the internal zone and form Riedel-type patterns ranging

Fig. 7. Scale independent schematic representation of the major structural features from the internal zone. MFS, Minas fault system; ST, transposed foliation; S, foliation; C, plane; SB, shear bands; X, shears; Y, planes; P, orientations; R1} shears.

in scale from 200 m wide (full internal zone) to a few metres (Fig. 7). Displacement sense on these faults can be determined by offsets of lithologically distinct fault blocks and ductile foliations. Both individual faults and the overall pattern demonstrate dextral displacement within the MFS (Fig. 6a). Large displacement faults with kilometres of offset (Donohoe & Wallace 1982) juxtapose lithologically distinct foliated fault blocks parallel to the MFS boundary. The latter faults define the largest scale Riedel pattern (Fig. 7) and as first-order faults are defined as the principal Y-shears, (Logan et al. 1979). Secondorder faults oblique to the MFS boundary, R r shears, are typically sigmoidal, curving asymptotically into the Y-faults. Displacement on Rrshears is synthetic to the displacement on the Y-faults. In competent quartz-rich fault rocks, only Y- and Rrfaults are developed. In addition to Y- and R!-faults, phyllitic fault rocks exhibit antithetic rotation of the pre-existing compositional foliation along Rrshears to form P-orientation fabrics, along which displacement can also be observed (P-shears). Isolated dextral faults that could not be unambiguously classified as Rrshears nevertheless strike obliquely to the shear zone boundary and are approximately parallel Rrorientations (Fig. 6a).

210

E. A. MACINNES & J. C. WHITE

Sinistral faults are less common and exhibit two orientation concentrations that dip steeply to the west and east (Fig. 6b). These strike at a high angle to the fault zone and correspond to regional sinistral faults active during the Mesozoic (Olsen & Schlische 1990). Other sinistral faults are overprinted by dextral faults; these are interpreted as Riedel X-shears (Logan et al. 1979) that are synthetic to the overall displacement. Normal and reverse faults are largely absent from the internal zone.

Fault zone kinematics Despite the geometric heterogeneity of the deformation, displacement sense and vorticity indicators (fold vergence, S-C fabrics, shear band displacements, Riedel fault patterns) consistently indicate dextral movement (preMesozoic) throughout the MFS. Of these, only fold and fabric development relate strictly to

flow kinematics, whereas shear bands and Riedel faults are mechanically controlled phenomena that accommodate the same overall kinematics. The large-scale partitioning of deformation into distinct external (contraction) and internal (shear) zones is comparable to models developed for transpression (Sanderson & Marchini 1984; Tikoff & Teyssier 1994; Jones & Tanner 1995). The abruptness of this transition within a crustal scale shear zone is remarkable (e.g. typically less than 50 m for transition from external to internal zones) and defines distinct deformation domains (marginal, core and shear band domains) within the same overall tectonic picture. The division of the internal or high-strain zone reflects increasing finite strain from margin to core of the main fault zone. The primary differences between the marginal and core domains are a discontinuous increase in the level of deformation, the strength and ubiquity of stretching lineations and the style of structures. The shear band domain reflects the most intense deformation in the internal zone in which extensional shear bands developed in order to accommodate displacements. The orientations of stretching lineations do not correspond to the calculated shear direction within the internal zone (Fig. 4a,b), consistent with triclinic transpression zones (Lin et al 1998; Jiang etal 2001). The shear direction is obtained by determining the intersection of the MFS trace (C-plane) and the vorticity normal section (VNS); (Robin & Cruden 1994); the VNS is the great circle normal to the line of intersection of the S-foliation and the MFS, where, at large strains, the average S-plane pole approximates the normal to the maximum principal finite

Fig. 8. Shear band domain, internal zone; (a) Dextral displacement by shear bands (top left to bottom right) offset the S-C foliation in pelitic black phyllite; (b) Black phyllite exhibits strong S-C anisotropy. Shears bands are discrete penetrative surfaces with displacement synthetic to the overall displacement direction, (c) Shear band in black phyllite with layering defined by micas, chlorite and titanium oxides.

KINEMATICS OF THE MINAS FAULT SYSTEM

strain axis (Lin & Williams 1992). The calculated shear direction throughout the internal zone is dextral with a small WSW plunge, although stretching lineations are, on average, steeper toward the SW; this relationship holds for both averaged regional orientation data (marginal and core domains) and more localized data (shear band domain). Poles to foliation are distributed along the vorticity normal section (VNS), whereas stretching lineations lie off of this plane and are interpreted to have migrated toward the down-dip direction (Lin et al 1998) of the shear plane as anticipated for triclinic transpression. The steep average plunge of stretching lineations relative to the shear direction implies a moderate ratio of y/8 less than 10 within the internal high-strain zone (Lin et al. 1998); that is, there has been a significant component of pure shear strain across the zone boundary in concert with MFS-parallel displacement. Evidence supporting this, in a zone that largely has the attributes of transcurrent shear, are shallowly plunging sheath folds that are shortened normal to the transposed foliation (Fig. 3e). The external (low strain) zone has kinematical attributes similar to the internal zone, but the significantly lower levels of finite strain enable fuller preservation of the temporal sequence of deformation. Several fold generations (Fig. 2b) have an approximately common NNE plunging fold axis, with extensional structures plunging shallowly NE (Fig. 4d). Folds and thrusts verge NW, suggesting an overall zone of contraction, in contrast to the dominant MFS-parallel shearing of the internal zone. The average axial plane of the latest folds (F4) trends in what would be an early S-foliation orientation (i.e. perpendicular to the instantaneous shortening direction) for dextral displacement prior to rotation of these material planes toward the bulk shear plane (MFS boundary) during accumulation of finite strains. In other words, the external zone appears to preserve the early stage of the folding-transposition cycle experienced most intensely within the internal high-strain zone where records of the initial stages of the cycle are more commonly lost because of the combined effect of the higher y/e ratio and the localization of strain in this zone. Although deformation has produced materially, temporally and spatially distinct features throughout the fault zone history, these same geometrically complex, overprinting relationships can be interpreted in terms of kinematically related conditions. Such behaviour is consistent with quasi-constant fault zone boundary constraints where the heterogeneity of

211

deformation within the fault zone arises from inherent contrasts in rock response, as opposed to fundamental variations in the boundary conditions (Jiang & White 1995; Jiang & Williams 1999), although the latter is not precluded.

Fault zone deformation mechanisms Overprint relationships demonstrate a general progression from ductile through semi-brittle to brittle structures suggesting a ductile to brittle fault zone evolution; however, there is no strong metamorphic evidence for significant crustal exhumation. Instead, a specific tendency towards either ductile or brittle deformation is related to several factors that include fault rock type, evolutionary history of structures, inferred deformation micromechanisms, accumulated finite strain and strain localization. Rock type and structural style have a firstorder association. For example, sheath folds are restricted to the very fine-grained carbonate-quartz-mica schist, whereas the quartz-rich L-S tectonite exhibits the most pervasive and penetrative extensional structures (lineations, boundinage, rodding) and the highest proportion of introduced quartz veins. Phyllites, with the highest proportion of primary layer silicates, provide excellent markers for multiple generations of folding and transposition and are the only rock unit in which shear bands, in combination with strong S-C fabrics are found. The partitioning of deformation found amongst these units can in turn be associated with contrasting deformation micromechanisms. Of the fault rocks, the carbonate-quartz-mica schist has a uniquely small grain size (<10 jam) with mutual pinning of grain boundaries by the contrasting mineral phases. There is a distinct absence of evidence of grain size reduction that is otherwise common throughout the fault rocks suggesting that the current small grain size reflects a fine-grained primary sedimentary rock. The fine-grained, equant grain texture is indicative of grain-size-sensitive (GSS) flow that appears to have been an important mechanism throughout the deformation of this unit. Although sheath fold formation is a function of the deformation path (Jiang & Williams 1999), the multiple degrees of freedom for grain-scale displacement provided by GSS flow favour such structures (Jiang & White 1995); that is, given an imposed kinematic condition that would favour sheath folds, there is no inhibition of that path at the micromechanical level. The L-S tectonite exhibits dislocation creep textures associated with cycles of deformationrecovery-recrystallization that produce S-C

212

E. A. MACINNES & J. C. WHITE

fabrics, a strong preferred crystallographic orientation and a penetrative quartz mineral lineation. The apparent concentration of quartz veining in this unit argues for temporal variations in fluid pressure and hydraulic fracturing that preferentially localize the veining in this unit. This might be explained by operation of a positive feedback system whereby work hardening of a initially low permeability, strong sandstone unit led to hydraulic fracturing and introduction of veins that enabled continuation of dislocation-mediated deformation (Kennedy & White 2001), with the cycle repeated many times. The smaller volumes of quartz veins in other units are inferred to arise because these units were more able to accommodate deformation under the imposed conditions by ductile flow without hardening to the point of brittle fracture. The phyllite unit, because of its Theologically contrasting compositional components (e.g. mica vs. quartz veins), is an excellent marker during the early stages of transposition. However, as finite strain increases, the silicate microfabric layer of the composite transposed foliation becomes strongly aligned with the shear planes. This high-strain fabric has a tendency to restrict generalized deformation by constraining displacement within the shear plane. Essentially there is development of a single macroscopic slip system that has the effect of work hardening for deformations that cannot be thus accommodated. In response to this restriction, shear bands develop by ductile yielding synthetic to the overall kinematics of the fault zone. Formation of shear bands softens the rocks in two ways. First, the shear bands themselves provide an additional macroscopic slip system. Secondly, the back rotation of material between shear bands during their displacement reorients the S-foliated phyllite into orientations at higher angles to the shear plane that will allow additional deformation in rock volumes between C-planes. The latter is not unlike a microscopic Bauschinger effect in experimental deformation where reorientation of crystals after an increment of strain can lead to subsequent deformation at lower flow stress.

Strain partitioning and localization The observed variations in the intensity of transposition, fabric development and overprinting establish an overall gradient of finite strain, with strain increasing from the external zone to the core of the internal zone. The phyllitic shear band domain is arguably the highest ductile strain portion of this part of the MFS; as well as

being embedded in the core domain of the internal zone, extensional shear bands are typically associated with high finite shear strain (Passchier 1984). In support of the latter interpretation, phyllite in the marginal domain has well preserved transposition features, and an absence of shear bands. With this increase in strain from edge to centre of the MFS, there is a related progressive localization of deformation, culminating in the discrete shear band domain; that is, domain widths decrease through the sequence external zone => marginal domain =$ core domain => shear band domain. The narrowing of definable domains (strain localization) and the concomitant increase of strain within the domains can be explained in terms of variations in strain rate throughout the deformation zone. In other words, because the accumulation of strain is distinctly heterogeneous, at least some of that finite strain variation is a function of strain-rate variation. The development of faster strain rates in order to accommodate imposed boundary condition displacements would impact both on work hardening rates and the degree of strain localization. At the limit, such strain-rate localization contributes to the transition from ductile to brittle structures. The similarities of the shear band domain fabrics and the Riedel fault pattern orientations (Fig. 7) supply evidence for the latter type of transition. The quasi-parallelism between C-S-shear band ductile fabrics and Y/P/RX brittle fabrics (Fig. 7) support the contention that they fulfil the same mechanical and kinematical purpose (Mawer & White 1987; White et al. 1980). Although the faults might simply reactivate preexisting ductile fabrics, the fact that they overprint all rock types favours their origin as kinematic equivalents of the shear band domain structures under conditions of extreme work hardening of the overall fault zone. Additionally, those cases where S-C fabrics overprint brittle deformation patterns demonstrate that there can be concomitant operation of largescale brittle faulting and quasi-ductile deformation. Ductile to brittle transitions are commonly the result of crustal exhumation or some other cause of crustal cooling. However, the MFS fault rocks do not show any explicit mineralogical evidence consistent with extensive uplift through the ductile-brittle transition. The kinematic analysis (Fig. 4) has demonstrated that the MFS is a thinning deformation zone in which strain is accommodated in progressively narrower volumes of rock; that is, the basin in which lower Carboniferous Horton Group

KINEMATICS OF THE MINAS FAULT SYSTEM

rocks were deposited has been condensed into a much narrower zone between the Meguma and Avalon terranes. Such thinning can be anticipated to give extensive hardening consistent with the progressive ductile localization and ductile-brittle transitions. In effect the MFS can be described as an initially diffuse zone of transpression comprising significant folding and thrusting typified by the external zone. With continued displacement, the MFS evolves towards more localized zones of concentrated deformation related to shear zone thinning and the accumulation of finite strain. Such a progressive evolution makes the definition of a discrete 'shear zone boundary' difficult in that the overall zone of deformation is implicated in the same kinematical response. It is not possible to demarcate 'deformed shear zone' and 'undeformed host rocks'. Instead, the MFS is described in terms of discontinuous transitions in finite strain and deformation style within a broad movement picture.

Conclusion The Minas fault system is a broad, initially diffuse zone of triclinic transpression that has evolved, with the accumulation of finite strain, into zones of distinct structural style and variation in finite strain. Despite the geometrical complexities observed, both large-scale and long-lived consistency amongst the kinematical indicators can be demonstrated when they are carefully discriminated. Within the MFS, there is large-scale partitioning of deformation into distinct external (contractional structures) and internal (shear) zones. With the increase in strain from external to internal zones there is progressive localization of deformation, culminating in the discrete shear band domain. The narrowing of these domains and the concomitant increase in strain can be explained as reflections of variation in strain rate across the fault zone. The faster strain rates would impact both on work hardening rates and/or the degree of localization that could contribute to the transition from ductile to brittle structures. The orientation similarities between the shear band domain and the Riedel shear patterns are consistent with transitions from ductile to brittle deformation associated with work hardening at constant crustal depth without the need to invoke crustal exhumation.

References ALSOP, G.I. 1993. Sequential generation of listric shear bands during protracted ductile thrusting within

213

the Ballybofey nappe, north-west Ireland. Irish Journal of Earth Sciences, 12,1-12. BERTHE, D., CHOUKROUNE, P. & JEGOUZO, P. 1979. Orthogneiss, mylonite and non-coaxial deformation of granites: the example of the South Amorican shear zone. Journal of Structural Geology, 1, 31-42. CAWOOD, P.A., DUNNING, G.R., Lux, D. & VAN GOOL, J.A.M. 1994. Timing of peak metamorphism and deformation along the Appalachian margin of Laurentia in Newfoundland: Silurian, not Ordovician. Geology, 22, 399-402. DALZIEL, I.W.D., DALLA SALDA, L.H. & GAHAGAN, L.M. 1994. Paleozoic Laurentia-Gondwana interaction and the origin of the Appalachian-Andean mountain system. Geological Society of America Today, 2, 237-241. DONOHOE, H.V. & WALLACE, P.I. 1982. Geological Map of the Cobequid Highlands. Colchester, Cumberland and Pictou counties, Nova Scotia. Nova Scotia Department of Mines and Energy, Map 82-06, Scale 1:50 000. GIBBONS, W., DOIG, R., GORDON, T, MURPHY, B., REYNOLDS, P. & WHITE, J.C. 1996. Mylonite to megabreccia: Tracking fault events within a transcurrent terrane boundary in Nova Scotia, Canada. Geology, 24, 411-414. 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. JIANG, D. & WILLIAMS, PR 1999. A fundamental problem with the kinematic interpretation of geological structures. Journal of Structural Geology, 21, 933-937. JIANG, D., LIN, S. & WILLIAMS, P.F. 2001. Deformation path in high-strain zones, with reference to slip partitioning in transpressional plate-boundary regions. Journal of Structural Geology, 23, 991-1005. JONES, R.R. & TANNER, P.W.G. 1995. Strain partitioning in transpression zones. Journal of Structural Geology, 17, 793-802. KENNEDY, L.A. & WHITE, J.C. 2001. Low-temperature recrystallization in calcite: Mechanisms and consequences. Geology, 29,1027-1030. KEPPIE, J.D. 1982. The Minas Geofracture. In: ST. JULIEN, P. & BELAND, J. (eds) Major structural zones and faults of the Northern Appalachians. Geological Association of Canada Special Paper, 24, 263-280. LIN, S., JIANG, D. & WILLIAMS, P.F. 1998. Transpression (or transtension) zones of triclinic symmetry: natural example and theoretical modelling. In: HOLDSWORTH, R.E.,

STRACHAN, R., DfiWEY, J.F.

(eds) Continental transpressional and transtensional tectonics. Geological Society, London, Special Publications, 135,41-57. LIN, S. & WILLIAMS, P.F. 1992. The geometrical relationship between the stretching lineation and the movement direction of shear zones. Journal of Structural Geology, 14, 491-497. LISTER, G.S. & SNOKE, A.W. 1984. S-C mylonites. Journal of Structural Geology, 6, 617-638.

214

E. A. MACINNES & J. C. WHITE

LOGAN, J.M., FRIEDMAN, M., HIGGS, N.G., DENGO, C. & SHIMAMOTO,T. 1979. Experimental studies of simulated gouge and their application to studies of natural fault zones. U.S. Geological Survey Openfile Report, 79-1239, 305-343. MAWER, C.K. & WHITE, XC. 1987. Sense of displacement on the Cobequid-Chedabucto fault system, Nova Scotia, Canada. Canadian Journal of Earth Sciences, 24, 217-223. MILLER, B. V., NANCE, R.D. & MURPHY, J.B. 1995. Kinematics of the Rockland Brook fault, Nova Scotia: implications for the interaction of the Meguma and Avalon Terranes. Journal of Geodynamics, 19, 253-270. MURPHY, J.B. 2000. Tectonic influence on sedimentation along the southern flank of the late Paleozoic Magdalen basin in the Canadian Appalachians: geochemical and isotopic constraints on the Horton Group in the St. Marys basin, Nova Scotia. Geological Society America Bulletin, 113, 997-1011. MURPHY, J.B. & NANCE, R.D. 1989. Model for the evolution of the Avalonian-Cadomian belt. Geology, 17, 735-738. MURPHY, J.B., KEPPIE, J.D. & NANCE, R.D. 1999. Fault reactivation within Avalonia: plate margin to continental interior deformation. Tectonophysics, 305,183-204. MURPHY, J.B., KEPPIE, J.D. STAGEY, J. & TRAINOR, R. 2001. Deciphering the Neoproterozoic history of the Hallow fault, Avalon terrane, mainland Nova Scotia. Journal of Structural Geology, 23, 1067-1077. OLSEN, RE. & SCHLISCHE, R.W 1990. Transtensional arm of the early Mesozoic Fundy rift basin: Penecontemporaneous faulting and sedimentation. Geology, 18, 695-698. PASSCHIER, C.W. 1984. The generation of ductile and brittle shear bands in a low-angle mylonite zone. Journal of Structural Geology, 6, 273-281. PIPER, DJ.W. 1994. Late Devonian-earliest carboniferous basin formation and relationship to plutonism, Cobequid Highlands, Nova Scotia. Geological Survey of Canada Paper, ID, 109-112. PIPER, DJ.W, DESSUREAU, G. & PE-PIPER. G. 1999. Occurrence of early Carboniferous high-Zr rhyolites, Cobequid Highlands, Nova Scotia: temperature effect of a contemporaneous mafic magma. The Canadian Minerolgist, 37, 619-634. PLATT, J.P. & VISSERS, R.L.M. 1980. Extensional structures in anisotropic rocks. Journal of Structural Geology, 2, 397^10.

ROBIN, P.-Y.F & CRUDEN, A.R. 1994. Strain and vorticity patterns in ideally ductile transpression zones. Journal of Structural Geology, 16,447-466. SANDERSON, D.J. & MARCHINI, WR.D. 1984. Transpression. Journal of Structural Geology, 6, 449-458. STEVENS, G.R. 1980. Mesozoic vulcanism and structure; Northern Bay of Fundy region, Nova Scotia: Field Trip Guidebook 8. Geological Association of Canada Annual Meeting, Halifax, Nova Scotia, 40pp. STEVENS, G.R. 1987. Jurassic basalts of northern Bay of Fundy region, Nova Scotia. Geological Society of America Centennial Field Guide - Northeastern Section, pp. 415-420. TIKOFF, B. & TEYSSIER, C. 1994. Strain modeling of displacement-field partitioning in transpressional orogens. Journal of Structural Geology, 16, 1575-1588. WALDRON, J.W.F, PIPER, DJ.W & PE-PIPER, G. 1989. Deformation of the Cape Chignecto pluton, Cobequid highlands, Nova Scotia: thrusting at the Meguma-Avalon boundary. Atlantic Geology, 25, 51-62. WEBSTER, T.L., MURPHY, J.B. & BARR, S.M. 1998. Anatomy of a terrane boundary: an integrated structural, geographic information system, and remote sensing study of the late Paleozoic Avalon-Meguma terrane boundary, mainland Nova Scotia, Canada. Canadian Journal of Earth Sciences, 35, 787-801. WHITE, S.H., BURROWS, S.E., CARRERAS, I, SHAW, N.D. & HUMPHREYS, FJ. 1980. On mylonites in ductile shear zone. Journal of Structural Geology, 2, 175-187. WILLIAMS, P.F., GOODWIN, L.B. & LAFRANCE, B. 1995. Brittle faulting in the Canadian Appalachians and the interpretation of reflection seismic data. Journal of Structural Geology, 17, 215-232. WILLIAMS, P.F. & PRICE, G.P. 1990. Origin of kink bands and shear-band cleavage in shear zones: an experimental study. Journal of Structural Geology, 12,145-164. YEO, G.M. & RUIXIANG, G. 1987. Stellarton graben: an Upper Carboniferous pull-apart basin in northern Nova Scotia. In: BEAUMONT, C. & TANKARD, A.J. (eds) Sedimentary basins and basin-forming mechanisms. Canadian Society of Petroleum Geologists, 12, 299-309.

Development of local orthorhombic fabrics within a simple-shear dominated sinistral transpression zone: the Arronches sheared gneisses (Iberian Massif, Portugal) M. FRANCISCO PEREIRA1 & J. BRANDAO SILVA2 Grupo de Tectonica, Centra de Geofisica de Evora, Departamento de Geociencias, Universidade de Evom,Apt.94, 7002-554 Evora, Portugal (e-mail: [email protected]) ^Departamento de Geologia,Faculdade de Ciencias, Universidade de Lisboa, C2-5°piso, Campo Grande, 1749-016 Lisboa, Portugal l

Abstract: The Coimbra-Cordoba shear zone (Iberian Massif), characterized by simpleshear dominated sinistral transpression, exposes several outcrops of strongly sheared peralkaline gneisses surrounded by mica schists and amphibolites. These gneisses are included in the Arronches Tectonic Unit, a thick unit of mylonitic rocks with a steep foliation and an associated gently plunging stretching lineation parallel to the fold axes. Strain partitioning is testified by widely spaced anastomosing shear bands around less-strained domains and by the existence of different shearing domains ranging from relatively 'lessstrained' and coarse-grained mylonites to highly strained and fine-grained ultramylonites. Three shearing domains defined by textural and structural changes resulted from progressive deformation and increasing strain, which leads to increased mylonitization of gneisses. This is revealed by the increased modal percentage of the matrix and the decreased percentage of porphyroclasts, accompanied by evolution from orthorhombic to monoclinic fabrics: Conjugate Shearing Domain (CSD), Intermediate Sinistral Domain (ISD), and Sinistral Domain (SD). This contribution shows that in a simple-shear sinistral dominated transpression zone with a well-developed and widespread monoclinic fabric, it is possible to find mechanical conditions to produce local orthorhombic fabrics. In the Arronches gneisses a local strain regime exists in apparent contradiction with the bulk deformation regime.

Structural analysis on mylonitic gneisses is widely used to characterize deformation in highstrain ductile shear zones. These strongly foliated and lineated rocks usually display fine-grained mylonitic textures, derived from an original coarser-grained rock as a result of distinct deformation mechanisms (e.g. White et al. 1980; Tullis et al. 1982; Knipe 1989; Passchier & Trouw 1996). Mylonites located along anastomosing narrow planar shear zones with associated strain partitioning, have been described over a wide range of scales (e.g. Carreras et al. 1980; Hanmer 1988; Hudleston 1999; Carreras 2001). They normally contain fabric elements with a monoclinic symmetry (Berthe et al. 1979; Platt & Vissers 1980; Gapais & White 1982; Simpson & Schmid 1983; Hanmer & Passchier 1991). An important feature of these ductile shear zones is the existence of strain partitioning into domains with different strain gradients (e.g. Bell 1981) leading to the existence of highly variable fabrics related to different strain states through which the rock passes during progressive deformation (e.g. Twiss & Moores 1992; Means 1994). This type of

partitioning related to progressive deformation may explain the local preservation of lessstrained rocks preserving earlier stages of formation of mylonitic fabrics, surrounded by highly intense stages of mylonitization with ultramylonites. The distinction of shearing domains where the imposed bulk strain path is simultaneously distributed in a coaxial component (orthorhombic fabric), mainly outside the shear zone, and a non-coaxial component (monoclinic fabric) as a result of increased strain within the shear zone (e.g. Wenk et al. 1987), constitutes a general assumption for shear zones that is questioned in this case of the Arronches gneisses. This study describes a variety of textures and structures that define distinct shearing domains formed simultaneously as a consequence of increased strain in the Arronches gneisses. When strain softening is introduced to these rocks, their behaviour is accompanied by the development of local orthorhombic fabrics (conjugate shear planes), as can be observed in directly comparable experiments with analogue material by Mancktelow (2002). Increasing strain softening promotes the development and

From: ALSOP, G. L, HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 215-227. 0305-8719/$15.00 © The Geological Society of London 2004.

216

M. F. PEREIRA & J. B. SILVA

propagation of shear planes and, as a consequence, an evolution from orthorhombic to monoclinic fabrics may occur. Considering our own observations, and comparing them with similar examples described by other authors (e.g. Gapais & Bale 1990; Goodwin & Williams 1996; Carreras 2001), we propose an interpretation for the geometric and kinematic conditions by which the local orthorhombic fabrics in the Arronches gneisses may have been formed.

The Arronches Tectonic Unit Regional setting The Arronches Tectonic Unit (ATU) is part of a major intra-continental simple-shear sinistral dominated transpression zone, known as the Coimbra-Cordoba Shear Zone (CCSZ; Burg et al 1981; Matte 1991). The ATU is a 0.5-5 km wide, 50 km long NW-SE trending sheared band located south of a major high-grade blastomylonitic unit (Fig.l) (Pereira 1999; Pereira & Silva 2002). The structure of the CCSZ is characterized by an asymmetric flower structure (Fig. 2A) (Pereira & Silva 2001) roughly parallel to the boundary between the Ossa-Morena and Central-Iberian Zones (Pereira 1999) that extends for more than 600 km in the SW Iberian Massif (Fig. 1A). The available geochronological data for the CCSZ (Garcia-Casquero et al. 1985; Quesada & Dallmeyer 1994; Ordones-Casado 1998; Eguiluz et al. 2000) document a complex and polymetamorphic history recorded on rocks of various ages. It is generally accepted that a sinistral transcurrent motion dominated during the Upper Palaeozoic (D2 Variscan orogeny: ~ 410-280 Ma), superposed on a previously deformed and metamorphosed Neoproterozoic-Lower Cambrian basement (D! Cadomian orogeny: ~ 650-520 Ma) intruded by Lower Palaeozoic rocks (Eguiluz et al 2000; Pereira & Silva 2002). The peralkaline rocks represent pre-Variscan intrusives, which have been associated with Ordovician-Silurian extensional episodes of intra-continental rifting (Lancelot & Alegret 1982; Eguiluz et al. 2000; Simancas et al. 2001).

Structure and metamorphism The ATU comprises medium-grade metamorphic rocks (T: 550-600 °C; P: 8-9 kbar): metasedimentary and metavolcanic rocks of Neoproterozoic and probably Lower Cambrian

age (Pereira 1999) and Ordovician intrusives (Gonc.alves 1971). Rock units exposed in the Arronches area include biotite-rich mica schists, garnet-rich mica schists, staurolite-garnet paragneisses, amphibolites, marbles, quartzites, garnet-rich amphibolites and alkali amphibolerich gneisses (Fig. 2). This NW-SE trending sigmoidal band (Fig. IB) contains rocks with a complex tectonometamorphic history with a penetrative deformation (D2) characterized by well-developed steep foliation and sub-horizontal lineation (S2-L2 fabric), superposed on earlier structures (DI). The steep mylonitic S2 foliation in the ATU strikes 330° dipping 60°-80° SW (Fig. 2B), parallel to the NW-SE orientation of the CCSZstrike. The D2 folds have S2 axial-plane parallel cleavage and their axes are aligned parallel to the NW or SE gently plunging stretching lineation (L2), defined by the growth of biotite, garnet, muscovite and hornblende.

Structural analysis of Arronches gneisses Peralkaline gneisses Peralkaline gneisses within the ATU form narrow bands 4-8 km long and 2-30 m wide (Fig. 2B). These rocks show typical intrusive relationships with respect to the surrounding deformed country rocks (Pereira 1999; Pereira & Silva 2002): dykes cutting across a former tectonic anisotropy on country rocks; and xenoliths of foliated country rocks inside the gneisses. During the Variscan deformation (D2), the peralkaline rocks and the previously metamorphosed country rocks were metamorphosed to amphibolite facies and retrogressed to greenschist facies. Original textures were overprinted by intense D2 progressive deformation during prograde and retrograde metamorphism, obscuring the geometric and kinematic characteristics of the magmatic fabrics. The field aspect of the peralkaline gneisses from Arronches is characterized by the presence of dark green and/or blue to black alkali-amphiboles (hastingsite, riebeckite) in a foliated pink to greyish fine-grained matrix (K-feldspar, quartz, biotite, amphibole, plagioclase and garnet; sphene, chlorite, zircon and opaques occur as accessories). The fine-grained facies (fine porphyritic) consists of stretched and fractured porphyroclasts of amphibole and feldspars (crystals <5 mm in size) surrounded by a felsic matrix (crystals <1 mm in size) defining a planar fabric parallel to the biotite alignment. The shear indicators are marked by G-type porphyroclasts

Fig. 1. (A) Location of the study area of Arronches (Portugal) as part of the Coimbra-Cordoba shear zone (CCSZ) near the Ossa-Morena (OMZ) and CentralIberian (CIZ) zones (Iberian Massif). (B) Simplified geological map of the Coimbra-Cordoba shear zone and adjacent margins (after Gongalves 1970; Pereira 1999).

218

M. F. PEREIRA & J. B. SILVA

Fig. 2. (A) Generalized cross-section through the CCSZ and adjacent margins (see Fig. 1 for location and map keys). ATU, Arronches Tectonic Unit; CCSZ, Coimbra-Cordoba shear zone; D b Cadomian structures; D2, Variscan structures. (B) General structures and lithology of the Arronches Tectonic Unit (after Pereira 1999). Three main types of shearing domains are exposed in outcrops of sheared peralkaline gneisses: a conjugate shearing domain (CSD), an intermediate sinistral domain (ISD) in the SE and a sinistral domain (SD) in the north, close to the high-grade blastomylonitic unit.

ORTHORHOMBIC FABRICS IN SHEARED GNEISSES

219

Fig. 3. (A) Detailed structural map of the outcrops where the Arronches and Caia rivers merge (for location see Fig. 2B). Examples of the CSD structures observed within: (B) Country rocks with metre-scale sinistral shear bands in mica schists and conjugate sets of faults in dykes of garnet-rich amphibolites and peralkaline gneisses; (C) Peralkaline gneisses with conjugate sets of centimetre-scale dextral and sinistral shear bands; (D) and (E) Peralkaline gneisses with a cross-cutting relationship between conjugate sets of centimetre-scale dextral and sinistral shear bands; (F) Mica schists with centimetre-scale dextral shear bands. Legend for the schematic stereogram: dark grey field, average strike of dextral shear planes; white field, average strike of sinistral shear planes.

of K-feldspar and alkali-amphiboles, and C'-type shear band cleavage. The coarse-grained facies (coarse porphyritic) consists of stretched crystals of alkali feldspars and dark amphiboles (2-20 mm in size) surrounded by a felsic matrix of quartz-feldspar (crystals <1 mm in size). The prominent sigmoidal geometry of the S-C structures is characterized by the development of bands a few millimetres thick with stretched feldspar and alkali-amphibole porphyroclasts separated by very thin biotite-rich planes. The most remarkable feature of the peralkaline gneisses is the heterogeneity of the strain distribution (strain partitioning) defined by the development of distinct associations of structures and mylonitization stages. Detailed investigation of well-exposed outcrops allows the distinction between three main types of shearing domains (Fig. 2B): a Conjugate Shearing Domain (CSD), an Intermediate Sinistral Domain (ISD) and a Sinistral Domain (SD). These shearing domains are described as individual and/or composite (cross-cutting one

another) occurrences, at different scales and with a heterogeneous distribution along normal and parallel profiles. Their description is based upon textural and structural changes from orthorhombic to monoclinic fabrics. This is due to a mylonitization gradient, as revealed by the increased modal percentage of matrix and decreased percentage of porphyroclasts.

Conjugate shearing domain (CSD) A symmetrical pattern of closely spaced millimetre- to centimetre-scale dextral (350°) and sinistral (110°) sets of subvertical shear bands occurs within a 40 m wide outcrop of coarse-grained gneisses (Fig. 3). Pervasive strike-slip sets of C-S planes are associated with a subvertical intersection lineation (Fig. 4). A superposed gently plunging weak stretching lineation (5°-10°/330°) is observed on discrete fine-grained centimetrewide shear bands (310°, 90°) with a sinistral sense of movement. Fine-grained gneisses are also present as metre-scale wide bands parallel to the

Fig. 4. Meso- and microscopic structures observed in peralkaline gneisses from the Arronches river (for location see Figs 2B and 3E) included in a conjugate shearing domain (CSD). Schematic block diagram showing the relationships between pervasive conjugate sets of shear planes (orthorhombic fabric): (A) dextral and (B) sinistral. (C) Surface vertical and normal to main foliation-trend with no evidence for tectonic movement. (D) and (E) A vertical lineation is observed in the foliation and in both sets of conjugate shear bands. There is no optical evidence that this lineation is a vertical stretching lineation. This linear fabric is interpreted as an intersection lineation between the main foliation and both sets of conjugate dextral and sinistral shear planes. Legend for the schematic stereogram: single black lines, average strike of dextral and sinistral shear planes; black dot, intersection lineation.

ORTHORHOMBIC FABRICS IN SHEARED GNEISSES

main foliation. Mesoscale D2 folds with axial planes parallel to the foliation show fold axes subparallel to the stretching lineation. Discrete mesoscale sinistral shear is also present in the surrounding garnet-rich mica schists with penetrative C'-type shear band cleavage. The mylonitized gneisses have a well-developed foliation where several shear sense indicators are present, showing a sinistral and dextral sense of shear (C-S fabric, a-type porphyroclasts) in sections normal to the foliation and parallel to the subhorizontal stretching lineation.

Intermediate sinistral domain (ISD) This pattern is associated with coarse to finegrained gneisses with a well-developed foliation (330°, 80°/240°) (Fig. 5). It is denned by millimetre-scale compositional variations with stretched feldspars and dark amphiboles, and a gently plunging stretching lineation (10°-15°/330°). The orientation of the foliation varies across a few metres due to rotations imposed by the development of close spaced, metre-scale sinistral shear bands (110°/90°) (Fig. 5B). Dextral shear bands are present forming conjugate planes (350°/90°), locally creating a symmetrical orthorhombic fabric (Fig. 5A).

Sinistral domain (SD) The SD gneisses are mainly fine-grained rocks with small 'augen' of amphibole and feldspar (Fig. 6A), along 3 cm to 20-50 cm narrow bands orientated parallel to the well-developed 280°trending foliation dipping 80° to the SW (Fig. 6B). A penetrative S-L fabric is present, defined by millimetre-scale compositional variations, as well as an alignment of mafic mineral (amphiboles + biotite) aggregates and strongly flattened and stretched minerals, with a gently plunging stretching lineation 10°-15° to 280° parallel to the fold axis. The S-L fabric is defined by strain shadows around porphyroclasts and by elongated mafic mineral aggregates lying in the plane of the foliation. Sheared asymmetric porphyroclasts and millimetre- to centimetre-scale C'-type shear band cleavage indicate a sinistral sense of movement.

The role of grain-size reduction and deformation mechanisms in the fabric formation Dynamic recrystallization significantly influences grain size distribution and affects fabric development during progressive deformation

221

(Tullis et al. 1982; Tullis & Yund 1985; Drury & Urai 1990). This is observed within the Arronches gneisses, where the development of discrete shear planes with a very strong planar fabric is characterized by small new recrystallized crystals that deflect the 330°-trending foliation on the CSD and ISD mylonites. These SD ultramylonites present a well-developed S-L fabric and microscale shear sense indicators with sinistral movement. Weak lineations are present where partial recrystallization along coarser grained old relic boundaries induces the growth of new grains. This observation is typical of the CSD mylonites where it is possible to identify an orthorhombic fabric as a result of the development of sets of penetrative conjugate shear bands. In these coarser grained peralkaline gneisses, dynamic recrystallization of a new mafic mineral phase with biotite and green amphibole along the S-C planes from both sets of conjugate shear bands also occurs. This mineral aggregate growth is associated with the development of penetrative subvertical conjugate sets of shear planes, whose superimposition defines a vertical intersection lineation. Fig. 7 illustrates fabrics and corresponding petrographic descriptions of the different investigated shearing domains. Textural analysis of the mylonitic gneiss allows the identification of different types of microstructures that demonstrate that the process of dynamic recrystallization, by grain boundary migration-recrystallization and subgrain rotation recrystallization, is linked to the development of strain softening. Serrated grain boundaries and formation of lobes into the adjacent grains in K-feldspar and quartz, which replaces hardened grains (relics of deformed old grains with deformation lamellae, undulose extinction and subgrains, as a response to recovery, as well as growth of twins in plagioclase by deformation twinning) by new undeformed grains, indicates grain boundary migrationrecrystallization. Polycrystalline feldspar and quartz aggregates, with transitions between grains surrounded by high angle boundaries and subgrains, constitute evidence for sub-grain rotation recrystallization.

Discussion and conclusions Dynamic recrystallization and development of fabric and texture Figure 8 illustrates the proposed development of the Arronches gneiss different textures and

Fig. 5. Detailed structural map illustrating the intermediate sinistral domain (ISD) observed in peralkaline gneisses from an outcrop located SE of Arronches (Fialha) (for location see Fig. 2B). Mesoscale structures from the subhorizontal sections are gently parallel to the stretching lineation and perpendicular to the foliation. (A) Orthorhombic fabric with a conjugate set of shear planes; (B) Monoclinic fabric of a sinistral shear band. Legend for the schematic stereogram: single black line, average strike of dextral shear bands; white field, average strike of sinistral shear planes.

ORTHORHOMBIC FABRICS IN SHEARED GNEISSES

223

Fig. 6. Meso- and microscopic structures from ultramylonites and peralkaline gneisses in a sinistral domain (SD). (A) Shear sense indicators of sinistral movement in amphiboles and feldspar. (B) Centimetre-scale interlayered bands of peralkaline gneisses, mica schists and amphibolites.

structures. In the Arronches peralkaline gneisses there is a strong relationship between the stage of mylonitization and the development of distinct fabrics. The data presented in this study allow the identification of an incremental gradient of mylonitization, related to the development of monoclinic fabrics from earlier local orthorhombic fabrics. The strong dynamic recrystallization of originally large porphyroclasts of K-feldspar, plagioclase, and alkali-amphiboles, which significantly exceed the size of the new-formed grains, is the reason for the development of fine-grained elongate aggregates composed of K-feldspar, quartz, plagioclase, mica, garnet and amphibole. The orientated growth and arrangement of these minerals was responsible for the well-developed S-L fabric and associated monoclinic fabric in the SD ultramylonites. Progressive stages of fabric development with increasing mylonitization are controlled by significant dynamic recrystallization. As more of the rock forms a fine-grained texture characterized by the growth of new grains and decreasing volume of porphyroclasts, there is a greater

tendency to observe the monoclinic symmetry (SD), which becomes more representative than the orthorhombic fabric (CSD). As a result of different stages of strain-softening, these peralkaline gneisses with similar original mineral and chemical composition have developed distinct types of fabric and textures. The monoclinic symmetry of the SD ultramylonites and other included fabric elements reflect high strain values of non-coaxial flow, associated with strong strain-softening (finegrained texture due to strong dynamic recrystallization). In contrast, the CSD mylonites, which present lower strain values, are relatively less deformed and have orthorhombic symmetry and coarse-grained texture, indicating the earlier stages of dynamic recrystallization.

Relative timing of formation of orthorhombic and monoclinic fabrics Observed overlap relations between distinct shearing domains indicates a local progressive change from orthorhombic to monoclinic fabric.

224

M. F. PEREIRA & J. B. SILVA

Fig. 7. Schematic representation of texture development for the Arronches peralkaline, with increasing dynamic recrystallization gneisses (given in percentage). Microfabric characterization, symmetry and corresponding shearing domain, are given. Legend for the schematic stereogram: white, average strike of sinistral shear planes.

This interpretation is based on the superimposition of opposite strain systems present in the CSD and ISD orthorhombic fabrics, where dextral and sinistral shear planes cross-cut one another without a consistent order (Figs 4 & 5), as well as a superimposition of the SD monoclinic fabric over the CSD and ISD orthorhombic fabrics (Figs 3, 4 & 5). Therefore, orthorhombic fabrics formed within the CSD and the ISD must be coeval with each other, and must be related to the non-coaxial bulk deformation responsible for the development of the SD monoclinic fabric. This study indicates that it is possible to find mechanical conditions to produce local

orthorhombic fabrics in a simple-shear dominated sinistral transpression zone, with a welldeveloped and widespread monoclinic fabric. In the Arronches gneisses, a local strain path exists which is in apparent contradiction to the bulk deformation regime of the CCSZ, where there is no evidence for a significant pure-shear component along more than 600 km.

The generation of sets of conjugate shear bands (orthorhombic fabric) Considering the ISD mylonites case (Fig. 5) we propose an interpretative geometric and

ORTHORHOMBIC FABRICS IN SHEARED GNEISSES

225

Fig. 8. Diagram depicting the effect of progressive deformation on the development of distinct shearing domains with respect to increasing strain, increased modal percentage of matrix (mylonitization) and evolution from orthorhombic to monoclinic symmetry. Legend for the schematic stereograms: dark grey field, average strike of dextral shear planes; white field, average strike of sinistral shear planes; single black line, average strike of dextral and sinistral shear planes; black dot, intersection lineation.

kinematic model to explain the generation of orthorhombic fabrics as the result of progressive deformation (Fig. 9). As with other examples described in the literature (e.g. Gapais & Bale 1990; Goodwin & Williams 1996; Carreras 2001), our observations show that the existence of conjugate sets of shear bands may be developed locally within simple-shear dominated zones of intense ductile deformation. Gapais & Bale (1990) described local symmetric patterns (dextral and sinistral shear bands) on mylonites passing to well-developed ultramylonites with monoclinic symmetry (discrete C'-type shear bands), in a major sinistral ductile deformation zone (St Cast shear zone, North Brittany). These local perturbations in the deformation regime occur close to a contact between a leucogranite and paragneisses, and relate to a single progressive ductile deformation with local strain conditions controlled by the emplacement of syntectonic

intrusive rocks. In our opinion, despite Gapais & Bale's (1990) detailed description of variation on orientations of C and S planes relative to the distance to the contact, their interpretation has neglected implications concerning foliation inflexions and local angular relations between rotated foliations and the local imposed compressive stress. These could constitute probable causes of mechanical instabilities responsible for the nucleation of local conjugate shear bands. Within a major dextral ductile shear zone (Marble Cove, Newfoundland), Goodwin & Williams (1996) have also recognized the existence of local orthorhombic fabrics (conjugate shear bands) within dominantly simple-shear domains where mylonites exhibit a well-developed monoclinic symmetry. These authors do not discuss variations of main foliation orientation or local angular relations between rotated foliations and the local imposed compressive stress. They interpret this orthorhombic fabric

226

M. F. PEREIRA & J. B. SILVA

Fig. 9. Schematic diagram illustrating a case of progressive deformation with development of structures associated with an intermediate sinistral domain (ISD). (A) The first structures to form were 80°-90°, subvertical C-type shear planes oblique to the foliation and subparallel to the remote CCSZ compressive stress (GI). Under continued deformation, the foliation gently curves (is apparently extended) into these local subvertical C-type shear planes, changing its orientation and becoming subjected to normal local compression °iLocai- (B) Local perturbations by deflection of the foliation provided conditions to initiate the formation of local conjugate sets of shear planes (dextral 350° and sinistral 90°) as a response to foliation-normal compression by the continued action of the remote compressive stress (o^).

as a consequence of a component of shortening at right angles to the main foliation, suggesting that the pure shear component was relatively strong with respect to simple shear during local development of such conjugate shear zones. The proposed mechanical behaviour of the Arronches gneisses with the formation of conjugate shear planes, suggests that the foliation was not passively sheared. Variation in foliation orientation relative to local compressive stress as a consequence of inflexions due to progressive deformation can induce formation of instabilities similar to those achieved by experiments upon analogue materials with anisotropies (e.g. Cosgrove 1989). This mechanical behaviour was noted by Carreras (2001), who described and discussed in detail a natural example of ductile shear zones in previously foliated rocks (Cap de Creus, Spain). Carreras (2001) shows that, depending on variations in the foliation orientation close to shear zones, mechanical conditions for development of local instabilities and generation of conjugate shear bands may occur.

The existence of orthorhombic symmetry is not an exclusive characteristic of shear zones associated with a significant pure-shear component, as it may occur locally within simpleshear dominated shear zones. The suggestions of two unknown referees, which helped to improve the manuscript, are gratefully acknowledged. Financial support was given by the Tectonic Group, Geophysical Centre of Evora budget for 2001-2002.

References BELL, T.H. 1981. Foliation development - the contribution, geometry and significance of progressive bulk inhomogeneous shortening. Tectonophysics, 75, 273-296 BERTHE, D., CHOUKROUNE, P. & JEGOUZO, P. 1979. Orthogneiss, mylonite and non-coaxial deformation of granites: the example of the South Armorican shear zone. Journal of Structural Geology, 1, 31-42. BURG, X, IGLESIAS, M., LAURENT, P., MATTE, PH. &

ORTHORHOMBIC FABRICS IN SHEARED GNEISSES RIBEIRO, A. 1981. Variscan intracontinental deformation: the Coimbra-Cordoba shear zone (SW Iberian Peninsula). Tectonophysics, 76, 161-177. CARRERAS, J., JULIVERT, M. & SANTANACH, P. 1980. Hercynian mylonite belts in the Eastern Pyrenees: an example of shear zones associated with late folding. Journal of Structural Geology, 2,5-9. CARRERAS, J. 2001. Zooming on Northern Cap de Creus shear zones. Journal of Structural Geology, 23,1457-1486. COSGROVE, J.W. 1989. Cleavage, folding and the finite strain ellipsoid. Proceedings of the Geological Association, 100, 461-479. DRURY, M.R. & URAI, J.L. 1990. Deformation-related recrystallization processes. Tectonophysics, 172, 235-253 EGUILUZ, L., GIL-IBARGUCHI, J.L, ABALOS, B. & APRAIZ, A. 2000. Superposed Hercynian and Cadomian orogenic cycles in the Ossa-Morena zone and related areas in the Iberian Massif. Geological Society America Bulletin, 112,1398-1413. GAPAIS, D. & WHITE, S.H. 1982. Ductile shear bands in a naturally deformed quartzite. Textures and Microstructures, 5,1-17. GAPAIS, D. & BALE, P. 1990. Shear zone pattern and granite emplacement within a Cadomian sinistral wrench zone at St Cast, N. Brittany. In: D'LEMOS, R.S., STRACHAN, R.A. & TOPLEY, C.G. (eds) The Cadomian Orogeny. Geological Society, London, Special Publications, 51,169-179. GARCIA CASQUERO, BOELRIJK, N. A.I.M., CHACON, J. & J.L. & PRIEM, H.N.A. 1985. Rb-Sr evidence for the presence of Ordovician granites in the deformed basement of the Badajoz-Cordoba belt, SW Spain. Geologisches Rundschau, 74, 379-384. GONCALVES, F. 1971. Subsidies para o conhecimento geologico do Nordeste Alentejano. PhD thesis. Memorias Serviqos Geologicos Portugal, 69,1-62. GOODWIN, L.B. & WILLIAMS, PR 1996. Deformation path partitioning within a transpressive shear zone. Marble Cove, Newfoundland. Journal of Structural Geology, 18, 975-990. HANMER, S. 1988. Great Slave Lake Shear Zone, Canadian Shield: reconstructed vertical profile of a crustal-scale fault zone. Tectonophysics, 149, 245-264. HANMER, S. & PASSCHIER, C.W. 1991. Shear sense indicators: a review. Geological Survey Canada, 90, 1-71. HUDLESTON, P.E. 1999. Strain compatibility and shear zones: is there a problem? Journal of Structural Geology, 21, 923-932. KNIPE, R.J. 1989. Deformation mechanisms - recognition from natural tectonites. Journal of Structural Geology, 11,127-146. LANCELOT, J.R. & ALLEGRET, A. 1982. Radiochronologie U/Pb de 1'orthogneiss alcalin de Pedroso (Alto Alentejo, Portugal) et evolution ante-Hercynienne de 1'Europe Occidentale. Neues Jahrbuch fiir Mineralogie Monatschefning, 9, 385-394. MANCKTELOW, N. 2002. Finite-element modelling of shear zone development in viscoelastic materials and its implications for localisation of partial

227

melting. Journal of Structural Geology, 24, 1045-1053. MATTE, P. 1991. Accrectionary history and crustal evolution of the Variscan belt of Europe. Tectonophysics, 196, 309-337. MEANS, W.D. 1994. Rotational quantities in homogeneous flow. Journal of Structural Geology, 16, 437-445. ORDONEZ-CASADO, B. 1998. Geochronological studies of the Pre-Mesozoic basement of the Iberian Massif: the Ossa Morena Zone and Allochthonous Complexes within the Central Iberian Zone. PhD thesis, ETH (12.940), Zurich. PASSCHIER, C.W. & TROUW, R.A.J. 1996. Microtectonics. Springer. PEREIRA, M.F. 1999. Caracteriza^ao da estrutura dos domfnios setentrionais da Zona de Ossa-Morena e seu limite com a Zona Centro-Iberica, no Nordeste Alentejano. PhD thesis (unpublished). Universidade de Evora, Portugal. PEREIRA, M.F. & SILVA, J.B. 2001. A new model for the Hercynian orogen of Gondwanan France and Iberia: discussion. Journal of Structural Geology, 23, 835-838. PEREIRA, M.F. & SILVA, J.B. 2002. NeoproterozoicPaleozoic tectonic evolution of the CoimbraCordoba shear zone and related areas of the Ossa-Morena and central-Iberian zones (Northeast Alentejo, Portugal). Comunicaqoes do Institute Geologico e Mineiro, 89, 112-124, Lisboa, Portugal. PLATT, J.P. & VISSERS, R.L.M. 1980. Extensional structures in anisotropic rocks. Journal of Structural Geology, 2, 397-410. QUESADA, C. & DALLMEYER, R.D. 1994. Tectonothermal evolution of the Badajoz-Cordoba shear zone (SW Iberia): characteristics and 40Ar/39Ar mineral age constraints. Tectonophysics, 231,195-213. SIMANCAS, J.F., MARTINEZ POYATOS, D., EXPOSITO, L, AZOR, A. & GONZALEZ LODEIRO, F. 2001. The structure of a major suture zone in the SW Iberian Massif: the Ossa-Morena/Central Iberian contact. Tectonophysics, 332, 295-308. SIMPSON, C. & SCHMID, S.M. 1983. An evaluation of criteria to determine the sense of movement in sheared rocks. Bulletin of the Geological Society of America, 94, 1281-1288. TULLIS, IT., SNOKE, A.W & TODD, V.R. 1982. Significance of petrogenesis of mylonitic rocks. Geology, 10, 227-230. TULLIS, J. & YUND, R.A. 1985. Dynamic recrystallization of feldspar: a mechanism for ductile shear zone formation. Geology, 13, 238-241. Twiss, R.J. & MOORES, E.M. 1992. Structural Geology. Freeman. WENK, H.R., TAKESHITA, T, BECHLER, E., ERSKINE, B.G. & MATTHIES, S. 1987. Pure shear and simple shear calcite textures. Comparison of experimental, theoretical and natural data. Journal of Structural Geology, 9, 731-746. WHITE, S.H., BURROWS, S.E., CARRERAS, I, SHAW, N.D. & HUMPHREYS, F.J. 1980. On mylonites in ductile shear zones. Journal of Structural Geology, 2, 175-187.

This page intentionally left blank

Deformation in a complex crustal-scale shear zone: Errabiddy Shear Zone, Western Australia S. A. OCCHIPINTI & S. M. REDDY Tectonics Special Research Centre, Department of Applied Geology, Curtin University, PO Box U1987, Perth 6845, Australia (e-mail: [email protected]; [email protected]) Abstract: Detailed mapping of four areas representing different geological units with varying formation histories within the crustal-scale Errabiddy Shear Zone shows an apparently simple temporal progression from foliation and mineral lineation development to folding and then to brittle deformation across the shear zone. However, in detail the structural evolution of the shear zone shows considerable complexity. The dominant foliation throughout the shear zone was formed in the upper greenschist to amphibolite facies during the 2000-1960 Ma Glenburgh Orogeny, which involved the accretion of the Archaean to Palaeoproterozoic Glenburgh Terrane onto the Archaean Yilgarn Craton and the subsequent formation of the Errabiddy Shear Zone. Orthorhombic kinematic indicators formed during the Glenburgh Orogeny as did the widespread mineral lineation. These fabrics were overprinted by a greenschist facies deformation and metamorphic event during the 1830-1780 Ma Capricorn Orogeny. During the Capricorn Orogeny mineral lineation development was rare, and mostly took place in high-Capricorn strain zones in areas where a pre-existing Glenburgh-aged mineral lineation was present. Such mineral lineations trend parallel to Capricorn-aged fold hinges. Regardless of the presence or absence of Capricornaged mineral lineations, dextral strike-slip kinematics and simple shear indicated by delta and sigma porphyroclasts, and displacement along detachment faults, are prevalent close to discrete shear zone boundaries, within the Errabiddy Shear Zone. However, between shear zone boundaries flattening and coaxial strain dominated during the Capricorn Orogeny. This difference in Capricorn Orogeny kinematics throughout the shear zone is caused by strain partitioning - although progressive deformation throughout the shear zone with dextral strike-slip faults overprinting older structures formed by pure shear also took place. These results suggest that analyses of small parts of shear zones may not give the complete history of an evolving transpressional shear zone because of the presence of strain partitioning and strain localization over time.

Crustal-scale shear zones are fundamental discontinuities that often are the sites of continental accretion, collision, extension and intraplate deformation. Such zones may accommodate deformation via simultaneous components of pure and simple shear (i.e. general shear). The distribution of strain in shear zones may vary spatially (strain localization) and also temporally and pure and simple shear deformation may be partitioned within the shear zone (Tikoff & Teyssier 1994; Jones & Tanner 1995; Dewey et al 1998; Lin & Williams 1998; Reddy et al. 1999, 2003). Once formed, shear zones are often zones of weakness (Holdsworth et al. 20010,/?) and may further deform by reactivation leading to complex polyphase deformation histories (Scrimgeour & Raith 2001; Hand & Buick 2001). Recognizing the range of complexity in shear zone evolution is difficult, as it requires quanti-

tative characterization of the spatial and ternporal distribution of deformation that can only be resolved by applying careful strain and kinematic analysis together with careful dating of geological structures. Due to the spatial and temporal localization of deformation, the formation of natural shear zones can only be evaluated when as much of the shear zone as possible is studied in detail. However, studying an entire crustal-scale shear zone is often difficult due to limitations in outcrop and accessibility and most detailed studies of shear zones focus on small parts or specific sections, assuming the results can characterize the development of the entire shear zone. In this paper we analyse four separate wellexposed areas within a crustal-scale shear zone (Errabiddy Shear Zone). The Errabiddy Shear Zone was chosen for this study for the following reasons:

From: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 229-248. 0305-8719/$15.00 © The Geological Society of London 2004.

230

S. A. OCCHIPINTI & S. M. REDDY

(1) it is a major crustal-scale shear zone, and site of accretion of an Archaean to Palaeoproterozoic terrane onto a stable Archaean Craton (Occhipinti et al 2004); (2) outcrops of different geological units are easily accessible throughout the shear zone; (3) preliminary regional-scale work suggested that metamorphic events can be correlated throughout the shear zone and felsic magmatic events can be correlated with adjacent terranes (Occhipinti et al. 1998, 2004); (4) a detailed study to the east of the Errabiddy Shear Zone suggested shear zone development during dextral transpression and possible subsequent reactivation (Reddy & Occhipinti 2004). Each of the four areas chosen for detailed mapping are situated in the ENE trending part of the shear zone. Each section consists of different lithological units of diverse geological age and origin. We examine in detail the geometries, kinematics and timing of structures within the four study areas by field mapping, and determining relationships between deformation fabrics, metamorphic mineral assemblages, and crosscutting igneous rocks. The study involved detailed analysis of relative age relationships and construction of younging tables and deformation networks in each of the following areas (Potts & Reddy 2000). Different structural nomenclature for each of the four areas has been used throughout the text for ease of comparison between the areas. This nomenclature is as follows: n (for gneiss), for the Archaean basement-granitic gneiss area; g (for granite) and b (for basement) for granite and felsic gneiss, respectively in the felsic gneiss (basement) and c. 1958 Ma granite of the Erong Shear area; s (for metasedimentary) for psammitic rocks in the psammitic gneiss area; and p (for pelite) for the migmatized pelite area. These data are used to characterize how deformation in a regional-scale shear zone developed, and to compare the results to previously completed regional-scale studies, which were accompanied by U-Pb SHRIMP dating of granites and metasedimentary rocks (Nelson 1997, 1998, 1999, 2000, 2001). By doing so we show that shear zones are the foci of the complicated formation of structures in response to both single and multiple deformation events. However, the development of structures and kinematics within a shear zone is strongly influenced by strain localization, and therefore the results of small-scale detailed studies may be scale dependent and may not be representative of the kinematics or formation of the entire shear zone.

The Errabiddy Shear Zone The Errabiddy Shear Zone is a curved shear zone that is more than 200 km long, up to 20 km wide and forms the northwestern margin of the Archaean Yilgarn Craton. The shear zone strikes NE in its western part, ENE in its central part, and NNE in its eastern part (Fig. 1), and is exposed in the Capricorn Orogen - a Palaeoproterozoic collisional belt between the Archaean Pilbara and Yilgarn Cratons in Western Australia (Tyler & Thorne 1990; Occhipinti et al 1998,2001,2004; Sheppard et al. 2002, 2004). The Errabiddy Shear Zone formed due to the collision and accretion of the Glenburgh Terrane onto the Archaean Yilgarn Craton during the 2000-1960 Ma Glenburgh Orogeny (Occhipinti et al. 2001; Fig. 2). It comprises several fault-bounded lenses of rock units that differ in type and age, and are heterogeneously deformed, which have been metamorphosed at medium- to high-grade. The Errabiddy Shear Zone was subsequently reactivated in the greenschist facies at 1830-1780 Ma during the Capricorn Orogeny (Sheppard & Occhipinti 2000; Occhipinti & Sheppard, 2001; Occhipinti et al. 2001; Sheppard et al. 2004), which may record the inter-cratonic collision of the Archaean Pilbara (to the north) and Yilgarn Craton Glenburgh Terrane (inset, Fig. 1; Tyler & Thorne 1990; Occhipinti et al 1998), or intracratonic shortening between the already joined, and relatively stable Archaean Pilbara and Yilgarn Cratons (Occhipinti et al 2004; Sheppard et al 2004). The Errabiddy Shear Zone comprises different lithological units of diverse geological age. These units include Archaean granitic gneiss basement, Palaeoproterozoic metasedimentary rocks, and Palaeoproterozoic granite, which formed in different geological environments at dissimilar crustal levels but have since been juxtaposed in the shear zone (Fig. 2). All of these rocks have been variably intruded by leucocratic pegmatites.

Archaean granitic gneiss (basement) A well foliated to banded granitic gneiss (the Warrigal Gneiss) dominantly comprising interleaved mesocratic and leucocratic granitic gneiss, outcrops in fault-bounded lenses throughout the Errabiddy Shear Zone. Locally this granitic gneiss is pegmatite banded. The granite protoliths to the gneiss comprise medium-grained, equigranular monzogranite, and is locally porphyritic. A part of one of these

CRUSTAL-SCALE SHEAR ZONE AT ERRABIDDY

231

Fig. 1. Simplified geological map (modified after Occhipinti et al 2004) of the Errabiddy Shear Zone and its components showing the location of the four mapped areas in this study (marked by star locations on map). Inset shows the Archaean to Palaeoproterozoic Glenburgh Terrane (to the north), the Archaean Yilgarn Craton (to the south), and the Archaean to Palaeoproterozoic Yarlarweelor Gneiss Complex (to the east).

fault-bounded lenses located approximately 7.5 km NE of Erong Homestead at MGA 470691E 7178872N was studied (Fig. 3). The granitic gneiss consists of quartz, feldspar, and mica (variable amounts of biotite and muscovite), and locally contains garnet. SHRIMP U-Pb zircon ages on four individual granitic components of the Warrigal Gneiss from throughout the Errabiddy Shear Zone (Nelson 2000, 2001; Occhipinti et al. 2001), indicating that its formation was complex and not confined to one intrusive event. Gneissic banding within the granitic gneiss is parallel to a well-developed foliation (locally mylonitic) and is broadly parallel to contacts with boundaries of supracrustal rocks. These include amphibolite schist and gneiss, calc-silicate gneiss, quartzite, and pegmatite. Amphibolite lenses within the granitic gneiss consist of both medium-grained, even-textured gneissic amphibolite (comprising plagioclase, hornblende, actinolite-tremolite) and mafic schist (comprising varying amounts of actinolitetremolite and chlorite). The calc-silicate gneiss

comprises varying amounts of quartz, clinopyroxene, actinolite-tremolite, calcite, and feldspar. The quartzite is interleaved on a 0.3-1 m-scale with the mylonitized granitic gneiss, and both form a narrow (up to 20 m wide) mylonite zone. Locally, amphibolite gneiss clearly cuts across layering in the granitic gneiss, indicating that the amphibolites were originally dykes. However, contacts between the amphibolite, calc-silicate and granitic gneiss are commonly either not exposed or appear to be tectonic. Massive pegmatite dykes cut the granitic gneiss, calc-silicate gneiss, amphibolite, and mylonite zone. The pegmatite edges are often broadly parallel to the well-developed foliation in the gneisses.

Deformation in Archaean granitic gneiss (basement) The Petter Bore area of the Errabiddy Shear Zone has been polydeformed, metamorphosed, and has undergone several periods of granite

232

S. A. OCCHIPINTI & S. M. REDDY

Fig. 2. Schematic diagram modified after Occhipinti et al (2004) illustrating the possible tectonic evolution of the southern Capricorn Orogen, and the origin and changing position of the mapped units in this study. These units were metamorphosed and deformed during the Glenburgh Orogeny, when the Archaean to Palaeoproterozoic Glenburgh Terrane was accreted onto the Archaean Yilgarn Craton to form the Errabiddy Shear Zone. The bottom inset of the Capricorn Orogen shows the possible effects of the c. 1800 Ma Capricorn Orogeny in the region, which is largely reflected by the current-day geometry of the southern Capricorn Orogen.

intrusion (Fig. 3). The earliest recorded structure is a well-developed foliation in the granitic gneiss (Sln). This was locally observed and is generally overprinted by D2n to form a composite Sin/S2n fabric (Si/2n, which is the most pervasive foliation in the area) (Fig. 5a). However, in low strain zones Sln can be observed as having developed subparallel to the axial surface of tight to isoclinal folds deforming thin (less than 3 cm wide) pegmatite dykes. The S1/2n foliation (Figs 3 & 5a) strikes east-west, subparallel to contacts between the different rock types (Fig. 3), and is steeply dipping (Fig. 5a). The S1/2n foliation is parallel to F2n fold axial surfaces that are developed throughout the area (Fig. 5a). The F2n folds are tight to isoclinal and have fold hinges parallel to L12n intersection lineations, that plunge moderately to steeply towards the west to NW (Fig. 5a). Locally, F2n folds and associated crenulation cleavages overprint Fjn isoclinal folds. Mineral lineations, defined by quartz, mica and amphibole are present on the S1/2n planar surfaces (Figs 5a & 7a), plunging shallowly to

steeply towards the west or east. Locally, quartz mineral aggregate lineations plunge steeply towards the NE, and shallowly to steeply towards the west or WNW, indicating that they have been re-orientated, or there are two groups of quartz lineations (Fig. 5a). Minor, gentle 1 m-scale folds (F3n) deform the S1/2n foliation and F2n folds (Fig. 6a). The axial planes of these folds are north to northwesterly striking and steeply dipping with subvertically plunging hinges. Eastnortheasterly striking vertically dipping detachments and quartz veins cut the S1/2n foliation surface, but are locally subparallel to it (Fig. 6a). Their relationship to the F3n folds is unknown. Other brittle and brittle-ductile features include dextral strike-slip, sinistral strike-slip and reverse faults. Kinematic indicators included brittle-ductile displacement of the S1/2n foliation, ductile displacement of S1/2n foliation, and S-C fabrics. Dextral strike-slip faults with no observed vertical slip trend either in an eastnortheasterly, southwesterly or northwesterly direction, whereas a measured reverse

Fig. 3. Geological map of Archaean basement granitic gneiss. Coordinates are specified by Universal Transverce Mercator (UTM) coordinates using the Map Grid of Australia (MGA) and Geocentric datum of Australia 1994 (GDA94).

234

S. A. OCCHIPINTI & S. M. REDDY

Fig. 4. Equal area stereonets of early structural fabrics (Regional D t ) observed in each of the mapped areas.

and sinistral strike-slip fault is easterly trending. The kinematics of several late fractures could not be ascertained. These are steeply dipping and northerly or westerly striking.

Felsic gneiss (basement) and c. 1958 Ma granite of the Erong Shear Area The Erong Shear Area is located approximately 7.0 km east of Erong Homestead at MGA 474078E 7172884N (Fig. 1). The main rock types

at this locality are weakly to well-foliated granite of the Bertibubba Supersuite (Occhipinti et al. 2001), foliated to mylonitized granite, psammite and quartzite. The granite of the Bertibubba Supersuite was dated by U-Pb SHRIMP zircon dating as c. 1958 Ma, approximately 800 m north of the map area (Occhipinti et al. 2001). The area can be divided into two structural domains (1 and 2). Domain 1 consists of weakly to well-foliated granite, which is monzogranitic in composition and ranges from porphyritic, to medium equigranular granite of 0.5-1 cm

CRUSTAL-SCALE SHEAR ZONE AT ERRABIDDY

Fig. 5. Equal area stereonets of main structures (Regional D2) observed in each of the mapped areas.

235

236

S. A. OCCHIPINTI & S. M. REDDY

Fig. 6. Equal area stereonets of brittle and brittle-ductile structures and F3 folds observed in the mapped areas.

CRUSTAL-SCALE SHEAR ZONE AT ERRABIDDY

237

grainsize, and fine-grained granite that forms 1-10 m-scale thick sheets intruded by minor pegmatite, and fine-grained granite. These rocks mainly consist of quartz, feldspar, and mica (biotite and/or muscovite). Domain 2 contains well foliated to mylonitized granite, metasedimentary rocks and quartz mylonite (henceforth Felsic gneiss), that generally comprises variable amounts of quartz, feldspar, and mica (biotite and/or muscovite). The contact between Domains 1 and 2 is a discrete high-strain zone.

Easterly to eastnortheasterly striking quartz veins cut the Slg foliation in Domain 1 (Fig. 8). These are typically less than 1 m wide and discontinuous. The Slg foliation in Domain 1 is locally folded about gentle, steeply plunging, upright, southeasterly striking folds. The relationship of these folds to ductile shear structures, faults, and quartz veins in the area are unknown.

Deformation in Domain 1

In Domain 2 the earliest recognizable structure is a foliation in the metasedimentary rock component of felsic gneiss, which may represent bedding or a tectonic foliation. Locally, shallow to moderately plunging isoclinal, small-scale Flb folds deform this early fabric (Figs 4a & 7e). The main regional foliation in Domain 2 (Slb) is schistose to gneissic, and subparallel to observed Flb fold axial surfaces (Fig. 4a). This foliation is variably defined by strongly recrystallized granoblastic quartz-feldspar domains and elongate aggregates of biotite or muscovite. Mineral lineations associated with S lb are defined by elongate quartz and quartz mineral aggregates, and elongate muscovite or biotite aggregates. These mineral lineations are well developed in quartz mylonite and the felsic gneiss and have variable orientations (Figs 4a & 7f) at least in part due to later folding (F2b). F2b folds are tight to isoclinal upright, with eastnortheasterly striking axial planes that typically dip steeply to the SE (Figs 5c & 7e, f). A variation in attitude of F2b fold axial surface orientations to northwesterly or northeasterly striking is due to the curviplanar nature of the folding indicated by hinges with strongly variable plunges and opposing plunge directions on folds with only slightly variable axial surface orientations. Thus the spread of F2b fold hinges is more likely to be due to F2b folds being noncylindrical, and strongly curvilinear. In areas of high D2b strain a well-developed crenulation to crenulation cleavage is present. In these areas feldspar is typically replaced by sericite and biotite may be partially replaced by chlorite. These crenulations dip towards the SSE and are similar to the average fold-axial surfaces of F2b folds in the region. Crenulations and small-scale very tight F2b folds are most common in Domain 2 along and adjacent to the contact with Domain 1, possibly indicating that this contact is a zone of later localized deformation. Northeasterly striking, steeply dipping quartz veins cut D2b structures in Domain 2. These veins are usually only a few metres long and less than 1 m wide. However, a few larger ones up to

The earliest recorded structure in Domain 1 is a pervasive Sig foliation that strikes ENE and dips moderately to steeply towards the SE or south (Figs 5b & 8). This foliation is defined by the alignment of variable amounts of deformed feldspar (variably sericitized), quartz, biotite, muscovite, and minor epidote, although epidote is often randomly orientated mainly replacing feldspar. Variably recrystallized aggregates of quartz make up domains that follow Slg. Mineral lineations on the Slg surface are rare (Fig. 5b). However, where present they are defined by aggregates of quartz or muscovite, and elongate feldspar porphyroclasts, now partly pseudomorphed by sericite. These mineral lineations show two preferred orientations plunging moderately to steeply to the south, or shallowly plunging to the east or west (Fig. 5b). Pegmatite and fine-grained leucogranite dykes within the medium-grained granite commonly strike subparallel to the Slg foliation. A few dykes that are oblique to Slg are folded into close to tight folds with their fold-axial surface parallel to Slg, although they are usually unfoliated. Localized shear bands, which overprint S lg , but are subparallel to Sig may have developed soon after the formation of this foliation. These shear bands lie approximately parallel to brittle-ductile reverse and normal shears (Figs 6b & 7c). These may have developed synchronously. Delta and sigma feldspar porphyroclasts (Figs 6b & 7b) associated with the brittle-ductile shears, indicate sinistral and dextral strike-slip shearing. Locally sinistral or dextral strike-slip movements have taken place with reverse faulting, with displacement along the faults being only 1-4 cm. Reverse faults are the most abundant fault type observed in Domain 1 and strike in a southwesterly to southerly or southeasterly direction (Fig. 6b). Only four normal faults were recognized and these have similar strikes and dips as the northeasterly striking reverse faults (Fig. 6b).

Deformation in Domain 2

238

S. A. OCCHIPINTI & S. M. REDDY

Fig. 7. (a) A strong quartz aggregate mineral lineation on the S1/2n foliation surface of a quartz mylonite in the Archaean granitic gneiss (basement) area, (b) A pegmatite dyke (?Capricorn-aged) cutting granitic gneiss and amphibolite, but trending subparallel to the S1/2n foliation in the Archaean granitic gneiss (basement) area, (c) Reverse fault offsetting aplitic granite vein in foliated porphyritic granite in Domain 1 of the Granitic gneiss c. 1958 Ma granite area. The S lg foliation is subparallel to the aplite vein, (d) Dextral strike-slip kinematics around deformed feldspar phenocrysts in Domain 1 of the Felsic gneiss - c. 1958 Ma granite area, (e) Flb fold refolded by upright steeply plunging F2b folds in quartzite in Domain 2 of the Felsic gneiss - c. 1958 Ma granite area, (f) Folded quartz aggregate mineral lineation around upright shallow plunging F2b fold in Domain 2 of the Felsic gneiss - c. 1958 Ma granite area.

CRUSTAL-SCALE SHEAR ZONE AT ERRABIDDY

239

Fig. 8. Geological map of the Felsic gneiss-Bertibubba Supersuite in the Erong Shear Area. Coordinates are specified by Universal Transverce Mercator (UTM) coordinates using the Map Grid of Australia (MGA) and Geocentric datum of Australia 1994 (GDA94).

30 m long and a few metres wide are present in the area.

metamorphosed and migmatized rocks of the Camel Hills Metamorphics, and is locally deformed with them in the greenschist facies.

Palaeoproterozoic metasedimentary rocks - Camel Hills Metamorphics

Deformation in psammitic gneiss

In the area studied, the Camel Hills Metamorphics comprises the Quartpot Pelite (Fig. 1), an intensely deformed and poly-metamorphosed metasedimentary succession of psammitic to pelitic gneiss, intercalated with discontinuous bands or pips of minor quartzite and calc-silicate gneisses, and actinolite-tremolite schist. In the central and eastern parts of the Errabiddy Shear Zone these rocks are variably migmatized. The Quartpot Pelite has been dated by the SHRIMP U-Pb zircon method from three localities within the Errabiddy Shear Zone (Occhipinti et al 2001, 2004). These showed that the pelite was mostly derived from 2550-2025 Ma rock, and was deformed and metamorphosed at medium to high grade and locally migmatized at c. 1960 Ma (Occhipinti et al. 2001, 2004). Leucocratic pegmatite intrudes

An area of mostly psammitic gneiss of the Camel Hills Metamorphics in the central to western part of the Errabiddy Shear Zone was mapped and has been subdivided into two domains based on their location within a larger scale fold: the Limb zone and the Hinge zone (Fig. 9). The earliest recorded structure in the Hinge and Limb Zones is a pervasive quartz-feldsparbiotite Sis foliation that is easterly to northerly striking and dips moderately to steeply to the north or west (Fig. 4b). This foliation is subparallel to compositional layering in metasandstone and quartz-rich lithologies. It is also subparallel to the contacts between different metasedimentary lithologies in the area (Fig. 9). In quartzofeldspathic rocks and quartzite the foliation is gneissic, whereas in more biotite-rich lithologies a schistosity is developed. The Sls foliation is defined by biotite and feldspar (in psammite)

240

S. A. OCCHIPINTI & S. M. REDDY

Fig. 9. Geological map of Psammitic Gneiss (Quartpot Pelite). Coordinates are specified by Universal Transverce Mercator (UTM) coordinates using the Map Grid of Australia (MGA) and Geocentric datum of Australia 1994 (GDA94).

and actinolite-tremolite (in para-amphibolite) and is much more difficult to recognize in the Hinge Zone because it has been intensely folded by F2s upright tight to isoclinal easterly striking folds, and in many cases a well-developed S2s crenulation cleavage is present. A well-developed mineral elongation lineation present on the Sls surface is composed of aggregates of mica and/or quartz. In the western part of the limb zone (Fig. 4b) this mineral lineation plunges moderately to the north, whereas in the remainder of the limb zone it plunges steeply to the north, oblique to the fold axis of subsequent F2s folds (Fig. 5d). Elsewhere in the limb zone mineral lineations are steeply plunging trending parallel to mineral lineations observed in the hinge zone, which plunge steeply to the north or south and trend parallel to hinges of small-scale F2s folds (Figs 4b & 5d). Small-scale easterly striking F2s folds that deform the Sls foliation are close to tight, upright, inclined folds. These structures typically plunge steeply towards the west; although in a

few cases small-scale folds are northwesterly or southwesterly plunging, larger scale folds show south vergence. During D2s retrogression of higher-grade metamorphic assemblages that formed during Dls in the greenschist facies took place. For example biotite and feldspar may be partly replaced by sericite and actinolite-tremolite (in para-amphibolite schists) is replaced by finegrained chlorite. Retrogression is better developed in higher strain D2s zones indicating a probable greenschist facies metamorphic event accompanied D2s. Brittle and brittle-ductile faults and fractures are present in both the Limb and Hinge Zones (Figs 6c, 9 & lOa). Sinistral faults are generally steeply dipping and are NE to easterly striking in the limb zone, whereas in the hinge zone they are mostly westerly to southwesterly striking. Dextral brittle and brittle-ductile faults and fractures are steeply dipping and easterly or northeasterly striking in the limb zone, and easterly to southeasterly striking in the hinge zone. These

CRUSTAL-SCALE SHEAR ZONE AT ERRABIDDY

241

Fig. 10. Field photographs of Palaeoproterozoic metasedimentary rocks of the Camel Hills Metamorphics. (a) Detachment in the limb zone cutting the Sls foliation which trends subparallel to F2s fold hinges in the psammitic gneiss area, (b) F2s fold that has been cut by a detachment fault in the hinge zone in the psammitic gneiss area, (c) A deformed boudin within Si/2p foliation in migmatized pelitic schist and gneiss, (d) Diatexite melt and biotite-rich restite folded into F3p fold in migmatized pelitic schist and gneiss.

structures trend subparallel to Sls in the F2s fold limb, and in the western part of the limb zone and in the hinge zone they trend subparallel to the F2s fold axial surfaces, even though they commonly cut F2s folds further accommodating north-south shortening across the zone (Fig. lOb). Northerly or northwesterly striking, steeply plunging gentle to open folds locally influence the orientation of both the F2s fold axial surfaces, the Sls foliation, and the mineral lineations in the limb zone. Late northwesterly to northeasterly striking, steep plunging kink bands or open folds cross cut F2s folds and earlier structures; however, their relationship to the fractures and brittle-ductile faults in the area is unknown.

Deformation in migmatized pelitic schist and gneiss The Pannikan Bore area of the Errabiddy Shear Zone comprises mainly pelitic to psammitic

components of the Quartpot Pelite intercalated with minor calc-silicate and quartzite. The area has been polydeformed, metamorphosed and migmatized, and has undergone distinct periods of granite intrusion and quartz vein formation. The area has been heterogeneously deformed (Fig. 11). The earliest recorded structure is a locally developed foliation, Si, in quartzite and calcsilicate gneiss boudins within the more pelitic and psammitic components of the Quartpot Pelite. This foliation is folded into tight to isoclinal folds within the quartzite and calc-silicate 'pips' (Fig. lOc). The axial surfaces of these folds strike subparallel to the foliation in the surrounding pelitic and psammitic rocks, which wraps around the pips. We interpret the dominant fabric throughout Area 4 as a composite Si/S2 fabric, which is denoted S1/2p (Fig. 4c). The S1/2p foliation is well developed throughout the area, and is typically a steeply dipping east to southeasterly striking fabric that ranges

242

S. A. OCCHIPINTI & S. M. REDDY

Fig. 11. Geological map of migmatized pelitic schist and gneiss-Quartpot Pelite. Coordinates are specified by Universal Transverce Mercator (UTM) coordinates using the Map Grid of Australia (MGA) and Geocentric datum of Australia 1994 (GDA94).

from schistose to gneissic (Fig. 4c). The foliation mostly consists of alternating bands of quartz, feldspar, sillimanite, and biotite. Stromatic migmatite veins which trend subparallel to S1/2p accentuate the foliation, although in areas of high degrees of melting S1/2p is often cut by melt veins. In areas where the rocks were not as intensely deformed during subsequent deformation events S1/2p is northerly striking and plunges moderately to the west within open fold hinge zones (Fig. 11). Rare, small (< 10 cm long), thin quartzite lenses are deformed into F2p folds. Upright to steeply inclined easterly to northnortheasterly striking F3p folds and D3p crenulations are present throughout the area (Fig. 5e). In areas of more intense D3p strain these folds are mostly tight to isoclinal and moderately to steeply plunging structures and the development of an S2/3p foliation is common (Figs 5e & lOd). Contrasting this where D3p strain was less intense (Fig. 11) the F3p folds are open and mostly, but not exclusively moderately, westerly plunging (Fig. 5e). South of the mapped area, and elsewhere in the Quartpot Pelite, F3p folds are locally commonly shallowly plunging to the east or west. Where an S3p or S2/3p foliation is developed, sillimanite and biotite are replaced by white mica and chlorite, garnet is replaced by chlorite,

or chloritoid and quartz has recrystallized. Fine mats of sericite also pseudomorph feldspar and sillimanite. This indicates that retrogression of the earlier higher-grade assemblages is spatially related to D3p strain. Thin, coarse-grained granite and pegmatite dykes that trend subparallel to orthogonal to F3p fold axial surfaces are present in the southern part of the area. These rocks are massive to locally weakly foliated. They commonly cut F3p folds and locally, they appear to be boudinaged parallel to F3p fold axial surfaces, indicating that they might be a range of ages. The Si/2p foliation, and F3p folds and pegmatites are all boudinaged. Easterly striking detachments appear to cut the boudinaged foliations within the area; however, they may have developed in response to brittle-ductile processes operating at about the same time as the boudinage, accommodating strain in slightly higher strain domains (Fig. 6d). Northerly to northeasterly striking, steeply plunging F4p kink folds deform F3p folds and the S1/2p foliation; however their relationship to boudinage throughout the area is unknown (Fig. 6d). Possible late, northeasterly striking faults are present in the middle of Domain 1 (Fig. 6d), although it is generally difficult to tell the sense of movement around the faults.

CRUSTAL-SCALE SHEAR ZONE AT ERRABIDDY

Discussion Structural geometry within the Errabiddy Shear Zone Structural elements across the four map areas can be correlated based on their metamorphic grade and overprinting relationships (Figs 4-6 & 12). The oldest fabric that can be correlated across the Errabiddy Shear Zone is the main foliation in all the areas, which is denoted Si (Fig. 4), which is parallel to fold axial surfaces of rare FI folds in metasedimentary components of felsic gneiss in the region. This foliation developed in the amphibolite facies and is east to eastnortheasterly striking, and steeply dipping. However, in areas of weakest subsequent deformation it is northwesterly to northeasterly striking. The orientation of mineral lineations on the Si surface is commonly highly variable both between and within the different areas. Whereas the trend of mineral lineations is significantly changeable in basement felsic gneiss and in psammitic gneiss (Quartpot Pelite), in granitic gneiss of the Warrigal Gneiss the mineral lineations are all similar, steeply plunging to the west. Mineral lineations that developed during a subsequent deformation on the 'Sig' foliation and on shear-band surfaces, slightly oblique to 'Sig' in granite of the Bertibubba Supersuite (Figs 5 & 12) also show a range in orientations and are shallow to steeply plunging. There is no evidence in the field that suggests variability of mineral lineation orientations 'between' each of the field areas is related to folding. The variability of orientations of mineral lineations of the same age arises for several reasons. In the basement felsic gneiss, variable mineral lineation trends are the effect of reorientation of originally shallowly plunging, possible northeasterly trending lineations by upright tight to isoclinal curviplanar folds. In psammitic gneiss of the Quartpot Pelite some variability is due to younger folding (moderately north plunging in the western part of the Limb zone to steeply north plunging elsewhere, including in the higher strain Hinge zone). However, the main change is from moderately north plunging to steeply north plunging mineral lineations. This directly corresponds to mineral lineations trending oblique to fold hinge lines in areas of lower D2 strain, and parallel to fold hinge lines in areas of higher D2 strain. In granite of the Bertibubba Supersuite D2 mineral lineations range from shallow to steeply plunging. The greenschist facies, geometrically similar regional upright tight to isoclinal folds and

243

crenulations denoted F2 (Figs 5 & 12), which deform Si and its related mineral lineation, contain variable fold hinge orientations throughout and within most areas of the Errabiddy Shear Zone. However, in granitic gneiss of the Warrigal Gneiss the fold hinge lines and related intersection lineations are all similar, steeply plunging to the west trending subparallel to mineral lineations on the Si surface. Variablity of fold hinge orientations throughout the Errabiddy Shear Zone could be due to a number of different reasons. For example the very strong variability in the basement felsic gneiss is due to the curvilinear nature of these folds, which is probably the effect of re-folding of different parts of a large-scale sub-horizontal fold; although they could also be in part constrictional folds with a component of vertical stretch. The difference in fold hinge orientation in the hinge and limb zones of the psammitic gneiss could largely be due to heterogeneous deformation around different layers of the large-scale south verging parasitic fold. Contrasting this, the variation in fold hinge orientations in the migmatized pelitic schist and gneiss is due to refolding about upright northnortheasterly striking folds in the area. Heterogeneous regional D2 deformation throughout the shear zone has also resulted in the development of open to close folds in areas of low D2 strain and crenulations and crenulation cleavages in areas of high D2 strain. The apparent parallelism of Dj mineral lineations and F2 fold hinges throughout the Errabiddy Shear Zone appears to correspond to areas of relatively higher D2 strain and could be due to the strong re-orientation and recrystallization of mineral lineations during the subsequent D2 deformation event (Figs 4, 5 & 12). Upright easterly to eastnortheasterly striking folds did not develop in the foliated c. 1958 Ma granite of the Bertibubba Supersuite. However, the 'Sig' foliation, which is the dominant structure throughout this domain, is subparallel to the fold axial surfaces of the tight to isoclinal F2 folds in the adjacent felsic gneiss, and also formed in the greenschist facies (Figs 5 & 12). This indicates these structures probably developed coevally. The shallow to steeply plunging D2 mineral lineations, which have a range of orientations and are related to the formation of the 'Sig' foliation (Fig. 12) in granite of the Bertibubba Supersuite are present both on the 'Slg' surface and on shear band surfaces, slightly oblique to 'Slg' (Fig. 4). There is no evidence in the field that suggests the variable orientations of these mineral lineations is related to folding. Detachments that cut the upright tight to isoclinal F? fold axial surfaces and S2 foliation

Fig. 12. Summary deformation networks from the four areas showing the general overprinting relationships observed within each of the areas and correlated between them. The relationship of observed structures to regional correlations (this paper and previous work) has also been shown. Explanations of the U-Pb SHRIMP igneous and metamorphic crystallization ages shown are in Occhipinti et al. (2004). Arrows indicate observed overprinting relationships. Deformation networks were constructed from four separate younging tables constructed in the field following Potts & Reddy (2000).

CRUSTAL-SCALE SHEAR ZONE AT ERRABIDDY

245

Fig. 13. Model summarizing the evolution of the Errabiddy Shear Zone, (a) shows the accretion of the Glenburgh Terrane onto the Yilgarn Craton during the Glenburgh Orogeny producing symmetric kinematics, and homogenous deformation and metamorphism, which can be traced throughout the shear zone (regional DJ; b and c shows two possible interpretations on the effects of the Capricorn Orogeny throughout the zone, (b) Dextral strike-slip movement concentrated along main bounding faults synchronous with both dextral and sinistral faults cutting regional D2 folds within the shear zone, (c) Dextral strike-slip faults along bounding faults synchronous with upright D2 folding between faults within the shear zone.

within the region but strike subparallel to them probably also formed during the D2 deformation event. The northerly striking open kinks and folds that deform the D2 structures with variable intensity throughout the region are related to a regional D3 deformation event.

Kinematic evolution of the Errabiddy Shear Zone Symmetrical kinematic indicators were associated with the main foliation surfaces associated with DX. Similarly, only symmetrical kinematic indicators were found to be associated with the upright tight to isoclinal easterly striking F2 folds in the areas studied. However, asymmetrical kinematic indicators such as delta and sigma feldspar porphyroclasts and shear bands are present in foliated granite of the Bertibubba Supersuite. These are sometimes associated with mineral lineations on the regional 'S2' surface, but are sometimes present where a mineral lineation is absent (Fig. 12). In the case of ductile kinematic indicators in the c. 1958 Ma granite of the Bertibubba Supersuite both dextral and sinistral ENE striking (along map strike; Fig. 8) strike-slip kinematic indicators are

present. However, dextral strike-slip movement indicators are most prevalent. The brittle detachments that trend subparallel to fold axial surfaces of the easterly striking upright folds and the regional S2 foliation throughout the shear zone commonly show both dextral and sinistral separations. Mineral lineations on the main foliation or shear surfaces associated with these detachments are usually absent, therefore making the kinematics difficult to interpret. Often these separations of opposing displacement sense are very near to each other (sometimes only a few metres) and suggest a bulk coaxial flow. The reverse and normal faults cut the along strike separations indicating they may have developed slightly later. However, the shear bands, deformed feldspar porphyroclasts, mineral lineations and brittle fault structures all developed in the greenschist facies and may have formed during the same progressive deformation event.

Temporal and tectonic evolution of the Errabiddy Shear Zone The Errabiddy Shear Zone is thought to have developed as a consequence of easterly or

246

S. A. OCCHIPINTI & S. M. REDDY

southeasterly directed collision and accretion of the Late Archaean to Palaeoproterozoic Glenburgh Terrane onto the Archaean Yilgarn Craton between 2000-1960 Ma ago and was probably a northnortheasterly or northerly trending shear zone (Fig. 2; Sheppard et al. 2004; Occhipinti et al. 2001, 2004). During this time one set of structures formed in the Errabiddy Shear Zone between 2000-1960 Ma (Occhipinti et al. 2004). Only symmetrical kinematic indicators (e.g. flattened feldspar porphyroblasts) were found to be associated with the formation of the pervasive ST foliation and the DI mineral lineation, which developed in the amphibolite or upper greenschist facies in each of the four mapped areas formed during the Glenburgh Orogeny (Fig. 13). The intrusion of the c. 1958 Ma granite of the Bertibubba Supersuite post dates this D! deformation event, as the granite does not contain the regional Si foliation. The eastnortheasterly striking, isoclinal to tight upright F2 folds present throughout the Errabiddy Shear Zone developed at greenschist facies conditions, suggesting that they formed during the 1830-1780 Ma Capricorn Orogeny (Sheppard & Occhipinti 2000; Occhipinti & Sheppard 2001; Occhipinti et al. 2001). The strike-slip, reverse and normal detachment faults that cut the F2 folds at low angles may have developed soon after their formation and may be part of the same progressive deformation event (Fig. 12). During the Capricorn Orogeny the Errabiddy Shear Zone was probably reorientated from a northnortheasterly, or northerly trend into its approximate current orientation. The F2 folds, which are the dominant structure to have developed during D2 in the Errabiddy Shear Zone are commonly associated with symmetrically flattened porphyroblasts, and where present mineral lineations around these folds mostly appear to have developed prior to their formation (Fig. 13). In addition the fold axial surfaces of F2 folds are subparallel to the bounding faults of the Errabiddy Shear Zone, suggesting that the shortened Si fabric was initially highly oblique to the shear zone boundary. In contrast to areas of F2 fold development, the S2 foliation in the c. 1958 Ma Bertibubba Supersuite granite is dominated by dextral kinematics, also recording some sinistral deformation. The high obliquity of the regional Si foliation to the shear zone boundary, prior to D2 deformation, and the prevalence of orthorhombic kinematics, suggests that general shear with shortening perpendicular to the shear zone boundaries dominated during D2 in much of the

Errabiddy Shear Zone. However, the prevalence of dextral strike-slip kinematics in the c. 1958 Ma granite of the Bertibubba Supersuite suggests that locally simple shear also took place during D2 (Fig. 13). Partitioning of simple shear deformation in the foliated granite of the Bertibubba Supersuite and synchronous pure shear processes dominated deformation in the basement felsic gneiss is a similar scenario to that presented recently by (Lin & Jiang 2001) who reported that pure shear dominated in a transpressional shear zone. No data in this study suggests that the variable orientation of F2 fold axial surfaces that mimic the regional curvature of the Errabiddy Shear Zone (inset, Fig. 1) is due to refolding after D2. Northerly striking F3 upright folds that developed after the Capricorn Orogeny only weakly deformed D2 structures and were not responsible for the formation of this curvature. This implies that reactivation did not influence the orientation of D2 structures around this bend in the shear zone but that a variation of the strain geometry during the regional D2 event operated throughout the shear zone, thus supporting that transpression is likely to have taken place during D2. Brittle vertical detachment faults with dextral displacement, which formed either late in D2, or during D3 and cut the S2 foliation in the foliated granite of the Bertibubba Supersuite, indicate that shear zone boundary-parallel shear took place in region. Brittle dextral separations were also common in the Archaean granitic gneiss (basement) area. The apparent rarity of dextrally displacing brittle structures in the other two areas studied might represent a sampling bias. The region of the Bertibubba Supersuite is the only area within this part of the Errabiddy Shear Zone to contain two different lithological units of probable differing geological age that have an exposed faulted boundary (Fig. 13), and the Archaean granitic gneiss area is situated only 200 m south of an inferred easterly trending boundary fault that separates it from the Quartpot Pelite (Fig. 1). Except for the case of the basement felsic gneiss-Bertibubba Supersuite area such boundaries have been inferred from aeromagnetic data and in the field usually consist of quartz veins and rubble that lie in between outcrop of different rock types (Sheppard and Occhipinti 2000). The development of dextral kinematics along and in close proximity to bounding faults within the shear zone implies that a relatively high simple shear component operated along these fault zones, whereas between these zones pure shear deformation dominated during D2. This is a scenario

CRUSTAL-SCALE SHEAR ZONE AT ERRABIDDY previously reported by Tikoff and Greene (1997), and more recently by Czeck and Hudleston (2003) in the Superior Province in Canada where strain partitioning between pure and simple shear components resulted in simple shear being concentrated within discrete shear zones, whereas the areas between these shear zones underwent deformation dominated by pure shear. To the east, the Kerba Shear Zone, which is a discrete shear zone that splays off the Errabiddy Shear Zone, contains kinematic indicators suggesting that the Archaean to Palaeoproterozoic Yarlarweelor Gneiss Complex and Palaeoproterozoic metasedimentary and metavolcanic rocks of the Bryah and Padbury basins that the shear zone separates were juxtaposed by dextral transpressional shear with only a minor dip-slip component (Reddy and Occhipinti 2004). This is further evidence that dextral transpression operated in the southern Capricorn Orogen during the greenschist facies D2 deformation (Fig. 13).

Conclusions This study illustrates that strain localization and kinematic partitioning within shear zones leads to differences in structural development within different parts of the shear zones. Therefore shear zone analysis requires a range of representative areas throughout the zone to be studied and correlated in order to understand the processes involved in the shear zone formation. Failure to do so could lead to misleading interpretations of shear zone development. The four areas mapped in detail record a range of structures associated with the initial and evolving development of a reactivated shear zone. Although there are some dissimilarities between the orientations and types of structures within each of the four areas there is also an overwhelming similarity between them, and the structural elements within the different parts of the shear zone can be correlated based on their geometries, metamorphic grade, and overprinting relationships. However, despite these correlations the data presented here suggest that the evolution of the Errabiddy Shear Zone is complex, both spatially and temporally. The shear zone evolved through two main orogenic events correlated with the Glenburgh and Capricorn Orogenies, which are separated by over 100 Ma. Only orthorhombic kinematic indicators were observed in fabrics that developed during the Glenburgh Orogeny and most of the mineral lineations observed during this study developed at this time. However, it is unlikely

247

that deformation during the Glenburgh Orogeny was solely by pure shear processes and the apparent lack of monoclinic structures that formed during this time may reflect the subsequent pervasive re-working of structures during the Capricorn Orogeny. Deformation during the Capricorn Orogeny was heterogeneous dextral transpression throughout the Errabiddy Shear Zone with components of both pure and simple shear operating at the same time but being partitioned throughout the area. Localization of simple shear and dextral strike-slip movements is likely to have occurred within the numerous discrete shear zones within the Errabiddy Shear Zone. These are responsible for juxtaposed different rock units of varying tectonic origin and geological age. Between these discrete shear zones pure shear processes dominated. We would like to thank Michael Wells and Martin Hand for very helpful reviews, and the Geological Survey of Western Australia for their ongoing logistical support. Thanks to Suzzanne Dowsett for helping with some of the figures, and Steve Sheppard and Ian Tyler for ongoing interesting geological discussions. This project has been funded by an Australian Research Council large grant No. A00106036 and a Curtin University of Technology Australian Postgraduate Award. This is TSRC publication number 235.

References CZECK, D.M. & HUDLESTON, PJ. 2003. Testing models for obliquely plunging lineations in transpression: a natural example and theoretical discussion. Journal of Structural Geology, 25, 959-982. DEWEY, J.F., HOLDSWORTH, R.E. & STRACHAN, R.A. 1998. Transpression and transtension zones. In: HOLDSWORTH, R.E., STRACHAN, R.A. & DEWEY, J.F. (eds) Continental transpressional and transtensional tectonics. Geological Society, London, Special Publications, 135,1-4. HAND, M. & BUICK, I.S. 2001. Tectonic evolutuion of the Reynolds-Anmatjira Ranges: a case study in terrain reworking from the Arunta Inlier, central Australia. In: MILLER, J.A., HOLDSWORTH, R.E., BUICK, I.S. & HAND, M. (eds) Continental reactivation and reworking. Geological Society, London, Special Publications, 184, 237-260. HOLDSWORTH, R.E., HAND, M., MILLER, J.A. & BUICK, I.S. 20010. Continental reactivation and reworking: an introduction. In: MILLER, J.A., HOLDSWORTH, R. E., BUICK, I. S. & HAND, M. (eds) Continental reactivation and reworking. Geological Society, London, Special Publications, London, 184,1-12. HOLDSWORTH, R.E., STEWARD, M., IMBER, J. & STRACHAN, R.A. 2001/?. The structure and rheological evolution of reactivated continental fault zones: a review and case study. In: MILLER, J.A., HOLDSWORTH, R.E., BUICK, I.S. & HAND, M. (eds)

248

S. A. OCCHIPINTI & S. M. REDDY

Continental reactivation and reworking. Geological Society, London, Special Publications, London, 184,115-137. JONES, R.R. & TANNER, P.W.G. 1995. Strain partitioning in transpression zones. Journal of Structural Geology, 17, 793-802. LIN, S. & JIANG, D. 2001. Using along-strike variation in strain and kinematics to define the movement direction of curved transpressional shear zones: An example from northwestern Superior Province, Manitoba. Geology, 29, 767-770. LIN, S., JIANG, D. & WILLIAMS, P.P. 1998. Transpression (or transtension) zones of triclinic symmetry: natural example and theoretic modilling. In: HOLDSWORTH, R.E., STRACHAN, R.A. & DEWEY,

IF. (eds) Continental transpressional and transtensional tectonics. Geological Society, London, Special Publications, 135, 41-57. NELSON, D.R. 1997. Compilation of SHRIMP U-Pb zircon geochronology data, 1996, Record 1997/2. Western Australia Geological Survey. NELSON, D.R. 1998. Compilation of SHRIMP U-Pb zircon geochronology data, 1997, Record 1998/2. Western Australia Geological Survey. NELSON, D.R. 1999. Compilation of geochronology data, 1998, Record 1999/2. Western Australia Geological Survey. NELSON, D.R. 2000. Compilation of geochronology data, 1999, Record 2000/2. Western Australia Geological Survey. NELSON, D.R. 2001. Compilation of geochronology data, 2000, Record 2001/2. Western Australia Geological Survey. OCCHIPINTI, S.A. & SHEPPARD, S. 20000. Errabiddy W.A. Sheet 2347. Western Australia Geological Survey. OCCHIPINTI, S.A. & SHEPPARD, S. 20006. Glenburgh, W.A. Sheet 2147. Western Australia Geological Survey. OCCHIPINTI, S.A. & SHEPPARD, S. 2001. Geology of the Glenburgh 1:100000 sheet. Western Australia Geological Survey, 37 pp. OCCHIPINTI, S.A., SHEPPARD, S., NELSON, D.R., MYERS, IS. & TYLER, I.M. 1998. Syntectonic granite in the southern margin of the Palaeoproterozoic Capricorn Orogen, Western Australia. Australian Journal of Earth Sciences, 45, 509-512. OCCHIPINTI, S.A., SHEPPARD, S., MYERS, IS., TYLER, I.M. & NELSON, D.R. 2001. Archaean and Palaeoproterozoic geology of the Narryer Terrane (Yilgarn Craton) and the southern Gascoyne Complex (Capricorn Orogen) - a field guide. Western Australia Geological Survey, Perth. OCCHIPINTI, S.A., MYERS, IS., SHEPPARD, S., TYLER, I.M. & SWAGER, C.P. 2002. Robinson Range, W.A. Western Australia Geological Survey.

OCCHIPINTI, S.A., SHEPPARD, S., PASSCHIER, C, TYLER, I.M. & NELSON, D. 2004. Palaeoproterozoic crustal accretion and collision in the southern Capricorn Orogen: The Glenburgh Orogeny. Precambrian Research, 128, 237-255. POTTS, G.I & REDDY, S.M. 2000. Application of younging tables to costruction of relative deformation histories - 1: Fracture systems. Journal of Structural Geology, 1473-1490. REDDY, S.M. & OCCHIPINTI, S.A. 2004. High-strain zone deformation in the southern Capricorn Orogen, Western Australia: Kinematics and age constraints. Precambrian Research, 128, 295-314. REDDY, S.M., WHEELER, I & CLIFF, R.A. 1999. The geometry and timing of orogenic extension: an example from the Western Italian Alps. Journal of Metamorphic Geology, 17, 573-589. REDDY, S.M., WHEELER, I, BUTLER, R.W.H., CLIFF, R.A., FREEMAN, S., INGA, S., PICKLES, C. & KELLEY, S.P. 2003. Kinematic reworking and exhumation within the convergent Alpine Orogen. Tectonophysics, 365, 77-102. SCRIMGEOUR, I. & RAITH, IG. 2001. High-grade reworking of Proterozoic granulites during Ordovician intraplate transpression, eastern Arunta Inlier, central Australia. In: MILLER, I A., HOLDSWORTH, R.E., BUICK, I.S. & HAND, M. (eds), Continental reactivation and reworking. Geological Society, London, Special Publications, 184, 261-287. SHEPPARD, S. & OCCHIPINTI, S.A. 2000. Geology of the Errabiddy and Landor 1:100 000 sheets. Western Australia Geological Survey, 37 pp. SHEPPARD, S., OCCHIPINTI, S.A. & TYLER, I.M. 2002. The relationship between the tectonic setting and composition of granitoid magmas, Yarlarweelor Gneiss Complex, Western Australia. Lithos, 66, 133-154. SHEPPARD, S., OCCHIPINTI, S.A. & TYLER, I.M. 2004. Palaeoproterozoic (2005-1970 Ma) granites related to convergence and collision of the Gascoyne Complex and Yilgarn Craton, Western Australia. Precambrian Research, 128, 257-277. TIKOFF, B. & TEYSSIER, C. 1994. Strain modeling of displacement-field partitioning in transpressional orogens. Journal of Structural Geology, 16, 1575-1588. TIKOFF, 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. TYLER, I.M. & THORNE, A.M. 1990. The northern margin of the Capricorn Orogen, Western Australia - an example of an early Proterozoic collision zone. Journal of Structural Geology, 12,685-701.

Strain and vorticity analysis of transpressional high-strain zones from the Virginia Piedmont, USA CHRISTOPHER M. BAILEY, BARBARA E. FRANCIS & ELEANOR E. FAHRNEY Department of Geology, College of William & Mary, Box 8795, Williamsburg, VA 23187 USA (e-mail: [email protected]) Abstract: Strain and vorticity analysis of two Late Palaeozoic high-strain zones from the southern Appalachian Piedmont indicates that these zones experienced general shear transpression with a monoclinic to triclinic symmetry. Granitic rocks in the Brookneal highstrain zone from the southwestern Virginia Piedmont were transformed into mylonites under greenschist facies conditions. Sectional strains, estimated from quartz grain shapes, in mylonites range from three to ten and three-dimensional fabrics record flattening strains. The mean vorticity number (Wm) estimated with the RJQ method ranges from 0.3 to 0.95. In the central Virginia Piedmont, lower amphibolite facies deformation in the Spotsylvania high-strain zone affected biotite gneisses, amphibolites, and granitic pegmatites. Minimum sectional strains, estimated from folded and boudinaged pegmatite dykes, of 8-20 are common and three-dimensional strains are dominantly constrictional. Porphyroclast hyperbolic distribution analysis of ultramylonites yields Wn values from 0.4 to 0.8. The kinematic significance of these transpressional high-strain zones is threefold: they record tens to hundreds of kilometres of strike-slip offset; 40 to 70% contraction normal to the zone; and significant orogen-parallel material elongation.

Discrete zones of high-strain are common features in erogenic belts and elucidating the tectonic significance of these zones requires knowledge of the deformation kinematics. Understanding the kinematics of flow in highstrain zones has long challenged structural geologists. Early kinematic models were twodimensional and evaluated inhomogeneous simple shear and/or volume change in parallelsided zones (Fig. la) (Ramsay & Graham 1970; Ramsay 1980). However, field geologists have recognized that not all high-strain zones experienced plane strain simple shear deformations with easily resolvable kinematics. During the last few decades many two- and three-dimensional flow models have been developed. A number of workers considered combinations of two-dimensional simple and pure shear deformation (Fig. Ib) (Ramberg 1975; Ghosh & Ramberg 1976; Simpson & De Paor 1993; Fossen & Tikoff 1993). Transpression, as defined by Sanderson & Marchini (1984), is simultaneous simple and pure shearing that shortens the deformation zone normal to its boundaries, stretches material in one direction parallel to the zone boundaries, and leads to flattening strains (Fig. Ic). The deformation symmetry in highstrain zones has traditionally been considered to be monoclinic, however a number of recent studies have modelled and/or documented tri-

Fig. 1. Kinematic models for deformation in highstrain zones: (a) simple shear; (b) simultaneous simple and pure shear; (c) transpression; and (d) triclinic shear. Note transpression and triclinic shear can have more complex geometries than those illustrated.

clinic deformation symmetries (Fig. Id) (Jones & Holdsworth 1998; Lin et al 1998; Jiang & Williams 1998). Applying flow models to naturally deformed rocks has been slower than the proliferation of new models because quantifying the finite deformation and rotational component of strain in mylonitic rocks is not trivial. However, in the

from: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 249-264. 0305-8719/$15.00 © The Geological Society of London 2004.

250

CM. BAILEY ETAL.

Fig. 2. Geological overview map of the Virginia Piedmont. BHSZ: Brookneal high-strain zone; MRHSZ: Mountain Run high-strain zone; SHSZ: Spotsylvania high-strain zone.

proper setting a number of techniques can be used to measure strain and vorticity in highly deformed rocks (e.g. Passchier 19880; Passchier & Urai 1988; Wallis 1992; Simpson & De Paor 1993; Tikoff & Fossen 1995). Quantitative strain studies of mylonitic rocks indicate that nonplane strains are common (Mitra 1979; Mawer 1983; Bailey et al 1994; Fletcher & Hartley 1994) and vorticity analyses consistently reveal that non-simple shear deformations are widespread (Simpson & De Paor 1997; Grasemann et al. 1999; Xypolias & Koukouvelas 2001; Bailey & Eyster 2003). Quantifying the strain geometry, strain magnitude, and vorticity is essential for accurately determining displacement across major tectonic boundaries, crustal thinning/ thickening during collision or rifting, and material flow in three-dimensions. The southern Appalachian Piedmont province is composed of distinct terranes, many of which are bounded by high-strain zones (Fig. 2). These zones of mylonitic rock are up to tens of kilometres wide and hundreds of kilometres in strike length. Piedmont high-strain zones are interpreted to record significant dextral transcurrent movement in the Appalachian hinterland during the Late Palaeozoic (Alleghanian) collision between Africa and North America (Bobyarchick 1981; Gates et al 1986; Secor et al. 1986; Gates & Glover 1989; Hatcher 1989, 2002). However, the absolute magnitude of dextral transcurrent slip across these zones is poorly known, although many Appalachian tectonic models require hundreds to thousands of kilometres of displacement across these structures (McKerrow & Ziegler 1972; Vauchez et al. 1987; Hatcher 1989). The

absence of piercing points and general unsuitability of Piedmont rocks for palaeomagnetic analysis makes displacement estimates difficult to obtain. The rationale for our study stems from the lack of quantitative structural observations concerning the kinematics of these zones. We estimate strain and vorticity for two Appalachian high-strain zones and apply those estimates towards quantifying the kinematics of Late Palaeozoic deformation in the Piedmont. Quantifying the total displacement, as well as shortening, across Late Palaeozoic high-strain zones is prerequisite to understanding the magnitude of Alleghanian collisional tectonics and reconstructing an accurate palaeogeography of the Appalachian orogen prior to the Alleghanian deformation.

Transpression and general shear Transpression is a widespread mode of deformation because oblique convergence at plate margins is common. As defined by Dewey et al. (1998) transpression is a strike-slip deformation that deviates from simple shear. Transpressional deformation models have become increasingly popular in the geological literature (e.g. Holds worth et al 1998). The transpressional model of Sanderson & Marchini (1984) has been modified by incorporating volume change, lateral stretch, strain asymmetry, and heterogeneity/strain partitioning into transpressional models (Fossen & Tikoff 1993; Robin & Cruden 1994; Tikoff & Teyssier 1994; Goodwin & Williams 1996; Jones et al 1997; Lin et al 1998). Fossen & Tikoff (1998) designated five types of transpression that differ in the amount of

TRANSPRESSION IN THE VIRGINIA PIEDMONT

251

Fig. 3. (a) General shear with constrictional, plane and flattening strains, (b) Transpressional and general shear flattening deformations. Transpressional deformation is a general shear deformation with strike-slip displacement.

extension and contraction in both vertical and horizontal directions. Some transpressional strains produce complex strain paths in which the principal axes of the strain ellipsoid switch and the elongation lineation develops normal to the transport direction (Sanderson & Marchini 1984; Tikoff & Greene 1997). General shear, as discussed by Simpson & De Paor (1993), is a two dimensional deformation that deviates from simple shear and includes both sub-simple shear and super-simple shear (De Paor 1983). Sub-simple shear describes a two-dimensional deformation that combines a simple shear (rotational) component and pure shear component, whereas super-simple shear describes a deformation whose rotational component is greater than that of simple shear (De Paor 1983). Super-simple shear deformations are considered rare in crustal settings and where present localized to grain-scale domains (Means 1981; Simpson & De Paor 1993; see however, Jiang & White 1995). A quantitative measure of the simple shear and pure shear components for a finite deformation is given by the mean kinematic vorticity number (Wm = Wnm of Passchier 1988&), where for pure shear Wm = 0 and for simple shear Wm = 1. An equal contribution of pure and simple shear occurs at a Wm value of 0.81. We use the term general shear for any deformation between simple and pure shear (0 < Wm< 1, sub-simple shear). General shear is useful term for describing any three-dimensional deformation that deviates from pure or simple shear regardless of whether the finite strain ellipsoid is prolate, plane, or oblate (Fig. 3a). General shear is not specific with regard to geographic coordinates (it may range from dip-slip dominated to strike-slip dominated), whereas transpressional strains produce

strike-slip offset/displacement and as such are linked to geographic coordinates (Fig. 3b). Kinematic vorticity and vorticity analysis The structural geology literature is replete with kinematic vorticity related terminology. Our intent is not to provide a comprehensive review of kinematic vorticity, but rather to clarify the meaning of different kinematic vorticity numbers (Wk, Wn, and Wm) and their application to naturally deformed rocks. Wk is a dimensionless quantity that relates the instantaneous rotation to the instantaneous stretching at a point (Truesdell 1953; Means et al 1980). Vorticity is partitioned into internal (shear induced) and external (spin with respect to an external reference frame) components. Geologists can rarely quantify rates of rotation versus stretching in deformed rocks; it is also uncommon to have knowledge of the external vorticity. Tikoff & Fossen (1995) transformed W^ (internal vorticity component) to represent three-dimensional finite deformation parameters, defined in terms of shear strains (y) and stretches (A:n), assuming a steady-state deformation. Passchier (1988«, b) defined Wn (neutral vorticity number for a twodimensional deformation) in terms of flow parameters, vorticity (W) and mean stretching rate (s). Wn is equivalent to Truesdell's Wk when the sectional volume change rate (a) is zero (Passchier 19880, b). Passchier (19880) introduced the mean vorticity number (Wm) that describes the bulk rotation of material lines coincident with the principal strain axes and is independent of deformation path. Most examples of geological vorticity analyses have used Wk or Wn, but in some instances it is Wm that has been measured. The structural geology community is still

252

CM. BAILEY ETAL.

Fig. 4. (a) Steady state general shear deformation and non-steady state deformation (pure shear followed by simple shear) both produce the same cumulative deformation (Rs = 14; 0 = 6°; Wm = 0.9). Both deformations are plane strain, (b) Rsl® diagram with deformation paths for examples illustrated in (a). Steady state general shear (solid line) and non-steady state (dashed line). Inset box illustrates Wn during deformation.

debating the merits and pitfalls of vorticity analysis. Jiang & White (1995) and Jiang & Williams (1999) emphasized that vorticity (as originally defined) is an instantaneous quantity and argued that the kinematic vorticity number for a three-dimensional, non-steady deformation is of limited value. During a deformation with a triclinic symmetry both the vorticity vector and vorticity normal section rotate, thus for finite deformations the vorticity vector and vorticity normal section do not necessarily coincide with fabric elements. For triclinic deformations, vorticity estimates made in relation to fabric elements (e.g. on lineation parallel/foliation normal sections) are likely to underestimate the rotational component of strain. Tikoff & Fossen (1995) noted that an estimate of Wk describes an infinite array of three-dimensional deformations and illustrate that two-dimensional vorticity estimates tend to overestimate the three-dimensional vorticity number (although generally by <0.1). Our perspective is

similar to Passchier (19880), in that a measure of rotational strain is, like finite strain and volume change, an important parameter needed to characterize cumulative deformation. We emphasize that vorticity estimates must be combined with an understanding of the strain symmetry and its three-dimensional geometry. Some features preserved in deformed rocks provide information about the rotational nature of strain. We use two methods to estimate vorticity: (1) the RS/Q method (Fossen & Tikoff 1993; Bailey et al 1999; Bailey & Eyster 2003) and (2) the PHD (porphyroclast hyperbolic distribution) method (Simpson & De Paor 1993, 1997). The RJ® method estimates the mean vorticity (Wm) and does not require the assumption of a steadystate deformation path (Fig. 4). This method may provide information about the deformation path if strain gradients are present (Bailey et al. in press). Wm is given by the relationship between the XZ strain ratio (,RS) and the angle (0) between the high-strain zone boundary and the long axis of the strain ellipsoid (Fig. 4). With increasing strain the long axis of the finite strain ellipsoid rotates towards the high-strain zone boundary and follows a unique path for different Wm values. This method is also valid for isovolumetric non-plane strain deformations because material movement in the Y direction effectively reduces (for flattening) or expands (for constriction) the area of the XZ section in a homogeneous way, but does not change the angular relation between the X axis and the high-strain zone boundary (Fig. 5, Table 1). The Rs/& method is most appropriate for 'low-strain' mylonites (Rs < 20) because at high finite strains it is difficult to measure strain accurately and the 0 angle becomes very small regardless of Wm. If the high-strain zone boundary is localized due to lithological or mechanical heterogeneities this method may not be valid; high-strain zones that affect 'isotropic' protoliths (e.g. massive plutonic rocks) are the best suited for this method. The PHD method is based on the rotation of rigid particles. During general shear deformation porphyroclasts with different aspect ratios may rotate forward or backwards. The angle (v) between the flow apophyses can be used to determine the kinematic vorticity number (Wn), where Wn = cos(v) (Bobyarchick 1986; Simpson & De Paor 1993). The foliation plane and the eigenvector separating the fields of forward and backward rotation are assumed to be the flow apophyses. This relationship is valid for plane strain deformations (Bobyarchick 1986; Tikoff & Fossen 1995). PHD analysis is appropriate for ultramylonites that record very high strains (Rs ^>20) and have

TRANSPRESSION IN THE VIRGINIA PIEDMONT

253

Fig. 5. (a) Plane strain deformation with Rs - 6 and 0 = 9.5° yields a Wk value of 0.80. (b) Flattening strain deformation (K = 0.5) with Rs = 6 and Q = 9.5° yields a Wk value of 0.79. Flattening strain produces a loss of area in the XZ section (Axz = 0.81) because of material elongation in the Indirection (&2). \|/, angular shear; y, shear strain; &i, k2, and k3 are the principal stretches of the pure shear deformation component, F, effective shear strain; yxz, shear strain of simple shear component. See Table 1 for deformation parameters and equations. Deformation in b represents mean values for the BHSZ.

numerous porphyroclasts separated from one another by a fine-grained matrix. The PHD method yields an estimate of Wn not Wm, because if vorticity changes during deformation the rotating porphyroclasts are likely to 'equilibrate' to the last increments of rotational strain. Porphyroclasts are more rigid than the enclosing matrix, but recrystallized tails indicate that porphyroclasts commonly change shape with deformation. If porphyroclast aspect ratios change dramatically during deformation the results of PHD analysis may not record Wn accurately. Despite these limitations, PHD analysis is useful because it allows for discrimation between pure shear dominated and simple shear dominated deformations. Brookneal high-strain zone The Brookneal high-strain zone (BHSZ) is up to 5 km wide, and located in the southwestern Virginia Piedmont (Figs 2 & 6). Gates (1987) suggested that the BHSZ forms the Taconic (Ordovician) suture between the exotic metavolcanic Carolina and Chopawamsic terranes to the east and a native Laurentian metasedimentary sequence to the west. Hibbard et al. (2001) placed the BHSZ in the Piedmont Zone, where it separates the enigmatic Smith River Allochthon to the west from the metavolcanic Chopawamsic (Milton) terrane to the east. Recent geothermochronology indicates that the

Table 1. Deformation parameters for plane strain and flattening strain deformations (Fig. 5) Parameter

Plane strain

Flattening strain

Yxz

2.450 1.000 0.408 6.0 1.0 1.0 9.5° 71° 2.89 1.78 1.00 0.56 1.62 1.53

2.204 1.235 0.367 6.0 0.5 0.81 9.5° 71° 2.89 1.59 1.24 0.51 1.47 1.55

wk

0.80

0.79

X Y Z R*z K AXZ 8 V Y *i k2 ^3

F

X, Y, and Z are the principal stretches of the finite strain ellipsoid. R^z, aspect ratio in XZ plane; K, \n(X/Y)/\n(Y/Z)\ Axz, area of XZ plane; 0, angle between high-strain zone boundary and Xaxis of strain ellipsoid, xj/, angular shear; y, shear strain (tan \j/). k\, k2, and k3, principal stretches of the pure shear component. F, effective shear strain (k3 tan \|/); yxz, shear strain of the simple shear component = [F (ln(A:1//c3)]/(A:1-/:3); Wk, kinematic vorticity number = yrz/(2[ln(A:1)2 + \n(k2)2 + ln(&3)2] + (y^)2)° 5. Equations from Fossen & Tikoff (1993) and Tikoff & Fossen (1995).

254

C.M.BAILEY£TAL.

Fig. 6. Geological map of the Brookneal high-strain zone from the southwestern Virginia Piedmont. Modified from Gates et al. (1986), Virginia Division of Mineral Resources (1993), and Henika (1997). Samples for this study were collected from a traverse along the Roanoke River.

Smith River Allochthon is an exotic terrane with respect to Laurentia (Hibbard et al. 2003). The Melrose granite is progressively transformed into mylonitic rocks within the BHSZ (Fig. 6). Traditional U-Pb zircon dating yielded a 512 ± 5 age for the Melrose granite (Gates et al. 1986), but recent a U-Pb ion probe age of 447 ± 4 Ma is consistent with an Ordovician age for this pluton (Wilson 2001). Metasedimentary rocks of the Late Ordovician Arvonia Formation and felsic metavolcanic rocks of the Ordovician Chopawamsic Formation are strongly deformed along the southeastern side of the BHSZ (Fig. 6). Gates etal. (1986) mapped the northwestern boundary of the BHSZ as striking about 040°. The southeastern boundary of the BHSZ is difficult to define, but highly strained metasedimentary and metvolcanic rocks extend a number of kilometres to the SE into the Chopawamsic terrane. Gates et al. (1986) reported dextral shear sense indicators in S-C mylonites and calculated a minimum displacement of 17 km (assuming simple shear). Regional cooling ages of 324 ± 3 Ma (Ar/Arhornblende) to 300 ± 5 Ma (Rb/Sr- biotite/whole rock), are consistent with an Alleghanian (Late Palaeozoic) age for ductile deformation (Glover et al. 1983; Gates et al. 1986). The Brookneal zone was reactivated during Triassic brittle normal faulting associated with the formation of the Dan River rift basin (Henika 1997). The Melrose granite is texturally homogeneous, medium- to coarse-grained, and composed of plagioclase, quartz, K-feldspar, biotite and titanite. In the BHSZ, mylonitic Melrose granite commonly displays two foliations (Type 1 S-C mylonite of Lister & Snoke 1984). The

main mylonitic S-foliation strikes NE/SW and dips moderately to the SE (Fig. 7a). A mineral elongation lineation is weakly to moderately developed and plunges very gently to the NE (Fig. 7a). C-bands have a more easterly strike (10°-25°) than the main mylonitic foliation (Fig. 7a). Centimetre-scale ultramylonite zones occur throughout the BHSZ. Dextral shear sense indicators (asymmetric porphyroclasts, mica fish, fractured feldspars, and S-C patterns) are common. Symmetric kinematic indicators occur on lineation/foliation normal faces, consistent with an overall monoclinic deformation symmetry. Igneous textures are well preserved along the undeformed western part of the Melrose pluton (Fig. 8a). Individual quartz grain shapes are commonly inequant, but Rf/Q analysis reveals no preferred grain shape orientation at the thin section to hand sample scale. Within the BHSZ, quartz is transformed into monocrystalline and polycrystalline lenses and ribbons with a strong crystallographic preferred orientation (Fig. 8b). Many quartz ribbons are recrystallized aggregates that are readily distinguished from feldspars and the mica-rich matrix. Feldspars are extensively fractured with quartz and epidote filled cracks, muscovite mantles along many grain boundaries, and commonly display some undulose extinction. The modal abundance of quartz, muscovite, epidote, biotite, chlorite, and titanite increases in mylonitic rocks compared to their protolith. C-bands are defined by finegrained quartz, muscovite, chlorite, and hematite. Quartz and feldspar microstructures are consistent with middle to upper greenschist facies deformation conditions (T = 400-450 °C). Strain was estimated from quartz grains in granitic tectonites (collected from a traverse along the Roanoke River, Fig. 6) using standard Rf/Q techniques with a hyperbolic stereonet (De Paor 1988; Bailey et al. 1994). Quartz grains shapes are not ideal strain markers as in both the undeformed granite and mylonitic rocks they are non-elliptical. At high strains some necking occurred and in those grains the aspect ratio yields a minimum strain estimate (Fig. 4b). However, Bailey etal. (1994) demonstrated that strain estimated from quartz grain shapes in granitic mylonites deformed at the greenschist facies is similar to strain measured using fractured feldspars. 32-84 grains were measured per sample and Rs values in XZ sections range from 3-10 (Fig. 7b). Three-dimensional strains for all samples plot in the field of apparent flattening on a Flinn diagram (Fig. 7c) (K - 0.1-0.7). The 0 angle was determined by using the S-C angle within the sample and also the angle between

TRANSPRESSION IN THE VIRGINIA PIEDMONT

255

Fig. 8. Photomicrographs of (a) undeformed Melrose granite, f: feldspar; q: quartz; x: igneous grain boundaries, cross polarized light, (b) mylonitic Melrose granite with quartz ribbon, f: feldspar; m: micaceous matrix; q: quartz; rq: recrystallized quartz, cross polarized light.

Fig. 7. (a) Equal area stereogram of fabric elements for samples used for strain and vorticity analysis in the Brookneal high-strain zone, (b) RS/Q diagram for mylonitic rocks from the Brookneal high strain zone. Note: samples tend to plot on the same Wm curve perhaps an indication of steady-state general shear. Error bars based on the uncertainty in the orientation of S planes, C planes, and zone boundary, (c) Logarithmic Flinn diagram for quartz grain shapes from the Brookneal high-strain zone. Even with 20% volume loss samples plot in the field of flattening.

the main mylonitic foliation (S) and the strike of the high-strain zone boundary as defined by Gates et al (1986). The angle between S-C fabrics decreased as the strain ratio increased (Fig. 7b). Wm values based on the S-C angle range from 0.8 to 0.95 (Fig. 7b). Wm values based on the orientation of the high-strain zone boundary at the map scale were consistently lower than the S-C method (Wm = 0.3-0.6). Minimum and maximum values for Wm were based on the uncertainty in the orientation of the S and C planes and our estimate of the precision with which the strike of the northwestern high-strain zone boundary is known (Fig. 7b). The discrepancy between methods may reflect the uncertainty in defining the boundary of the high-strain zone at map scale. Multiple sets of Csurfaces can be generated during a single progressive deformation (Platt 1984; Passchier 1984; Alsop 1993), however in BHSZ mylonites derived from granitic protoliths there is only one set of C-surfaces that display a consistent orientation. There is a gradual transition from undeformed to weakly mylonitized granite along the

256

CM. BAILEY ETAL.

western boundary of the BHSZ and the exact orientation of the boundary is not clear. Both methods indicate that rocks in the BHSZ experienced general shear deformation not simple shear as suggested by Gates et al. (1986).

Spotsylvania high-strain zone The Spotsylvania high-strain zone (SHSZ) forms the boundary between the Early Palaeozoic Chopawamsic terrane and the Mesoproterozoic Goochland terrane in the central Virginia Piedmont (Fig. 2). This boundary was originally recognized as a sharp geophysical (magnetic and radiometric) boundary (Neuschel 1970). To the SW, the SHSZ appears to connect with the Hyco shear zone, a component of the Alleghanian central Piedmont shear zone, a signified boundary traceable for over 500 km in the southern Appalachians (Hibbard et al 1998; Wortman et al 1998). Hibbard et al (1998) interpreted the Hyco zone in southern Virginia to be a ductile thrust that emplaced the Carolina terrane over the Chopawamsic terrane. Farrar (1984), Pratt et al (1988), Glover et al (1989) interpreted the Spotsylvania zone as a significant thrust fault (not a suture) along which amphibolite/granulite facies rocks of the Goochland belt were emplaced to the NW in the Late Palaeozoic. In north-central Virginia, Pavlides et al (1980) interpreted the Spotsylvania zone to be a 2-3 km wide zone of predominantly brittle en-echelon faults. In central Virginia, Marr (1991) reported tectonic melange zone within the SHSZ and suggested it may represent a suture. Bourland (1976) and Spears & Bailey (2002) recognized brittle fault rocks in the SHSZ and interpreted these to have formed during Mesozoic reactivation of the Palaeozoic high-strain zone. Most previous studies are regional in scope and the kinematic and temporal history of the SHSZ is poorly understood. Rocks in the Chopawamsic terrane to the NW of the SHSZ include gneissic and schistose rocks derived from mafic to felsic volcanic rocks, quartzite, granitic gneiss, and pegmatite. Coler et al (2000) obtained Ordovician U-Pb ages for metavolcanic rocks from the Chopawamsic terrane in central Virginia. An areally extensive biotite-rich gneiss crops out SE of the SHSZ. At many locations the gneiss is pelitic (with kyanite ± sillimanite) and may correlate with the Maidens gneiss of the Proterozoic Goochland terrane (Farrar 1984; Aleinikoff et al 1996). The biotite-rich gneiss is interlayered with amphibolite-rich gneiss and intruded by a suite of granodioritic pegmatites. The SHSZ is about 15 km wide comprising

Fig. 9. Geological map of the Spotsylvania highstrain zone in the central Virginia Piedmont. GCT: Goochland Courthouse; GT: George's Tavern. Modified from Virginia Division of Mineral Resources (1993). Samples for this study were primarily collected from traverses along the James and South Anna Rivers.

heterogeneously deformed mylonitic rocks that typically lacks distinct boundaries (Fig. 9). Its northwestern boundary is defined by a geophysical lineament at the contact between amphibolerich gneisses to the NW and mylonitic rocks generally derived from granitic protoliths to the SE. The Lakeside high-strain zone, located five kilometres to the NW of the geophysical lineament, is a 1-2 km wide, NE/SW striking belt of mylonitic rocks similar to the SHSZ (Bourland 1976; Spears & Bailey 2002). Gneissic rocks to the SE of the SHSZ are strongly deformed many kilometres into the Goochland terrane. Mylonitic biotite schist, granitic mylonite, biotite-rich ultramylonite, amphibolite, and protomylonitic pegmatite are common rocktypes in the SHSZ. Foliation in the SHSZ strikes NE/SW and generally dips moderately to gently SE (Fig. lOa). Lineations (elongation and mineral) plunge shallowly to the NE and SW in the plane of the foliation. Shear sense indicators (asymmetrical porphyroclast tails and boudins) from

TRANSPRESSION IN THE VIRGINIA PIEDMONT

257

Fig. 10. (a) Equal area stereogram of fabric elements for the Spotsylvania high-strain zone, (b) Hyperbolic stereograms of forward and backward rotated porphyroclasts in ultramylonitic rocks from the Spotsylvania high-strain zone. Wn values indicate general shear.

Fig. 11. (a) Dextral asymmetry in pegmatitic mylonite, outcrop surface approximately parallel to lineation, normal to foliation, (b) Polished slab of porphryoclastic mylonite with dextral asymmetry, slab approximately parallel to lineation, normal to foliation, (c) Subhorizontal outcrop surface with boudinaged pegmatite dyke in biotite-bearing mylonitic gneiss. Knife is approximately 10 cm in length, (d) Granitic mylonite with feldspar (f) porphyroclast with myrmekite (m) and quartz ribbon composed of an aggregate of dynamically recrystallized (rq) grains with a strong crystallographic preferred orientation, cross polarized light, (e) Polished slab of ultramylonite with forward (f) and backward (b) rotated feldspar porphyroclasts, overall sense of shear is dextral, slab parallel to lineation, normal to foliation.

surfaces normal to foliation and parallel to lineation are consistent with dextral strike-slip displacement across the SHSZ (Fig. 11). Pegmatite dikes are commonly folded and boudinaged in a geometry consistent with bulk constrictional

strain (K = 1.2-5) (Figs lie & 12). Folded dykes are asymmetrical and generally verge to the NW. The geometry of asymmetrical structures, both parallel and normal to the elongation lineation, is consistent with at least a modest triclinic

258

CM. BAILEY ETAL.

estimates probably overestimate the pure shear component.

Interpretations and discussion Brookneal high-strain zone

Fig. 12. Idealized block diagram illustrating fabrics in the Spotsylvania high-strain zone. Triclinic deformation symmetry with strong component of dextral shear. Material elongation to the NE/SW with an overall constrictional strain.

deformation symmetry. At some locations dykes are boudinaged in two directions, but constrictional strains are more common. Minimum sectional strains, estimated from boudinaged and folded dikes on lineation parallel surfaces, range from 8-20. Feldspar porphyroclasts, pegmatitic boudins, and amphibolite boudins are superficially similar to clasts or blocks in a melange, but exhibit consistent dextral asymmetries and at many locations occur as tabular bodies with a pinch and swell character. Quartz grains in mylonitic rocks from the SHSZ are completely recrystallized, straight extinction is common, and strong crystallographic preferred orientations are well developed (Fig. lid). Feldspar porphyroclasts display core and mantle structures and strong undulose extinction. In mylonites and ultramylonites, myrmekite and flame perthite are localized along high-strain grain boundaries (Fig. lid). Synkinematic metamorphic minerals include biotite, garnet, epidote, and staurolite. Microstructures preserved in mylonitic rocks from the SHSZ are consistent with deformation conditions in the upper greenschist to lower amphibolite facies (450-500 °C). The rotational component of strain in the SHSZ was estimated using the PHD method on porphyroclastic ultramylonites (Fig. lie). Ultramylonites were slabbed on faces parallel to the lineation and normal to the foliation and the orientation and aspect ratio of forward and backward rotated grains plotted on a hyperbolic net. The smallest hyperbola separating the fields of backward rotated grains from forward rotated grains range from 40° and 68° (n - 6), yielding Wn values of 0.8 to 0.4, indicating general shear deformation (Fig lOb). Given the triclinic deformation symmetry in the SHSZ these vorticity

Mylonitic rocks from the BHSZ experienced general shear flattening strains (Wm - 0.8-0.95, C-S method: Wm = 0.3-0.6, HSZB-S method). XZ strain ratios would have to be two or three times larger to plot on the curve for simple shear (Wm = 1) (Fig. 7b). Strain estimates based on quartz grain shapes may not represent the bulk rock strain, but are not off by a factor of two or three. Mineralogical changes accompanied mylonitization in the BHSZ, and Gundersen & Gates (1995) estimated volume losses of 0-20% associated with deformation of the Melrose granite. Even with 20% volume loss, Brookneal mylonites experienced true flattening strains (Fig. 7c). At different strain values Brookneal mylonites record similar Wm values (Fig. 7b) a pattern consistent with a steady state general shear deformation path. We use mean values for vorticity (Wm = 0.8), XZ strain (Rs = 6), and three-dimensional strain (K = 0.5) for the BHSZ to model the kinematics (Fig. 5b; Table 1). This model assumes a monoclinic deformation symmetry with no volume loss. In the XZ section, flattening strain produces a sectional area loss (Fig. 5). Integrating shear strain (y = 2.89) over the width of the highstrain zone (about 5 km) yields a displacement estimate of 14.5 km. This estimate is an absolute minimum because it does not incorporate displacement associated with narrow ultramylonite zones or brittle faults and rocks to the SE of the BHSZ are penetratively deformed. General shear deformation in the BHSZ requires significant thinning (about 50%) normal to the highstrain zone and stretching parallel to the zone boundary (Fig. 5b). Data from the Brookneal zone fits well with the Type C transpressional model (Fig. 13) of Fossen & Tikoff (1998) or 'lengthening-widening shear' of Tikoff & Fossen (1999). Type C transpression leads to flattening strains with vorticity numbers less than one, vertical foliations with horizontal lineations, and material elongation parallel to the high-strain zone boundary. Mylonitic rocks from the BHSZ record general shear, true flattening strains, and have subhorizontal elongation lineations. Foliations in the BHSZ are not vertical (Fig. 7a). However, during mylonitization these rocks were more steeply dipping as the BHSZ is in the hanging wall of a Mesozoic listric normal fault and

TRANSPRESSION IN THE VIRGINIA PIEDMONT

259

Fig. 14. Schematic cross sections illustrating dextral transpressive deformation and development of steep foliation during the Late Palaeozoic. Listric normal faulting during the Mesozoic produces sedimentfilled half grabens and rotates Palaeozoic foliations to more gentle dips in the hanging wall block.

Spotsylvania high-strain zone

Fig. 13. Transpressional models of Fossen & Tikoff (1998). Mylonitic rocks from the Brookneal highstrain zone are similar to Type C transpression and those from the Spotsylvania high-strain zone are similar to Type E transpression. High-strain zone boundaries strike NE/SW.

horizontal axis rotation due to slip on this fault shallowed pre-existing steep foliations (Fig. 14). Synrift sedimentary rocks in the Dan River basin (immediately west of the BHSZ, Fig. 6) dip up to 45° west. Restoration of Mesozoic sedimentary rocks to their depositional orientation brings the mylonitic foliation in the BHSZ to dipsof60°-80°SE(Fig. 14).

The Spotsylvania high-strain zone is a zone of dextral transpressional deformation with an overall triclinic symmetry and constrictional strain. Type A & E transpressions of Fossen & Tikoff (1998) produce constrictional strains, however only Type E transpression leads to subhorizontal lineations (Fig. 13). Deformation in the SHSZ may approximate Type E transpression, but the same finite strain patterns could be created from superposed deformations. For instance, a Type B or C transpressional flattening followed by thrust loading could produce prolate strains. A number of workers have suggested that the boundary between the Chopawamsic and Goochland terranes is a thrust. However, mylonitic rocks from the SHSZ preserve no kinematic evidence of thrusting, nor are multiple overprinting fabrics present. Furthermore, rocks in both the Chopawamsic and Goochland terranes are at similar metamorphic grade; if thrusting occurred along the SHSZ it was prior to Alleghanian metamorphism and ductile deformation. The SHSZ is located, at some locations, in the hanging wall of listric Mesozoic normal faults and mylonitic foliations may have been steepened by horizontal axis rotations similar to that proposed for the BHSZ. We use values for vorticity (Wm = 0.4-0.8), XZ strain (7?s = 8-20), and three-dimensional strain (K = 1.8) to model the kinematics of deformation in the SHSZ (Fig. 15). Our models assume a monoclinic deformation symmetry and no volume loss. In the XZ section, constrictional strain produces a sectional area gain (Fig. 15). Shear strain estimates from these models range from 2.5 to 9.6. All models produce thinning (40 to 70%) normal to the high-strain zone and up to 500% stretching parallel to the zone boundary (Fig. 15). These deformation models can be used

260

CM. BAILEY ETAL.

Fig. 15. Kinematic models for the SHSZ. All models are monoclinic, isovolumetric, constrictional (K - 1.8) deformations. Deformation illustrated in two-dimensions, however constrictional strains produce a sectional area gain (A^z > 1). Note shortening normal to the deformation zone and significant zone boundary stretching in all models.

to restore the Goochland terrane to its paleogeographic position prior to dextral transpression. Consider George's Tavern (GT) in the Chopawamsic terrane and Goochland Courthouse (GCH) in the Goochland terrane (Figs 9 & 16). These locations are separated by the approximately 15 km wide SHSZ. Keeping George's Tavern fixed, Goochland Courthouse is retrodeformed to a predeformation position 80 to 300 km NE of its present location (Fig. 16). The greatest displacements are produced by the high-strain, pure shear dominated deformation (Model 4: Rs = 20, Wm = 0.4). Although the overall shear strain for Model 4 is less than for Model 2 (Rs = 20, Wm = 0.8), the pure shear dominated deformation produces the most zone normal thinning which results in significant offset when Goochland Courthouse is restored to its predeformation geometry (Figs 15 & 16). Displacement estimates are minimum values because strains calculated from boudinaged and folded dykes are minimum values and ultramylonite layers are likely to record strain ratios much greater than 20:1. In summary, the Goochland terrane, relative to the more western elements in the Virginia Piedmont, experienced significant southwestern translation during the Alleghanian orogeny.

Strain compatibility Non-plane strain, general shear deformations lead to compatibility problems with rigid wall

rocks because material is typically shortened normal to the zone and stretched parallel to zone boundaries. Bulk compatibility in general shear zones may be accomplished by: (1) deformable wall rocks; (2) sectional area changes along curved, non-parallel sided boundaries; (3) arrays of anastomosing high-strain zones; and (4) the development of discrete faults or dilational gaps (Simpson & De Paor 1993; Hudleston 1999; Bhattacharyya & Hudleston 2001). Away from Piedmont high-strain zones, rocks are penetratively deformed (albeit at lower strains) in complex patterns (Gates 1987; Gates & Glover 1989; Spears & Bailey 2002). In the Chopawamsic terrane, Goodman et al. (2001) recognized domains of constrictional and flattening strains with strike-parallel elongation lineations. The entire Chopawamsic terrane, between the BHSZ and the SHSZ, experienced elongation that is parallel and kinematically similar to deformation in the high-strain zones (Goodman et al 2001; Spears & Bailey 2002). With the exception of the northwestern boundary of the BHSZ in the Melrose pluton, the BHSZ and SHSZ have gradational boundaries with their wall rocks, such that there is a transition from high-strain to lower strain domains. Bulk compatibility is maintained between Piedmont high-strain zones and their wall rocks because deformation in both domains is kinematically related and occurs across a distinct strain gradient. Stretching faults occur in flowing rock bodies in which wall rocks lengthen or

TRANSPRESSION IN THE VIRGINIA PIEDMONT

261

Fig. 16. Present day map with the 'palaeogeographic'position of Goochland Courthouse prior to dextral transpressive deformation in the SHSZ. Goochland Courthouse retrodeformed using kinematic models in Fig. 15. Piedmont to the NW of the SHSZ is assumed to be fixed. Kinematic models require 80 to 300 km of SW displacement of the Goochland terrane relative to Piedmont terranes to the NW.

shorten while slip accumulates (Means 1989). Kinematically, Piedmont high-strain zones and their surrounding lower strain domains behaved as positive (material lengthening) stretching faults during deformation.

Tectonic significance of Piedmont high-strain zones Dextral transpressive high-strain zones in the Appalachian hinterland form a megaduplex structure (Costain et al 1987; Gates et al 1988; Hatcher 1989) active between about 320 and 280 Ma (Dallmeyer et al 1986; Horton et al 1987; Gates & Glover 1989; Wortman et al 1998). Strike-slip offset was important in the Piedmont, but the kinematics of deformation in the BHSZ and SHSZ reveals that material was also elongated parallel to the strike of the Appalachian orogen (NE-SW). Although collisional orogens are commonly viewed as thickening the crust and shortening the crust normal to the orogen, Piedmont high-strain zones demonstrate the significance of the three-dimensional (non-plane) strains and orogen-parallel material elongation. An unresolved controversy in southern Appalachian tectonics concerns the origin of the Goochland terrane. Mesoproterozoic rocks and A-type Neoproterozoic granitoids of the

Goochland terrane are similar to native Laurentian rocks in the Virginia Blue Ridge and have led many workers to conclude that the Goochland terrane is of Laurentian origin (Farrar 1984; Glover et al 1989; Aleinikoff et al 1996; Owens & Tucker 2000). Others suggest the Goochland terrane may be of peri-Gondwanan affinity and was accreted to Laurentia during the Appalachian orogen (Rankin et al 1989; Hibbard & Samson 1995). The Goochland terrane is currently separated from the Virginia Blue Ridge by the Mountain Run, Brookneal/Shores, and Spotsylvania high-strain zones (Fig. 2). All of these structures record late Palaeozoic dextral transpression (Gates et al 1986; Bobyarchick 1999; this work). If displacement on the Mountain Run and Brookneal/Shores high-strain zones are of similar magnitude to that of the SHSZ (over 100 km), the Goochland terrane was located at least 300 km NE of the Virginia Blue Ridge prior to the Alleghanian orogen. Tectonic models directly linking the Goochland terrane to the Virginia Blue Ridge are untenable because of significant dextral translation across Alleghanian high-strain zones. Quantitative understanding of the kinematics does not resolve whether the Goochland is a native Laurentian or exotic terrane, but does place meaningful limits on its pre-Alleghanian position in the Appalachian orogen. If the

262

C.M.BAILEYCTAL.

Goochland terrane is Laurentian, it originated somewhere between the Pennsylvania reentrant and the New York promontory, not outboard of the Virginia Blue Ridge. Conclusions Although naturally deformed rocks experience complex strain paths that may not be fully recorded in rock structures, strain and vorticity analyses provide reasonable limits on the nature of deformation in high-strain zones. Analysis of two Palaeozoic high-strain zones in the Virginia Piedmont reveals that they experienced nonplane strain general shear (Wm
Geological Society of America with Programs, 31, 59. BAILEY, CM., MAGER, S.M., GILMER, A.K. & MARQUIS, M.N. in press. Monoclinic and triclinic high-strain zones: examples from the Blue Ridge province Virginia. Journal of Structural Geology. BHATTACHARYYA, P. & HUDLESTON, P. 2001. Strain in ductile shear zones in the Caledonides of northern Sweden: a three-dimensional puzzle. Journal of Structural Geology, 23,1549-1565. BOBYARCHICK, A.R. 1981. The eastern Piedmont fault system and its relationship to Alleghanian tectonics in the southern Appalachians. Journal of Geology, 89, 335-347. BOBYARCHICK, A.R. 1986. The eigenvalues of steady state flow in Mohr space. Tectonophysics, 122, 35-51. BOBYARCHICK, A.R. 1999. Kinematics of the Mountain Run fault zone, Virginia. Geological Society of America with Programs, 31, 3, 6. BOURLAND, W.C. 1976. Tectogenesis and metamorphism of the Piedmont from Columbia to Westview, Virginia, along the James River. M.S. thesis, Virginia Polytechnic and State University. COLER,

D.G.,

WORTMAN, G.L.,

SAMSON, S.D.

&

HIBBARD, J.P. 2000. U-Pb geochronologic, Nd isotopic, and geochemical evidence for the correlation of the Chopawamsic and Milton Terranes, Piedmont Zone, Southern Appalachian Orogen. Journal of Geology, 108, 363-380.

COSTAIN, J.K., CORUH, C. & HATCHER, R.D. JR. 1987.

Geophysical signature of an inclined strike-slip duplex in the southeastern United States. Geological Society of America with Programs, 19, 7, 628. DALLMEYER, R.D., WRIGHT, IE., SECOR, D.T. JR. & SNOKE, A.W. 1986. Character of the Alleghanian orogeny in the southern Appalachians, Part II. Geochronogical constraints on the tectonothermal evolution of the eastern Piedmont in South Carolina. Geological Society of Bulletin, 96, 1329-1344. DE PAOR, D.G. 1983. Orthographic analysis of geological structures -1. Deformation theory. Journal of Structural Geology, 5, 255-278. DE PAOR, D.G. 1988. Rf/<|> strain analysis using an orientation net. Journal of Structural Geology, 10, 323-333. DEWEY, J.F., HOLDSWORTH, R.E. & STRACHAN, R.A. 1998. Transpression and transtension. In: HOLDSWORTH, R.E., STRACHAN, R.A. & DEWEY, J.F. (eds) Continental transpressional and transtensional tectonics. Geological Society, London, Special Publications, 135,1-14. FARRAR, S.S. 1984. The Goochland granulite terrane; remobilized Grenville basement in the eastern Virginia Piedmont. Geological Society of America Special Paper, 194, 215-227. FLETCHER, J.M. & BARTLEY, J.M. 1994. Constrictional strain in a non-coaxial shear zone: implications for fold and rock fabric development, central Mojave metamorphic core complex, California. Journal of Structural Geology, 16, 555-570. FOSSEN, H. & TIKOFF, B. 1993. The deformation matrix for simultaneous simple shearing, pure shearing and volume change, and its application to

TRANSPRESSION IN THE VIRGINIA PIEDMONT transpression-transtension tectonics. Journal of Structural Geology, 15, 413-422. FOSSEN, H. & TIKOFF, B. 1998. Extended models of transpression and transtension, and application to tectonic settings. In: HOLDSWORTH, R.E., STRACHAN, R.A. & DEWEY, IF. (eds) Continental transpressional and transtensional tectonics. Geological Society, London, Special Publications, 135,15-34. GATES, A.E. 1987. Transpressional dome formation in the southwest Virginia Piedmont. American Journal of Science, 287, 927-949. GATES, A.E. & GLOVER, L. 1989. Alleghanian tectonothermal evolution of the dextral transcurrent Hylas zone, Virginia Piedmont, USA. Journal of Structural Geology, 11, 407-419. GATES, A.E., SIMPSON, C. & GLOVER, L. 1986. Appalachian Carboniferous dextral strike-slip faults: an example from Brookneal, Virginia. Tectonics, 5,119-133. GATES, A.E., SPEER, J.A. & PRATT, T.L. 1988. The Alleghanian southern Appalachian Piedmont: A transpressional model. Tectonics, 7,1307-1324. GHOSH, S.K. & RAMBERG, H. 1976. Reorientation of inclusions by combination of pure and simple shear. Tectonophysics, 34,1-70. GLOVER, L., SPEER, A., RUSSELL, G.S. & FARRAR, S.S. 1983. Ages of regional metamorphism and ductile deformation in the central and southern Appalachians. Lithos, 15, 223-245. GLOVER, L., EVANS, N.H., WEHR, F. & PATTERSON, J. 1989. Tectonics of the Virginia Blue Ridge and Piedmont. IGC Field Trip Guide, T363, 88 pp. GOODMAN, M.C., DUBOSE, J.K., BAILEY, CM. & SPEARS, D. 2001. Petrologic and structural analysis of the Columbia Pluton, Central Virginia Piedmont. Southeastern Section Geological Society of America Abstracts with Programs, 32, 33. 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. GRASEMANN, B., FRITZ, H. & VANNAY, J.-C. 1999. Quantitative kinematic flow analysis from the Main Central Thrust Zone (NW-Himalaya, India): implications for a decelerating strain path and the extrusion of orogenic wedges. Journal of Structural Geology, 21, 837-853. GUNDERSEN, L.C.S. & GATES, A.E. 1995. Mechanical response, chemical variation, and volume change in the Brookneal and Hylas shear zones, Virginia. Journal of Geodynamics, 19, 231-252. HATCHER, R.D. JR. 1989. U.S. Appalachians synthesis. In: HATCHER, R.D. JR, THOMAS, W.A. & VIELE, G.W. (eds) The Appalachian-Ouachita orogen in the United States. Geological Society of America, The Geology of North America, F-2, 511-535. HATCHER, R.D. JR. 2002. Alleghanian (Appalachian) orogeny, a product of zipper tectonics: Rotational transpressive continent-continent collision and closing of ancient oceans along irregular margins. In: MARTINEZ, C., HATCHER, R.D. JR., ARENAS, R. & DIAZ GARCIA, F. (eds) Variscan-Appalachian dynamics: the building of the late Paleozoic basement. Geological Society of America Special Paper, 364,199-208. HENIKA,W.S. 2001. Geologic map of the RoanokeBO X

263

60 minute quadrangle, Virginia. Virginia Division of Mineral Resources Publication 148. HIBBARD, J.P. & SAMSON, S.D. 1995. Orogenesis exotic to the lapetan cycle in the southern Appalachians. In: HIBBARD, IP, VAN STALL, C.R. & CADWOOD, PA. (eds) Current perspectives in the Appalachian-Caledonian orogen. Geological Society of America Special Paper, 241,191-205. HIBBARD, J.P, SHELL, G.S., BRADLEY, PI, SAMSON, S.D. & WORTMAN, G.L. 1998. The Hyco shear zone in North Carolina and southern Virginia: implications for the Piedmont zone-Carolina zone boundary in the southern Appalachians. American Journal of Science, 298, 85-107. HIBBARD, IP, STEWART, K.G. & HENIKA, W.S. 2001. Framing the Piedmont Zone in North Carolina and Southern Virginia. Geological Society of America Southeastern Section Field Trip Guide, 1-26. HIBBARD, IP, TRACY, R.I & HENIKA, W.S. 2003. The Smith River Allochthon: a southern Appalachian peri-Gondwanan terrane emplaced directly on Laurentia. Geology, 31,123-126. HOLDSWORTH, R.E., STRACHAN, R.A. & DEWEY, IF. (eds) 1998 Continental transpressional and transtensional tectonics. Geological Society, London, Special Publications, 135, 339 pp. HUDLESTON, PI 1999. Strain compatibility and shear zones: is there a problem? Journal of Structural Geology, 21, 923-932. HORTON, J.W. JR., SUTTER, IF, STERN, T.W.

& MlLTON,

D.I 1987. Alleghanian deformation, metamorphism, and granite emplacement in the central Piedmont of the southern Appalachians. American Journal of Science, 287, 635-660. JIANG, D. & WHITE, 1C. 1995. Kinematics of rock flow and the interpretation of geological structures, with particular reference to shear zones. Journal of Structural Geology, 17, 1249-1265. JIANG, D. & WILLIAMS, P.F. 1998. High-strain zones: a unified model. Journal of Structural Geology, 20, 1105-1120. JIANG, D. & WILLIAMS, P.F. 1999. A fundamental problem with the kinematic interpretation of geological structures. Journal of Structural Geology, 21, 933-937. JONES, R.R. & HOLDSWORTH, R.E. 1998. Oblique simple shear in transpression zones. In: Continental transpressional and transtensional tectonics. In: HOLDSWORTH, R.E., STRACHAN, R.A. & DEWEY, IF. (eds) Continental transpressional and transtensional tectonics. Geological Society, London, Special Publications, 135, 35-40. JONES, R.R., HOLDSWORTH, R.E. & BAILEY, W. 1997. Lateral extrusion in transpressive zones: the importance of boundary conditions. Journal of Structural Geology, 19,1201-1217. LIN, S., JIANG, D. & WILLIAMS, P.F. 1998. Transpression (or transtension) zones of triclinic symmetry: natural example and theoretical modelling. In: HOLDSWORTH, R.E., STRACHAN, R.A. & DEWEY, IF. (eds) Continental transpressional and transtensional tectonics. Geological Society, London, Special Publications, 135, 41-57. LISTER, G.S. & SNOKE, A.W. 1984. S-C mylonites. Journal of Structural Geology, 6, 617-638.

264

CM. BAILEY ETAL.

MARK, J.D. JR. 1991. The Ca Ira melange; indicator of a major suture in the Piedmont of Virginia. Geological Society of America Abstracts with Programs, 23(1), 62. MAWER, C.K. 1983. State of strain in a quartzite mylonite, Central Australia. Journal of Structural Geology, 5, 401-409. MCKERROW, W.S. & ZIEGLER, A.M. 1972. Paleozoic Oceans. Nature: Physical Sciences, 240, 92-94. MEANS, W.D. 1981. The concept of steady-state foliation. Tectonophysics, 78,179-199. MEANS, W.D. 1989. Stretching faults. Geology, 17, 893-896. MEANS, W.D., HOBBS, B.E., LISTER, G.S. & WILLIAMS, P.P. 1980. Vorticity and non-coaxiality in progressive deformations. Journal of Structural Geology, 2, 371-378. MITRA, G. 1979. Ductile deformation zones in Blue Ridge basement rocks and estimation of finite strain. Geological Society of America Bulletin, 90, 935-951. NEUSCHEL, S.K. 1970. Correlations of aeromagnetics and aeroradioactivity with the lithology in the Spotsylvania area, Virginia. Geological Society of America Bulletin, 81, 3575-3589. OWENS, B.E. & TUCKER, R.D. 2000. Late Proterozoic plutonism in the Goochland terrane: Laurentian or Avalonian connection? Geological Society of America with Programs, 32(2), 65. PASSCHIER, C.W. 1984. The generation of ductile and brittle shear bands in a low-angle mylonite zone. Journal of Structural Geology, 6, 273-281. PASSCHIER, C.W. 19880. Analysis of deformation paths in shear zones. Geologische Rundschau, 77, 309-318. PASSCHIER, C.W. 19886. The use of Mohr circles to describe non-coaxial progressive deformation. Tectonophysics, 149, 323-338. PASSCHIER, C.W. & URAI, J.L. 1988. Vorticity and strain analysis using Mohr diagrams. Journal of Structural Geology, 10, 755-763. PAVLIDES, L., BOBYARCHICK, A.R. & WIER, K.E. 1980. Spotsylvania Lineament of Virginia. U.S. Geological Survey Professional Paper, 1175, 73. PLATT, J.P. 1984. Secondary cleavages in ductile shear zones. Journal of Structural Geology, 6, 439-442. PRATT, T, CORUH, C, COSTAIN, J. & GLOVER, L. 1988. A geophysical study of earth's crust in central Virginia: implications for Appalachian crustal structure. Journal of Geophysical Research, 93, 6649-6667. 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. RAMSAY, J.G. & GRAHAM, R.H. 1970. Strain variation in shear belts. Canadian Journal of Earth Sciences, 7, 786-813. RANKIN, D.W., HALL, L.M., DRAKE, A.A. JR., GOLDSMITH, R., RATCLIFFE, N.M. & STANLEY, R.S. 1989. Proterozoic evolution of the rifted margin of Laurentia. In: HATCHER, R.D.,THOMAS, W. A. & VIELE, G.W. (eds) The Appalachian-Ouachita orogen in the United States. Geological Society of America, The Geology of North America, v. F-2, pp. 10-42.

ROBIN, P.-Y.F. & CRUDEN, A.R. 1994. Strain and vorticity patterns in ideally ductile transpressive zones. Journal of Structural Geology, 16,447-466. SANDERSON DJ. & MARCHINI, W.R.D. 1984. Transpression. Journal of Structural Geology, 6, 449-458. SECOR, D.T. JR., SNOKE, A.W., BRAMLETT, K.W. COSTELLO, O.P. & KIMBRILL, O.P. 1986. Character of the Alleghanian orogeny in the southern Appalachians: Part I. Alleghanian deformation in the eastern Piedmont of South Carolina. Geological Society of America Bulletin, 97,1319-1328. SIMPSON, C. & DE PAOR, D.G. 1993. Strain and kinematic analysis in general shear zones. Journal of Structural Geology, 15,1-20. SIMPSON, C. & DE PAOR, D.G. 1997. Practical analysis of general shear zones using the porphyroclast hyperbolic distribution method: an example from the Scandinavian Caledonides. In: SENGUPTA, S. (ed.) Evolution of Geological Structures in Microto Macro- Scales, 169-184. SPEARS, D.B. & BAILEY, C.M. 2002. Geology of the central Virginia Piedmont between the Arvonia syncline and the Spotsylvania high-strain zone. Virginia Geological Field Conference Guidebook, 32, 35 pp. TIKOFF, B. & FOSSEN, H. 1995. The limitations of threedimensional kinematic vorticity analysis. Journal of Structural Geology, 17,1771-1784. TIKOFF, B. & FOSSEN, H. 1999. Three-dimensional reference deformations and strain facies. Journal of Structural Geology, 21,1497-1512. TIKOFF, B. & GREENE, D. 1997. Stretching lineations in transpressional shear zones. Journal of Structural Geology, 19, 29-40. TIKOFF, B. & TEYSSIER, C. 1994. Strain modeling of displacement field 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. VAUCHEZ, A., KESSLER, S.F. LECORCHE, J.-P. & VILLENEUVE, M. 1987. Southward extrusion tectonics during the Carboniferous Africa-North America collision. Tectonophysics, 142, 317-322. VIRGINIA DIVISION OF MINERAL RESOURCES. 1993. Geologic Map of Virginia. Virginia Division of Mineral Resources, scale 1: 500,000. WALLIS, S.R. 1992. Vorticity analysis in metachert from the Sanbagawa belt, SW Japan. Journal of Structural Geology, 14, 271-280. WILSON, J. 2001. U/Pb zircon ages of plutons from the Central Appalachians and GIS-based assessment of plutons with comments on their regional tectonic significance. M.S. thesis, Virginia Tech, 121 pp. WORTMAN, G.L., SAMSON, S.D. & HIBBARD, J.P. 1998. Precise timing constraints on the kinematic development of the Hyco shear zone: implications for the central Piedmont shear zone, southern Appalachian orogen. American Journal of Science, 298,108-130. XYPOLIAS, P. & KOUKOUVELAS, I.K. 2001. Kinematic vorticity and strain rate patterns associated with ductile extrusion in the Chelmos Shear Zone (External Hellenides, Greece). Tectonophysics, 338, 59-77.

Constraints on kinematics and strain from feldspar porphyroclast populations SCOTT GIORGIS & BASIL TIKOFF Department of Geology and Geophysics, University of Wisconsin - Madison, 1215 W. Dayton St, Madison, WI, 53706, USA (e-mail: [email protected]; [email protected]) Abstract: We develop a method for constraining the kinematics and finite strain of deformation in shear zones based on a three-dimensional numerical model of the rotation populations of rigid clasts. The results of the model are characterized in terms of a fabric ellipsoid, which is directly measurable from field data. Fabric ellipsoids measured from populations of prolate clasts have anisotropies that increase steadily and plateau; the shape of the fabric ellipsoid becomes increasingly more prolate with progressive deformation. The behaviour of populations of oblate clasts is much more complex because the stability of individual oblate clasts depends on their aspect ratio and the vorticity of deformation. Populations of oblate clasts may produce fabric ellipsoids with oscillating anisotropies and shapes if their aspect ratio is low enough for a continuous rotation. For either prolate or oblate clasts, the maximum anisotropy that a fabric ellipsoid will reach is governed by the aspect ratio of the individual clasts of that population. The theoretical maximum anisotropy is achieved when all of the clasts are perfectly aligned. The shape of the fabric ellipsoid, in conjuncture with the anisotropy, can be used to constrain the vorticity and finite strain of deformation. The numerical model suggests that there is no consistent relationship between the asymmetrical orientation of a population of rigid markers and the simple shear component of deformation. Therefore, the asymmetrical alignment of a population of porphyroclasts is not a reliable shear sense indicator. Additionally, there is no direct correlation between the fabric ellipsoid and the strain ellipsoid. Model results are applied to shape preferred orientation data collected from a feldspar megacrystic granite in the western Idaho shear zone (USA). Three-dimensional fabric ellipsoids are calculated from two-dimensional sectional measurements of oblate-shaped, unmantled, potassium feldspar porphyroclasts. Comparison of these data with the results of the numerical model suggests that transpressional deformation had an intermediate angle of oblique convergence (30°-60°). This implies that deformation in the western Idaho shear zone was characterized by a large component of convergent motion.

The development of a shape preferred orientation of porphyroclasts in deforming materials is commonly approached by quantifying the rotation rate of rigid markers. Jeffery (1922) outlined a quantitative analysis of the rotation of a rigid body in a viscously deforming Newtonian fluid. Subsequent studies based on the work of Jeffery (1922) have examined rigid clast rotation within the framework of a variety of two-dimensional deformation models: pure shear (Gay 19680,&), simple shear (Gay 19680), and a combination of pure and simple shear (Ghosh & Ramberg 1976). The rotation of spheroidal and triaxial clasts were also examined in the context of three-dimensional deformation (Willis 1977; Freeman 1985; Passchier, 1987; Jezek^ al 1994, 1996). Analogue experimentation has shown Jeffery model rotation to be a robust description of rigid body rotation in two dimensions (e.g.

Ghosh & Ramberg 1976; Mancketelow et al 2002) and three dimensions (e.g. Willis 1977; Arbaret et al 2000). There are several assumptions common to all models based on Jeffery's mathematics: the markers are perfectly rigid; the matrix is a Newtonian fluid; and the markers are not interacting with other markers or the boundaries of the deforming material. If these assumptions are applicable to rigid clasts (such as feldspar porphyroclasts) in naturally deformed rocks, numerical modelling indicates that the rate of rotation of an isolated clast is a function of the aspect ratio of the clast, the orientation of the clast relative to the flow, the vorticity of deformation, and the total amount of finite strain accumulated in the system (Ghosh & Ramberg 1976; Willis 1977; Freeman 1985; Passchier 1987; Jezek et al. 1994,1996). Numerical modelling of

From: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. I W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 265-285. 0305-8719/$15.00 © The Geological Society of London 2004.

266

S. GIORGIS & B. TIKOFF

populations of rigid porphyroclasts demonstrates that these same factors control the development of shape preferred orientation in both two dimensions (e.g. Ildefonse & Fernandez 1988) and three dimensions (e.g. Jezek et al 1994,1996). Within numerical models, the intensity of the shape preferred orientation of a population of markers is most often quantified by examining the angular relationship between linear features. In two dimensions, this linear feature is usually the long axis of the porphyroclast (e.g. Tikoff & Teyssier 1994); for spheroidal particles (i.e. an ellipsoid with two axes of equal length and one unique axis) in three dimensions, the orientation of the unique axis of the spheroid is typically used (e.g. Passchier 1987; Jezek et al 1996). In either two or three dimensions, a quantitative measure of the degree of preferred orientation of a population of linear features can be obtained using a normalized orientation tensor method (see Harvey & Laxton 1980). Twodimensional data, such as the distribution of the orientation of the long axes of porphyroclasts, is readily collected from outcrop faces or polished slabs. Therefore, two-dimensional models of clast preferred orientation are easily applied to field studies because the modelled preferred orientations are directly comparable to geological data. A direct correlation between three-dimensional models of rigid clast rotation and geological data has proved more elusive. It is generally not feasible to gather geological information that describes the orientation of an individual porphyroclast in three dimensions. Even if simplifying assumptions are made, knowledge of the exact orientation of the unique axis of an individual marker is not possible at an outcrop. Two methods around this problem are threedimensional X-ray tomographic imaging (CT scanning) for small particles and excavation of rigid markers in a removable matrix, such as pebbles in glacial till. However two-dimensional sectional data from several faces can be gathered using image analysis techniques. These data can be combined and described by an ellipsoid (e.g. Owens 1984; Robin 2002). The resultant fabric ellipsoid provides three-dimensional information about the distribution in shape and orientation of the sampled population of porphyroclasts (Robin 2002). We present a model which tracks the progressive development of the fabric ellipsoid measured from a population of rigid, spheroidal particles subjected to a transpressional deformation. This method has the advantage of

examining the effect of particle shape, vorticity, and finite strain on the development of a factor that is calculated directly from field data. We then present fabric ellipsoid data collected from a potassium feldspar megacrystic orthogneiss within a transpressional shear zone, the western Idaho shear zone (USA). The results of the model are then used to place constraints on the bulk vorticity and total finite strain accumulated within the western Idaho shear zone.

Forward model of clast rotation Jeffery (1922) outlined a series of equations that describe the rotation of rigid ellipsoids in a viscously deforming Newtonian medium. Jezek et al. (1994, 1996) solved these equations for a transpressional deformation model described by the velocity gradient tensor:

where ex, ey, and £v are the rates of stretching along the x, y, and z directions, respectively, and 7 is the rate of simple shearing in the positive x direction (Fig. 1). In a transpressional deformation, as described by Sanderson and Marchini (1984), there is shortening in the y direction (€y<0), elongation in the z direction (£ z >0), simple shearing in the x direction (7^0), and no pure shearing in the x direction (ex = 0). We consider only the special case of an ellipsoid where two of the three axes have equal lengths (spheroids). Consequently, the individual rigid particles are either perfectly prolate or oblate. Spheroidal particles require only two Euler angles to describe any three-dimensional rigid body rotation. The principal axes of the rotating ellipsoid are denoted by a, b, and c, where a - b ^ c. In the case of spheroidal particles, the equations are written in closed form solution (Jezek et al 1996), with respect to the rotation rate:

STRAIN FROM FELDSPAR PORPHYROCLASTS

267

wise direction and 0 describes the rate of rotation around the x axis in a counterclockwise direction (Fig. 1). 0 describes the 'plunge' of the unique axis of the clast, while 0 describes the trend (Fig. Ic). £x, 8y, Sv and 7 are components of the velocity gradient matrix (Fig. la). BI is given by:

where c is the length of the unique axis of the spheroid and b is the length of the two equal axes. Bj is a function of the axial ratio of the ellipsoid and was proposed by Bretherton (1962). For an infinitely oblate object (i.e., a material plane) #/ —> +1, whereas for an infinitely prolate object (i.e. a material line) Bj —> -1. Thus the rate of rotation of a rigid particle in a viscously deforming medium is a function of the orientation of the particle, the deformation boundary conditions, and the axial ratio of the particle (e.g. Ghosh & Ramberg 1976; Passchier, 1987; Jezek et al 1994,1996). Equations 2 and 3 calculate rates of rotation, whereas only the finite orientation of particles can be measured from field data. To express these rates in terms of finite amounts of rotation, we assume homogeneous flow and steady state deformation. Consider the deformation matrix (the position gradient matrix Fin the continuum mechanics literature; Malvern 1969):

Fig. 1. (a) Block diagram of transpressional deformation modified from Sanderson & Marchini (1984). a is the angle of oblique convergence in transpression. (b) The coordinate system of the numerical model. £x, £ v , and 8z are the rates stretching of the coaxial components along the coordinate axes, and 7 is the simple shear rate component that acts in the xy plane. Note for the special case of transpression there is no pure shearing in the x-direction (i.e. £x - 0). (c) The coordinate system for the unique axis of either an oblate clast or a prolate clast. The angle 0 describes the 'trend' on the unique axis, while 9 describes the 'plunge' of the unique axis.

0 and 0 are the Euler angles that express the initial position of the clast. 0 describes the rate of rotation around the z axis in a counterclock-

which describes any monoclinic deformation where A:x, A:y, and kz are the pure shear components and y is the simple shear component (Tikoff & Fossen 1993). For a steady-state deformation, Merle (1986) showed that the rates of pure shearing and simple shearing associated with this deformation could be expressed as a time-independent solution using the relations:

The above terms describe deformation for a unit time (t). Following Tikoff and Fossen (1993) the finite deformation (£>) can be subdivided into n

268

S. GIORGIS & B. TIKOFF

equal increments to create a matrix which expresses each incremental step of deformation (Ancr.):

If the total finite deformation is subdivided into a sufficiently large number of increments, then Dincr describes the infinitesimal displacement of material points for that deformation. Thus by choosing a large number of steps and assuming steady-state kinematics, the relations in Eqn 6 can be used to describe the rates of deformation associated with an infinitesimal amount of strain for a unit of time:

Substituting these terms into Eqns 2 and 3 allows us to solve for a finite amount of rotation for each increment of deformation as a function of the orientation of the ellipsoid prior to that increment of deformation. The accuracy of the solution depends on choosing appropriately small deformation increments (i.e. a large number of increments) for a given finite strain. Once the amount of rotation is known, it is straightforward to calculate the new position of the particle. Both an incremental *-axis rotation (
Because the theoretical clasts are treated as rigid markers, they undergo no internal strain as a result of deformation. Rigid markers do translate during deformation, in addition to rotating, a process which leads to tiling (e.g. Ildefonse et al 1992; Tikoff & Teyssier 1994). We neglect this translational component, as we are principally interested in clast orientation.

Three-dimensional description of clast orientation The above theory of clast rotation allows one to calculate the three-dimensional orientation of any spheroidal particle subjected to transpressional deformation. However, feldspar clasts and other geological markers, are generally only visible on two-dimensional outcrop faces. Consequently, in order to compare modelled results with field data, we must describe the intersection of the three-dimensional particle with a twodimensional planar surface. This procedure involves using elliptical integrals. The surface of an ellipsoid is approximated using three parametric equations:

where a, b, and c are the lengths of the principal axes of the ellipsoid; A is defined over the interval [0, 2n] and p over the interval [0, n] (e.g. Bowman 1961). The matrix / contains all of points that define the surface of an ellipsoid. The number of points defining the surface of the ellipsoid is a function of the number subdivisions of A and p. After an increment of deformation, the new orientation of the surface of the ellipsoid (/') is

This equation provides a quantitative description of the position of the surface of the ellipsoid (i.e. the orientation and shape of the ellipsoid). Equation 15 implies that the *-axis rotation occurs first followed by the z-axis rotation. Thus although this order of operations accurately describes the final orientation of the unique axis after rotation, it does not describe the actual

STRAIN FROM FELDSPAR PORPHYROCLASTS

269

Fig. 2. Construction of the fabric ellipsoid, (a) After each increment of deformation in the model, each clast in the population is cut by three orthogonal planes, (b) The expression of the population of clasts on each of these planes is used to calculate a three-dimensional fabric ellipsoid, (c) Field data is collected in a similar manner. Two-dimensional sectional data is collected on faces orientated at high angles to each other and subsequently combined into a three-dimensional fabric ellipsoid.

rotation path. An accurate description of the rotational path relies on choosing a sufficiently large number of increments to describe the deformation. To draw the most direct correlation between the results of the numerical model and data collected in the field, we calculate the two-dimensional expression of the surface of the ellipsoid on three orthogonal sections. The elliptical (two-dimensional) expression of the ellipsoidal surface (three-dimensional) is calculated for the three orthogonal planes (Fig. 2). On each surface three pieces of information are necessary to define the ellipse: the orientation of the long axis; the length of the long axis; and the length of the short axis. Using the xy plane as an example, all of the

points in the /' matrix are surveyed to determine which sets of points have z coordinates equal to zero (i.e. lie in the xy plane). Of the points that meet these criteria, the point that is the farthest distance from the origin is at the tip of the long axis. The orientation of the long axis is calculated using the inverse tangent function:

The length of the long axis is calculated using the distance formula:

The length of the short axis of the ellipse is calculated by surveying the same set of points in the

270

S. GIORGIS & B. TIKOFF

xy plane and determining which one is closest to the origin. In a similar manner the orientation and dimensions of the ellipse on other sections are determined. The process outline above is repeated for each clast after each increment of deformation. Thus on a single face there will be a series of ellipses; the aspect ratio and orientation of the long axis of each ellipse is described using a two dimensional inverse shape matrix (Robin 2002). Once all of the ellipses on each face (XY, YZ, and XZ) have been described as inverse shape matrices, all of these data can be averaged into a three dimensional inverse shape matrix; this matrix describes the shape and orientation of the fabric ellipsoid. See Robin (2002) for a detailed explanation. Using this approach, we investigate the evolution of the fabric ellipsoid throughout a single progressive deformation.

Discussion of model results The numerical model of rigid clast rotation that we present examines how the shape preferred orientation of a population of clasts develops during a transpressional deformation. To understand how a population of clasts rotates, it is critical to understand how individual clasts in that population behave. Thus we will begin our discussion with a brief description of how single clasts rotate. See Passchier (1987) for a more detailed description of the behaviour of single, spheroidal clasts.

Rotation of single clasts The rotation of individual perfectly linear and perfectly planar markers (material lines and planes) is generally well understood (e.g. Jeffery 1922; March 1932; Flinn 1962; Passchier 1987; Fossen et al 1994). The rate and direction of rotation of a material line or plane is a function of its orientation in space and the vorticity of deformation. We use kinematic vorticity (Wk) to describe the relative contributions of pure shear and simple shear, with Wk - 0 implying pure shear deformation and Wk = I implying simple shear deformation (Truesdell 1953). Stable orientations of material lines and planes are governed by the orientation of the flow apophyses, which characterize the particle flow in a deforming medium (Bobyarchick 1986; Passchier 1990). In transpression, the divergent flow apophysis is in the z direction (the fabric attractor), the convergent flow apophysis is in the xy plane (the fabric repellor; Fig. 1), whereas the intermediate flow apophysis is parallel to the *-axis (Fossen et al. 1994). A material line is simply an infinitely

prolate ellipsoid, whereas a material plane is an infinitely oblate ellipsoid. Therefore, the stable and metastable orientations of prolate and oblate single ellipsoids and populations of ellipsoids are also dictated by the orientation of the flow apophyses (Passchier 1987).

Single prolate clasts Most material lines rotate towards the fabric attractor (z-axis), however those lines that are orientated in the xy plane are 'metastable' and rotate into parallelism with the intermediate flow apophysis (Fig. 3a; Fossen et al. 1994). In transpression, prolate ellipsoids generally follow spiral rotational paths that are different from those paths followed by material lines (Fig. 3c, e; e.g. Freeman 1985). However, prolate ellipsoids and material lines have the same stable destination. Most rotate into parallelism with the divergent flow apophysis (Fig. 3; e.g. Passchier 1987). Prolate ellipsoidal particles with their long axes parallel to the xy plane are metastable. These markers may reach a stable orientation parallel to the shear plane (long axis parallel to the .x-axis, the intermediate flow apophysis) during deformation, depending on their aspect ratios and the kinematics of flow. In general, small aspect ratios and large simple shear components of deformation facilitate continued rotation (e.g. Passchier 1987; Wallis 1993).

Single oblate clasts The behaviour of oblate ellipsoids is generally more complex than that of prolate ellipsoids or material planes. The behaviour of material planes is usually described in terms of the rotational path of the poles to those planes (e.g. Fossen et al. 1994). Similarly, the behaviour of oblate spheroids is discussed in terms of the rotational path of the short axis of the ellipsoid (e.g. Passchier 1987). Material planes and oblate ellipsoids that are parallel to the xy plane (i.e. the pole or short axis is vertical) are metastable and will not rotate during transpressional deformation (Fig. 3b; Passchier 1987; Fossen et al 1994). All other material planes rotate into parallelism with the shear plane in transpression (Fig. 3b; Fossen et al. 1994). The short axes of oblate ellipsoids follow spiral rotational paths until the short axis is horizontal (Fig 3d; Freeman 1985). Whether or not an oblate particle becomes trapped in the shear plane is governed by the vorticity of the deformation and the aspect ratio of the particle (Fig. 3f; Ghosh & Ramberg 1976; Passchier 1987; Wallis 1993). In

STRAIN FROM FELDSPAR PORPHYROCLASTS

271

Fig. 3. Lower hemisphere projection stereonet plots of the rotational paths of deforming material lines (a), poles to material planes (b), the long axes of prolate rigid ellipsoids (c,e), and the short axes of oblate rigid ellipsoids (d,f) in dextral transpression. Although stable orientations exists for elongate markers (c,d) and more equant prolate markers (e), there are no stable orientations for oblate clasts with the lower aspect ratios (f). a & b modified from Fossen et al. (1994).

a pure shear deformation, all oblate ellipsoidal particles become trapped on the shear plane. With increasing simple shear components added to a deformation, the more equant clasts begin to rotate. Consequently, particles with increasingly larger aspect ratios are able to rotate out of the shear plane with increasingly large simple shear components (Passchier 1987; Wallis 1993). This behaviour will provide an important constraint when attempting to interpret kinematic

vorticity from naturally deformed populations of oblate clasts.

Rotation of populations of clasts Quantifying the orientation of populations of clasts in aggregate relies on methods other than measuring the orientation of individual clasts. Two-dimensional studies of shape preferred orientation commonly use the normalized

272

S. GIORGIS & B. TIKOFF

orientation tensor method of Harvey and Laxton (1980) to quantify the degree of preferred orientation of a population (e.g. Ildefonse & Fernandez 1988; Benn & Allard 1989; Tikoff & Teyssier 1994). The degree of shape preferred orientation is related to deformation quantities such as the ratio of finite strain (Tikoff & Teyssier 1994). Extending this same idea from two dimensions to three dimensions requires the quantification of 3D shape preferred orientation and 3D finite strain. As discussed above, shape preferred orientation is quantified using the fabric ellipsoid (Robin 2002). Calculation of the fabric ellipsoid yields the magnitude and orientation of three eigenvectors, similar to those that describe the finite strain ellipsoid. A variety of parameters based on ratios of eigenvalues have been used to quantify the anisotropy and shape of ellipsoids. Jelinek (1981) proposed two useful parameters: the corrected degree anisotropy degree (Pj) and the shape factor (T). Pj is given by:

kinematics. In all cases, three-dimensional finite strain is measured using the octahedral strain factor (Es) of Hossack (1968). The shape of the bulk fabric ellipsoid must be distinguished from that of the individual markers. The anisotropy and shape factor apply to the aggregate, and not an individual rigid marker whose shape remains constant throughout deformation. Maximum anisotropy is reached when all of the markers are perfectly aligned. For example, a perfect alignment of a population of oblate ellipsoids with aspect ratios of 1.8:1.8:1 results in a maximum anisotropy of 1.97, whereas a population of oblate ellipsoids with aspect ratios 2.3:2.3:1 results in a maximum anisotropy of 2.62. Thus, populations with higher aspect ratio clasts can achieve higher degrees of anisotropy. Any deviation from perfect alignment reduces the bulk anisotropy of the population. We will refer to the maximum possible anisotropy that a population of identically shaped clasts can reach as its 'theoretical maximum anisotropy'. In terms of shape fabric, prolate markers that are perfectly parallel will produce a prolate fabric and parallel oblate markers produce an oblate fabric. Populations of prolate or oblate markers can cause either prolate or oblate fabric ellipsoids, depending on where /1? /2, and /3 are the natural logs of the their orientation, and both cause spherical fabric magnitudes of the maximum (a), intermediate ellipsoids if orientated randomly. (b), and minimum (c) axes of the fabric ellipsoid The progressive development of the fabric and /= (/1 + /2+/3) /3 (Jelinek 1981). Pj varies ellipsoid is investigated by examining a popufrom one to infinity and is similar in mathemati- lation of 124 rigid ellipsoids. Each individual cal formulation to the better-known octahedral rigid ellipsoid in the population has the same shear strain used in Hsu plots (Hossack 1968). A aspect ratio (Bi). The most intuitive initial conPj value of one indicates that the ellipsoid has no dition for the orientation of these 124 ellipsoids anisotropy (i.e. it is a sphere). Increasing values is a truly isotropic distribution of the unique axes of Pj indicate greater deviations from a sphere, (Fig. 4). Turner and Weiss (1963) suggest that and a value of infinity describes a material line although truly isotropic primary fabrics are most or a material plane. likely rare, primary fabrics that are randomly The shape factor (T) ranges from -1 to +1, is distributed in a statistically isotropic manner mathematically similar to Lode's parameter may be more common. Therefore we examine used in Hsu plots, and is defined with the same the initial condition of two random populations terms as used for Pj: with uniform distributions in addition to the truly isotropic case. Studies of both single rigid clasts and populations of rigid clasts have demonstrated that the (Jelinek 1981). A perfectly prolate ellipsoid (i.e. rate of rotation of a clast is based on its oriena = b < c) has a T value equal to -1; a perfectly tation, aspect ratio, and the vorticity of the oblate object (i.e. a = b > c) has a rvalue of +1; deformation (e.g. Ghosh & Ramberg 1976; a sphere has a lvalue equal to zero. If a > b > c Jezek et al 1994,1996). The effect of orientation then T lies between 0 and +1; alternatively if a < (Fig. 4), aspect ratio (Fig. 5), and vorticity (Fig. b
STRAIN FROM FELDSPAR PORPHYROCLASTS

273

Fig. 4. (a) Stereonet plot of the initial orientation of an isotropic population and two random distributed, statistically isotropic populations of rigid clasts. Each lineation marks the long axis of a prolate clast or the short axis of an oblate clast. The orientation of random populations was determined using a random number generator, (b) The effect of initial orientation on the progressive development of the anisotropy and shape of the fabric ellipsoid. No significant difference is observed in marker behaviour, except at low strains, for the different populations.

anisotropy of the fabric ellipsoid for initially isotropic populations versus initially random populations of prolate markers (Fig. 4). Slight variation in the shape of the fabric ellipsoid occurs, but the general pattern is the same (Fig. 4). Variation of the aspect ratio of the ellipsoids that constitute the population produces more significant differences (Fig. 5). In each case the initial population is isotropic and subjected to the same vorticity of deformation. An increase in aspect ratio (i.e. an increase in the absolute value of BI) yields an increase in the maximum degree of anisotropy that the population reaches. Populations of prolate rigid markers reach a stable position parallel to the divergent flow

apophysis (z-axis; Fig. 3c,e). Once the long axis of an individual clast is orientated parallel to the z-axis it is only affected by rotation about the zaxis. A prolate marker in such an orientation presents a circular section to the xy plane, therefore z-axis rotation produces no change in the orientation of that clast. Once all of the prolate markers are orientated parallel to the z-axis then the population reaches its maximum anisotropy; the absolute value of this anisotropy is dictated by the aspect ratio of the clasts. Prolate markers with larger aspect ratios rotate faster for any given vorticity of deformation. Therefore, it takes lower amounts of finite strain for high aspect ratio populations to reach their maximum anisotropy than for lower aspect ratio populations (Fig. 5).

274

S. GIORGIS & B. TIKOFF

Fig. 5. The effect of aspect ratio on the progressive evolution of the anisotropy and shape of the fabric ellipsoid. The horizontal axis measures three-dimensional strain in the form of octahedral shear strain. The dark black line indicates the degree of anisotropy and the thinned dashed line is the theoretical maximum anisotropy for any population of markers, with values on the left side of the graphs. The thick dashed line represents the shape factor. More elongate clasts, indicated by higher Bretherton numbers, have a higher theoretical anisotropy. Prolate marker and highly elongate oblate markers asymptotically approach the theoretical maximum anisotropy. Populations of less elongate oblate markers develop cyclical fabrics. A more dramatic effect for populations of oblate markers is the development of cyclical variations in shape. In general, the fabric ellipsoid varies between oblate and prolate shapes, with larger oblate values and longer wavelengths occurring for more elongate clasts. Populations of prolate markers show an initial oblate fabric, which quickly becomes prolate with higher finite strains.

STRAIN FROM FELDSPAR PORPHYROCLASTS

275

Fig. 6. The effect of flow geometry (vorticity) on the progressive development of the anisotropy and shape of the fabric ellipsoid. Same representation as Fig. 5. Populations of prolate markers show a well-behaved pattern, in which the theoretical maximum anisotropy and prolate fabrics are asymptotically approached. Populations of oblate markers with large simple shear components never approach a theoretical maximum anisotropy and the anisotropy development is partly cyclic. Populations of oblate markers show cyclical fabric development of oblate vs. prolate fabric, with shorter wavelengths correlating with increasing simple shear components. Those prolate clasts which are initially orientated with their long axes in the xy plane are in a metastable position. They rotate into the shear

plane but remain horizontal and parallel to the *-axis. This orientation is at 90° to most of the clasts, which are aligned parallel to the z-axis.

276

S. GIORGIS & B. TIKOFF

Consequently, clasts in the metastable position reduce the strength of the anisotropy of the fabric ellipsoid, and prevent the population from reaching its theoretical maximum anisotropy. Varying the vorticity of deformation for a prolate population with a constant aspect ratio has very little effect on the development of the fabric ellipsoid (Fig. 6). Prolate markers have the same stable orientation regardless of the vorticity of the deformation. A greater simple shear component (i.e. larger Wk) leads to irregularities as the anisotropy of the fabric ellipsoid increases. However, the anisotropy of each population reaches a maximum limit with approximately the same amount of deformation.

Populations of oblate clasts Similar to populations of prolate markers with a constant aspect ratio and vorticity, there is no significant difference in the anisotropy of the fabric ellipsoid for initially isotropic population versus initially random populations (Fig. 4). Changes in the aspect ratio of oblate objects produce more complicated patterns. This complexity arises because not all oblate markers reach a stable orientation (Fig. 3f). This lack of a stable orientation for oblate markers is the source of the oscillating degrees of preferred orientation described by Jezek et al. (1996). For a given vorticity, clasts below a critical aspect ratio can rotate 360° (e.g. Ghosh & Ramberg 1976; Passchier 1987; Wallis 1993). Once below this threshold, the amount of deformation needed to rotate a clast out of the shear plane is a function of the aspect ratio: higher aspect ratio clasts need more finite strain to rotate out of the shear plane. An increase in the aspect ratio of a clast orientated parallel to the shear plane leads to a decrease in the rate of z-axis rotation, which approaches zero at some critical aspect ratio. Therefore, the number of oscillations in anisotropy and shape that a population of clasts experiences for a given amount of finite strain is a function of the vorticity of the deformation and the aspect ratio of the clast population. As with a prolate population, an oblate population of clasts fails to reach its theoretical maximum anisotropy (Fig. 5). Two factors prevent oblate populations from achieving their maximum anisotropy. First, oblate markers which are orientated parallel to the xy plane (i.e. vertical pole) are metastable. Note that in the initial isotropic population (Fig. 4) there are no objects in this orientation. Secondly, because clasts are free to rotate 360° around the z-axis, all clasts are not orientated parallel to the shear plane during an increment of deformation.

Changes in the oscillation pattern of the anisotropy and shape of the fabric ellipsoid are observed for variations in the vorticity of the deformation imposed on a population of oblate markers (Fig. 6). For a constant aspect ratio, an increase in the pure shear component (i.e. lowering Wk) reduces the frequency of oscillations and increases the maximum anisotropy that a population reaches. Pure shear traps the clasts parallel to the shear plane. Therefore, with decreasing amounts of simple shear, the rate of rotation of clasts in the shear plane approaches zero. This increased stability from the pure shear component leads to an increase in the degree of anisotropy of the fabric ellipsoid with decreasing Wk values. For the case of oblate clasts with aspect ratios of 1.75:1.75:1 in a pure shear dominated transpressional deformation (Wk = 0.15 or 0.31), all of the clasts in the population become perfectly aligned and the degree of anisotropy of the fabric ellipsoid reaches it theoretical maximum (Fig. 6). The shape of the fabric ellipsoid also varies consistently with changes in the vorticity of the deformation. Shape variations have the same wavelength as anisotropy variations and are apparent even where the anisotropy oscillations have low amplitudes (Figs 5 & 6). For a population of oblate ellipsoids, transpression with a large component of pure shear (i.e. low Wk) results in highly aligned populations that reach a stable orientation. This deformation history yields a fabric ellipsoid that is oblate and remains oblate throughout deformation (Fig. 6). Transpression with a large simple shear component (i.e. high Wk) yields non-stable oscillating fabric ellipsoids; the shape of these fabric ellipsoids ranges from prolate to oblate (Fig. 6). Thus the presence of a prolate fabric ellipsoid measured from a population of oblate markers requires a significant simple shear component to the deformation and can be used to constrain the lower bounds of possible Wk values.

Orientation of fabric ellipsoid with respect to shear sense There is no simple relation between the orientation of the fabric ellipsoid and the overall sense of shear determined by the simple shear component of deformation. Shear sense indicators are best defined on the plane that is orientated normal to the vorticity vector, which is the horizontal plane (xy) in transpression. Intuitively, the distribution of a population of elliptical markers in this plane (xy) should have an asymmetry which is orientated synthetically

STRAIN FROM FELDSPAR PORPHYROCLASTS

277

with respect to the simple shear component of deformation (Fig. 7a). Figure 7b illustrates the progressive orientation of the two-dimensional expression of a population of oblate ellipsoids in the horizontal plane of dextral transpressional deformation. At various points during the progressive deformation the average orientation of the population, corresponding to the long axis of the fabric ellipsoid in the horizontal plane, may be orientated synthetically or antithetically to the simple shear component of deformation. In fact, any orientation for the long axis of the fabric ellipsoid is possible under these conditions. In three dimensions, this relationship applies to all oblate ellipsoids and to prolate ellipsoids that are stuck in the metastable xy plane. As relatively few prolate ellipsoids will lie in the xy plane, this effect may be negligible for prolate markers. However, this effect will be significant for oblate clasts with a low enough aspect ratio to rotate in a given flow (i.e. Wk value). Consequently, this effect suggests that fabric orientation as described by populations of rigid markers does not provide a good shear sense indicator for transpressional deformation.

Fabric ellipsoid versus the finite strain ellipsoid Numerical modelling suggests that there is a complex relationship between the development of the fabric ellipsoid and the strain ellipsoid. A direct correlation between the increase in anisotropy of the fabric ellipsoid with increasing amounts of finite strain only holds true at lower strains. At higher strains the aspect ratio of the clasts and the vorticity of the deformation can either cause the anisotropy of the fabric ellipsoid to plateau or oscillate (Figs 5,6 & 8a). The finite strain ellipsoid, however, steadily increases in magnitude and maintains a consistent shape (Fig. 8b). Therefore, there is no direct correlation between the shape and anisotropy of the fabric ellipsoid and the finite strain ellipsoid. Fig. 7. (a) Schematic diagram of a population of ellipses orientated synthetic to the simple shear component of deformation, (b) Rose diagram plots of the orientation of a population oblate clasts (n = 124) as expressed on the horizontal plane (xy) in dextral transpression. The average orientation of the clasts, a proxy for the orientation of the fabric ellipsoid in the xy plane, is both synthetic (Es = 3,5,10) and antithetic (Es = 1,8) to the simple shear component of deformation. In all cases the outer circle of the rose diagrams is 30%.

Application to the western Idaho shear zone, USA Introduction The western Idaho shear zone is a mid-crustal exposure of an intra-arc shear zone located in the Cordillera of the western United States (Fig. 9; McClelland et al 2000). This high strain zone

278

S. GIORGIS & B. TIKOFF

Fig. 8. (a) Progressive evolution of the shape (T, the tangential component) and anisotropy (P/, the radial component) of the fabric ellipsoid calculated from a population of oblate clasts (aspect ratio of 1.75:1.75:1) subjected to transpressional deformation (Wk = 0.66). A slightly prolate fabric with higher anisotropy alternates with a oblate fabric with lower anisotropy at high finite strains, (b) Progressive evolution of the finite strain ellipsoid shape (Lode's parameter, the tangential component) and magnitude (octahedral shear strain, the radial component) for the same deformation. The finite strain ellipsoid lies within the flattening field and the octahedral strain continues to accumulate with higher finite strains. Comparison of the graphs indicates a lack of correlation between the fabric ellipsoid and the finite strain ellipsoid.

Fig. 9. Location of the western Idaho shear zone in the North American Cordillera. The shear zone parallels the Sr 0.706 line, which marks the boundary between cratonic North America and the accreted terranes to the west. CA, California; ID, Idaho; MT, Montana; NV, Nevada; OR, Oregon; WA, Washington; WY, Wyoming.

developed in the Late Cretaceous and deformed a series of mid- to Late Cretaceous granitic intrusive suites (Manduca et al 1993; Fig. lOa). The Payette River tonalite (92.3 ±1.2 Ma, U-Pb zircon, McClelland pers. comm. 2003) is the westernmost unit of the Idaho Batholith suite and is the youngest unit deformed by the shear zone (Manduca et al. 1993). A subvertical, north-south striking foliation with a down dip lineation characterizes the fabric of the western Idaho shear zone. The presence of rare shear

sense indicators on the lineation normal face, in addition to the foliation and lineation orientations, indicate that this shear zone was a predominately dextral, transpressional feature (McClelland et al. 2000; Tikoff et al 2001). The central intrusive suite, the Little Goose Creek Complex (105.9 ± 2.0 Ma, U-Pb zircon, McClelland pers. comm. 2003) contains most of the western Idaho shear zone and consists predominantly of a mylonitic potassium feldspar megacrystic orthogneiss (Manduca et al. 1993). The K-feldspar porphyroclasts (1-5 cm in length) are contained in a matrix of fine grained (<0.5-1.5 mm) plagioclase, quartz, and biotite (Fig. 11). In thin section, individual quartz grains demonstrate blocky extinction and show no evidence of subgrain formation. Quartz-quartz grain boundaries are characterized by 120° junctions or bulbous geometries suggesting high rates of recovery and grain boundary migration, respectively. This suite of quartz microstructures is characteristic of high temperature deformation (Regime 3; Hirth & Tullis 1992). The granitic units deformed by the western Idaho shear zone do not contain piercing points, finite strain markers, or vorticity markers; thus very little is known about the magnitude of finite strain or the vorticity of deformation. We apply the numerical model discussed above to the feldspar megacryst populations of the Little Goose Creek Complex to gain some insight into the magnitude of finite strain and the kinematics of deformation.

STRAIN FROM FELDSPAR PORPHYROCLASTS

279

Fig. 10. (a) Geological map of the study area. The 4 km width of the western Idaho shear zone is shown across the bottom of the map with a dark line. The shear zone is contained almost entirely within the Little Goose Creek Complex (modified from Manduca 1988). (b) Distribution of the long axis of the fabric ellipsoid in the xy plane measured from the feldspar megacrysts of the Little Goose Creek Complex. The length of each bar is proportional to the anisotropy measured at that station. A filled bar indicates that the fabric elliopsoid had an oblate shape; a hollow bar indicates a prolate fabric ellipsoid.

Assumptions Application of the theoretical model to field data requires that all of the assumptions behind the theoretical model are reasonable approximations of the deformation. The assumptions underlying our model of rigid clast rotation are: (1) deformation is monoclinic, homogeneous,

and kinematic-steady state (i.e. Wk does not vary through time); (2) the clasts are perfectly rigid, perfectly oblate ellipsoids, and do not interact; and (3) the grain size of the matrix should be significantly finer grained than the clasts being modelled in order to facilitate homogeneous flow, as noted by Passchier (1987). The validity of each of these assumptions with respect to the

280

S. GIORGIS & B. TIKOFF

Fig. 11. (a) An outcrop photo showing the K-feldspar megacrysts of the Little Goose Creek Complex. The outcrop face is perpendicular to foliation, (b) A polished slab of the Little Goose Creek Complex, (c) A thinsection photomicrograph of the same feldspar crystal shown in (b). Note the lack of a recrystallized mantle, (d) A close up of the rim of the feldspar crystal shown in (b) and (c). The matrix material is much finer grained than the feldspar porphyroclast and maintains a consistent size moving away from the margin of the crystal.

feldspars of the Little Goose Creek Complex and deformation in the western Idaho shear zone is discussed below. The homogeneity and steady versus nonsteady nature of deformation is best assessed where a strain gradient is present. The fabric of the youngest unit that intrudes the western Idaho shear zone, the Payette River Tonalite, records a solid-state strain gradient. The western side of this unit is characterized by a strong solidstate foliation, whereas the eastern side is characterized by weak magmatic fabrics. Across this strain gradient, foliation is consistently north-south striking and steeply east dipping with a down dip lineation. Once the effects of overprinting Miocene extensional deformation are removed, the foliation is subvertical and north-south striking with a subvertical lineation (Tikoff et al 2001). Subvertical foliations and lineations are characteristic of monoclinic transpressional flow (e.g. Tikoff & Green 1997). The consistent fabric patterns across the strain gradient suggest steady-state flow, but do not conclusively rule out non-steady behaviour. In the absence of further information to constrain the deformation path of the western Idaho shear zone, steady-state kinematics is the simplest and

most reasonable assumption. On the outcropscale, high strain zones are observed locally where sufficient markers are present. However, they are generally difficult to recognize in the porphyritic granite. We assume that the vorticity in those high strain zones, if present, is similar to the vorticity of the shear zone as a whole. To a first order, these observations are consistent with a monoclinic, homogeneous, and steady state transpressional deformation. The feldspars of the Little Goose Creek Complex are elliptical to tabular in cross section and show remarkably little evidence of microstructural scale deformation (Fig. 11). These porphyroclasts are much coarser grained than the matrix and typically have not developed mantles (Fig. 11). Due to their undeformed nature, we interpret these grains to be primary phenocrysts that behaved as rigid markers during solid-state deformation. Individually, these megacrysts are roughly oblate in shape and the cross sectional aspect ratios measured on three nearly orthogonal outcrop planes are on average 1.8 ± 0.2 (Table 1). Therefore, on average this population of grains is within approximately ±10% of being perfectly oblate. Studies of feldspar megacrystic granites in the

281

STRAIN FROM FELDSPAR PORPHYROCLASTS Table 1. Summary of feldspar shape preferred orentation data Average aspect ratio Station 101-03 101-7 101-8 101-9 101-10 101-19 101-57 101-60 101-60.5 101-61 101-62 101-64 101-66 101-68 101-69 101-70 IJ99-106 IJ99-108 IJ99-109 IJ99-110 IJ99-111 IJ99-112 IJ99-113

n

Pj

T

% Error

Face A

FaceB

FaceC

Modelled Es

76 88 60 84 70 88 83 105 108 61 86 63 37 70 67 92 126 114 132 132 132 132 132

1.20 1.63 1.31 1.58 2.83 1.50 1.73 1.71 1.62 2.05 1.36 1.74 1.81 1.60 1.36 1.40 1.26 2.51 1.41 1.86 1.79 1.25 1.27

0.11 0.04 -0.37 -0.67 -0.85 -0.20 -0.12 0.63 0.16 0.06 0.74 0.30 -0.35 0.13 0.11 0.35 0.15 0.67 -0.20 0.75 0.08 0.17 -0.22

21.30 17.64 22.60 17.04 18.45 19.68 21.43 14.68 19.82 16.27 21.56 20.10 19.53 22.17 24.88 19.06 22.82 15.71 22.94 16.73 19.96 25.48 25.63

1.70 1.66 1.79 1.72 1.54 1.83 1.98 1.76 2.04 1.87 1.91 1.83 2.07 1.89 2.05 1.67 1.67 1.60 1.80 1.59 1.86 1.94 1.78

1.56 1.97 1.99 1.72 1.92 2.10 2.51 2.00 2.04 2.04 1.78 1.96 1.62 2.21 1.68 1.81 1.76 2.61 1.95 1.96 2.01 1.98 2.06

1.51 1.45 1.41 1.35 1.43 1.68 1.62 1.53 1.40 1.41 1.25 2.05 2.09 1.75 1.82 1.41 1.85 2.34 2.01 1.79 1.84 1.88 2.01

1.0 4.6 1.6 3.7 2.9 7.7 6.6 4.3 2.0 7.7 7.7 4.0 2.0 2.1 1.4 2.2 7.7 7.7 1.4 1.5

Average:

1.81

1.97

1.69

4.0

n, number of grain measured; Pj, corrected degree of anistropy; T, shape factor; % Error, Frombis norm error (Robin 2002); Face A, lineation normal; Face B, lineation parallel & foliation normal; Face C, foliation parallel; Es, octahedral strain.

Sierra Nevada have yielded similar aspect ratios (approximately 1.7; Lockwood 1975; Tikoff & Teyssier 1994). Thus, although the individual clasts are triclinic in detail, characterizing the population as perfectly oblate ellipsoids is a fair approximation. Lastly, clast-clast interaction was avoided by rejecting closely spaced or tiled clasts during data collection for the shape preferred orientation analysis.

Quantification of field data Due to alpine glaciation, there is excellent exposure in the mountains of this part of western Idaho. Outcrops commonly weather along three faces: a shallowly west dipping surface that is often glacially polished (horizontal plane); a steeply east dipping face (foliation plane); and an east-west, vertical joint set (vertical plane). These faces are almost ideally orientated with respect to foliation and lineation. The first face is orientated parallel to foliation, the second parallel to lineation, normal to foliation, and the third is normal to both foliation and lineation.

Weathering has produced a high degree of colour contrast between the K-feldspar megacrysts and the matrix on the horizontal plane, creating a face that is ideally suited for image analysis. The vertical plane and foliation plane are typically fresher faces with a much lower contrast between the feldspars and the matrix, making image analysis more difficult. The presence of suitable vertical planes and foliation faces is the limiting factor in collecting orientation data in the Little Goose Creek Complex. Shape preferred orientation of K-feldspar megacrysts was determined following the techniques originally suggested by Launeau et al. (1990) and later modified by Davis (1997). Images of each of these three faces were collected using a portable video camera. On each face, the strike and dip of that face was marked on the outcrop. The camera was orientated normal to the surface and 15-100 feldspar grains were measured per face. Only isolated grains were measured to eliminate the effect of grain-grain interaction or tiling. The horizontal

282

S. GIORGIS & B. TIKOFF

plane typically provided the highest number of grains to measure, whereas the lower contrast on the other faces resulted in fewer measurable grains. Each image was then imported onto a computer and analysed using NIH Image software (v. 1.62). NIH Image was used to measure the length of the long axis and the length of the short axis of each feldspar grain. The orientation (rake) of the long axis of each grain was measured relative to the strike of the face. A fabric ellipsoid was then calculated from these sectional data. These calculations were performed using the software program 'Ellipsoid 2001' written by Launeau and Robin, which uses the mathematics outlined in Robin (2002). Results The numerical model constrains both the magnitude of finite strain and the vorticity of deformation in the western Idaho shear zone. The measured anisotropies and shape factors of the feldspar fabric ellipsoids are compared to the modelled values for an oblate clast with an aspect ratio of 1.75: 1.75: 1 (hereafter abbreviated as an aspect ratio of 1.75). A cross sectional aspect ratio of 1.8 for the feldspar megacrysts of the Little Goose Creek Complex translates into a theoretical maximum anisotropy of 1.98. Examination of the fabric ellipsoid data (Table 1) indicates that three stations (IJ99-108, 101-10, and 101-61) have anisotropies that are significantly higher than this value. These values require that many of the individual clasts have aspect ratios greater than 1.8. The larger average aspect ratios measured at these stations indicate that it is inappropriate to apply the results of a numerical model based on an aspect ratio of 1.75. Station IJ99-108, for example, has an average aspect ratio of 2.61, which is significantly higher than 1.8. Consequently, these outliers are not considered in our analysis of the bulk deformation of the shear zone. There is no clear gradient in anisotropy across the western Idaho shear zone (Fig. lOb). Moreover, the projection of the fabric ellipsoid onto the horizontal plane is not parallel to the north-south orientated trace of the field foliation (Fig. lOa). An explanation for this discrepancy lies in the what defines the field foliation: micas and elongate quartz grains. The shape of mica grains is much closer to that of a material plane and therefore the micas will become trapped in the shear plane (Fig. 3b). The ribbon shape of the quartz grains indicates that these grains are passive markers and therefore may approximate some portion of the orientation and shape of the

finite strain ellipsoid. For high strains in transpressional deformation, a flattening shape subparallel to the shear zone boundaries is expected, which is consistent with the observed quartz fabrics. In contrast, the fabric recorded by the feldspar megacrysts is less restricted because of the oblate and less elongate nature of the feldspars. Consequently, depending on the aspect ratio of the oblate markers and the vorticity of the deformation, the field foliation and the fabric ellipsoid may or may not be parallel. The feldspar megacryst populations do not have a consistent sense of asymmetry with respect to the shear zone boundary (Fig. lOb). The highest anisotropies are orientated within 30° of the shear zone boundary, but there is no clear pattern otherwise. These data are consistent with our modelling, which predicts that the trace of the fabric ellipsoid will rotate in the horizontal plane (xy). These observations are consistent with the interpretation of the western Idaho shear zone as a transpressional shear zone. We will now consider how these data can be used to constrain the vorticity of deformation by examining three scenarios. First, transpressional deformation may have a large pure shear component (i.e. Wk < 0.31; Fig. 6). In this case, Pj values from all stations are very close to or equal to the theoretical maximum anisotropy and the shape fabric is oblate. In the second scenario, transpression has a large simple shear component (i.e. Wk > 0.66; Fig. 6). This case is distinguished by maximum Pj values that are significantly lower than the theoretical maximum anisotropy and shape fabrics that are mixed prolate and oblate shapes. The last possibility occurs for intermediate values of Wk (0.31 < Wk < 0.66; Fig. 6). These cases are characterized by anisotropies close to the theoretical maximum and shape factors with mixed prolate and oblate fabrics. The data from the western Idaho shear zone indicate intermediate values of Wk for the deformation. The Pj values of some stations approach 1.98, but many are significantly below that value (Table 1). The shape fabrics are prolate for approximately one third of the stations and oblate for the other two thirds. Additionally, the dispersed distribution of orientations suggest that this is not pure shear dominated deformation (Fig. lOb). Recall that the pure shear component of deformation tends to trap oblate clasts parallel to the shear plane; thus for the average orientation of clasts to be both synthetic and antithetic to the shear sense (Fig. 7) the clasts must be free to rotate 360°. Taken together, all of these observations suggest that the value of Wk is

STRAIN FROM FELDSPAR PORPHYROCLASTS

283

The simple shear component, applied across the initial width of the zone translates to about 50 km of lateral offset. A pure shear strain of 14 integrated across the shear zone implies a large amount of vertical extrusion. This component suggests that the exposed shear zone is a deepcrustal expression of a transpressional flower structure. The pure shear component also indicates a large (92%) amount of shortening normal to the shear zone, which is equivalent to 50 km of contraction.

Conclusions Fig. 12. Plot of anisotropies measured from field data against modelled anisotropies for a transpressional deformation with a Wk = 0.5. Error bars reflect the misfit between the calculated fabric ellipsoid and the 2D measured sectional data (see Robin 2002).

between 0.31 and 0.66. Thus the angle of oblique convergence in transpression would be between 30° and 60° (Fossen et al 1994). With no further constraints on vorticity we will analyse the case in which Wk = 0.5. For transpression this value indicates approximately equal components of pure shear and simple shear, or an angle of oblique convergence of 45° (Fossen et al. 1994). With an estimate of the vorticity of deformation it is possible to constrain the magnitude of finite strain using a plot of Pj against Es (Fig. 12; Table 1). This plot yields an average octahedral shear strain of about 4.0. Applying the same data to a Pj vs. Es graph for an angle of oblique convergence of 30° yields an average Es of about 3.2; an angle of oblique convergence of 60° yields an average Es of about 4.8. We emphasize that these are minimum estimates; the larger Pj values could indicate significantly higher strains given the cyclic nature of fabric formation. We use these estimates to make preliminary tectonic estimates using the kinematic information recorded by the western Idaho shear zone. We will assume that the western Idaho shear zone is characterized by a transpressional deformation with a Wk = 0.5 and Es = 4. If strain accumulated by material steady-state deformation (i.e. the Wk does not change during deformation), this corresponds to a simple shear component (y) of about 2.5 and a pure shear component (k) of 14. The western Idaho shear zone is approximately 4 km wide in the area immediately north of McCall, Idaho (Fig. 10).

Model conclusions The shape of the fabric ellipsoid is a function of the aspect ratio of the individual clasts and the vorticity of deformation. A theoretical maximum anisotropy exists for any population of markers and its value depends on the aspect ratio of the individual markers. The theoretical maximum anisotropy is equal to the shape of a single clast (in a population composed of clasts with identical aspect ratios) and occurs when all of the markers in the sampled population are perfectly aligned. Populations of prolate spheroids rotate into a stable orientation parallel to the divergent flow apophysis and therefore will not produce oscillating fabrics regardless of the vorticity. Populations of oblate spheroids with an aspect ratio below some critical value for a given vorticity of deformation produce fabric ellipsoids that oscillate with respect to both anisotropy and shape. Consequently, perfect alignment of oblate ellipsoids with low aspect ratios is achieved only with large amounts of pure shear. The relationship between the measured anisotropy, the maximum theoretical anisotropy, and the shape of the fabric ellipsoid provides a qualitative estimate of the relative contributions of pure shear and simple shear to the deformation history. Populations of oblate spheroids subjected to transpression with a large simple shear component produce fabric ellipsoids with maximum anisotropies well below the theoretical maximum and shapes fabrics that oscillate from prolate to oblate. Transpression with a large pure shear component results in anisotropies that approach the theoretical maximum and shape fabrics that are only oblate. Transpression with sub-equal amounts of pure shear and simple shear produces intermediate results: the anisotropy of the

284

S. GIORGIS & B. TIKOFF

fabric ellipsoid can be near that of the theoretical maximum, but the shapes of the fabric ellipsoid range from prolate to oblate.

Field conclusions The anisotropy, shape, and orientation of fabric ellipsoids measured from populations of potassium feldspar porphyroclasts are heterogeneous across the western Idaho shear zone. No discernable gradient in the anisotropy of feldspar fabrics occurs across the shear zone. Both sythetic and antithetic senses of asymmetry of the clast populations relative to the shear zone boundary are observed in the western Idaho shear zone, which is consistent with the numerical simulation. An intermediate vorticity for the western Idaho shear zone is suggested by two observations: the obliquity between the orientation of field foliation and the orientation of the fabric ellipsoid foliation; and the high degrees of anisotropy of the fabric ellipsoids (many close to the theoretical maximum for clasts with an aspect ratio of 1.75:1.75:1) and the mixed prolate and oblate shape fabrics. These observations are consistent with a highly variable orientation of the fabric ellipse within the horizontal plane. Using an intermediate value for the vorticity, comparison of the numerical model results to the field data allows a first-order tectonic reconstruction of the western Idaho shear zone. The zone is interpreted to have experienced approximately 50 km of lateral offset, and a significant amount of contractional deformation and vertical extrusion.

mplications for shear zone studies If rigid oblate markers are present, the anisotropy and shape factor of the fabric ellipsoid can be used to constrain the vorticity of a shear zone. Given an estimate of vorticity, the minimum amount of finite strain can be constrained using the anisotropy of a fabric ellipsoid measured from either prolate or oblate populations of clasts. The asymmetrical distribution of populations of porphyroclasts may be orientated at any angle with respect to the shear zone boundary. Therefore pophyroclasts without tails are an unreliable shear sense indicator. There is no direct correlation between the shape and anisotropy of the fabric ellipsoid and strain ellipsoid.

We would like to thank John Gillaspy who collected the initial three-dimensional feldspar orientation data set. Reviews by C. Passchier and D. Jiang improved both the quality and the clarity of the manuscript. The research was supported by NSF grant EAR 0001092.

References ARBARET, L., FERNANDEZ, A., JEZEK, J., ILDEFONSE, B., LAUNEAU, P. & DIOT, H. 2000. Analogue and numerical modeling of shape fabrics: application to strain and flow determination in magmas. Transactions of the Royal Society of Edinburgh: Earth Sciences, 90, 97-109. BENN, K. & ALLARD, B. 1989. Preferred mineral orientations related to magmatic flow in ophiolite layered gabbros. Journal of Petrology, 30, 925-946. BOBYARCHICK, A.R. 1986. The eigenvalues of steady flow in Mohr space. Tectonophysics, 122, 35-51. BOWMAN, F. 1961. Introduction to elliptic functions, with applications. New York, Dover, 115 pp. BRETHERTON,FP. 1962. The motion of rigid particles in a shear flow at low Reynolds number. Journal of Fluid Mechanics, 14, 284-301. DAVIS, M.R. 1997. Anisotropy of magnetic susceptibility and fabric analysis in granitoids: the Cascade Lake shear zone, Sierra Nevada, California. M.S. Thesis, University of Minnesota, 95 pp. FLINN, D. 1962. On folding during three-dimensional progressive deformation. Quarterly Journal of the Geological Society of London, 118, 358-428. FOSSEN, H., TIKOFF, B. & TEYSSIER, C.T. 1994. Strain modeling of transpression and transtensional deformation. Norsk Geologisk Tidsskrift, 74, 134-145. FREEMAN, B. 1985. The motion of rigid ellipsoidal particles in slow flows. Tectonophysics, 113, 163-183. GAY, N.C. 19680. Pure shear and simple shear deformation of inhomogeneous viscous fluid. 1. Theory. Tectonophysics, 5, 211-234. GAY, N.C. 19686. The motion of rigid particles embedded in a viscous fluid during pure shear deformation of the fluid. Tectonophysics, 5, 81-88. GHOSH, S.K. & RAMBERG, H. 1976. Reorientations of inclusion by combinations of pure shear and simple shear. Tectonophysics, 17,133-175. HARVEY, PK. & LAXTON, R.R. 1980. The estimation of finite strain from the orientation distribution of passively deformed linear markers: eigenvalue relationships. Tectonophysics, 70, 285-307. HIRTH, G. & TULLIS, J. 1992. Dislocation creep regimes in quartz aggregates. Journal of Structural Geology, 14,145-159. HOSSACK, J. 1968. Pebble deformation and thrusting in the Bygdin Area (Norway). Tectonophysics, 5, 315-339. ILDEFONSE, B. & FERNANDEZ, A. 1988. Influence of the concentration of rigid markers in a viscous medium on the production of preferred orientations. An experimental contribution, 1. Noncoaxial strain. Bulletin of Geology of the University of Uppsala, N.S., 14, 55-60.

STRAIN FROM FELDSPAR PORPHYROCLASTS ILDEFONSE, B., SOKOUTIS, B. & MANCKTELOW, N.S. 1992. Mechanical interactions between rigid particles in a deforming matrix. Analogue experiments in simple shear flow. Journal of Structural Geology, 14,1253-1266. JEFFERY, J.B. 1922. The motion of ellipsoidal particles immersed in viscous fluid. Proceeding of the Royal Society of London, A102,161-179. JELINEK, V. 1981. Characterization of the magnetic fabrics of rocks. Tectonophysics, 79, T63-T67. JEZEK, J., MELKA, R., SCHULMANN, K. & VENERA, Z. 1994. The behavior of rigid triaxial particles in viscous flows - modeling of fabric evolution in a multiparticle system. Tectonophysics, 299, 165-180. JEZEK, J., SCHULMANN, K. & SEGETH, K. 1996. Fabric evolution of rigid inclusions during mixed coaxial and simple shear flows. Tectonophysics, 257, 203-221. LAUNEAU, P., BOURCHEZ, J.L. & BENN, K. 1990. Shape preferred orientation of object populations: automatic analysis of digitized images. Tectonophysics, 180, 201-211. LOCKWOOD, J.P. 1975. Mount Abbot Quadrangle, Central Sierra Nevada, California - Analytical data. U.S. Geological Survey Professional Paper, 744C. MALVERN, L.E. 1969. Introduction to the mechanics of a continuous medium. Prentice Hall, Inc., Englewood Cliffs, New Jersey, 713 pp. MANCKTELOW, N.S., ARBARET, L. & PENNACCHIONI, G. 2002. Experimental observations on the effect of interface slip on rotation and stabilization of rigid particles in simple shear and a comparison with natural mylonites. Journal of Structural Geology, 24, 567-585. MANDUCA, C.A. 1988. Geology and geochemistry of the oceanic arc-continental boundary in the western Idaho Batholith near McCall Ph.D. Thesis, California Institute of Technology, Pasadena, 272 pp. MANDUCA, C.A., KUNTZ, M.A. & SILVER, L.T. 1993. Emplacement and deformation history 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. MARCH, A. 1932. Mathematische Theorie der Regelung nach der Korngestalt bei affiner Deformation. Zeitschrift Kristallographic, 81, 285-297. MCCLELLAND, W.C., TIKOFF, B. & MANDUCA, C.A. 2000. Two-phase evolution of accretionary margins: examples from the North American Cordillera. Tectonophysics, 326, 37-55.

285

MERLE, O. 1986. Patterns of stretch trajectories and strain rates within spreading-gliding nappes. Tectonophysics, 124, 211-222. OWENS, W.H. 1984. The calculation of a best-fit ellipsoid from elliptical sections on arbitrarily orientated planes. Journal of Structural Geology, 6, 571-578. PASSCHIER, C.W. 1987. Stable positions of rigid objects in non-coaxial flow - a study in vorticity analysis. Journal of Structural Geology, 9, 679-690. PASSCHIER, C.W. 1990. Reconstruction of deformation and flow parameters from deformed vein sets. Tectonophysics, 180,185-199. ROBIN, P.F. 2002. Determination of fabric and strain ellipsoids from measured sectional ellipses theory. Journal of Structural Geology, 24, 531-544. SANDERSON, D.J. & MARCHINI, W.R.D. 1984. Transpression. Journal of Structural Geology, 6, 449-458. TIKOFF, B. & FOSSEN, H. 1993. Simultaneous pure and simple shear: the unifying deformation matrix. Tectonophysics, 217, 267-283. TIKOFF, B. & GREEN, D. 1997. Stretching lineations in transpressional shear zones. Journal of Structural Geology, 19, 29-40. TIKOFF, B. & TEYSSIER, C. 1994. Strain and fabric analysis based on porphyroclast interaction. Journal of Structural Geology, 16, 477-491. TIKOFF, B., KELSO, P., MANDUCA, C, MARKLEY, M.J. & GILLASPY, J. 2001. Lithospheric and crustal reactivation of an ancient plate boundary: the assembly and disassembly of the Salmon River suture zone, Idaho, USA. In: HOLDSWORTH, R.E., STRACHAN, R.A., MAGLOUGHLIN, IF. & KNIPE, R.J. (eds) The nature and tectonic significance of fault zone weakening. Geological Society, London, Special Publications, 186, 213-231. TRUESDELL, C. 1953. Two measures of vorticity. Journal of Rational Mechanical Analysis, 2, 173-217. TURNER, F.J. & WEISS, L.E. 1963. Structural analysis of metamorphic tectonites. McGraw-Hill, New York, 545 pp. WALLIS, S. 1993. Vorticity analysis and recognition of ductile extension in the Sanbagawa belt, SW Japan. Journal of Structural Geology, 17, 1077-1093. WILLIS, D.G. 1977. A kinematic model of preferred orientation. Geological Society of America, 88, 883-894.

This page intentionally left blank

Strain removal within the Hercynian Shear Belt of Central Brittany (western France): methodology and tectonic implications C. GUMIAUX, J. P. BRUN & D. GAPAIS Geosciences Rennes, UMR 6118 CNRS, Campus de Beaulieu, Universite de Rennes 1, 35042 Rennes cedex, France (e-mail: [email protected]) Abstract: Central Brittany comprises a major shear belt formed during the Hercynian Orogeny. In this paper, we present a model for the restoration of the belt, based on an exhaustive compilation of available structural data. The model uses a geostatistical analysis of cleavage orientation data. The analysis shows that the eastern part of Central Brittany has been deformed by bulk dextral strike-slip, a feature that validates previous structural interpretations. It also provides explanations for: the strain patterns observed within and around the restored domain; the nature of domain boundaries and of associated kinematics; and the occurrence of local superimposed deformations.

Central Brittany is bounded by the North Armorican Shear Zone (NASZ) to the north, and by the South Armorican Shear Zone (SASZ) to the south (Fig. 1). The SASZ (Jegouzo 1980) is one of the main tectonic features of the Hercynian belt of Western Europe that separates two domains of contrasted structural style and deformation history (Fig. 1). Central Brittany is mostly composed of Late Proterozoic to Upper Palaeozoic sediments affected by anchizonal to greenschist metamorphism, and deformed during Carboniferous times (see Le Corre et al 1991, and references therein). Deformation, which is closely related to shearing along the South Armorican Shear Zone, corresponds to a regional-scale dextral wrenching (Gapais & Le Corre 1980). Granitic plutons, that were emplaced during regional shearing (Hanmer etal 1982; Vigneresse & Brun 1983; Vigneresse 1987), are mylonitized along the South Armorican Shear Zone (Fig. 1), producing the C/S fabrics first identified and described by Berthe etal (19790, b). To the south, the domain of South Brittany displays a long and complex tectonic history. Crustal thickening and subsequent thermal relaxation led to partial melting from Late Devonian to Carboniferous times (Le Metour & Audren 1977; Brown & Dallmeyer 1996). As a result of crustal thickening, high-pressure metamorphic rocks (including blueschists and eclogites) were exhumed during early Carboniferous times (Bosse et al. 2000, 2002; Le Rebel etal. 2002). Then, synconvergence extension during the Upper Carboniferous was accommodated locally by detachment zones allowing the melted lower crust to rise up, forming migmatite domes (Gapais et al 1993).

Fig. 1. Major structural-metamorphic domains of the Armorican Massif, resulting from contrasted evolutions during Hercynian collision. SASZ and NASZ are South and North Armorican Shear Zone, respectively.

The present paper gives first a summary of the large data base of structural observations and measurements collected in Central Brittany by a number of workers during the last 30 years. A geostatistical analysis of cleavage data is carried out in order to constrain a regional-scale strain model. A simple shear model is then used to restore the bulk deformation in the Eastern part of Central Brittany. Geological and tectonic implications, within and at the boundaries of the restored domain, are discussed.

From: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 287-305. 0305-8719/$15.00 © The Geological Society of London 2004.

288

C. GUMIAUX ET AL.

Fig. 2. Simplified geology of Central Brittany (cross-section from Le Corre 1977).

Geological setting Lithologies The stratigraphy of Central Brittany (Fig. 2) starts with the Brioverian, made of monotonous pelites and sandstones, and interpreted as the erosional product of the nearby relief of the Cadomian chain to the north (Le Corre 1977; Chantraine et al. 1994). It is Upper Proterozoic to Cambrian in age (Chantraine et al 1994). According to Guillocheau and Rolet (1982) and Robardet et al (1994), the history of subsequent Palaeozoic sedimentation can be summarized as follows. Lower Ordovician sandstones, representing a marine transgression, lie unconformably on top of the Brioverian. Up to the Middle Devonian, marine sedimentation was

continuous, with alternating deposits of pelites and sandstones. During the Upper Devonian, sedimentation is characterized by rapid vertical and horizontal variations, with local erosion, and formation of independent sub-basins during lowermost Carboniferous times. During the Carboniferous, the sedimentation was restricted to the borders of Central Brittany, at the vicinity of the main fault zones (mainly in the basins of Laval and Chateaulin, Fig. 2). Several syntectonic plutons were emplaced along the major Armorican shear zones (Fig. 2), most Rb/Sr whole-rock ages ranging between 340 and 290 Ma (Vidal 1976; Peucat et al 1979; Bernard-Griffiths et al 1985) (see also compilation by Le Corre et al 1991). Along the SASZ, most pluton shapes are of laccolith type (Vigneresse 1978) and emplaced at shallow levels in the

STRAIN REMOVAL IN CENTRAL BRITTANY

upper crust (Hanmer & Vigneresse 1980; Vigneresse & Brun 1983; Roman-Berdiel et al 1997). Calc-alkaline granites of Ordovician age (Lanvaux orthogneiss) (Vidal 1972; Vidal 1980; Jegouzo et al 1986) occur along the northern branch of the SASZ (Cogne 1974; Bouchez & Blaise 1976), and are recognized further east by gravimetric data below the Mesozoic cover of the Paris Basin (Weber 1967).

Structures The two main shear zones bounding Central Brittany, the SASZ and the NASZ, display the same right lateral sense of shear (Berthe et al. 1979Z>; Watts & Williams 1979; Regnault 1981), but differ in terms of geometry and amount of displacement. Both branches of the SASZ have sharp, fairly linear traces. Bulk displacements reach values of around 40-50 km along the northern branch (Jegouzo & Rosello 1988), and more than 100 km along the southern branch. The NASZ has an irregular trace. Displacement is at a minimum south of St Brieuc Bay and increases westward and eastward from this point (Watts & Williams 1979). The eastern and western parts of Central Brittany display slightly different deformation histories. The boundary between these two domains corresponds to the linear zone of the Montagnes Noires (Fig. 2). In the eastern part of Central Brittany, the deformation is marked by upright folds with subhorizontal axes (Fig. 2). Buckling is mostly controlled by the Armorican Quartzite Formation, a thick sandstone unit (up to 600 m) of Early Ordovician age (Le Corre 1978). The amplification of folds is limited and the envelope of fold trains is roughly horizontal (Fig. 2). This is consistent with only small variations of vertical strain across the deformed domain. An axial plane cleavage is associated with the folds, and strain markers (e.g. conglomerates) indicate a subhorizontal principal stretch parallel to fold axes (Le Theoff 1977; Le Corre 1978). At the vicinity of the northern branch of the SASZ, folds are reversed with south-dipping axial planes (Le Corre 1978) (Fig. 2). They are locally associated with thrusts (Dubreuil 1987; Faure & Cartier 1998) (Fig. 2). Based on maps of illite crystallinity and cleavage types, Le Corre (1978) has argued that cleavage development and granite intrusions were at least partially synchronous, and therefore that the deformation was Carboniferous in age (Le Corre 1978). The partial synchronism between regional deformation and pluton emplacement is also demonstrated by the existence of cleavage

289

triple points in the vicinity of plutons (Hanmer & Vigneresse 1980; Hanmer et al 1982). In the western part of Central Brittany, the early stages of deformation are represented, to the north, by the development of flat-ramp-type thrust systems (Babin etal 1975; Darboux 1991) on which upright folding associated with dextral wrenching is superimposed (Darboux 1991). To the south, the Chateaulin Basin was interpreted as a pull-apart and (or) an extensional basin during the Lower Carboniferous (Darboux et al 1977; Rolet & Thonon 1979; Guillocheau & Rolet 1982; Darboux 1991). However, the basin is bounded to the south by the transpressional shear zone of the Montagnes Noires (Fig. 2), with northward verging thrust faults affecting the Carboniferous sediments (Darboux & Le Gall 1988; Plusquellec et al 1999). The presentday shape of the basin is that of an overturned syncline, with an axial plane dipping to the SSE and a southern overturned limb overridden by the Montagnes Noires (Darboux 1991; Plusquellec et al 1999). Within the basin, cleavage trajectories trend mostly parallel to the syncline limbs, and stretching lineations plunge close to the SSE direction of cleavage dip. The restoration model presented in the present paper provides an explanation of these apparently conflicting features.

Kinematic data Cleavage and finite strain ellipsoid The rocks of Central Brittany show pervasive moderate deformation across the whole area. The axial-plane cleavage developed during folding is roughly vertical and strikes east-west in the central part of the fold belt (Fig. 3) (Le Corre 1977,1978). In this area, finite strains have been estimated from measurements of pebble shapes in deformed conglomerates, assuming no volume change during progressive deformation (Le Theoff 1977). Data show that the direction of the principal stretch A! is subhorizontal and parallel to the fold axes. They further emphasize that the direction of the intermediate principal strain axis A^ is roughly vertical throughout the area, (Le Theoff 1977), with a mean value of around one (Fig. 4). Le Theoff thus inferred overall plane strain (K = 1) in the horizontal plane.

Deformation regime A number of arguments demonstrate that the deformation involves a component of dextral

290

C. GUMIAUX^r^L.

Fig. 3. Set of cleavage directional data used for geostatistical analysis and the -strain model in Central Brittany. Frequency histogram shows the distribution of azimuths for all data. (Sources of data: Barbotin 1987; Berthe & Brun 1980; Dadet et al 1995; Darboux 1991; Dubreuil et al 1989; Fourniguet & Trautmann 1985; Hanmer & Vigneresse 1980; Herrouin et al. 1989; Herrouin & Rabu 1990; Hirbec 1979; Janjou et al. 1998; Le Corre 1978; Lopez-Munoz 1981; Pivette 1978; Plaine 1976; Plaine et al. 1984; Quete et al. 1981; Regnault 1981; Sagon 1976; Trautmann 1988; Trautmann et al. 1994.)

strike-slip (Gapais & Le Corre 1980): vertical cleavage and horizontal stretching lineation (Le Theoff 1977; Le Corre 1978), en echelon patterns of fold axes, dextral shear bands in conglomerates and sandstones, asymmetric quartz fabrics (Gapais 1979), and elliptical mapshapes of plutons with NE-SW trending long axes (Brun 1981; Vigneresse & Brun 1983; RomanBerdiel ef a/. 1997). Within the area covered by Le Theoff's data (Fig. 3), a mathematical removal of ductile strains was performed by Percevault & Cobbold (1982). Their computation involved a finite element model based on the division of the study area (about 40 by 50 km) into 110 elements within which homogeneous strain was assumed. Strain was removed from each element, and the best restored shape of the area was obtained by minimizing gaps and overlaps between elements. The restoration has argued for bulk dextral simple shear, with an overall

N120°-125°-striking shear direction and a mean y value of around 1 (Fig. 5). However, the restoration also suggests the occurrence of slight variations in the amount of shear strain (Fig. 5). These results are validated by all structural observations which indicate dextral shear within the area. However, a simple analysis of structural data is not sufficient to discuss the way a simple shear model can be extended regionally in terms of shear direction and shear intensity. The first analysis and interpretation of cleavage trajectories throughout Central Brittany was made by Barbotin (1987), who proposed a model of annular homogeneous dextral simple shear at regional scale, controlled by the arcuate shape of the southern branch of the SASZ. This model proposes that changes in cleavage directions are continuous throughout Central Brittany. However, the occurrence of transverse discontinuities, like the Montagnes Noires Shear Zone that separates domains with different

STRAIN REMOVAL IN CENTRAL BRITTANY

291

Data processing

Geostatistical analysis of cleavage directions

Fig. 4. Frequency diagram showing principal strain values estimated from strain analysis of deformed conglomerates (Le Theoff 1977) (strain measurements were made within the area outlined on Fig. 3). Bulk strain is of plane-strain type.

Carboniferous sedimentation histories, suggests that strain variations may not be so continuous across Central Brittany. To analyse the consistency of the cleavage data set available in Central Brittany, and to constrain the way a simple shear model can be extended regionally, we have performed a geostatistical analysis of cleavage directions.

The geology of Central Brittany has been studied extensively for more than a century (e.g. Barrois 1930). Structural studies provide a very large quantity of cleavage orientation data. We have digitized about 1800 of these data throughout Central Brittany (Fig. 3). Because of overall plane-strain in map view and of overall subvertical cleavage attitudes, spatial variations in orientations can be reasonably reduced to an analysis of cleavage azimuths. The quality of such a data set depends on different factors, in particular the degree of heterogeneity of the measurement density, and that of data sources. In the study area, the number of data points is large, but the density of measurements is highly variable (Fig. 3). This relates to poor outcrop conditions, in particular in the Brioverian. Data sources are also very heterogeneous in type and age (e.g. geological maps, PhD theses, articles) and quality. This can be minimized by a preliminary sorting (elimination of doubtful data). In addition, errors in location and direction of individual measurements are variable (initial documents of various scales and various quality), and can be amplified through the digitalization procedure. A geostatistical study was performed on this cleavage directional data for the Central Brittany shear belt. The method is based on a probabilistic approach developed by Matheron (1955,1962) for mining industry purposes. Variogram analysis, which is the first step of the approach, gives an evaluation of the natural spatial correlation between data values. In particular, it allows the distinction between significant signal and background, and separation of regional tendencies from local

Fig. 5. Strain restoration of the area shown in Fig. 3 (finite element model, Percevault & Cobbold 1982). The rectangle superimposed on the restored state transforms into a quasi-parallelogram in the present-day deformed state. The model is thus compatible with bulk dextral simple shear along the N123° direction. The non-linearity of deformed parallelogram sides suggests the occurrence of small shear strain gradients (modified after Percevault & Cobbold 1982).

292

C. GUUIAUX ET AL.

Fig. 6. (a) Omni-directional variogram calculated for cleavage directions in Central Brittany. The sill value is around 450, the nugget effect equals 350 square degrees and the range distance is around 37 km. (b) Directional variogram calculated along a N30°-N40° strike. Diagram shows a periodical trend with a wavelength value of about 40-45 km. Heavy curve is best-fit exponential function. See Appendix for further explanations.

perturbations (e.g. competency-induced cleavage refraction, or granitic intrusions). Principles of the method are described in detail in the Appendix, and only the results of the analysis are given here. Variograms have been calculated up to a maximum distance (dmax) of 100 km between individual data points, which corresponds roughly to the width of Central Brittany. The corresponding best value for the range of class of distance (Ad) is 1 km. The omni-directional variogram (Fig. 6a) shows a large 'nugget effect' of around 350 square degrees. This expresses a significant variability in cleavage directions, even for closely spaced locations, i.e. less than 1 km, which corresponds to the chosen range of classes of distance. The variogram reveals a spatial correlation between data pairs, up to a maximum distance of 37 km (range value r). Thus, variations in cleavage directions become statistically independent for distances larger than 35-40 km, which means that a unique and regular regional trend of directional variations does not hold for Central Brittany. On variograms, the occurrence of a periodic trend reveals a periodic spatial distribution of data values. For our data set, directional variograms show periodic trends when the chosen computation direction approaches NE-SW. This is best expressed for the N30°-N40° directional variogram (Fig. 6b), where the amplitude/wavelength ratio of the curve is at its maximum. A characteristic undulation of cleavage direction is thus highlighted along the N30°-N40° direction (i.e. perpendicular to a N125° trend). According to the variogram, the wavelength of this cleavage undulation can be estimated at around 40-45 km (Fig. 6b).

A model of cleavage trajectories A second step in statistical modelling consists in the interpolation of cleavage directions in order

to calculate a continuous cleavage trajectory map. The method used here is the kriging interpolation (Cressie 1991; Wackernagel 1998). It involves a moving average technique that estimates the value of a variable at a given location, from values measured at adjacent locations (Appendix). The interdependence between individual data points is directly expressed by a variogram, as follows. (1) A best-fit exponential variogram is computed from the experimental one, so that continuous equivalent functions can be used during interpolation between data points. (2) In this study, the search radius value applied during interpolation is the value of the range in the variograms. (3) The 'nugget effect' value is added to the exponential function of spatial correlation, in order to eliminate the non-regional part of the data set during interpolation. Thus, only regional tendencies are taken into account for the interpolation model and resulting strain trajectories. Simple kriging interpolation of cleavage directions is performed for each node of a regular grid covering the whole area (with a spacing of 500 m). Details for the calculation processing in the particular case of directional data are summarized in the Appendix. Results are presented on a contoured map (Fig. 7a) and on a cleavage trajectory map (Fig. 7b). To obtain this map, trajectory traces are drawn from regularly spaced points along a diagonal of the grid. Step by step from each initial trace along the diagonal, each trajectory propagates in the direction computed at the reached grid node, until reaching the edge of the domain. Computed directions vary from N 68° to N 133°. The minimum values are localized in the westernmost part of Central Brittany (Fig. 7a). The

STRAIN REMOVAL IN CENTRAL BRITTANY

293

¥ig. 7. (a) Cleavage orientation contours computed from kriging interpolation. A NW-SE-striking band, 40-45 km in width, is underlined by alternating alignments of lows and highs, (b) Regional cleavage trend trajectories constructed from interpolated cleavage directions.

largest values lie along the eastern border of the domain (Fig. 7a). With the technique used, the maximum range of interpolated values is much less than that of the initial directional data (compare histograms on Figs 3 and 8). In addition, computed trajectories can be significantly oblique with respect to

local neighbouring original data. This is observed around granites or along main shear zones (e.g. Montagnes Noires) (compare Figs 3 and 7b), illustrating that the high value of the 'nugget effect' is taken into account during interpolation, leading to substantial filtering of the data in order to image regional trends.

294

C. GUMIAUX£7ML.

Fig. 8. Domainal distribution of interpolated cleavage directions outlined by a correlation between peaks of the multimodal orientation distribution and particular geographic zones of Central Brittany.

Domainal distribution of cleavage directions Lateral variations of the mean regional cleavage trend are clearly seen on the trajectory map. A general increase in cleavage azimuth values occurs from west to east. This feature was already noted by previous authors for cleavage attitudes (Barbotin 1987), as well as for major structures (Gapais 1980). However, considering the contoured map (Fig. 7a), the variations in cleavage direction do not seem to be continuous, but rather organized into domains separated by tightened direction contours. In order to detect and characterize possible domains of cleavage preferred orientation, a frequency counting has been calculated for the values of all the nodes of the computed grid (Fig.

7a). The histogram shows a multi-modal distribution with three well-marked maxima (N 77°, N 86° and N 97°), and two secondary ones for highest direction values (N 112° and N 118°) (Fig. 8). From this, we infer that direction variations are not continuous across the area. Best-fit Gaussian functions are reported on the histogram and cutting values are deduced from local lows on the multi-peak Gaussian curve (vertical lines crossing the diagram, Fig. 8). These values are then used to separate the corresponding zones on the trajectory map (Fig. 8). The organization of spatial variations in large and continuous area confirms the presence of different domains. Domain boundaries actually match zones of high gradients on the interpolated directional map (Fig. 7a). After merging the two easternmost zones associated with the

STRAIN REMOVAL IN CENTRAL BRITTANY

lowest maxima, four main domains are recognized in Central Brittany (Fig. 8), as follows. (1) To the west, domain A lies south of the Leon metamorphic block (Fig. 1). In this domain, the cleavage strikes dominantly parallel to the South Leon Shear Zone (Fig. 7b) that separates the Leon block from western Central Brittany. (2) Domain B comprises the Chateaulin Carboniferous basin and most of the Hercynian granites (Fig. 8; see also Fig. 2). The mean cleavage direction is subparallel to the N80°-N 85° striking Montagnes Noires Shear Zone (Fig. 7b). (3) Domain C is the widest domain. It includes most of eastern Central Brittany, where lithologies are essentially folded Brioverian and pre-Carboniferous Palaeozoic sediments (Fig. 2). In this domain, the average cleavage direction strikes N97° (Fig. 8). However, cleavage trajectories show local deflections due to an alignment of lower directional values along NW-SE-trending linear anomalies (Figs 7a and b). This bandshaped structure is about 40 to 45 km in width (Fig. 7b). This value, as well as the NE-SW trend of the anomalies, is compatible with the wavelength revealed by the directional variogram analysis (Fig. 6). (4) Domain D is localized at the eastern edge of Central Brittany and shows the highest azimuth values of cleavage directions. In this area, geological contours show some bending towards NW-SE directions (Fig. 2). This feature could result from a late and local deformation associated with dextral strike-slip along a NNW-SSE directed fault situated slightly east of the area, below the Mesozoic cover of the Paris basin (Parthenay fault, Rolin & Colchen 2001). Restoration of eastern Central Brittany The above statistical analysis shows that cleavage directions are distributed into domains. It follows that a homogeneous simple shear model (e.g. Barbotin 1987) does not easily account for the cleavage pattern. On the other hand, the geostatistical analysis highlights a large domain C where cleavage directions are nearly constant, with continuous trajectories. The NW-SEtrending anomalies in cleavage directions observed within domain C are parallel to the shear direction deduced from the restoration of Percevault & Cobbold (1982). In the light of this model, the anomalies can be interpreted in terms of shear strain gradients perpendicular to

295

a regional NW-SE shear direction. From these, and the fact that deformation is continuous throughout the domain, we infer that the simple shear model developed by Percevault & Cobbold (1982) can be applied throughout the whole domain C. To the east, in domain D, the cleavage appears slightly bended during late deformation stages. To the west, the Montagnes Noires Shear Zone cuts across domain B. Thus, eastern and western boundaries of domain C do not coincide with localized fault zones. An extension of the simple shear model to the whole eastern Central Brittany is therefore supported by geological data. A detailed statistical analysis of the interpolated cleavage map yields a best fit around N123° for the direction of cleavage anomalies in domain C. This direction is parallel to the trace of the S ASZ in its central part, south of the study area (Vannes area, Fig. 7a). A N123°-directed simple shear was therefore chosen for the restoration. In such a model, the amounts of shear y can be directly calculated from the angle 6 between cleavage directions and the N123° direction of regional simple shear (tan 20 - 2/y, Ramsay & Graham 1970). At the scale of eastern Central Brittany, the mean calculated y value is of 1.5, slightly higher than that of about 1 proposed by Percevault & Cobbold (1982) for their study area. This y value yields a mean axial ratio of the finite strain ellipsoid of about 4 (Ramsay, 1967, eqn 3-67, p. 85), which is consistent with strain field data (Fig. 4) (Le Theoff 1977). The discrepancy between mean y values deduced from our study and from that of Percevault & Cobbold (1982) could be related to the differences in restoration methods, the fact that the area restored by Percevault & Cobbold (1982) overlaps a NW-SE-striking band with strain intensities lower than the mean value and differences in size of the restored domains. To restore the area, we have estimated a profile of y values from changes in cleavage attitudes perpendicular to the shear direction (Fig. 9). In addition, significant kinematic constraints at domain boundaries are found along the NASZ. Along the western part of the NASZ, granites emplaced during Carboniferous dextral shearing (Figs 2 and 9) show strong mylonitization. This is observed in the St Renan granite, the westernmost intrusion along the NASZ (Fig. 2) (Gore & Le Corre 1987), and dated around 330-340 Ma (see Le Corre et al 1991). Moving eastward, syntectonic granites show weaker internal deformations and no or limited offsets along the NASZ (Montcontour granite, Fig. 9). In this area, pre-Hercynian lithologies are only

296

C. GUMIAUXETAL.

Fig. 9. Inversion of the deformation in eastern Central Brittany by N123° sinistral simple shear. The geological map in the present-day situation (bottom) is from Chantraine et al (1996). Across-strike variations of y values applied for restoration (middle) are calculated from variations of interpolated cleavage directions.

weakly reorientated towards the shear direction, with no offset (Carrie et al. 1979; Chantraine et al. 1996). Moreover, the trace of the NASZ is interrupted in this zone on the recent 1:1 000 000 geological map of France (Chantraine et al. 1996). In contrast, large strains are observed further east along the NASZ (Regnault 1981; Gapais & Cobbold 1987, fig. 15a; Paris & Dadet 1988). From these, we infer that a pinning point of limited dextral displacement can be fixed along the NASZ, just west of the Moncontour granite (Fig. 9).

Validation of the model and regional implications The restoration applied using the above boundary conditions (Fig. 9) has several implications, both within the restored domain and at its boundaries, as discussed below.

Within the restored area At a regional scale, the attitude of folds is compatible with the strain field resulting from the

computed deformation model (Fig. 10). In the area, folding is controlled by early buckling of the thick layer of Armorican Quartzite (crosssection, Fig. 2) (Le Corre 1978). Figure lOa shows a qualitative theoretical distribution of poles to bedding affected by incipient buckling associated with 123°N strike-slip directed, i.e. with a NNW-SSE-striking direction of incremental shortening. We chose an initial elliptical shape of the pole distribution to take into account some primary scattering around the vertical, and the doubly-plunging attitude of fold axes. Indeed, folds are en-echelon type as a result of strike-slip deformation. Assuming the bedding planes rotate passively during deformation, the attitude of their poles after a shear strain y of 1.5 is shown on Fig. lOb, together with poles to bedding measured in Central Brittany within the Armorican Quartzite (Le Corre 1978). This diagram underlines that, despite the certain occurrence of mechanical effects, the overall attitude of regional fold axes and the amounts of fold tightening are compatible with the restoration model. In Central Brittany, the Palaeozoic sedimentary sequence shows evidence for local

STRAIN REMOVAL IN CENTRAL BRITTANY

297

Fig. 10. Comparison of poles to bedding observed in Central Brittany with a model of N123°-directed simple shear with a y value of 1.5. (a) Initial model distribution of pole scattering after small NNW-SSE shortening, (b) Distribution of model poles to bedding planes passively rotated during simple shear. The distribution is consistent with field data (poles to bedding within the Armorican Quartzite; Le Corre 1978).

erosion events and (or) lack of deposition during the Upper Devonian, followed by localized Carboniferous basins (Guillocheau & Rolet 1982; Robardet et al 1994). If these features can be attributed to early vertical differential motions, the onset of deformation within Central Brittany appears relatively well constrained, around 360-350 Ma. Available geochronological data on syntectonic plutons emplaced along the northern branch of the

Fig. 11. Restoration of some outlines of syntectonic plutons located along the northern branch of the SASZ, further evidence that they were emplaced during progressive shearing.

SASZ yield ages around 340-330 Ma (see Le Corre et al (1991) and references therein), i.e. during the course of regional shearing. Figure 11 shows the outline of these granites in map view (Pontivy and Lizio granites). These plutons are

298

CGUMIAUXEr^L.

Fig. 12. Implications of the restoration of simple shear in eastern Central Brittany, with particular reference to its western boundary. Section across the Chateaulin basin is after Darboux (1991).

asymmetrical in shape as a result of dextral shearing. Assuming that they were initially symmetrical, their present-day geometry implies a mean shear strain y around 0.9, lower than the total regional shear strain y of 1.5. This is compatible with late synshearing emplacement, which is in agreement with the conclusions of Vigneresse & Brun (1983).

At the boundaries of the restored area Restoration of simple shear throughout the eastern part of Central Brittany has several implications for kinematics along its boundaries

(Figs 9 & 12). Along the NASZ, east of the pinning point, overlaps occur between the restored boundary of Central Brittany and the present-day NASZ boundary (Fig. 9). The NASZ in the area shows combined dextral wrenching and SSE-directed thrusting (Carrie et al 1979; Regnault 1981; Paris & Dadet 1988). Shear zones in the area are locally marked out by Carboniferous sediments (e.g. Regnault 1981; Paris & Dadet 1988). Similar overlaps between restored and present-day geometries are observed for the southern boundary of eastern Central Brittany, along the northern branch of the SASZ (Fig. 9). This is consistent

STRAIN REMOVAL IN CENTRAL BRITTANY

with field data that locally indicate combined north-directed thrusting and dextral wrenching (Le Corre 1978; Dubreuil 1987; Faure & Cartier 1998). Evidence of northward thrusting components also occurs further south (Fig. 2) (Ledru etal. 1986). To the west, the Montagnes Noires Shear Zone marks the boundary between eastern Central Brittany and western Central Brittany (Figs 1, 7b, 8 & 12). Because of the NW-SEdirected simple shear and the occurrence of a pinning point along the NASZ, the restoration implies some scissor-type movement of the fault, with clockwise rotation from an initial NNE-SSW direction (Figs 9 and 12). This is compatible with transpressive motion along the Montagnes Noires, a feature indeed shown by field data (Darboux & Le Gall 1988; Fig. 12). The Montagnes Noires zone bounds the Carboniferous Chateaulin basin to the South. This basin is affected by NNW-verging thrusts (Castaing et al 1987; Darboux & Le Gall 1988), some being interpreted as coeval with dextral strike-slip (Darboux 1991), this latter becoming dominant during late deformation stages (Darboux 1991). Evidence for substantial amount of thrusting of eastern Central Brittany over the Chateaulin basin is further attested by metamorphic conditions observed south of the Montagnes Noires. Indeed, a local domain of relatively high-grade Hercynian metamorphism (e.g. occurrence of kyanite) is documented south of the Montagnes Noires (see Hanmer et al. 1982, and references therein), roughly in the area between its present-day position and its restored one (Fig. 12). Thus, the different signatures of cleavage trajectories in this triangular area south of domain B with respect to domain C (Fig. 8), might thus be associated with differential exhumation during regional shearing. Classical interpretations for the formation of the Chateaulin basin involved either extension (Darboux 1991), or pull-apart associated with dextral strike-slip (Darboux et al 1977; Rolet & Thonon 1979; Guillocheau & Rolet 1982; Castaing et al. 1987). However, several lines of evidence suggest that the Carboniferous sedimentation in the Chateaulin basin could have occurred in a compressional context. These are: evidence for vertical differential movements since the Late Devonian; an increase in deposit depth from NW to SE, toward the Montagnes Noires (see Darboux 1991; Fig. 12); the onset of deformation in eastern Central Brittany probably around 360-350 Ma; and the overlap between ages of syntectonic granites (some being Visean in age), and sedimentation in the

299

basin (up to Namurian in age). From the above arguments and those implied from restoration, we infer that the Chateaulin basin is a syntectonic transpressive basin. Furthermore, the restoration emphasizes that the two main structural features of western Central Brittany, i.e. dextral strike-slip and NNW-verging thrusting, could be explained by a single progressive deformation event. Indeed, field data show that strike-slip structures often post-date those related to thrusting (Darboux 1981; Darboux 1991). This probably occurred during clockwise rotation of the Montagnes Noires: the early attitude of the fault, at high angle to the incremental shortening direction would have favoured thrusting components; whereas the progressive rotation of the fault toward the direction of the western part of the S ASZ would favour dextral strike-slip components. The nature of domain boundaries with regard to the consistency between geological data and outcomes of our restoration was analysed. This indicated that the best candidate for a bulk direction of simple shearing was NW-SE. Major shear zones oblique with respect to this direction are likely to have been inherited discontinuities. The inherited character of the NASZ has been already argued (e.g. Watts & Williams 1979; Ballevre et al. 2001). After restoration, the Montagnes Noires shear zone strikes about 30°N (Figs 9 & 12). This direction joins that of major shear zones of Cadomian age located in the Cadomian domain of North Brittany (Figs 9 & 12) (see Brun et al. 2001). The northern branch of the SASZ is marked out by the pre-collision Lanvaux granite and suggests a pre-Hercynian discontinuity. More generally, I-type granitic intrusions, which imply substantial mantle contributions, occur along and (or) at the vicinity of the eastern part of the SASZ (e.g. Moelan orthogneiss), of the Montagnes Noires Shear Zone (e.g. Rostrenen Massif), and of the NASZ (e.g. Huelgoat Massif). Implications of our restoration can be extended further to the southern branch of the SASZ. Its western segment is part of the northern branch, as shown by its east-west orientation. Its southeastern part, which strikes N130°-N140°, is outlined locally by eclogitic rocks (Fig. 1) (Godard 2001), and is a good candidate to represent a preserved remnant of a major suture zone (Colchen & Rolin 2001). The segment linking the western and eastern parts of the shear zone, which strikes N120°-N125° (Figs 1 & 12), is likely to result from strain localization associated with the Carboniferous dextral shearing directed 123°N, reactivating earlier fault segments.

300

C. GUMIAUX ET AL.

Conclusions The above study leads to the following conclusions: (1) The geostatistical method can provide a powerful tool for structural data analysis. In this example, the method enabled the separation of local and distant features, and the extraction of a regional signal from the set of available orientation data. (2) The interpolation method permitted computation of cleavage trajectories and provided constraints for the restoration model of Central Brittany. (3) The method showed that Central Brittany was marked by a domainal distribution of cleavage directions, and that a single restoration model could not be applied to the entire area. (4) It also showed that deformations in eastern Central Brittany could be interpreted in terms of regional dextral simple shear, moderately heterogeneous, along a N120°-N125° direction, and with a mean y of 1.5. This is consistent with the previous restoration model performed by Percevault & Cobbold (1982). The simple shear model argued by these authors on a restricted area can thus be extended to a large part of Central Brittany. (5) Kinematic implications of the restoration model within and around the restored area are validated by field data, and bring substantial light on their tectonic interpretation. (6) According to the restoration model, major shear zones oblique to the regional shear direction (Montagnes Noires Shear Zone,

Appendix: principles of the geostatistical analysis Variogram computation and interpretation For many regional variables, that are functions defined for a spatial or time continuum, statistics have shown that measured values for two close data points (e.g. 1 and 5; Fig. Al) are more similar than for two distant ones (e.g. 1 and 9; Fig Al). For a given data set, the problem is to determine the maximum distance for which data are interdependent, and how this dependence decreases with increasing distance. The function linking values between data points is not a deterministic one. However, a natural regionalized

northern branch of the SASZ, NASZ) should show thrust components, as is indicated by field data. (7) The model also emphasizes that these major shear zones oblique to the regional shear direction are most probably inherited structures. Among these, the Montagnes Noires Shear Zone appears as a Cadomian structure rotated during regional shearing. (8) Concerning this particular domain boundary, the restoration model implies scissortype motions involving substantial thrusting components. Consistently, geological and structural data show that this shear zone acted as a dextral transpressional zone during Carboniferous times (Darboux & Le Gall 1988). Scissor-type motion of the Montagnes Noires Shear Zone during regional shearing may also account for structural patterns observed in the Chateaulin basin, marked by dominant NNW-directed thrusting followed by dominant ENE-WSW strike-slip. This restoration model shows that both thrusts and dextral strike-slip shearing can be integrated into a progressive deformation model involving the development of the Carboniferous Chateaulin basin in a transpressive context. (9) This restoration model shows that the southern branch of the South Armorican Shear Zone, which is the most prominent structure of the Hercynian belt of Brittany, is not a simple curved shear zone associated with the Hercynian collision, but is most probably made of inherited segments and of new segments formed during late orogenic stages. S. Treagus and R. Strachan are thanked for their constructive reviews.

phenomenon should not be entirely random. Kriging interpolation is based on a theoretical probabilistic model. Each piece of data of the regionalized function is considered as a possible variable of a random function attached to the location for the considered data point. The problem is then reduced to the study of spatial variations for the parameters defining these random functions. The variability of these parameters is estimated by computing variograms. The half square of the differences between values for a given pair of data points is calculated. For directional data, the difference is defined modulus 180°, i.e. for algebraic values in a given spatial reference:

STRAIN REMOVAL IN CENTRAL BRITTANY

301

Four important parameters can be deduced from an experimental variogram (Fig. A2b):

Fig. Al. Example of a data set used for variogram computation, here for i = 1 and/ = ( 2 , . . . , 10).

The variogram value is then:

and this result is expressed as a function of the vector h separating the two considered points i and /:

It follows that r(-h) = F(h). A variogram is thus an even function and is only dependent on the separating distance d = Ihl between data points, and not on their spatial location. In practice, the calculation of F^ is carried out for each of the data pairs constituting the set of n data, i.e. for / and/ = (1, 2 , . . ., n). Because of the great number of data pairs, even for a reasonable number of n data points, the variogram is broken down into classes of distance (//k) of width (Ad), where the result is given as a mean (Fig. A2a). The final equation for the computation of the experimental variogram becomes a function of these classes:

where i and / correspond to data points separated by a distance Ihl e //k, and nk is the number of data pairs fitting the class //k.

(1) Variograms often reach a constant value with increasing distance. This sill value is the half of the bulk variance for the whole data set (i.e. the square of the standard deviation). (2) The distance for which the variogram reaches the sill value (or 95 % of it for the exponential model) is called the range r. It indicates the maximum distance of mutual dependency between values measured at two distinct locations. The range value will be used as the search radius value during kriging interpolation. (3) The 'nugget' value, defined as the variogram ordinate at the origin, gives the constant part of the variance across the considered domain. This parameter is not a regionalized one; in the probabilistic model, it is considered as the purely random part of the random functions for the entire space. With respect to data analysis, it can be considered as the error value, including measurement and computation ones, or as very local effects, below the computation resolution. This latter is equal to the distance class range value. For short distances, it is a function of the minimum sampling distance in the data set. (4) Finally, the trend of the variogram indicates the way in which the similarity between data values decreases with increasing distances. It can be fitted with a theoretical variogram function that is classically chosen as exponential (present case), spherical, gaussian or power function (Fig. A2a). The corresponding theoretical covariance function will be used to determine the weighting values to apply during kriging interpolation (Fig. A2b). Regionalized variables can show different spatial distributions along different directions in the considered space. To analyse such possible anisotropies, the variogram can be sliced into classes of orientations relative to the vector h. During computation, only the data pairs for which the orientation of the separating vector is included in the range of the considered class are retained for the variogram calculation, as shown on Fig. A3. We obtain as many directional variograms as the number of orientation classes, the number chosen being large enough to reveal the anisotropy, but not too large, so that data pairs involved in the calculation are kept to a minimum.

302

C. GUMIAUX ET AL.

Fig. A2. Example of variogram construction, (a) Example of experimental variogram cloud and adjusted exponential theoretical function, (b) Theoretical variogram showing major attached statistical parameters.

Kriging interpolation of directional data via direction cosines Directional data are particular because they are circular variables, with a periodicity. Thus, within a given reference frame, a direction defined by an angle is expressed modulo 360° (e.g. a = a ± 360°). Because of this, such variables cannot be processed by standard statistical techniques (e.g. direct mean calculation). For statistical analysis, each piece of data can be considered as a unit vector characterized by its direction cosines in the chosen reference frame, each direction cosine being defined as the vector coordinates along each reference axis. Cleavage directions are expressed modulo 180° and therefore the use of direction cosines for data treatment cannot be applied directly on azimuths defined, for example, from 0° to 180°. First the data is transformed from a 180° range to a 360° range by doubling the algebraic directional values. Then, the data are treated in terms of direction cosines (Upton & Fingleton 1989). Final results are obtained by halving the solutions for doubled values. This computation technique must be used to interpolate between directional data. The kriging interpolation of the direction cosines is performed individually, and the resulting calculated angles are halved to obtain the final interpolated direction. This approach ensures the stability of the result, especially around the

reference direction, but also for a data set covering the entire 180° range of non-orientated directions.

Fig. A3. Principle of data selection to compute an orientated variogram.

STRAIN REMOVAL IN CENTRAL BRITTANY

References BABIN, C, DARBOUX, J.R., DUEE, G., GRAVELLE, P., MORZADEC, P., PLUSQUELLEC, Y. & THONON, P. 1975. Tectoniques tangentielles et tectoniques superposees dans le Devonien de la rade de Brest (Nord-Finistere). Comptes Rendus de VAcademic des Sciences de Paris - D, 280, 259-262. BALLEVRE, M., LE GOFF, E. & HEBERT, R. 2001. The tectonothermal evolution of the Cadomian belt of northern Brittany, France: a Neoproterozoic volcanic arc. Tectonophysics, 331,19-43. BARBOTIN, E. 1987. Trajectoires de deformation finie et interpretation cinematique. Modeles numeriques et application a des exemples regionaux (Bretagne Centrale etAlpes Penniques). Troisieme cycle universitaire thesis, Rennesl, Rennes. BARROIS, C. 1930. Les grandes lignes de la Bretagne, Livre Jubilaire. Centenaire de la Societe geologique de France, Paris. BERNARD-GRIFFITHS, J., PEUCAT, J.J., SHEPPART, S. & VIDAL, P. 1985. Petrogenesis of Hercynian leucogranites from the southern Armorican Massif: contribution of REE and isotopic (Sr, Nb, Pb and O) geochemical data to the study of source rock characteristics and ages. Earth and Planetary Science Letters, 74, 235-250. BERTHE, D. & BRUN, J.P 1980. Evolution of folds during progressive shear in the South Armorican Shear Zone, France. Journal of Structural Geology, 2,127-133. BERTHE, D., CHOUKROUNE, P. & GAPAIS, D. 19790. Quartz fabrics and progressive gneissification of granites by simple shear - example of the South Armerican Shear Zone. Bulletin de Mineralogie, 102, 265-272. BERTHE, D., CHOUKROUNE, P. & JEGOUZO, P. 1979ft. Orthogneiss, mylonite and non coaxial deformation of granites: the example of the South Armorican Shear Zone. Journal of Structural Geology, 1, 31-42. BOSSE, V., BALLEVRE, M. & VIDAL, O. 2002. Ductile thrusting recorded by the garnet isograd from the blueschist-facies metapelites of the He de Groix, Armorican Massif, France. Journal of Petrology, 43, 485-510. BOSSE, V., FERAUD, G, RUFFET, G, BALLEVRE, M., PEUCAT, J.-J. & DE JONG, K. 2000. Late Devonian subduction and early-orogenic exhumation of eclogite-facies rocks from the Champtoceaux Complex (Variscan belt, France). Geological Journal, 35, 297-325. BOUCHEZ, XL. & BLAISE, J. 1976. Une structure hercynienne liee a un accident ductile: 1'anticlinal de Lanvaux - les-Ponts-de-Ce, aux environs d'Angers (massif Armoricain). Bulletin de la Societe Geologique de France, 18,145-157. BROWN, M. & DALLMEYER, R. D. 1996. Rapid Variscan exhumation and the role of magma in core complex formation: southern Brittany metamorphic belt, France. Journal of Metamorphic Geology, 14, 361-379. BRUN, J. P. 1981. Instabilites gravitaire et deformation de la croute continentale: application au developpe-

303

ment des domes et des plutons. Troisieme cycle universitaire thesis, Rennes 1, Rennes. BRUN, J.-P, GUENNOC, P., TRUFFERT, C. & VAIRON, J. 2001. Cadomian tectonics in northern Brittany: a contribution of 3-D crustal-scale modelling. Tectonophysics, 331, 229-246. CARRIC, G, CHANTRAINE, J., DADET, P., FLAGEOLLET, J.C.,SAGON,J.P&TALBO,H. 1979. Geological map - Moncontour (1/50 000). BRGM, Orleans. CASTAING, C., ROLET, J., CHEVREMONT, P., LE CALVEZ, J. Y. & THONON, P. 1987. La region de Huelgoat (Finistere central) dans le contexte geodynamique armoricain. Geologic de la France, 1, 23-36. CHANTRAINE, I, AUTRAN, A. & CAVELIER, C. 1996. Carte Geologique de la France; 1/1000000 (French Geolgical Map). BRGM, Orleans. CHANTRAINE, J., CHAUVEL, J.J. & RABU, D. 1994. The Cadomian Orogeny in the Armorican Massif Lithostratigraphy. In: KEPPIE, ID. (ed.) PreMesozoic Geology in France and related areas. Springer-Verlag, Berlin. COGNE, J. 1974. Le Massif Armoricain. In: DEBELMAS, J. (ed.) Geologic de la France, pp. 105-161. COLCHEN, M. & ROLIN, P. 2001. La chaine hercynienne en Vendee. Geologic de la France, 1-2, 53-86. CRESSIE, N.A.C. 1991. Statistics for spatial data. John Wiley & Sons, New York. DADET, P., HERROUIN, Y, BARDY, P., LEBRET, P.,TRAUTMANN, F. & CARN, A. 1995. Geological map (1/50 000) - Pipriac (387). BRGM, Orleans. DARBOUX, J.R. 1981. Caracterisation du regime cisaillant de la deformation hercynienne dans les Monts d'Arree (Massif Armoricain, France). Comptes Rendus de VAcademic des Sciences de Paris, II, 1497-1500. DARBOUX, J.R. 1991. Evolution tectonosedimentaire et structuration synmetamorphe des zones externes du segment Hercynien Ouest-Europeen. Etat thesis, Bretagne Occidentale, Brest. DARBOUX, J.R. & LE GALL, B. 1988. Les Montagnes Noires: cisaillement bordier meridional du bassin carbonifere de Chateaulin (Massif Armoricain, France). Caracteristiques structurales et metamorphiques. Geodinamica Ada, 2,121-133. DARBOUX, J.R., GRAVELLE, M., PELHATE, A. & ROLET, J. 1977. L'evolution tectonique de la terminaison occidentale du domaine centre-armoricain au Devonien et au Carbonifere. Comptes Rendus de I'Academie des Sciences de Paris - D, 284, 1151-1154. DUBREUIL, M. 1987. Le bassin en decrochement de Saint-Julien-de-Vouvantes-Angers (Carbonifere inferieur du Sud-Est du Massif armoricain). Bulletin de la Societe Geologique de France, 3, 215-221. DUBREUIL, M., CAVET, P., BLAISE, J., ESTEOULECHOUX, J., GRUET, M. & LARDEUX, H. 1989. Geological map (1/50 000) - St-Mars-la-Jaille. BRGM, Orleans. FAURE, M. & CARTIER, C. 1998. Deformations ductiles polyphasees dans 1'antiforme orthogneissique de St-Clement-de-la-Place (Unite de Lanvaux, Massif Armoricain). Comptes Rendus de I'Academie des Sciences de Paris, 326, 795-802.

304

C. GUMIAUX ETAL.

FOURNIGUET, J. & TRAUTMANN, F. 1985. Geological map (1/50 000) - Redon (419). BRGM, Orleans. GAPAIS, D. 1979. Deformation progressive d'un quartzite dans une zone plissee (segment hercynien de Bretagne centrale). Bulletin de Mineralogie, 102, 249-264. GAPAIS, D. & COBBOLD, P.R. 1987. Slip system domains. II: Kinematic aspects of fabric development in polycristalline aggregates. Tectonophysics, 138, 289-309. GAPAIS, D., LAGARDE, J. L., LE CORRE, C, AUDREN, C., JEGOUZO, P., CASAS SAINZ, A>& VAN DEN DRIESSCHE, J. 1993. La zone de cisaillement de Quiberon: temoin d'extension de la chaine varisque en Bretagne meridionale au Carbonifere. Comptes Rendus de VAcademie des Sciences, II, 1123-1129. GAPAIS, D. & LE CORRE, C. 1980. Is the Hercynian belt of Brittany a major shear zone? Nature, 288, 574-576. GODARD, G. 2001. Les Essarts eclogite-bearing Complex (Vendee). Geology of France 'Special Vendee', 1-2,19-51. GORE, B. & LE CORRE, C. 1987. Cinematique hercynienne du cisaillement nord-armoricain a la bordure du granite syntectonique de Saint Renan-Kersaint (Finistere). Bulletin de la Societe Geologique de France, 3, 811-819. GUILLOCHEAU, F. & ROLET, J. 1982. La Sedimentation Paleozoi'que Ouest-Armoricaine. Bulletin de la Societe Geologique et Mineralogique de Bretagne, 14, 45-62. HANMER, S. & VIGNERESSE, J.L. 1980. Mise en place de diapirs syntectoniques dans la chaine hercynienne: exemple des massifs leucogranitiques de Locronan et de Pontivy (Bretagne Centrale). Bulletin de la Societe Geologique de France, 22, 193-202. HANMER, S., LE CORRE, C. & BERTHE, D. 1982. The role of Hercynian granites in the deformation and metamorphism of Brioverian and Palaeozoic rocks of Central Brittany. Journal of the Geological Society of London, 139, 85-93. HERROUIN, Y., DADET, P., GUIGUES, I, LAVILLE, P. & TALBO, H. 1989. Geological map (1/50 000) - Bain de Bretagne (388). BRGM, Orleans. HERROUIN, Y. & RABU, D. 1990. Geological map (1/50 000) - Chateaubriant (389). BRGM, Orleans. HIRBEC, Y. 1979. Le complexe basique de Belle-Isle-enTerre (Cotes du Nord). Sa place dans revolution geodynamique du nord du Massif Armoricain. Troisieme cycle universitaire thesis, Rennes 1, Rennes. JANJOU, D., LARDEUX, H., CHANTRAINE, J., CALLIER, L. & ETIENNE, H. 1998. Geological map (1/50 000) Segre (422). BRGM, Orleans. JEGOUZO, P. 1980. The South Armorican Shear Zone. Journal of Structural Geology, 2, 39-47. JEGOUZO, P., PEUCAT, J.-J. & AUDREN, C. 1986. Caracterisation et signification geodynamique des orthogneiss calco-alcalins d'age ordovicien de Bretagne meridionale. Bulletin de la Societe Geologique de France, 2, 839-848. JEGOUZO, P. & ROSELLO, E.A. 1988. La branche nord du cisaillement sud-armoricain (France): un essai devaluation du deplacement par 1'analyse des

mylonites. Comptes Rendus de VAcademie des Sciences de Paris, 2,1825-1831. LE CORRE, C. 1977. Le Brioverien de Bretagne centrale: essai de synthese lithologique et structurale. Bulletin du B.R.G.M. Section I, 3, 219-254. LE CORRE, C. 1978. Approche quantitative des processus synschisteux. L'exemple du segment Hercynien de Bretagne Centrale. Etat Thesis, Rennes 1, Rennes. LE CORRE, C, AUVRAY, B., BALLEVRE, M. & ROBARDET, M. 1991. Le Massif Armoricain. Scientific Geological Bulletin, 44, 31-103. LE HEBEL, F, VIDAL, O., KIENAST, J.R. & GAPAIS, D. 2002. Les 'Porphyroides' de Bretagne meridionale: une unite de HP-BT dans la chaine hercynienne. Comptes Rendus Geoscience, 334, 205-211. LE METOUR, J. & AUDREN, C. 1977. Relations structurales entre 1'orthogneiss ordovicien de Roguedas et son encaissant migmatitique. Consequences sur 1'age des evenements tectonometamorphiques en Bretagne meridionale. Bulletin de la Societe Geologique et Mineralogique de Bretagne, 9,113-123. LE THEOFF, B. 1977. Marqueurs ellipso'idaux et deformation finie. Troisieme Cycle Universitaire Thesis, Rennes 1, Rennes. LEDRU, P., MAROT, A. & HERROUIN, Y. 1986. Le synclinorium de Saint-Georges-sur-Loire: une unite ligerienne charriee sur le domaine centre armoricain. Decouverte de metabasite a glaucophane sur la bordure sud de cette unite. Comptes Rendus de VAcademie des Sciences, 303-11, 963-968. LOPEZ-MUNOZ, M. 1981. Analyse structurale de la partie centrale du synclinorium de Saint-Julien de Vouvantes et de 1'axe Lanvaux-les-Ponts-de-Ce (Massif Armoricain). Bulletin de la Societe Geologique et Mineralogique de Bretagne - serie C, 13,117-123. MATHERON, G. 1955. Application des methodes statistiques a 1'evaluation des gisements. Annales des Mines, 144, 50-75. MATHERON, G. 1962. Traite de Geostatistique Appliquee. Technip, Paris. PARIS, F. & DADET, P. 1988. Geological map - Combourg (1/50 000). BRGM, Orleans. PERCEVAULT, M.-N. & COBBOLD, P.R. 1982. Mathematical removal of regional ductile strains in Central Brittany: evidence for wrench tectonics. Tectonophysics, 82, 317-328. PEUCAT, J.-J., CHARLOT, R., MIDFAL, A., CHANTRAINE, J. & AUTRAN, A. 1979. Definition geochronologique de la phase bretonne en Bretagne Centrale. Etude Rb/Sr de granites du domaine centre armoricain. Bulletin du BRGM, 1, 349-356. PIVETTE, B. 1978. Le synclinorium de Saint-Georges sur Loire, Massif Armoricain. Sa place dans revolution geodynamique de la Bretagne meridionale au Paleozoique. Troisieme cycle universitaire thesis, Rennes 1, Rennes. PLAINE, J. 1976. La bordure sud du synclinorium paleozo'ique de Laval Troisieme cycle universitaire thesis, Rennes 1, Rennes. PLAINE, I, HALLEGOUET, B. & QUETE, Y. 1984.

STRAIN REMOVAL IN CENTRAL BRITTANY Geological map (1/50000) - Questembert (418). BRGM, Orleans. PLUSQUELLEC, Y., ROLET, J. & DARBOUX, J.R. 1999. Geological map - Chdteaulin (1/50,000). BRGM, Orleans. QUETE, Y, PLAINE, J. & HALLEGOUET, B. 1981. Geological map (1/50 000) - Malestroit (386). BRGM, Orleans. RAMSAY, J.G. 1967. Folding and fracturing rocks. McGraw-Hill. RAMSAY, J.G. & GRAHAM, R.H. 1970. Strain variations in shear belts. Canadian Journal of Earth Sciences, 7, 786-813. REGNAULT, S. 1981. Stratigraphie et structure du Paleozoi'que dans le Menez-Belair occidental (Synclinorium median armoricain). Bulletin de la Societe Geologique et Mineralogique de Bretagne -C, 13,1-111.

305

TRAUTMANN, F. 1988. Geological map (1/50000) Nozay (420). BRGM, Orleans. TRAUTMANN, E, BECQ-GIRAUDON, J.F & CARN, A. 1994. Geological map (1/50000) - Janze (353). BRGM, Orleans. UPTON, G.J.G. & FINGLETON, B. 1989. Spatial Data Analysis by Example, vol. 2. John Wiley & Sons, New York. VIDAL,P 1972. L'axe granitique Moelan-Lanvaux (sud du Massif Armoricain): mise en evidence par la methode Rb-Sr de trois episodes de plutonisme pendant le Paleozoi'que Inferieur. Bulletin de la Societe Geologique et Mineralogique de Bretagne, 4, 75-89. VIDAL, P. 1976. L'evolution polyorogenique du Massif Armoricain. Apport de la geochronologie et de la geochimie isotopique du strontium. Troisieme cycle universitaire thesis, Rennes 1, Rennes. ROBARDET, M., BONJOUR, J.L., PARIS, E, MAORZADEC, VIDAL, P. 1980. L'evolution polyorogenique du Massif P. & RACHEBOEUF, PR. 1994. Ordovician, SilArmoricain. Apport de la geochronologie et de la urian, and Devonian of the Medio-North-Armorgeochimie isotopique du strontium. Memoires de ican Domain. In: KEPPIE, J.D. (ed.) Pre-Mesozoic la Societe Geologique et Mineralogique de Bregeology in France and related areas. Springertagne, 21,1-162. Verlag, Berlin. VIGNERESSE, J.L. 1978. Gravimetrie et granites armoriROLET, J. & THONON, P. 1979. Mise en evidence de trois cains. Structure et mise en place des granites Hercomplexes volcano detritiques d'age Devonien cyniens. Troisieme cycle universitaire thesis, inferieur a moyen, Strunien et Viseen inferieur Rennesl, Rennes. sur la bordure nord du bassin de Chateaulin VIGNERESSE, J.L. & BRUN, J.P. 1983. Les leucogranites (feuille Huelgoat 1/50000, Finistere). Impliarmoricains marqueurs de la deformation cations paleogeographiques et tectoniques. regionale: apport de la gravimetric. Bulletin de la Bulletin du BRGM, 1, 303-315. Societe Geologique de France, 25, 357-366. ROLIN, P. & COLCHEN, M. 2001. Les cisaillements her- WACKERNAGEL, H. 1998. Multivariate geostatistics. Springer, Berlin. cyniens de la Vendee au Limousin. Geologic de la France, 1-2,15-44. WATTS, M.S. & WILLIAMS, G.D. 1979. Faults rocks as ROMAN-BERDIEL, T, GAPAIS, D. & BRUN, J.-P. 1997. indicators of progressive shear deformation in the Granite intrusion along strike-slip zones in Gingamp region, Brittany. Journal of Structural experiment and nature. American Journal of Geology, 1, 323-332. Sciences, 297, 651-678. WEBER, C. 1967. Le prolongement des granites de SAGON, J.P 1976. Contribution a I'etude geologique de Lanvaux d'apres la gravimetric et Faeromagla partie orientale du bassin de Chateaulin. Stratinetisme. Memoires du BRGM, 52, 83-90. graphie, volcanisme, metamorphisme et tectonique. Etat thesis, Paris 6, Paris.

This page intentionally left blank

Strain and deformation history in a syntectonic pluton. The case of the Roses granodiorite (Cap de Creus, Eastern Pyrenees) J. CARRERAS, E. DRUGUET, A. GRIERA & J. SOLDEVILA Departament de Geologia, Universitat Autdnoma de Barcelona, 08193 Bellaterra, Barcelona, Spain (e-mail: [email protected]) Abstract: The Roses granodiorite is a Variscan stock with well developed syn- and postmagmatic deformation structures that crops out in the Pyrenean Axial Zone. Analysis of structures reveals a continuous deformation history during and after magma cooling. The deformation history is divided on the basis of mechanical behaviour into two stages: an early one with the development of magmatic structures and a late stage with the development of mylonitic fabrics along shear zones. Both stages are separated in time by the emplacement of aplite-pegmatite dykes. Time of dyke emplacement is thought to coincide with a sudden change in rheology of the granodiorite. The abundance of quartz dioritic enclaves permits the use of shape analysis to characterize the magmatic fabric as a homogeneous deformation. Later solid-state deformation led to the development of an inhomogeneous deformation pattern with different sizes of anastomosing shear zones wrapping around lozenge-shaped domains. The displacement/width ratio measured in shear zones ranges between one and two orders of magnitude. The Roses granodiorite is thought to be a synkinematically emplaced stock which records a continuous deformational history with two distinct deformation stages, both recording bulk finite strains of similar order of magnitude but with a marked difference in finite strain distribution.

Granitoid batholiths and stocks are abundant in the Variscan basement of the Pyrenees (Fig. 1). Although these were initially referred to in the literature as Variscan late-tectonic intrusions (e.g. Maladeta Massif; Zwart 1979), more recent studies have revealed the syntectonic nature of many of these plutons (e.g. Bassies Granite, Gleizes et al 1991). Such studies have been carried out both on the internal fabrics, by the use of the AMS method (Leblanc et al 1996; Gleizes et al. 19980), and on the structures present in the country rock aureoles (Evans et al. 1998). Thus, most Variscan plutons in the Pyrenees are now well characterized by means of deformational features and relative time of intrusion, being generally accepted as emplaced during a main Variscan deformational event. Furthermore, some of these syntectonic plutons are also affected by shear belts developing mylonite bands (Fig. 1). The age and geotectonic significance of the mylonite belts is still under debate (Guitard 1970; Carreras etal 1980; Lamouroux et al. 1980; Saillant 1982; Delaperriere et al 1994). The Roses and Rodes massifs are two synkinematically emplaced stocks of mainly granodioritic composition, located on the Cap de Creus Peninsula, which forms the easternmost outcrop of the Palaeozoic basement in the Pyrenean Axial Zone (Fig. 1). The Roses stock, elongated in a NW-SE direction, was emplaced

into low grade Cambro-Ordovician metasediments, developing a narrow contact aureole of spotted phyllites and hornfelses. The enclosing metasediments exhibit a polyphase history with two main deformation events. The first is responsible for a layer-subparallel penetrative cleavage referred to as the regional foliation. The second is represented by an inhomogeneously distributed crenulation cleavage which postdates the contact metamorphism. This deformation history is revealed by contact metamorphic porphyroblasts that grew over the main foliation, but show a crenulation cleavage related to the late folding phase wrapping around them. The emplacement of the Roses granodiorite started before the late folding phase. Folds related to the crenulation in the enclosing metasediments were contemporaneous with shear zone development in the granodiorite (Carreras & Losantos 1982). These two distinct types of structures, folds in the metasediments and shear zones in the granodiorite, possibly reflect how the two lithological units with different rheological properties responded to deformation. The orientation of all Variscan structures in the Roses granodiorite is affected by a local Alpine overturning. This overturning occurs all along the southern border of the Pyrenean Axial Zone (Vergely 1970; Munoz et al 1986) and

From: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 307-319. 0305-8719/$15.00 © The Geological Society of London 2004.

308

J. CARRERAS ETAL.

Fig. 1. Sketch of the main lithological units and structures in the Variscan of the eastern Pyrenees and location of the Roses granodiorite.

causes originally NE-dipping dextral shear zones to appear as sinistral ones (Carreras 2001).

Progressive development of structures in the Roses granodiorite Like some other Variscan granitoid massifs of the Pyrenean basement (e.g. Querigut Massif; Marre 1973 and Bassies granodiorite; Gleizes et al. 1991), the Roses granodiorite shows an early homogeneous magmatic fabric and a network of low-temperature shear zones (Fig. 2) that gave rise to inhomogeneous mylonitization at low greenschist facies, presumably of Variscan age (Carreras & Losantos 1982). This was followed by late cataclasis developed in narrow bands of millimetre thickness (Simpson et al. 1982). A range of magmatic-state/pre-full crystallization to solid-state/crystal plastic strain fabrics developed throughout the cooling history from hightemperature to the low-temperature regimes, in a manner similar to that proposed by Gapais (1989) in a general model. The existence of this continuous gradation of structures and the difficulty in establishing a clear distinction between the early synmagmatic structures and the fabrics related to the late

shear zones will be discussed below. This will be followed by an analysis of the magmatic fabrics and the structures related to late solid-state shearing. For the sake of simplicity and objectivity, these analyses will be presented separately, using the presence of a swarm of aplite-pegmatite dykes to distinguish between the early structures predating dyke emplacement and those affecting the dykes (Fig. 3).

Magmatic fabric and enclaves The oldest tectonic structure in the granodiorite is a pre-full crystallization fabric (Hutton 1988) or magmatic fabric (Paterson et al. 1989), defined by a preferred orientation of subhedral feldspar, sometimes with tiling between pairs of crystals, along with a weaker alignment of mafic minerals (biotite and amphibole). Additional evidence of magmatic flow is the preferred orientation of elongated enclaves (Fig. 4a) and the presence of schlieren layering in the granodiorite. These are all indicative of synmagmatic deformation. Synmagmatic foliations trend E-W to NW-SE (Figs 3 & 5) in a vaguely curved disposition. The magmatic fabric postdates the regional foliation in the enclosing sediments as is

STRAIN ANALYSIS IN A SYNTECTONIC PLUTON

309

Fig. 2. Geological setting of the Roses granodiorite and the studied area, (a) Structural map of the studied area showing traces of the main shear zones. Location of the area is shown in Fig. 1. (b) Stereoplot showing the orientation of the poles to the mylonite foliation and the associated stretching lineation. (c) Section across the Roses granodiorite.

evidenced from the presence of foliated metasedimentary xenoliths in the granodiorite. The Roses granodiorite is characterized by an abundance of enclaves (Figs 4a, b, c, & 5), most of them microquartz dioritic, with abundant mafic minerals (biotite and amphibole). There is also a small proportion of metasedimentary

xenoliths. Enclave distribution is inhomogeneous, with the presence of some dismembered synplutonic microquartz dioritic sheets. The enclaves are predominantly flattened and show a marked preferred orientation subparallel to the magmatic foliation (Fig. 4a, c). This foliation exhibits little or no deflection around

Fig. 3. Schematic qualitative model of the structural history of Roses granodiorite. The progressive development of structures from high to low temperature can be divided in two major stages with regard to the time of dyke emplacement. Stereoplots of orientation of different structures are also shown.

STRAIN ANALYSIS IN A SYNTECTONIC PLUTON

311

Fig. 4. Mesoscopic scale structures in the Roses granodiorite. (a) Preferred orientation of elongate microquartz dioritic enclaves of a former synmagmatic dyke, (b) Straight aplite vein cutting across the magmatic fabric marked by orientation of enclaves, (c) Syntectonically emplaced leucocratic dyke forming open folds with axial planes parallel to the magmatic fabric, (d) Microgranite with igneous texture emplaced along a curved shear zone, (e) Leucocratic dyke of aplite in the granodiorite cut by sinistral shear zones (f) Brittle-ductile transition conjugate fractures cutting across mylonites.

Fig. 5. Pre-dyke finite strains obtained from two-dimensional analysis of enclave shapes in sections close to the XZ plane, in a domain without significant late shearing. For each locality, the mean strain axial ratio (RS) is shown with an ellipse showing mean orientation and shape of the strain ellipse. Location is shown in Fig. 2.

STRAIN ANALYSIS IN A SYNTECTONIC PLUTON

313

Fig. 6. Part of an extremely elongated enclave of quartz diorite with an axial ratio of about 150:1 from the inner part of a shear zone.

elongated enclaves, from which it is envisaged that the enclaves were slightly more competent during development of the magmatic fabric (Gay 1968; Paterson & Miller 1998). Later solid-state deformation would probably have caused a further decrease in viscosity ratio, in agreement with the work by Tobisch & Williams (1998).

Leucocratic dykes In most domains the magmatic fabric is cut by straight granophyric aplite-pegmatite veins and dykes (Fig. 4b) that vary in size from a few millimetres to about one metre in width and tens of metres in length. They form a swarm of dykes of predominant NNE-SSW original orientation (Fig. 3), cross-cutting the magmatic fabric and the elongated enclaves at a high angle (70-90°). Thus, dykes are presumed to occur in tension fractures related to the same stress/strain field that produced the magmatic foliation. Furthermore, the dykes are slightly folded locally, with fold axial planes parallel to the earlier magmatic fabric (Fig. 4c), suggesting that high-temperature solid-state deformation occurred with a similar orientation in the granodiorite during and after dyke emplacement. Dykes do not show significant refraction when passing through host/enclave interfaces (Fig. 4b). The emplacement of dykes into a network of brittle fractures affecting the granodiorite is presumed to have been coeval with a sudden change in rheology of the granodiorite stock. The granodiorite, at a submagmatic stage with an overpressured residual leucogranitic melt, would have experienced a 'melt-enhanced embrittle-

ment' (Hollister & Crawford 1986; Davidson et al 1994; Brown & Solar 1998; Handy et al 2001).

Shear zones and associated mylonites Post-dyke deformation is mainly related to the development of shear zones, affecting both granodiorites and leucocratic dykes (Figs 4e & 6). In their present orientations shear zones form a NW-SE anastomosing network with predominant steep dips (Figs 2 & 3) and dominantly sinistral strike-slip components. Associated mylonite bands range in thickness from a few millimetres to more than 500 m. In contrast with magmatic structures, which are widely distributed in the entire stock, the shear zones represent a highly inhomogeneous deformation leaving different-sized elongate domains nearly untouched by mylonitization (Simpson et al. 1982). Movement direction is shown by the disposition of the stretching lineation (Figs 2b & 3). The sense of movement can be depicted easily from the marginal obliquity of the mylonitic foliation and the offset of aplitic dykes in appropiate sections (Fig. 4e). In addition there are abundant shear-band structures with dispositions always coherent with the depicted sense of shear. The anastomosing and fan-like pattern, with dominant sinistral shear zones and less abundant dextral ones, could represent a conjugated system. However, shear directions do not form two maxima but a single one coinciding with the intersection direction of differently orientated planes (Figs 2b & 3). Thus, in a stereoplot, poles of foliation planes define a great circle with its

314

J.CARRERASETAL.

pole coinciding with the stretching lineation. This geometrical relation appears to be a common feature of shear belts and is analogous to that described by Ramsay and Allison (1979) in the Maggia Nappe. Most mylonites developed under greenschist facies metamorphic conditions, assisted by strong quartz recrystallization and new growth of chlorite, albite, white mica and epidote. Some broad shear zones contain bands, ranging up to a few metres in width, where the mylonites are completely depleted of quartz, and albitechlorite mylonites form. This mineralogical and chemical transformation of the mylonites may have occurred along zones where fluid was channelled. These zones are probably related to the intrusion of quartz veins and dykes contemporaneous to mylonitization. Although nearly all shear zones formed in the solid-state and under greenschist facies conditions, a peculiar type has been observed. This consists of a complex network of centimetrethick shear zones with fine-grained granitic isotropic material injected along them (Fig. 4d). This particular type of shear zone is considered to represent the earliest stages of localized deformation before the complete crystallization of the granodiorite, and is presented as another argument for the transition from magmatic to solid-state deformation during the cooling of the stock.

fabrics and the superimposed effects of later mylonitization developed under greenschist facies conditions. In addition, rotation and thinning of aplite-pegmatite dykes by shearing and related offset enables shear strain determinations across shear zones. Furthermore, the outcrop conditions along the coastal fringe located in Fig. 2a, with abundant sections close to plane view, enable the establishment of strain profiles across the described structures and the determination of the kinematic pattern associated with shearing.

Deformation predating dykes

Strain analysis has been performed using the Rf/Q technique (Lisle 1985) for enclave populations in different locations. By measuring shape ratio and orientation of the enclaves one can infer information about the total strain and put some constraints on the deformation history. Previously, mafic enclaves have been used for quantitative estimation of finite strain (Ramsay & Rubber 1983; Mutton 1988; Williams and Tobisch 1994; Tobisch & Williams 1998; Wenk 1998). However, the reliability of enclaves as strain markers has been questioned because of the existence of some variables that cannot be related directly to deformation processes (see discussions in Paterson & Vernon 1995 and Tobisch & Williams 1998). Among these variables the most important are: enclave shapes and orientation may reflect other processes Late brittle fractures besides deformation during emplacement, such At the terminations of shear zones, but also as ascent and chamber boundary processes; and cutting across ductile shear zones (Fig. 4f), very the possible contrast in rheology between narrow fractures with associated cataclasites are enclaves and host rocks. present. Although brittle fractures on shear The prevalent subhorizontal sections were zone tips occur in association with shear zone not adequate to perform an exhaustive 3D propagation (Simpson 1983), the cross-cutting analysis. Instead, a general qualitative examinones represent the latest structures formed in ation of enclave population on the entire the granodiorite. These fractures commonly enclave-rich area was made first, from which it form conjugate sets but, in contrast with preced- was established that the XZ section of most ing ductile shear zones, conjugate brittle frac- enclaves is roughly sub-horizontal. After that, a tures always occur at an acute angle to the three-dimensional analysis from three different compressional field, with principal compression sections at one location was performed to estiaxis in an orientation close to north-south. Note mate the orientation and shape of the strain that the final orientation of the compressional ellipsoid. This gave a flattening strain ellipsoid direction is similar and coherent with the orien- with Rxz approximately 3, Ryz about 2.5, a tation required to develop the initial magmatic gently east-dipping XZ plane of finite strain and fabric, direction of dyke emplacement and the east-plunging maximum elongation (X). Then, based on the qualitative assessment and the later shear zones. result from three-dimensional analysis, an extensive two-dimensional strain analysis was Structure and strain profiles carried out on the prevalent sub-horizontal secThe abundance of micro-quartzdioritic enclaves tions, close to the XZ plane. Each location corpermits the use of shape analysis to characterize responds to a homogeneous domain covering a the magmatic to high-temperature solid-state surface ranging from a few square metres up to

STRAIN ANALYSIS IN A SYNTECTONIC PLUTON

about 100 m2, where undeformed dykes provide evidence for the absence of post-dyke strains. In each locality, between 40 and 70 enclaves were measured. The metasedimentary xenoliths and the dismembered microquartz dioritic sheets have been excluded from the shape analysis. Enclaves with unusual shapes and orientations with regard to the main distribution have also been excluded, as they can yield erroneous results. Rf/(j> analyses show a pre-dyke finite strain characterized by axial ratios ranging between 1.1 and 4. Figures 5 and 7 show shapes and orientations of pre-dyke finite strains in two different domains. These finite strains represent the sum of a large magmatic strain and a minor high temperature solid-state deformation. The results probably underestimate the total pre-dyke strain, due to: (1) the presumed viscosity contrast (albeit low) between enclaves and granodiorite; (2) the analysed section, which is not perfectly parallel to the XZ plane of the finite strain ellipsoid; and (3) the earliest deformation in the magmatic history may not have been, or only slightly recorded by the enclaves (Davidson et al. 1994). Not withstanding these problems, the strain measurements using enclave shape give an acceptable estimate of finite strain gradients undergone by the granodioritic stock. Bulk results obtained from different localities indicate that there is a slight inhomogeneity in strain and that this is of low intensity but widespread across the entire stock. Extrapolating the observed mean strain values to the entire pluton, the pre-dyke synmagmatic deformation related to emplacement of the stock represents a bulk horizontal shortening of about 33%.

315

ation patterns following a profile which is close to the one expected for simple shear. The shear zone profile analyses (Fig. 8) reveal a high inhomogeneity of strain at different scales. A mean shear strain value from each profile has been calculated by means of the total displacement/total width relationship. Furthermore, an analysis of the displacement/width ratio of shear zones is constant on shear zones of different size and ranges between one and two orders of magnitude (Fig. 9). Although the point distribution in this graph has a similar slope to the plot presented in Mitra (1979), significant differences exist concerning the intersecting point along the displacement axis. In the case of the Roses granodiorite, displacement in each shear zone is generally at least ten times its width, and therefore about one order of magnitude greater than in the relationship shown by Mitra (1979). Averaging the total displacement/total width relationships obtained from different profiles, and evaluating the total width of shear zones versus the total width of unsheared rocks in the area, a bulk shear strain of 1.4 was estimated. This corresponds to a finite strain axial ratio of 3.3. Extrapolating the shear strain values to the entire stock, by means of the movement along the described network of shear zones, we infer a post-dyke bulk horizontal shortening of about 45%. Discussion and conclusions

The Roses case study shows that enclaves and shear zone profiles are powerful tools for the comparison of deformation during two stages of the stock history. Although the magmatic fabric is less conspicuous than the mylonitic one, it Deformation postdating dykes appears that, considering average values, the In broad mylonitic bands, strain analysis was magnitude of horizontal shortening accommoperformed using enclave shape and orientation dated during the early pre-dyke stage is lower (Fig. 7). Enclaves reflect the variable degree of but of similar order of magnitude to the shortdeformation due to the high inhomogeneity of ening accommodated during solid-state mylonistrain distribution. In sections closely parallel to tization. If we superpose the post-dyke the finite XZ plane, RS values greater than 10 shortening (45%) on the pre-dyke shortening are common, with some enclaves reaching (33%), the granodiorite recorded a bulk horivalues up to RS = 150 (Fig. 6). zontal shortening of about 60%. This value is In thin shear zones, the best and most reliable slightly smaller than expected if the two superstrain profiles (Fig. 8) were obtained by applying posed deformations were coaxial. the Ramsay & Graham (1970) method for shear These two deformation events developed strain determination. This technique is based on under different temperature conditions and are the combined use of sigmoidal pattern of clearly separated in time by the emplacement of mylonitic foliation and change in orientations aplite-pegmatite dykes. However, the presence and offsets of dykes, assuming a simple shear of high-temperature fabrics related to open model for the shear zones. This assumption is folding of synkinematic dykes and the presence supported by the fact that the shear zones are of shear zones with igneous material injected localized along narrow, discrete bands, with foli- along them, indicate the continuity of

316

J. CARRERAS ET AL.

Fig. 7. Structural map and strain analysis of pre-dyke structures and post-dyke shear zones along a coastal section east of the Roses Lighthouse. In both cases strain analysis was performed using the Rf/Q technique for enclave populations. For each locality, the mean strain axial ratio (RS) is shown with an ellipse showing mean orientation and shape of the strain ellipse. Location is shown in Fig. 2.

deformation from the magmatic to the mylonitic stages, with the orientation of the regional shortening direction remaining fairly constant. Furthermore, the preferred disposition of dykes, at high angle to the magmatic fabric, suggests that dyke emplacement occurred along tension fractures compatible with the strain field active during preceding and subsequent events. In this way, there is no need for an interkine-

matic regime or a shift in tectonic regime from compressional to extensional to explain dyke emplacement. The structures observed in the Roses granodioritic stock developed during syntectonic cooling and reveal a continuity of deformation from high- to low-temperature conditions. A change of rheology from high- to low-temperature regime is capable of explaining the

STRAIN ANALYSIS IN A SYNTECTONIC PLUTON

317

Fig. 8. Shear strain analysis across three sections in areas affected by inhomogeneous post-dyke shearing. Mean shear strain values have been calculated using total width versus total displacement relationships. All sections are plane-view.

observed structures and strain pattern. At high temperature, strain is distributed more homogeneously, whereas the progressive temperature drop induces a inhomogeneous strain pattern due to high strain localization along shear zones. Two critical points in the rheological history of the stock can be identified: (1) The first one took place at high temperature when the dykes intruded. At this stage, the granodioritic melt was close to completion of crystallization and probably below the critical fraction for magmatic flow. Thus, this first critical point would not be exactly contemporaneous with the 'rheologically critical melt percentage', theoretically and experimentally inferred as a sudden change of rocks strength caused by the crossing of a critical volume fraction of melt (Arzi 1978; Van der Molen & Paterson 1979; Dell' Angelo & Tunis 1988; Rutter & Neumann 1995; Vigneresse and Tikoff 1999). Most probably, high fluid pressures, induced by the presence of water saturated melts, favoured brittle tension fractures filled with

the residual melts at this stage ('melt enhanced embrittlement'). Furthermore, a strain rate increase could also facilitate local brittle failure. (2) The second critical point corresponds to the low temperature ductile-brittle transition in granitic rocks when shearing is replaced by discrete faults. Most of the overall deformation was accommodated during the ductile stages (magmatic fabric, high-temperature solid-state deformation and shear zones), whereas deformation accommodated during brittle stages was negligible relative to the total deformation. Although no conclusive evidence for the strain regime responsible for synmagmatic deformation has been found in Roses, a dextral transpressional regime is inferred from correlation with the nearby northern Cap de Creus tectonometamorphic belt, where the kinematics are well established and documented (Carreras 2001; Druguet 2001). Moreover, this is also in accord with the work by Gleizes et al (19985) that indicates a dextral transpression as a main tectonic

318

J. CARRERAS £7,4L.

Fig. 9. Diagram of displacement against width measured in different sized shear zones. The displacement-width relationship line by Mitra (1979) has been drawn for comparison.

setting for the emplacement of granitoid batholiths in the Variscan of the Pyrenees. This work was financed by the BTE2001-2616 project (M.C.Y.T). The manuscript has benefited greatly from thoughtful reviews by B. Miller and S. Johnson. We thank P. D. Bons for his kind comments and help with English. We also thank I. Tribe and C. Simpson for careful review and constructive comments on a previous version of the manuscript. The stereographic analysis was done with Stereonet, a program by R. W. Allmendinger.

References ARZI, A. 1978. Critical phenomena in the rheology of partially melted rocks. Tectonophysics, 44, 173-184. BROWN, M. & SOLAR, G.S. 1998. Granite ascent and emplacement during contractional deformation convergent orogens. Journal of Structural Geology, 20,1365-1393. CARRERAS, J. 2001. Zooming on Northern Cap de Creus shear zones. Journal of Structural Geology, 23,1457-1486 CARRERAS,!. & LOSANTOS,M. 1982. Geological setting of the Roses granodiorite, (E-Pyrenees, Spain). Ada Geologica Hispanica, 17, 211-217. CARRERAS, J., JULIVERT, M. & SANTANACH, P. 1980. Hercynian Mylonite Belts in the Eastern Pyrenees: an example of shear zones associated with late folding. Journal of Structural Geology, 2, 5-9. DAVIDSON, C, SCHMID, S.M. & HOLLISTER, L.S. 1994. Role of melt during deformation in the deep crust. Terra Nova, 1,133-142. DELAPERRIERE, E., DE SAINT BLANQUAT, M., BRUNEL,

M. & LANCELOT, J. 1994. Geochronoloige U-Pb sur zircons et monazites dans le massif du Saint Barthelemy (Pyrenees, France): discussion des ages des evenements varisques et pre-varisques. Bulletin de la Societe Geologique de France, 165, 101-112. DELL'ANGELO, L.N. & TULLIS, J. 1988. Experimental deformation of partially melted granitic aggregates. Journal of Metamorphic Geology, 6, 495-515. DRUGUET, E. 2001. Development of high thermal gradients by coeval transpression and magmatism during the Variscan orogeny: insights from the Cap de Creus (Eastern Pyrenees). Tectonophysics, 332, 275-293. EVANS, N.G., GLEIZES, G, LEBLANC, D. & BOUCHEZ, J.L. 1998. Syntectonic emplacement of the Maladeta granite (Pyrenees) deduced from relationships between Hercynian deformation and contact metamorphism. Journal of the Geological Society, London, 155, 209-216. GAPAIS, D. 1989. Shear structures within deformed granites: Mechanical and thermal indicators. Geology, 17,1144-1147. GAY, N.C. 1968. Pure shear and simple shear deformation of inhomogeneous viscous fluid. 2. The determination of the total finite strain in rocks from objects such as deformed pebbles. Tectonophysics, 5, 295-302. GLEIZES, G, LEBLANC, D. & BOUCHEZ, J.L. 1991. Le pluton granitique de B assies (Pyrenees ariegoises): zonation, structure et mise en place. Comptes Rendus de VAcademie des Sciences (Paris), 312, 755-762. GLEIZES, G, LEBLANC, D., SANTANA, V., OLIVIER, P. & BOUCHEZ, J.L. 19980. Sigmoidal structures featuring dextral shear during emplacement of the Hercynian granite complex of Cauterets-Panticosa,

STRAIN ANALYSIS IN A SYNTECTONIC PLUTON Pyrenees. Journal of Structural Geology, 20, 1229-1245. GLEIZES, G., LEBLANC, D. & BOUCHEZ, J.L. 19986. The main phase of the Hercynian Pyrenees is a dextral transpression. In: HOLDSWORTH, R.E, STRACHAN, R.A & DEWEY, IF. (eds) Continental transpressional and transtensional tectonics. Geological Society, London, Special Publications, 135, 267-273. GUITARD, G. 1970. Le metamorphisme hercynien mesozonal et les gneiss oeilles du massif du Canigou, (Pyrenees Orientales). Memoir-es du Bureau de Recherches Geologiques et Minieres, 63, 353 pp. HANDY, M.R., MULCH, A., ROSENAU,M. & ROSENBERG, C.L. 2001. The role of fault zones and melts as agents of weakening, hardening and differentiation of the continental crust: a synthesis. In: HOLDSWORTH, R.E., STRACHAN, R.A., MAGLOUGHLIN, J.F. & KNIPE, R.J. (eds) The nature and tectonic significance of fault zone weakening. Geological Society, London, Special Publications, 186, 305-332. HOLLISTER, L.S. & CRAWFORD, M.L. 1986. Meltenhanced deformation: a major tectonic process. Geology, 14, 558-561. HUTTON, D.H.W. 1988. Granite emplacement mechanisms and tectonic controls: inferences from deformation studies. Transactions of the Royal Society of Edinburgh, 79, 245-255. LAMOUROUX, C, SOULA, J.C., DERAMOND, J. & DEBAT, P. 1980. Shear zones in the granodiorite massifs of the Central Pyrenees and the behaviour of these massifs during the Alpine orogenesis. Journal of Structural Geology, 2, 49-53. LEBLANC, D., GLEIZES, G, Roux, L. & BOUCHEZ, J.L. 1996. Variscan dextral transpression in the French Pyrenees: new data from the Pic des TroisSeigneurs granodiorite and its country rocks. Tectonophysics, 261, 331-345. LISLE, R.J. 1985. Geological strain analysis: a manual for the Rf/(/) technique. Pergamon, Oxford. MARRE, J. 1973. Le complexe eruptif de Querigut. Petrologie, Structurologie, cinematique de mise en place. These Toulouse, 543 pp. MITRA, G. 1979. Ductile deformation zones in Blue Ridge basement rocks and estimation of finite strains. Geological Society of America Bulletin, 90, 935-951. MUNOZ, J.A., MARTINEZ, A. & VERGES, J. 1986. Thrust sequences in the Spanish eastern Pyrenees. Journal of Structural Geology, 8, 399-405. PATERSON, S.R. & MILLER, R.B. 1998. Stoped blocks in plutons: paleo-plumb bobs, viscometers, or chronometers? Journal of Structural Geology, 20, 1261-1272. PATERSON, S.R. & VERNON, R.H. 1995. Bursting the

319

bubble of ballooning plutons: A return to nested diapirs emplaced by multiple processes. Geological Society of America Bulletin, 107,1356-1380. PATERSON, S.R., VERNON, R.H. & TOBISCH, O.T. 1989. A review of criteria for the identification of magmatic and tectonic foliations in granitoids. Journal of Structural Geology, 11, 349-363. RAMSAY, J.G & ALLISON, 1.1979. Structural analysis of shear zones in an Alpinised Hercynian granite. Schweizerische Mineralogische und Petrographische Mitteilungen, 59, 251-279. RAMSAY, J.G. & GRAHAM, R.D. 1970. Strain variations in shear belts. Canadian Journal of Earth Sciences, 7, 786-813. RAMSAY, J.G. & HUBBER, M.I. 1983. The techniques of modern structural geology, Vol. 1: Strain analysis. Academic Press, London. RUTTER, E.H. & NEUMANN, D.H.K. 1995. Experimental deformation of partially molten westerly granite under fluid-absent conditions, with implications for the extraction of granitic magmas. Journal of Geophysical Research, 100 (B8), 15697-15715. SAILLANT, J.P. 1982. La faille de Merens (Pyrenees Orientales) microstructures et mylonites. These 3eme cycle, 297 pp, Univ. Paris VII. SIMPSON, C. 1983. Displacement and strain patterns from naturally occurring shear zone terminations. Journal of Structural Geology, 5, 497-506. SIMPSON, C., CARRERAS, J. & LOSANTOS, M. 1982. Inhomogeneous deformation in Roses granodiorite. Acta Geologica Hispanica, 17, 219-226. TOBISCH, O.T. & WILLIAMS, Q. 1998. Use of microgranitoid enclaves as solid state strain markers in deformed granitic rocks: an evaluation. Journal of Structural Geology, 20, 727-743. VAN DER MOLEN, I. & PATERSON, M.S. 1979. Experimental deformation of partially-melted granite. Contributions to Mineralogy and Petrology, 70, 299-318. VERGELY, P. 1970. Etude tectonique des structures pyreneenes du versant sud des Pyrenees orientales. These 3eme cycle, Faculte Sciences, Universite de Montpellier. VIGNERESSE, J.L. &TiKOFF,B. 1999. Strain partitioning during partial melting and crystallizing felsic magmas. Tectonophysics, 312,117-132. WENK, H.R. 1998. Deformation of mylonites in Palm Canyon, California, based on xenolith geometry. Journal of Structural Geology, 20, 559-571. WILLIAMS, Q. & TOBISCH, O.T. 1994. Microgranitoid Enclaves shapes and magmatic strain histories: Constraints from drop deformation theory. Journal of Geophysical Research, 99, 24359-24368. ZWART, HJ. 1979. The geology of the central Pyrenees. Leidse Geologische Mededelingen, 50,1-74.

This page intentionally left blank

Shear zones and metamorphic signature of subducted continental crust as tracers of the evolution of the Corsica/Northern Apennine orogenic system G. MOLLI1'2 & R. TRIBUZIO3'4 Dipartimento di Scienze della Terra, Universitd di Pisa, Via S. Maria 53, 1-56126 Pisa, Italy (e-mail:[email protected]) 2 CNR Istituto di Geoscienze e Georisorse,Via G. Moruzzi, 1-56124 Pisa, Italy 3 Dipartimento di Scienze della Terra, Universitd di Pavia, Via Ferrata 1, 1-27100 Pavia, Italy (e-mail: [email protected]) 4 CNR Istituto di Geoscienze e Georisorse, Sezione di Pavia, Via Ferrata 1, 1-27100 Pavia, Italy l

Abstract: This paper focuses on new data concerning the deformation and metamorphic history of continental margins of the Mesozoic Ligurian Tethys. In the Tenda massif (NE Corsica), a slice of the European-Iberian continental margin, contractional shear zones show HP/LT metamorphic assemblages and top-to-the-west kinematics. These shear zones are overprinted by greenschist facies exhumation-related structures showing top-to-the-SW sense of transport and then top-to-the-NE extensional shearing. The presence of HP/LT metamorphism, together with the kinematics of syncontractional shear zones, supports the classic view of Cretaceous-Eocene east-vergent 'alpine-subduction' during the early evolution of the Corsica belt. By taking into account structural and metamorphic data on Tuscan continental units belonging to the other side (Adria margin) of the former Mesozoic Ligurian ocean, we ascribe the Corsica/Northern Apennine system to a polycyclic orogen.

Analyses of crustal scale shear zones and mesoto microscale kinematic criteria were first applied to unravel plate tectonic scale reconstructions in alpine Corsica, and to support a model of intraoceanic subduction blocked by underthrusting of the continental crust beneath the oceanic lithosphere (e.g. Mattauer et al. 1981; Gibbons & Horak 1984; Harris 1985; Warbourton 1986). Recently, however, these tenets have been challenged, and the importance of exhumation-related greenschist structures and kinematics have been explored (Jolivet et al. 1990,1998; Lahondere 1991; Daniel et al. 1996), resulting in uncertainty about the early contractional history (Jolivet et al. 1998; Rossetti et al. 2002). This debate has important implications for the development of the Corsica/Northern Apennine orogenic system (review in Alvarez 1991). Most recent studies (e.g. Lahondere et al. 1999; Padoa & Durand Delga 2001; Bortolotti et al. 2001; Faccenna et al. 2001; Rossetti et al. 2002), with only a few exceptions (Cello et al. 1996; Doglioni et al. 1998; Malavielle et al. 1998; Michard et al. 2002), follow the proposal of Principi & Treves (1984) in considering the evolution of alpine Corsica and Northern Apennines as the

development of an accretionary wedge formed by a continuous Cretaceous to Oligocene westdipping subduction of the Ligure-Piemontese ocean beneath the Corsica/European continental crust, fitting into a monocyclic-type doublevergent orogenic model (Fig. la). In this model, western 'Hercynian' Corsica is considered as the backstop of the Apennine wedge. Some of the ophiolitic units of alpine Corsica are regarded as the deeper part of the accretionary complex, backthrust on the Corsican crust by corner flow (Cowan & Silling 1978) to the rear of the Apennine accretionary system. This model does not take into account previous studies, which considered the early structures in the alpine Corsica units to be produced by an east-dipping 'alpine' subduction (Fig. Ib) (Mattauer & Proust 1975; Caron 1977; GLOM 1977; Mattauer et al. 1981; Faure & Malavielle 1981; Gibbons et al. 1986; Bezert & Caby 1988; Jolivet etal. 1990). We have tackled the problem by studying the larger slices of continental crust that make up alpine Corsica i.e. the Tenda unit. In order to constrain the early deep kinematics and the possible tectonic setting of the Tenda unit, we have focused our attention mainly on shear

From: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 321-335. 0305-8719/$15.00 © The Geological Society of London 2004.

322

G. MOLLI & R. TRIBUZIO

Fig. 1. Tectonic models proposed for the Corsica/Northern Apennine region: (a) continuous Cretaceous-Oligocene/Miocene west-dipping Apenninic subduction (e.g. Principi & Treves 1984; Jolivet et al. 1998; Rossetti et al. 2002); (b) Alpine-type east vergent intraoceanic subduction (Mattauer et al. 1981) followed by Middle Eocene continental collision (e.g. Gibbons & Horak 1984; Warbourton 1986).

zone structures, peak P-T conditions and metamorphic gradients. Furthermore, we have considered the available data on Adria-derived continental nappes (i.e. Tuscan metamorphic units) of the inner Northern Apennine, to analyse the character of the Corsica/Northern Apennine orogenic system from the Cretaceous to Late Oligocene.

Tectonic setting of alpine Corsica Corsica is located in the northern part of the western Mediterranean and is subdivided into two principal geological domains: a western area mainly formed by Late Hercynian granitoids and minor relicts of host rock-basement; and an eastern area characterized by continental and oceanic-derived units deformed during the Alpine orogeny (alpine Corsica) (Fig. 2). Corsica and Sardinia are traditionally regarded as a microblock welded to southern France-northern Iberia until the Middle Oligocene, when rifting and then drifting in the Ligure-Provencal basin took place (Guegen et al 1997; Carminati et al. 1998; Speranza et al 2002; Rollet et al 2002 and references therein). The western Hercynian domain is correlated with the Maures-Esterel basement in southern France, whereas alpine Corsica is regarded as the southern prolongation of the western Alps,

Fig. 2. (a) Tectonic setting of Corsica within the western Mediterranean, (b) Geological setting of the studied areas within the Corsica/Northern Apennine framework.

through the Western Liguria and Voltri Group Alpine units (Durand Delga 1984 and references therein). Corsica (Fig. 3) comprises a series of largescale nappes and can be subdivided into four major composite units (Nardi 1968; Caron 1977; Mattauer et al 1981; Durand Delga 1984; Warbourton 1986). From bottom to top these units are: (1) the autochthon 'Hercynian' Corsica; a mainly undeformed series of the Corsica plate (basement rocks and their sedimentary cover up to Eocene); (2) the deformed part of the Corsica continental margin, i.e. the Tenda massif and the more internal Serra di Pigno/Farinole units (mainly Hercynian granitoids with Permian to Mesozoic volcano-sedimentary and sedimentary cover);

SHEAR ZONES AND METAMORPHISM

323

1984; Lahondere 1991) and by a (retrograde) epidote-blueschistfaciesmetamorphism (Harris 1985; Fournier et al 1991; Lahondere 1991; Caron 1994). No general agreement exists regarding the metamorphic evolution of the Tenda unit. Peak pressure conditions are particularly uncertain. Gibbons & Horak (1984) gave pressure estimates of 0.6-0.9 GPa, based on the celadonite content of phengites. In contrast, Lahondere (1991), Jolivet etal (1998), and Lahondere et al (1999) give values lower than 0.5 GPa, suggesting that the Tenda massif was not involved in the subduction process (Tribuzio & Giacomini 2002).

Structural and metamorphic history of the Tenda massif Geological outline

Fig. 3. (a) Tectonic map of northern Corsica showing the main tectonic units, (b) Geological section through northern Corsica (modified after Jolivet et al 1990; Dalian & Puccinelli 1995; Malavielle et al. 1998).

(3) the 'Schistes Lustres' composite nappe formed by ophiolitic sequences (mantle ultramafics, gabbros, pillow lavas and associated Jurassic to Cretaceous metasediments); and (4) the Balagne/Nebbio/Macinaggio system, i.e. ophiolitic and continental units of internal 'Ligurian-Adria' affinity. These uppermost nappes and their contact with the underlying units are unconformably overlain by Early Miocene sediments (e.g. the St Florent limestone, Dalian & Puccinelli 1995; Ferrandinieffl/. 1998). The external autochthon basement and cover are basically unmetamorphosed or only locally affected by subgreenschist facies recrystallization along fault zones (Egal 1992). The uppermost Balagne/Nebbio/Macinaggio units, show prehnite/pumpellyite facies assemblages in basic rocks. The 'Schistes Lustres' nappe and the Serra di Pigno/Farinole gneissic units are characterized by relict eclogitic association (Dal Piaz & Zirpoli 1979; Caron et al. 1981; Pequignot et al

The Tenda massif (Figs 3 & 4) is exposed for nearly 200 km2 from St Florent-Ostriconi southward to Ponte Leccia, and is characterized by 280-300 Ma old Palaeozoic granitoids, mainly amphibole/biotite granodiorites and leucomonzogranites (Rossi et al 1994). In the southern part, the massif is intruded by a gabbroic complex (the Bocca di Tenda gabbro), which shows well preserved magmatic features (Ohnestetter & Rossi 1985). This complex was dated by single zircon lead evaporation at 274 ± 4 Ma (Lahondere et al 1999) and consists mainly of olivine-gabbronorites, gabbronorites and hornblende-bearing diorites/tonalites (Ohnestetter & Rossi 1985; Tribuzio etal 2001). Granitoids and gabbros are cross-cut by doleritic and peralkaline rhyolite dykes, which both show chilled margins against the host rocks (Tribuzio et al 2001; Tribuzio & Giacomini 2002). Remnants of a pre-Carboniferous basement occur locally in the southern and northwestern side of the massif (Delcey & Meunier 1966; Nardieffl/. 1978; Jourdan 1988; Rossi ef al 1994). The basement rocks are associated with a Permian volcano-sedimentary cover, well exposed in the northern part of the massif (Agriates's Desert). A thin Mesozoic metasedimentary cover (Mattauer et al 1981; Durand Delga 1984) crops out at the eastern and southern sectors of the massif. These sediments are mainly represented by Triassic quarzites and micaceous arkoses, Jurassic dolomitic and calcitic marble, and probable Cretaceous polygenic metaconglomerates, principally formed from basement-derived debris with subordinate carbonatic clasts (Jourdan 1988). The entire

324

G. MOLLI & R. TRIBUZIO

Fig. 4. Geological map of the Tenda massif with the location of relict HP shear zones and related kinematics, (a) our data (b) after Mattauer et al. (1981) and Jourdan (1988). Map compiled after this work, Delcey & Meunier (1966), Nardi et al. (1978), Jourdan (1988), Dalian & Puccinelli (1995), Rossi et al (1994) and Lahondere et al. 1999.

of the autochthon basement on the Tenda massif. Slickensides on minor fault planes testify to the predominant sinistral strike-slip character of this fault zone, which at least locally appears Deformation history of the Tenda massif to be reworked with normal movement The Tenda massif represents a regional open (Jourdan 1988; Daniel et al. 1996). In the central/southern Tenda (south of antiform (Fig. 3b), bounded by high angle faults in its western and eastern sides. This antiform Urtaca), the regional scale antiform is formed by has an amplitude of nearly 10 km in the northern kilometre scale antiform/synform pairs, locally sector of the massif, with its western limb associated with a steep crenulation to disjunctive deformed by a fault zone in which Tenda- cleavage observable in orthogneiss, as well as in derived rocks (basement and cover) and cover metasediments. The development of these autochthon series (basement and cover) are fold structures, showing local en-echelon trends, cataclastically deformed. Splays of this fault can be related to the activity of wrench faulting zone locally produce east-vergent overthrusting (whose age can be constrained as post-Eocene cover section shows the same metamorphic imprint as the basement (Mattauer et al. 1981).

SHEAR ZONES AND METAMORPHISM

325

Fig. 5. (a) Casta granodiorite, with synmagmatic deformed mafic xenolite. (b) Mylonitic orthogneiss, bounding the Casta granodiorite, with solid state deformed mafic xenolite. (c) Mylonitic blueschist facies metagranite; shear bands and asymmetrical porphyroclasts define a top towards 270° shearing, (d) Photomicrograph of blueschist granitic mylonite. (e) Asymmetrical porphyroclast with synkynematic Na-amphibole growth (detail of (d)). (f) Photomicrograph of mylonitic metadolerite with a well defined shear band system.

and pre-Burdigalian) affecting the boundaries of the Tenda massif (Maluski et al 1973; Waters 1990; Lahondere et al 1999). The geometry of deformation inside the Tenda massif is controlled by the predominance of granitoids deformed at deep structural levels (Ramsay & Allison 1979; Choukroune & Gapais 1983; Gapais et al 1987). A heterogeneous deformation pattern characterizes the entire massif with domains without fabric (i.e. isotropic) or with only a magmatic grain-shape

fabric locally preserved (e.g. the Casta granite, Fig. 5a) surrounded by mylonitic orthogneisses and/or mylonites (Fig. 5b). Foliated granitoids with S-L fabrics show the development of blueschist and/or greenschist facies minerals. The dominant fabric at regional scale is represented by greenschist facies (GS) tectonites. Based on kinematic criteria (stretching/mineral lineations and the associated shear sense), locally controlled by direct overprint relationships, we subdivided the GS structures

326

G. MOLLI & R. TRIBUZIO

Fig. 6. (a) Examples of the GS2 high strain zone with shear band systems indicating top-to-the-NE orthogneiss east of M. Buggentone, southeastern Tenda. (b) Shear bands in metasedimentary cover near S. Florent (RN D41) East Tenda Shear zone, (c) Conjugate system of centimetre-scale shear zones related to the GS2 deformation stage. North of M. Astu, central Tenda. (d) Example of orthogneiss with top-to-the-west shear bands (RN D62 south F.di Poragghia). (e) Asymmetrical boudinage of aplitic dyke indicating top-to-the-west shearing in orthogneiss south of Bocca di Tenda. All examples are observed in XZ sections.

into two groups related to distinct stages of deformation. The younger GS2 stage is characterized by a predominant top-to-the-NE shearing (Figs 6a,b, & 7a) localized in the eastern part of the massif (East Tenda Shear zone, ETSZ, of Jolivet et al

1990; Daniel et al 1996, but see also Waters 1990) towards the contact with the overlying Schistes Lustres. In the northern Tenda Massif where the ETSZ was first recognized, it comprises an inhomogeneous shear zone formed by several large scale shear bands up to 100 m thick

SHEAR ZONES AND METAMORPHISM

Fig. 7. Equal area projection (lower hemisphere) of foliation planes (open circles) and stretching/mineral lineations (black circles), (a) Greenschist facies shear zones GS2 top-to-the-NE (in the ETSZ) extensional stage (mean lineation direction 12/048). (b) Greenschist facies shear zones GS1 top-to-the-west stage (mean lineation direction 6/068). (c) High pressure/low temperature shear zones blueschist stage (mean lineation direction 2/090).

affecting granitoids and their metasedimentary cover (Jolivet et al 1990; Fournier et al. 1991;

327

Daniel et al. 1996). In the central-southern Tenda (south of Sorio village), the ETSZ consists of at least two significant high strain zones, several metres thick, separated by low strain domains in which subhorizontally crenulated orthogneiss can be recognized. Inside the massif, centimetre-scale shear zones with top-to-theNE kinematics can be related to the same deformation stage of the ETSZ. Conjugate shear zones associated with top-to-the-NE and top-tothe-SW shear sense (Fig. 6c) are present locally and may be related to partitioned coaxial strain far from the ETSZ (cf. Daniel et al 1996). Available radiometric data (Ar/Ar on phengites) constrain the age of the GS2 stage to between 35 and 25 Ma (Maluski 1977; Carpena et al. 1979; Brunet et al. 2000). The ETSZ has been interpreted by Jolivet et al. (1990) and Daniel et al. (1996) as a major extensional shear zone related to crustal-scale thinning bounding the Tenda massif, which is considered to be a metamorphic core complex. However, the pervasive GS1 deformation stage on the massif scale is associated with an ENE/WSW stretching lineation (Fig. 7b) and top-to-the-west kinematics (Fig. 5c,d,e). During this stage, westward thrusting of the Tenda unit over the western and external 'autochthon' units was realized, as observed in the Urtaca tectonic window (see also Nardi et al. 1975; Jourdan 1988 and discussion below). The heterogeneous deformation pattern (Fig. 4) that characterizes the Tenda massif allows us to recognize and analyse domains of mylonitic orthogneisses and mylonitized doleritic and peralkaline rhyolitic dykes (Fig. 5c,d,f) with a well preserved HP/LT fabric that was unaffected by greenschist facies static and/or dynamic retrogression. The HP microfabric in mylonitized granitoids (Fig. 5c,d,e) and rhyolitic dykes show the typical features of quartz-feldspathic mylonites, suggesting that bulk ductile flow during HP metamorphism was accommodated through dislocation creep of quartz. The HP relict structural domains are characterized by east-west orientated stretching and/or mineral lineations defined by sodic amphibole and are associated with a top-to-the-west shear sense (Fig. 7c). These kinematic indicators were observed at kilometre distances (Fig. 4), implying a regional significance to this HP east-west trend and thus strongly constraining the deep deformation history of the Tenda unit. In contrast with previous descriptions (Mattauer et al 1981; Gibbons & Horak 1984), we found HP relicts (structures and/or relict minerals) not only in the eastern upper part of the massif towards the contact of Schistes Lustres,

G. MOLLI & R. TRIBUZIO

328

Table 1. Significant HP/LT mineral assemblages in different rock types in the Tenda massif (thin lines represent minerals that are only present locally) HIGH PRESSURE ASSEMBLAGES

Qtz-diorites, tonalites and dolerites

Granitoids

Peralkaline rhyolites

Gabbronorites

Na-amphibole (riebeckite-ferroglaucophane) Phengite (Si = 3,5-3,6 apfu) Epidote Chlorite Albite K-feldspar Quartz Na-Cpx (Jd up to 46 mol %) Al-poor (minor than 1,2 apfu) hornblende Titanite Calcite

but distributed throughout the structural thickness of the unit. Where they were still recognizable, for example in the southern area around the Bocca di Tenda gabbro, the geometries of the shear zones fitted the anastomosed distribution pattern expected for crustal scale thrusting.

Metamorphic history of the Tenda massif The wide variety of rocks with different compositions (expecially in the southern part of the massif where the Bocca di Tenda gabbro crops out) allowed the peak metamorphic conditions at 0.8 GPa/300 °C to 1.1 GPa/500 °C to be determined (Tribuzio & Giacomini 2002). The rocks evolved from the gabbroic sequence (quartz diorite/tonalites), basalt doleritic dykes and granitoids are characterized by epidoteblueschist assemblages as they show the coexistence of riebeckite/ferroglaucophane, epidote, celadonite-rich phengite (Si = 3.5-3.6 apfu) and albite (Table 1). The per alkaline rhyolite dykes show an unusual metamorphic paragenesis, defined by jadeite-bearing (up to 46 mol%) aegirine, riebeckite, celadonite-rich phengite (Si = 3.5-3.6 apfu), quartz, albite and K-feldspar. The Mg-rich rocks (olivine gabbronorites to gabbronorites) are characterized by the absence of blue amphibole. Deformed gabbronorites show the development of a mineral association that can be related to the epidote-amphibolite facies, as it displays the coexistence of Al-poor horneblende (Al < 1.2 apfu), albite, epidote and celadonite-rich phengite (Si = 3.5 apfu).

The occurrence of epidote-amphibolite facies assemblages in Mg-rich rocks allow the peak P-T metamorphic conditions to be constrained at 1.0 ± 0.1 GPa and 450 ± 50 °C. These values attest to a geothermal gradient (dT/dP) of 10/13 °C km"1, thus suggesting a subductionrelated tectonic setting (Tribuzio & Giacomini 2002) of 'slow-type' (see discussion below). The age of//P/Lrmetamorphism in the Tenda massif is not well defined. A separate of celadonite-rich phengite (Si = 3.5 apfu) from a deformed granitoid of the Northern Tenda massif has yielded a discordant 39Ar/40Ar spectrum that regularly increases during step-heating, from about 25 Ma to 47 Ma (Brunet et al 2000). This might suggest that the high-pressure metamorphism had a minimum age of 47 Ma. Phengite compositions in deformed GS2 granitoids (Si = 3.2-3.3 apfu) suggest a decompression at pressure lower than 0.5 GPa (Fig. 8). The decompression was most likely coupled with a temperature decrease, as suggested by the amphibole compositional variations (e.g. outward decrease of Al, Na and Ti) in deformed gabbronorites. Ar/Ar investigations on white micas from deformed granitoids show that the youngest greenschist facies recrystallization occurred at around 25 Ma (Brunet etal 2000). In addition, phengitic micas with intermediate Si compositions have given values of 35-37 Ma (Brunet et al 2000). We therefore consider the 47 to 37 Ma period as related to the cooling and concomitant exhumation of the Tenda unit, in agreement with the work of Malavielle et al. (1998).

SHEAR ZONES AND METAMORPHISM

Fig. 8. Inferred pressure-temperature-time path of the Tenda massif. Gin-out taken from Maresch (1977), the reaction pumpellyite + chlorite —> actinolite + epidote after Liou et al (1983), lawsonite-clinozoisite transition after Barnicoat & Fry (1986), the lower stability limit of barroisite from Ernst (1979), oligoclase in reaction from Maruyama et al (1983), the reaction curve for Na-clinopyroxene (Jd46) + quartz -> albite was calculated with the 3.1 version of THERMOCALC program (Holland & Powell 1998, and references therein). The reaction curve for Mg-phengite (Si = 3.6 apfu) —> quartz + Kfeldspar + phlogopite + H2O) is after Massone & Szpurska (1997). Ar/Ar data after Brunet et al. (2000), Jourdan (1988) and Maluski (1977). Fission track apatite ages (Ap FT) after Janki et al (2000) and Cavazza et al (2001).

To sum up, three major deformation events are recorded in large scale shear zone development of the Tenda massif: (1) early stages of deformation under HPILT metamorphic conditions, recorded by localized shear zones showing top-to-trie-west kinematics. The age of this event is possibly older than 47 Ma; (2) syncontractional exhumation related to greenschist-facies retrogression and with westward thrusting (GS1). This history can be constrained between 47 and 37 Ma; (3) top-to-the-NE extensional shearing (GS2), between 35 and 25 Ma old, with partial reactivation and overprinting of previous fabrics.

Structural and metamorphic evolution of the inner Tuscan metamorphic units A recent synthesis of data and interpretations of the regional geology of the Northern Apennines is reported in Carmignani et al (1995), Jolivet et

329

al (1998) and Cerrina Feroni et al (2002). The former western margin of the Adria plate is exposed below the remnants of the Ligurian accretionary wedge on the Thyrrenian side of the Northern Apennines (Ligurian and sub-Ligurian units). This margin is represented by different thrust sheets forming the so-called Tuscan units (Elter 1975). Part of these continental-margin units, mainly associated with cover sequences of Triassic/Late OligoceneMiocene age (e.g. the Tuscan nappe), remained at high structural levels during the whole Apennine history. Other portions of the same continental margin were more deeply underthrusted and are now exposed in tectonic windows (Tuscan metamorphic units, see Fig. 2b) below the overlying lower grade composite nappesystem (Ligurian accretionary wedge units and Tuscan nappe). Some of the Tuscan metamorphic units show high-pressure greenschist-facies peak assemblages (Mg-chloritoid and kyanite in metapelites) that developed at 0.6-0.8 GPa and 400-500 °C (Massa unit in the Alpi Apuane region, Franceschelli et al 1986; Jolivet et al 1998; Molli et al 20000, b). In southern Tuscany, high-pressure/low-temperatureassemblages (Fe-Mg carpholite in metapelites) were recognized. Estimated pressure and temperature conditions vary from 0.6-1.0 GPa and 350-380 °C for the M.Leoni/Monticiano Roccastrada in the Montagnola Senese area (Giorgetti et al 1998) to 1.0-1.2 GPa and 350-420 °C for the Verrucano of Monte Argentario (Theye et al 1997; Jolivet et al. 1998). In these continental units, the structural history is characterized by an early generation of syn- to late-peak metamorphic structures related to contraction and nappe stacking, deformed by younger exhumationrelated structures (Molli et al 2000a,5 and references therein). The early stages of contractional deformation, 27-20 Ma ago (Kligfield etal. 1986; Deino et al 1992; Brunet et al 2000) resulted in SW to NE directed overthrusts and NE-facing tight to isoclinal recumbent folds at regional scale with flat-lying axial surfaces. These structures are traditionally related to the underthrusting of the Adria continental margin following the west-vergent subduction of the relict of the Ligurian Tethys ocean (Carmignani et al 1978; Carmignani & Kligfield 1990).

Discussion The deformation of continental (and oceanic) crust is characteristically heterogeneous in nature and localized linked faults and shear zone systems represent a common tectonic setting

330

G. MOLLI & R. TRIBUZIO

(Rutter et al. 2001). The results of our analyses in the Tenda massif show how a slice of continental crust was internally deformed during a subduction/exhumation cycle. The tectonic history has been unravelled by studying incomplete shear zone reactivation, associated with an increase in partitioning and localization of the strain during the exhumation, these have been constrained by metamorphic assemblages, kinematics (shear direction and sense of transport), overprinting relationships and radiometric ages. The results of this study carries implications for the tectonic evolution of the Corsica/Northern Apennine system with regard to monocyclic or polycyclic orogenic processes. This study shows that the Corsican crust of the Tenda massif: underwent peak metamorphism in epidote/blueschist facies at about 1 GPa and 450 °C, possibly earlier than 47 Ma; was deformed during subduction with top-to-thewest kinematics; and presently crops out as a core of antiformal stack overthrusted on autochthonous Corsica basement. These features support an east-dipping 'alpine' subduction of the Corsica basement. However, the inner Northern Apennine ophiolitic units (Gorgona, Roselle, Argentario, Giglio) and continental Adria-derived units underwent highpressure/low-temperature metamorphism or high-pressure greenschist facies metamorphism at 27-20 Ma. This supports a polycyclic development of the Corsica/Northern Apennine orogenic system as suggested by Boccaletti et al. (1971), Elter & Pertusati (1973), Dal Piaz (1974), Reutter et al. (1978), Doglioni et al (1998) and Michard et al. (2002). For the Corsica/Apennine area we suggest the following tectonic evolution (Fig. 9): (1) a Cretaceous stage of east-dipping intraoceanic subduction beneath the Nebbio microcontinent, which was related to ESE movement of the Iberian plate. This is recorded in the Corsican Schistes Lustres by development of eclogites dated at 84 ± 5 Ma (Lahondere & Guerrot 1997). A low to very low geothermal gradient for these ophiolitic units is shown by local occurrence of eclogites with coexisting almost pure jadeite and quartz (Caron et al. 1981; Lahondere 1988) and by lawsonite-bearing eclogites (Pequignot et al 1984; Lahondere 1991; Padoa 2001), which we ascribe to a fast and steep subduction; (2) (Fig. 9a) the subduction pulled down the thinned Corsican continental margin starting from the Late Cretaceous, as testified by 65 Ma old ages of continental derived

eclogites (Farinole-Serra di Pigno unit, Brunet et al 2000). The continental margin was progressively subducted, reducing the subduction rate and increasing the geothermal gradient (Tribuzio and Giacomini 2002); (3) (Fig. 9b) the involvement of thick continental crust (Tenda massif) blocked the subduction during Paleocene-Middle Eocene. Due to buoyancy forces and shearing, the Corsican crust failed and started to be exhumed, whilst break-off and detachment of subducted oceanic lithosphere took place (Van den Beukle 1992; Von Blackenburg and Davis 1995; Chemenda et al 1996; Malavielle et al 1998; Tribuzio and Giacomini 2002). The presence of a still open oceanic domain eastward of the Corsican accretionary wedge (central Tuscany and southern Italy) allowed subduction flip and the beginning of westward subduction to drive the early development of the Apennines (Fig. 9c). This event was followed by the Oligocene-Miocene calc-alkaline volcanism on the western side of Sardinia-southern Corsica, the associated back-arc rifting (30-21 Ma), the formation of oceanic crust (21-16 Ma) in the Ligure-Provencal basin, and the rotation of the Sardinian-Corsica block far from Iberia/Europe mainland (Scandone 1979; Rehault et al 1984; Serri et al. 1993; Speranza et al 2002 and references therein). The exhumation of the Tenda unit is considered as mainly related to a syncontractional tectonic history (Fig. 9b) during the early Alpine cycle, and only the latest stages (realized within the back-arc extensional/transtensional frame) are related to the Apennine subduction (Fig. 9c, d), in agreement with Malavielle etal (1998). In conclusion, our data on shear zones and the metamorphic signature of the Tenda Massif reaffirm alpine Corsica as a case study for intraoceanic subduction blocked by the underthrusting of the continental crust, as suggested more than 20 years ago by Mattauer et al (1981). Our ongoing work is focused on quantitative analyses of structures in terms of volume of deformed domains, evolution of their rheology from rock microfabrics and deformation/metamorphism/rock strength relationships. P. Elter and A. Puccinelli are thanked for their discussion and suggestions on the geology of Corsica. M. Krabbendam and M. Sandiford provided useful comments and suggestions during their review of the submitted manuscript. Editorial handling and remarks

SHEAR ZONES AND METAMORPHISM

331

Fig. 9. Tectonic evolution of Alpine Corsica within the framework of the Corsica/Northern Apennine erogenic system, (a) Subduction of the distal Corsican margin and eclogite metamorphism (Late Cretaceous), (b) Involvement of thick Corsican crust (Tenda), blocking of subduction, slab-break-off below 50 km and flip of subduction polarity, (c) Oligocene regional upper plate crustal extension related to Apenninic subduction; (d) Oligocene/pre-Early Miocene wrench and then late normal faulting possibly connected with Corsica/Sardinia rotation and the beginning of a retreat of Apenninic subduction zone. of I. Alsop are also greatly appreciated. This work was supported by the Universities of Pavia and Pisa, CNR and COFIN funds. G.M is an external collaborator of

RETREAT project, Continental Dynamics Program of National Science Foundation (publ. N°3). This paper is dedicated to the memory of our friend Graziano Plesi.

332

G. MOLLI & R. TRIBUZIO

References ALVAREZ, W. 1991. Tectonic evolution of the CorsicaApennines-Alps region studied by the method of successive approximations. Tectonics, 10,936-947. BARNICOAT, A.C & FRY, N. 1986. High-pressure metamorphism of the Zermatt-Saas ophiolite zone. Journal of the Geological Society, London, 143, 607-618. BEZERT, P. & CABY, R. 1988. Sur 1'age post-bartonien des evenements tectono-metamorphiques alpines en bourdure orientale de la Corse cristalline. Bulletin de la Societe Geologique de France, 8, 965-971. BOCCALETTI, M., ELTER, P. & GUAZZONE, G.J.P. 1971. Plate tectonic models for the development of the western Alps and Northern Apennines. Nature, 234,108-111. BORTOLOTTI, V, FAZZUOLI, M., PANDELI, E., PRINCIPI, G, BABBINI, A. & CORTI, S. 2001. Geology of central and eastern Elba island, Italy. Ofioliti, 26, 97-150. BRUNET, C, MONIE, P., JOLIVET, L. & CADET, IP. 2000. Migration of compression and extension in the Thyrrhenian Sea, insights from 40Ar/39Ar ages on micas along a transect from Corsica to Tuscany. Tectonophysics, 321,127-155. CARMIGNANI, L. & KLIGFIELD, R. 1990. Crustal extension in the Northern Apennines: the transition from compression to extension in the Alpi Apuane Core Complex. Tectonics, 9,1275-1303. CARMIGNANI, L., GIGLIA, G & KLIGFIELD, R. 1978. Structural evolution of the Apuane Alps: an example of continental.margin deformation in the Northern Apennine. Journal of Geology, 86, 487-504. CARMIGNANI, L., DECANDIA, F.A., DISPERATI, L., FANTOZZI, PL., LAZZAROTTO, A., LIOTTA, D. & OGGIANO, G 1995. Relationships between the Tertiary structural evolution of the SardiniaCorsica-Provengal Domain and the Northern Apennines. Terra Nova, 7,128-137. CARMINATI, E., WORTEL, M.J.R., SPAKMAN, W. & SABADINI, R. 1998. The role of slab detachment process in the opening of the western-central Mediterranean basins: some geological and geophysical evidence. Earth and Planetary Science Letters, 160, 651-665. CARON, J.M. 1977. Lithostratigraphie et tectonique des Schistes Lustres dans les Alpes Cottiens septentrionales et en Corse orientale. Memoire Universite Louis Pasteur Strasbourg, 326 pp. CARON, J.M. 1994. Metamorphism and deformation in Alpine Corsica. Schweizerische Mineralogische und Petrographische Mitteilungen, 7,105-114. CARON, J.M., KIENAST, J.R. & TRIBOULET, C. 1981. High-pressure/low-temperature metamorphism and polyphase Alpine deformation at Sant'Andrea di Cotone (Eastern Corsica). Tectonophysics, 78, 419-451. CARPENA, J., MAILLHE, D., NAESER, C.W. & POUPEAU, G 1979. Sur la datation par traces de fission d'une phase tectonique d'age Eocene superieur en Corse. Compte Rendu Academic Science Paris, serie II, 289, 829-832.

CAVAZZA, W., ZATTIN, M., VENTURA, B. & ZUFFA, G.G 2001. Apatite fission-track analysis of Neogene exhumation in northern Corsica (France). Terra Nova, 13, 51-57. CELLO, G, INVERNIZZI, C. & MAZZOLI, S. 1996. Structural signature of tectonic process in the Calabrian Arc, southern Italy: Evidence from oceanic-derived Diamante-Terranova unit. Tectonics, 15,187-200. CERRINA FERONI, A., MARTELLI, L., MARTINELLI, P., OTTRIA, G. & CATANZARITI, R. 2002. Carta Geologico-Strutturale dell'Appennino EmilianoRomagnolo-Note Illustrative. Regione EmiliaRomagna, Selca Firenze. CHEMENDA, A.I., MATTAUER, M. & BOKUN A.N. 1996. Continental subduction and a mechanism for exhumation of high-pressure metamorphic rocks: new modelling and field data from Oman. Earth and Planetary Science Letters, 143,173-182. CHOUKROUNE, P. & GAPAIS, D.C. 1983. Strain pattern in Aar granite (Central Alps): orthogneiss developed by bulk inhomogeneous flattening. Journal of Structural Geology, 5, 411-418. COWAN, D. & SILLING, R.M. 1978. A dynamic scaled model of accretion at trenches and its implications for the tectonic evolution of subduction complexes. Journal of Geophysical Research, 83, 5389-5396. DAL PIAZ, GV. 1974. Le metamorphisme alpine de haute pression et basse temperature dans 1'evolution structurale du bassin ophiolitique alpinoapenninique. le partie. Bollettino della Societd Geologica Italiana, 93, 437-468. DAL PIAZ, GV. & ZIRPOLI, G. 1979. Occurrence of eclogites relics in the ophiolite nappe from Marine d'Albo, Northern Corsica. Neues JahrbuchfurMineralogie, 3,118-122. DALLAN, L. & PUCCINELLI, A. 1995. Geologia della regione tra Bastia e St-Florent (Corsica Settentrionale). Bollettino della Societd Geologica Italiana, 114, 23-66. DANIEL, J.M., JOLIVET, L., GOFFE, B. & POINSOTT, C. 1996. Crustal-scale strain partitioning: footwall deformation below the Alpine Oligo-Miocene detachment of Corsica. Journal of Structural Geology, 18,1841-1859. DEINO, R., KELLER, IV.A., MINELLI, G. & PIALLI, G 1992. Datazioni Ar/Ar del metamorfismo dell'unita di Ortano-Rio Marina (Isola d'Elba). Studi Geologici Camerti, Vol. Spe. CROP 01, 187-192. DELCEY, R. & MEUNIER, R. 1966. Le massif du Tenda (Corse) et ses bourdes: la serie volcano-sedimentaire, les gneiss et les granites; leurs rapports avec les schistes lustres. Bulletin de la carte Geologique de la France, 278, 237-251. DOGLIONI, C., MONGELLI, F. & PIALLI, G. 1998. Boudinage of the Alpine belt in the Apenninic back-arc. Bollettino della Societd Geologica Italiana, 118, 75-89. DURAND-DELGA, M. 1984. Principaux traits de la Corse Alpine et correlations avec les Alpes Ligures. Memorie della Societd Geologica Italiana, 28,285-329. EGAL,E. 1992. Structures and tectonic evolution of the

SHEAR ZONES AND METAMORPHISM external zone of Alpine Corsica. Journal of Structural Geology, 14,1215-1228. ELTER, P. 1975. Introduction a la geologic de 1'Apennin septentrional. Bulletin de la Societe Geologique de France, 7, 956-962. ELTER, P. & PERTUSATI, PC. 1973. Considerazioni sul limite Alpi-Appennino e sulle sue relazioni con 1'arco delle Alpi Occidentali. Memorie delta Societd Geologica Italiana, 12, 359-375. ERNST, W.G. 1979. Coexisting sodic and calcic amphiboles from high-pressure metamorphic belts and the stability of barroisitic amphibole. Miner alogical Magazine, 43, 269-278. EVANS, M. 1986. Phase relationships of epidoteblueschists. Lithos, 25, 3-23. FACCENNA, C, BECKER, T.W., LUCENTE, F.P., JOLIVET, L. & ROSSETTI, F. 2001 History of subduction and back-arc extension in Central Mediterranean. Geophysical Journal International, 145, 809-820. FAURE, M. & MALAVIELLE, J. 1981. Etude structural d'un cisallement ductile: le chariage ophiolitique Corse dans la region de Bastia. Bulletin de la Societe Geologique de France, 23, 335-343. FERRANDINI, M., FERRANDINI, J., LOYE-PILOT, M.D., BUTTERLIN, J. CRAVETTE, J. & JANIN, M.C. 1998. Le Miocene du Bassin de Saint-Florent (Corse): Modalites de la transgression du Burdigalien Superior et mise en evidence du Serravalien. Geobios, 31,125-137. FOURNIER, M., JOLIVET, L., GOFFE, B. & DUBOIS, R. 1991. The Alpine Corsica metamorphic core complex. Tectonics, 10,1173-1186. FRANCESCHELLI, M., LEONI, L., MEMMI, I. & PUXEDDU, M. 1986. Regional destribution of Al-silicates and metamorphic zonation in the low-grade Verrucano metasediments from the Northern Apennines, Italy. Journal of Metamorphic Geology, 4, 309-321. GAPAIS, D., BALE, P., CHOUKROUNE, P., COBBOLD, PR., MAHJOUB, Y. & MARQUER, D. 1987. Bulk kinematics from shear zone patterns: some filed examples. Journal of Structural Geology, 9, 635-646. GIBBONS, W. & HORAK, J. 1984. Alpine metamorphism of Hercynan hornblende granodiorite beneath the blueschist facies schistes lustres nappe of NE Corsica. Journal of Metamorphic Geology, 2, 95-113. GIBBONS, W., WATERS, C & WARBOURTON, J. 1986. The blueschist facies schistes lustres of Alpine Corsica: A review. Geological Society of America Memoir, 164, 301-331. GIORGETTI, G, GOFFE, B., MEMMI, I. & NIETO, F. 1998. Metamorphic evolution of Verrucano metasediments in northern Apennines: new petrological constraints. European Journal of Mineralogy, 9, 859-873. GLOM. 1977.1 complessi ofiolitici e le unita cristalline della Corsica alpina. Ofioliti, 2, 265-324. GUEGUEN, E., DOGLIONI, C. & FERNANDEZ, M. 1997.

Lithospheric boudinage in the western Mediterranean back-arc basin. Terra Nova, 9,184-187. HARRIS, J. 1985. Progressive and polyphase deformation of the Schistes Lustres in Cap Corse,

333

Alpine Corsica. Journal of Structural Geology, 7, 637-650. HOLLAND, T.J.B. & POWELL, R. 1998. An internally consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology, 16, 309-344. JAKNI, B., POUPEAU, G, SOSSON, M., Rossi, P., FERRANDINI, J. & GUENNOC, P. 2000. Denudations cenozoiques en Corse: une analyse thermochronologique par traces de fission sur apatites. Comptes Rendus de I'Academic des Sciences Paris, 331, 775-782. JOLIVET, L., DUBOIS, R., FOURNIER, M., MICHARD, A. & JOURDAN, C. 1990. Ductile extension in Alpine Corsica. Geology, 18,1007-1010. JOLIVET, L., FACCENNA, C., GOFFE, B., MATTEI, M., ROSSETTI, F, BRUNET, C., STORTI, F, FUNICIELLO, R., CADET, J.P., D'AGOSTINO, N. & PARRA,T. 1998. Midcrustal shear zones in postorogenic extension: Example from the northern Tyrrhenian Sea. Journal of Geophysical Research, 103 (B6), 12123-12160. JOURDAN, C. 1988. Balagne orientale et massif du Tenda (Corse septentrionale): etude structurale, interpretation des accidents et des deformatios, reconstritutions geodynamiques. These de 3 erne cycle, Universite de Paris-Sud, Orsay. KLIGFIELD, R., HUNZIKER, J, DALLMEYER, R.D. & SCHAMEL, S. 1986. Dating of deformational phases using K-Ar and 40Ar/39Ar techniques: results from the Northern Apennines. Journal of Structural Geology, 8, 781-798. LAHONDERE, D. 1988. Le metamorphisme eclogitique dans les orthogneiss, et les metabasites ophiolitiques de la region de Farinole (Corse). Bulletin de la Societe Geologique de France, 4, 579-585. LAHONDERE, D. 1991. Les schistes blues et les eclogites a lawsonite des unites continentals et oceanique de la Corse alpine: Nouvelles donnee petrologique et structurales. (Corse). These de 3 erne cycle, Universite de Montpellier II, France, 262 pp. LAHONDERE, D. & GUERROT, C. 1997. Datation Nd-Sm du metamorphisme eclogitique en Corse alpine: un argument pour 1'existence, au Cretace superieur, d'une zone de subduction active localisee le long du block corse-sarde. Geologic de la France, 3, 3-11. LAHONDERE, D., Rossi, P. & LAHONDERE, J.C. 1999. Structuration alpine d'une marge continentale externe: le massif du Tenda (Haute-Corse). Implications geodynamiques au niveau de la transversale Corse-Apennines. Geologic de la France, 4, 27-44. Liou, J.G., KIM, H.S. & MARUYAMA, S. 1983. Prehniteepidote equilibria and their petrologic applications. Journal of Petrology, 24, 321-342. MALAVIELLE, J., CHEMENDA, A. & LARROQUE, C. 1998. Evolutionary model for Alpine Corsica: mechanism for ophiolite emplacement and exhumation of high-pressure rocks. Terra Nova, 10, 317-322. MALUSKI, H. 1977. Application des methods de datation 39Ar/40Ar aux mineraux des roches cristallins perturbes par les evenementes thermiques et

334

G. MOLLI & R. TRIBUZIO

tectoniques en Corse. These de 3 erne cycle, Universite Montpellier, France. MALUSKI, H., MATTAUER, M. & MATTE, P. 1973. Sur la presence de decrochement alpins en Corse. Comptes Rendus de I'Academie des Sciences Paris, serie D, 276, 709-712. MARESCH, W.V. 1977. Experimental studies on glaucophane: an analysis of present knowledge. Tectonophysics, 43,109-125. MARUYAMA, S., SUZUKI, K. & Liou, J.G. 1983. Greenschist-amphibolite transition equilibria at low pressures. Journal of Petrology, 24, 583-604. MASSONE, H.J. & SZPURSKA, Z. 1997. Theromodynamic properties of white micas on the basis of high-pressure experiments in the systems K2OMgO-Al2O3-SiO2-H2O and K2O-FeO-Al2O3Si02-H20. Lithos, 41, 229-250. MATTAUER, M. & PROUST, F. 1975. Sur quelques problemes generaux de la chaine alpine en Corse. Bulletin de la Societe Geologique de France, 18, 1177-1178. MATTAUER, M., FAURE, M. & MALAVIELLE, J. 1981. Transverse lineation and large-scale structures related to Alpine obduction in Corsica. Journal of Structural Geology, 3, 401-409. MICHARD, A., CHALOUAN, A., FEINBERG, H., GOFFE, B. & MONTIGNY, R. 2002. How does the Alpine belt end between Spain and Morocco? Bulletin de la Societe Geologique de France, 173, 3-15. MOLLI, G, CONTI, P., GIORGETTI, G, MECCHERI, M. & OESTERLING, N. 2000. Microfabric studies on the deformational and thermal history of the Alpi Apuane marbles (Carrara marbles), Italy. Journal of Structural Geology, 22,1809-1825. MOLLI, G, GIORGETTI, G. & MECCHERI, M. 2000. Structural and petrological contraints on the tectono-metamorphic evolution of the Massa Unit (Alpi Apuane, NW Tuscany, Italy). Geological Journal, 35, 251-264. NARDI, R. 1968. Le unita alloctone della Corsica e la loro correlazione con le unita delle Alpi e dell' Appennino. Memorie della Societd Geologica Italiana, 1, 323-344. NARDI, R., PUCCINELLI, A. & VERANI, M. 1978. Carta geologica della Balagne "sedimentaria" (Corsica) alia scala 1:25.000 e note illustrative. Bollettino della Societd Geologica Italiana, 97, 3-22. OHNESTETTER, M. & Rossi, P. 1985. Reconsitution d'une paleochambre magmatique expectionalle dans le complexe basique-ultrabasique du Tenda. Comptes Rendus de I'Academie des Sciences Paris, serie II, 300, 853-858. PADOA, E. 2001. Lawsonite-bearing eclogites from the Northern Castagniccia area (Alpine Corsica, France). 3° Forum FIST, Abstract volume, 561-563. PADOA, E. & DURAND-DELGA, M. 2001. L'unite du Rio Magno en Corse Alpine: elements des ligurides de 1'Appennin septentrional. Comptes Rendus de I'Academie des Sciences Paris, 333, 285-293. PEQUIGNOT, G, LARDEAUX, J.M. & CARON, J.M. 1984. Recrystallization d'eclogites de basse temperature dans les metabaltes corse. Comptes Rendus Academic des Sciences Paris, serie II, 299,871-874.

PRINCIPI, G. & TREVES, B. 1984. II sistema corso-appenninico come prisma d'accrezione. Riflessi sul problema generale del limite Alpi-Appennini. Memorie della Societd Geologica Italiana, 28, 549-576. RAMSAY, J.G. & ALLISON, J. 1979. Structural analysis of shear zones in an Alpinised Hercynian granite. Schweizerische Mineralogische und Petrographische Mitteilungen, 59, 251-279. REHAULT, J.P, MASCLE, J. & BOILLOT, G. 1984. Evolution geodynamique de la Mediterranee depuis 1'Oligocene. Memorie della Societd Geologica Italiana, 27, 85-96. REUTTER, K.J., GUNTHER, K. & GROSCURTH, J. 1978. An approch to the Geodynamics of the CorsicaNorthern Apennines double orogene. In: CLOSS, H., ROEDER, D. & SCHMIDT, K. (eds) Alps, Apennines, Hellenides, IUGG Scientific Report, 299-311. ROLLET, N, DEVERECHERE, J., BESLIER, M.-O., GUENNOC, P., REHAULT, J.-P, SOSSON, M. & TRUFFET, C. 2002. Back arc extension, tectonic inheritance, and volcanism in the Ligurian Sea, Western Mediterranean. Tectonics, 21,1-23. ROSSETTI, F, FACCENNA, G, JOLIVET, L., GOFFE, B. & FUNICIELLO, R. 2002. Structural signature and exhumation P-T-t paths of the blueschist units exposed in the interior of the Northern Apennine chain, tectonic implication. Bollettino della Societd Geologica Italiana, volume speciale n.l, 20,829-842. Rossi, P., COCHERIE, A. & LAHONDERE, D. 1992. Relations entre les complex mafiques-ultramafiques et le volcanisme andesitique stefanopermian de Corse occidentale, temoins des phenomenes d'amicissement crustal neovarisques. Comptes Rendus de I'Academie des Sciences Paris, serie II, 315, 1341-1348. Rossi, P., LAHONDERE, J.C., LUCH, D., LOYE-PILOT, M.D. & JACQUET, M. 1994. Carte geologique de la France a 1/50.000, feuille Saint-Florent. BRGM. RUTTER, E.H., HOLDSWORTH, R.E. & KNIFE, E.J. 2001. The nature and tectonic significance of fault-zone weakening: an introduction. In: HOLDSWORTH, R.E., STRACHAN, R.A., MAGLOUGHLIN, J.F. & KNIPE, R.J. (eds) The nature and tectonic significance of fault zone weakening. Geological Society, London, Special Publication, 186, 65-84. SCANDONE, P. 1979. Origin of the Thyrrenian Sea and the Calabrian arc. Bollettino della Societd Geologica Italiana, 98, 27-34. SERRI, G, INNOCENTI, F. & MANETTI, P. 1993. Geochemical and petrological evidence of the subduction of delaminated Adriatic continental lithosphere in the genesis of the Neogene-Quaternary magmatism of central Italy. Tectonophysics, 223,117-147. SPERANZA, F, VILLA, I.M., SAGNOTTI, L., FLORINDO, F, COSENTINO, D., ClPOLLARI, P. & MATTEI, M. 2002.

Age of the Corsica-Sardinia rotation and LigureProvencal Basin spreading: new paleomagnetic and Ar/Ar evidence. Tectonophysics, 6571, 112-133. THEYE, T, REINHARDT, J., GOFFE, B., JOLIVET, L. &

SHEAR ZONES AND METAMORPHISM BRUNET, C. 1997. Ferro and magnesiocarpholite from Mt Argentario (Italy): first evidence for high-pressure metamorphism of the metasedimentary Verrucano sequence, and significance for P-T path reconstruction. European Journal of Mineralogy, 9, 859-873. TRIBUZIO, R. & GIACOMINI, F. 2002. Blueschist facies metamorphism of peralkaline rhyolites from the Tenda crystalline massif (northern Corsica): evidence for involvement in the Alpine subduction event? Journal of Metamorphic Geology, 20, 513-526. TRIBUZIO, R., PATTI, G. & DALLAI, L. 2001. Interaction between mantle and continental crust liquids in the Bocca di Tenda Gabbroic complex. Journal of Conference Abstracts, EUG XI, 817.

335

VAN DEN BEUKEL J. 1992. Some thermomechanical asepcts of the subduction of continental lithosphere. Tectonics, 11, 316-329. VON BLACKENBURG, F. & DAVIES, J.H. 1995. Slabbreakoff: a model for syncollisional magmatism and tectonics in the Alps. Tectonics, 14,120-131. WARBOURTON, J. 1986. The ophiolite-bearing Schistes Lustres nappe in Alpine Corsica: A model for emplacement of ophiolites that have suffered HP/LT metamorphism. Geological Society of America Memoir, 164, 313-331. WATERS, CN. 1990. The Cenozoic tectonic evolution of alpine Corsica. Journal of the Geological Society, London, 147, 811-824.

This page intentionally left blank

Crenulation-slip development in a Caledonian shear zone in NW Ireland: evidence for a multi-stage movement history D. M. CHEW1, J. S. DALY2, M. J. FLOWERDEW3, M. J. KENNEDY2 & L. M. PAGE4 5 l Departement de Mineralogie, Universite de Geneve, Rue des Maraichers 13, CH-1205 Geneve, Switzerland (e-mail: [email protected]) ^Department of Geology, University College Dublin, Belfield, Dublin 4, Ireland ^British Antarctic Survey, c/o NERC Isotope Geosciences Laboratory, Kingsley Dunham Centre, Keyworth, Nottingham NG12 5GG, UK ^Laboratory of Isotope Geology, Vrije Universiteit, De Boelelaan 1085,1081 HV Amsterdam, the Netherlands 5 Present address: Department of Geology, Lund University, Solvegatan 13, 223 62 Lund, Sweden Abstract: In Scotland and Ireland, a Laurentian passive margin sequence, the Dalradian Supergroup, was deformed during the c. 470-460 Ma Grampian orogeny, resulting in the formation of crustal-scale recumbent nappes. In Ireland, this passive margin sequence is in general bounded to the SE by the Fair Head-Clew Bay Line (FHCBL), a segment of a major lineament within the Caledonides. Adjacent to the FHCBL, Dalradian metasediments in two separate inliers have undergone post-Grampian strike-slip movement, with the initially flat-lying Grampian nappe fabric acting as a decollement-like slip surface in both cases. As the orientation of these foliation slip surfaces was oblique to the local shear plane in both inliers, displacement along these pre-existing foliation surfaces was also accompanied by crenulation slip. However, the crenulation-slip morphologies produced imply the opposite sense of movement in the two inliers. 40Ar-39Ar dating of muscovite defining the crenulation-slip surfaces indicates that post-Grampian dextral displacement took place along the FHCBL at 448 ± 3 Ma. A subsequent phase of sinistral movement along the FHCBL took place at c. 400 Ma, based on previously published Rb-Sr muscovite ages for synkinematic pegmatites. The kinematic information obtained from crenulationslip morphologies combined with geochronology can thus be used to constrain the reactivation history of a major crustal-scale shear zone.

Deforming materials are seldom isotropic, and hence anisotropy plays an important role in partitioning strain in shear zones. Common examples of anisotropy encountered in midcrustal shear zones include planar elements such as sedimentary layering or foliations. Such planar anisotropies should act as decollementlike surfaces during shear deformation when they are suitably orientated (i.e. subparallel to the shear plane). Renewed movement along pre-existing foliations (foliation reactivation) is thus likely to be a feature of many mid-crustal shear zones. Kinematic models have been used (e.g. Dennis & Secor 1987) to predict the structural features produced when slip occurs along foliation surfaces which are oblique to the walls of a shear zone. Oblique slip produces a displacement component normal to the zone wall, which is inconsistent with plane strain, simple shear

deformation (Ramsay & Graham 1970). In order to maintain the initial thickness of the shear zone and thus preserve a simple shear deformation path, crenulation slip has been interpreted to compensate for this normal displacement component (Dennis & Secor 1987). This paper presents structural and 40Ar-39Ar and Rb-Sr isotopic data from a major shear zone within the Caledonides of NW Ireland, the Fair Head-Clew Bay Line. Detailed field mapping has demonstrated that the main regional foliation developed within Dalradian Supergroup metasediments adjacent to the Fair Head-Clew Bay Line is used as a slip surface within this shear zone, where it is accompanied by crenulation slip. The age of the main regional foliation is also well constrained by previous geochronological studies based immediately outside this shear zone (e.g. Flowerdew et al. 2000; Chew et al. 2003), where it is unaffected by later

From: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 337-352. 0305-8719/$15.00 © The Geological Society of London 2004.

338

D.M. CHEW ETAL.

Fig. 1. (a) Regional geology of NW Ireland displaying localities referred to in the text, (b) Location map of Ireland within the Caledonides.

deformation (i.e. shearing). Isotopic dating of both the crenulation-slip fabrics and the reactivated foliation within the shear zone enable individual phases of reactivation to be constrained temporally. The crenulation morphologies predicted by the model described above can thus be used to identify not only shear sense on a major crustal-scale shear zone, but also to establish the timing of movement.

Geological significance of the Fair Head-Clew Bay Line In NW Ireland, a Laurentian passive margin sequence, the Neoproterozoic-Cambrian Dalradian Supergroup (Fig. 1), was deformed during the c. 470-460 Ma Grampian orogeny (Friedrich et al 19990, b\ Flowerdew et al 2000). This orogenic episode is believed to be related to the collision of the Laurentian margin with an outboard oceanic arc and associated forearc ophiolite (Dewey & Shackleton 1984), which resulted in metamorphism and the production of crustalscale recumbent nappes within the Dalradian sequence.

The Dalradian Supergroup in NW Ireland (with the exception of the allochthonous Connemara terrane) is bounded to the SE by the Fair Head-Clew Bay Line (Fig. 1). This structure is believed to be equivalent to the Highland Boundary Fault of Scotland and the Baie Verte-Brompton Line of Newfoundland and as such is a significant lineament within the Caledonides. It is believed to represent the original collisional suture between the deformed and metamorphosed Laurentian margin sequences and the outboard oceanic arc (Dewey & Shackleton 1984). The Fair Head-Clew Bay Line itself is defined by a conspicuous magnetic lineament (Max & Riddihough 1975) from northeastern Ireland (Fig. 1) to the north shore of Clew Bay on the western Irish coast (Fig. 2a), whereas the main surface expression of the Fair Head-Clew Bay Line is a fault zone which in general lies about 10 km to the south of the magnetic lineament (Fig. 1). Both are often collectively referred to as the Fair Head-Clew Bay Line (Ryan et al 1995). Throughout most of the southeastern margin of the Dalradian Supergroup in Scotland and Ireland, ductile structures related to movement

REACTIVATION OF A CALEDONIAN SHEAR ZONE

339

Fig. 2. (a) Regional geology of Co. Mayo displaying localities referred to in the text, (b) North-south crosssection (X-Y) across the NW Mayo inlier displaying major F2 folds, (c) NW-SE cross section across the Central Ox Mountains inlier adapted from MacDermot et al. (1996) and displaying major F3 folds.

along the Highland Boundary Fault or the Fair Head-Clew Bay Line are rarely exposed. However, in Dalradian Supergroup metasediments on Achill Island in NW Mayo and in the Central Ox Mountains (Fig. 1), ductile structures which can be related to strike-slip motion along the Fair Head-Clew Bay Line (FHCBL) are clearly observed.

Evidence for strike-slip motion along the FHCBL in NW Mayo (South Achill) Achill Island is situated on the southern margin of the NW Mayo inlier (Figs 1 & 2a). The NW Mayo inlier preserves an excellently exposed transect from Laurentian basement, the Annagh Gneiss Complex (Daly 1996), through presumed para-autochthonous (Winchester 1992) Laurentian cover, the Dalradian Supergroup, to outboard oceanic elements located immediately to the south of the FHCBL (e.g. the Clew Bay Complex; Williams et al. 1994). Polyphase deformation is pervasive throughout the Dalradian Supergroup outcrop in Ireland and Scotland. In the NW Mayo inlier, a D! deformation event is responsible for the bulk of the high-strain observed, and is associated with the development of tectonic slides and locally-developed isoclinal Fj folds (Kennedy 1980). The D2 deformation event in the NW Mayo inlier is the main nappe-forming phase. The D2 nappes in general plunge gently east and 'root' in the basement core, the Annagh Gneiss Complex (Fig. 2b). Adjacent to, and directly

above the Annagh Gneiss Complex, the D2 nappes are upward-facing (Fig. 2b); to the south of this 'root zone' recumbent D2 folds face south (Fig. 2b; Kennedy 1980) and are rotated into a downward-facing orientation approaching the southern margin of the inlier (Fig. 2b; Chew 2003). The production of downward-facing D2 folds and the associated S2 strike-swing (Fig. 2a) in the southern part of the inlier has been attributed to either later modification of the D2 nappe pile by an east-west dextral shear zone running along the north margin of Clew Bay (Sanderson et al. 1980; Chew 2003), or a steep zone of D2 dextral transpression contemporaneous with the development of flat-lying D2 nappes to the north (Harris 1993,1995). The model of Sanderson et al. (1980) assumed that the S2 nappe fabric was progressively rotated into the proposed shear zone rather than new fabrics forming. However, detailed mapping of the southern part of Achill Island (South Achill) and the island of Achill Beg (Fig. 3) on the southern margin of the inlier demonstrates that the S2 nappe fabric is modified by 03 structures consistent with dextral strike-slip displacement. The northern boundary of the shear zone to the north is not sharply defined, but D3 structures gradually become less abundant to the north of Ashleam Bay in South Achill (Fig. 3). Two discrete elements have been recognized in the D3 deformation episode, asymmetrical buckle folds with axial planes anticlockwise to the S2 foliation, and extensional crenulation cleavages which cut the S2 foliation in a clockwise sense.

340

D.M. CHEW^r^L.

Fig. 3. Geological map of South Achill and Achill Beg.

REACTIVATION OF A CALEDONIAN SHEAR ZONE

341

Fig. 4. Angular relationships predicted between the shear zone wall and the foliation-slip and crenulation-slip surfaces, (a) Reverse-slip crenulations. (b) Normal-slip crenulations. Reprinted from Journal of Structural Geology, 9, Dennis & Secor: A model for the development of crenulations in shear zones with applications from the southern Appalachian Piedmont, pp. 809-817. Copyright 1987, with permission from Elsevier.

Crenulation-slip morphologies produced by oblique foliation-slip The kinematic models of Dennis & Secor (1987, 1990) predict the crenulation morphologies that develop in order to compensate for the displacement component of foliation slip normal to the shear zone wall. In a dextral shear zone, when the pre-existing foliation is at an acute, clockwise angle to the shear zone wall (Fig. 4a), movement away from the shear zone wall due to oblique foliation slip is compensated by reverseslip crenulations (RSC), which transfer slip up to 'higher' foliation planes. When the slipping foliation is at an acute anticlockwise angle to the shear zone wall in a dextral shear zone (Fig. 4b), movement normal to the shear zone wall is compensated by normal-slip crenulations (NSC).

Asymmetrical buckle folds (reverse-slip crenulations) The most common crenulation-slip morphologies in South Achill and Achill Beg are asymmetrical buckle folds (Fig. 5a). The strike of the F3 axial planes makes an angle of approximately 27° with the strike of the S2 foliation in an anticlockwise direction (Fig. 6a). F3 fold axes plunge moderately to the NE (Figs 3,6a). The largest F3 folds have wavelengths of only a few metres, and typically the smaller F3 folds can be observed to 'root' in the S2 foliation surface. These features are typical of the reverse-slip crenulations of Dennis & Secor (1987). With progressive dextral

shear, the S3 foliation which initiates anticlockwise to S2 is progressively brought into parallelism (Fig. 5a). This is particularly evident where there is a large competence contrast across a bedding surface. Relatively rigid psammitic layers respond to D3 shear by folding with S3 usually oblique to S2. If the S3 foliation continues out into an adjacent graphitic pelite layer, then commonly the S3 foliation swings clockwise into parallelism with S2 and hence the weak graphitic pelite layers are accommodating the bulk of the displacement. The weak graphitic pelite layers also often display evidence of slip along the S2 foliation surfaces (Fig. 5b). On a vertical surface, the S3 foliation commonly rotates into the vertical parallel to S2, with a down-to-the-south shear sense.

Extensional crenulation cleavages (normalslip crenulations) Extensional crenulation cleavages (Platt & Vissers 1980) are relatively common within pelitic lithologies in the South Achill sequence. They make an angle of approximately 29° (Fig. 6b) with the S2 foliation in a clockwise direction, and consistently give a dextral sense of shear (Fig. 5c). Identical in style to the normal-slip crenulations of Dennis & Secor (1987), they are believed to be broadly contemporaneous with the F3 asymmetrical buckle folds based on the absence of apparent overprinting relationships. The orientation of the earlier S2 foliation controls the morphology of the later crenulations

342

D. M. CHEW ETAL.

Fig. 5. (a) Asymmetrical buckle folds (F3), interpreted as reverse slip crenulations (RSC). Later rotation of S3 cleavage is due to progressive dextral shear. Hammer 40 cm long. South Achill Dalradian [L69329495]. (b) Graphitic pelite illustrating decollement-like slip along the 82 foliation surface. Plane-polarized light, scale bar 1000 urn. South Achill Dalradian [L69039540]. (c) D3 dextral shear bands cutting the S2 foliation, interpreted as normal slip crenulations (NSC). Hammer 40 cm long. South Achill Dalradian [L69029544]. (d) Sinistral extensional crenulations affecting the composite S2/S3 foliation, interpreted as normal slip crenulations (NSC). Coin 2.2 cm in diameter. Central Ox Mountains Dalradian [G323026]. (e) Photograph of a polished rock slice used for 40Ar-39Ar in situ laserprobe dating of muscovite defining both the S2 and S3 foliations. Sample DC-79, scale bar 1000 um. South Achill Dalradian [L69089511].

(e.g. RSC vs. NSC). On a vertical surface, the extensional shear bands give a down-to-thesouth shear sense. Orientation of the D3 dextral shear zone From the angular relationships proposed by Dennis & Secor (1987), the shear zone wall is

expected to lie between the axial planes of the asymmetrical buckle folds (Fig. 6a) and the extensional shear bands (Fig. 6b). At localities which display both RSC and NSC fabrics, the dominant slip foliation (S2) modified by the RSC makes a very small clockwise angle with the S2 foliation affected by the NSC, similar to the geometry predicted by Fig. 4. However, on a

REACTIVATION OF A CALEDONIAN SHEAR ZONE

343

Fig. 6. (a) Stereographic plot of the orientation of RSC-related structures in South Achill (F3 fold hinges and poles to the S3 foliation) along with the mean orientation of the S2 foliation, (b) Rose diagram illustrating the mean orientation of NSC-related structures in South Achill (dextral shears measured on horizontal surfaces). The mean orientation of the S2 foliation is also illustrated, (c) Stereographic plot of the orientation of the preexisting foliation-slip surface in South Achill (poles to the S2 foliation), (d) Stereographic plot of the orientation of the pre-existing foliation-slip surface in the Central Ox Mountains (poles to the composite S2/S3 foliation), along with the mean orientation of NSC-related structures (sinistral extensional crenulation cleavages). regional scale this relationship is not apparent and hence the dominant slip foliation (S2) data for both the RSC and NSC sets are presented together (Fig. 6c). Combining data from both horizontal and vertical surfaces, the shear zone

would therefore be an approximately east-west trending, sub-vertical structure, similar to the geometry proposed by Sanderson et al. (1980). No L3 elongation lineations have been observed and hence although shear sense has been

344

D. M. CHEW ETAL.

determined reliably from the crenulation-slip morphologies on both horizontal and vertical surfaces, the exact shear direction remains uncertain.

Evidence for strike-slip motion along the FHCBL in the Central Ox Mountains

Glennawoo Slide and the Callow Shear Zone (Taylor 1969). In the Callow Loughs region (Fig. 7), the Lough Talt and Glennawoo Slides may be adequately represented as discrete shear zones, but the Callow Shear Zone is significantly wider (more than 750 m across strike) and is regarded as a substantial mylonite belt. Both the Glennawoo Slide and the Callow Shear Zone display abundant evidence of sinistral extensional crenulation cleavage development (Fig. 7).

The Central Ox Mountains inlier consists of a sequence of Dalradian metasediments (Long & Max 1977; Alsop & Jones 1991) intruded by a Caledonian granite, the Ox Mountains granodi- Extensional crenulation cleavages (normalorite (Fig. 2a, c). The intrusion age of the Ox slip crenulations) Mountain granodiorite has proved controversial in the past (e.g. Kennan 1997), and it is discussed In the Callow Loughs region, the main foliation in detail later. This granodiorite is intruded into is close to the vertical and trends NE (Figs 6d the core of a significant upright D3 antiform & 7). It is regarded as a composite S2/S3 fabric as which trends NE-SW, subparallel to the length S2 and S3 are usually coplanar (Jones 1989; Macof the inlier (Fig. 2c; Taylor 1969). High strain Dermot et al. 1996) and the S3 fabric can only be zones are well developed on the limbs of the conclusively identified when F3 folds are main D3 antiformal structure, are parallel to the present. F3 folds plunge gently to the NE and vertical, axial planar S3 fabric and kinematic verge towards the Ox Mountains granodiorite indicators such as rotated porphyroblasts and (Fig. 7). Extensional crenulation cleavages make extensional crenulation cleavages display abun- an angle of approximately 28° (Fig. 6d) with the dant evidence for sinistral shear (e.g. Hutton & main (composite S2/S3 foliation) in an anticlockDewey 1986; Hutton 1987; Jones 1989; McCaf- wise direction, and consistently give a sinistral frey 1992, 1994). These shear zones have been sense of shear (Fig. 5d). On a vertical surface, regarded as contemporaneous with the develop- the extensional shear bands give a down to the ment of the main E>3 antiform and the Central south shear sense. Ox Mountains has thus been regarded a trans- , The composite (S2/S3) main foliation is usually pressive sinistral shear zone during D3 (Hutton defined by muscovite, chlorite and equigranular & Dewey 1986; Hutton 1987; Jones 1989; quartz grains with interlobate grain boundaries. McCaffrey 1992,1994). These quartz grains are typically of the order of The Ox Mountains granodiorite itself also dis- 100-200 um in diameter and display undulose plays abundant evidence of sinistral strike-slip extinction. MP3 porphyroblasts of albite and deformation. The main solid-state foliation is almandine garnet overgrow the main foliation subvertical, strikes NE-SW and is accompanied and staurolite is locally developed (MacDermot by a stretching lineation which plunges gently to et al. 1996). Along the extensional crenulation the NE or SW (McCaffrey 1992, 1994). NNE cleavage planes, quartz has undergone signifitrending sinistral S-C fabrics are commonly well cant dynamic recrystallization. It is conspicudeveloped within the granodiorite and are sub- ously finer grained (10-20 um) than that parallel to the asymmetrical extensional crenu- defining the composite S2/S3 foliation, and has a lation cleavages developed in the country rock weak shape-preferred orientation (aspect ratios (Hutton & Dewey 1986; McCaffrey 1992,1994). of up to 3:1). Additionally, phyllosilicates within The Ox Mountains granodiorite has thus been the shear bands (chlorite and muscovite) are regarded as being emplaced synkinematically significantly finer grained than those defining the with respect to D3 sinistral transpressive defor- S2/S3 foliation. The extensional crenulation mation in the country rock (Hutton & Dewey cleavage planes are short and anastomosing, commonly rooting in the pre-existing composite 1986; Jones 1989; McCaffrey 1992,1994). S2/S3 foliation, and are very similar in morphology to the normal-slip crenulations (NSC) of Dennis & Secor (1987). High strain zones (tectonic slides) in the

Central Ox Mountains Three high strain zones are particularly well developed on the SE flank of the Central Ox Mountains inlier - the Lough Talt Slide, the

Isotopic dating of crenulation-slip surfaces Many deformed rocks contain evidence (e.g. multiple fabrics) for having experienced more

REACTIVATION OF A CALEDONIAN SHEAR ZONE

Fig. 7. Geological map of the Callow Loughs area in the Central Ox Mountains.

345

346

D.M. CHEW ETAL.

than one deformation event. The timing of growth of multiple fabrics can be constrained by dating fabric-forming minerals which crystallize below their closure temperatures (e.g. Cliff 1985). One of the most common fabric-forming minerals is muscovite, and Rb-Sr dating and 40 Ar-39Ar laserprobe dating of multiple generations of muscovite has been employed in many studies (e.g. Mtiller et al 1999). Both the Rb-Sr and Ar-Ar systems have shown that muscovite grown below its closure temperature during deformation (e.g. in mylonites) may record the age of neocrystallization (e.g. Dunlap 1997; Freeman et al. 1997, Miiller et al. 1999). Additionally, samples containing earlier generations of white mica (e.g. as porphyroclasts or older foliations) record the crystallization age of these early fabrics where they have not been rejuvenated by later deformation (West & Lux 1993; Freeman et al. 1997, Miiller et al. 1999). In this study, fabric-forming muscovite has been dated using the 40Ar-39Ar and Rb-Sr systems in order to constrain both the age of shearing and the age of the reactivated foliationslip surface within samples that have undergone shear-related deformation along the FHCBL. The age of the pre-existing foliation which is exploited by shearing along the FHCBL is also well constrained by previous geochronological studies based immediately outside of this shear zone (e.g. Flowerdew et al. 2000; Chew et al. 2003), where it is unaffected by later deformation. Thus the age of this foliation in undeformed samples and in samples where it has been reactivated due to subsequent shearing can be compared.

Pre-existing constraints on the age of the foliation-slip fabric in both inliers Recent geochronological studies in the Dalradian of NW Ireland (Flowerdew et al. 2000; Chew et al. 2003) has shown that main foliation development in both South Achill and the Central Ox Mountains inlier is Grampian (c. 470-460 Ma) in age in samples which are unaffected by later shearing along the FHCBL. Knowledge of the age of this main foliation in samples unaffected by later deformation is essential, as it enables us to assess whether the age of the reactivated foliation is partially reset when it is rejuvenated by later shearing along the FHCBL. Four Rb-Sr S2 muscovite ages from the Dairadian of South Achill range from 460-458 Ma (± 7 Ma), and four 40Ar-39Ar S2 muscovite step-

heating plateaux from the same samples range from 463-457 Ma (± 4 Ma) (Chew et al. 2003). Two Rb-Sr S2 muscovite ages of 460 ± 7 Ma and 461 ± 7 Ma have been obtained from the low greenschist facies Clew Bay Complex and probably record crystallization (Chew et al. 2003). As this outboard terrane (Figs 2 & 3) is in structural continuity with the Dalradian Supergroup (Chew 2003), it is thought likely that the c. 460 Ma Rb-Sr S2 muscovite ages for the South Achill Dalradian also record crystallization during the Grampian orogeny, particularly as the peak metamorphic temperature in South Achill is unlikely to have exceeded 450 °C (Chew et al. 2003). The marked similarity with the S2 muscovite 40Ar-39Ar step heating data may imply that the Ar-Ar system is recording either rapid cooling or crystallization, despite growing c. 50 to 100 °C above its closure temperature (c. 350-400 °C; Wijbrans & McDougall 1988). The possibility of extraneous Ar contamination is discussed in detail later. One 40Ar-39Ar hornblende age of 467 ± 3 Ma and one Rb-Sr muscovite age of 472 ± 8 Ma have been obtained from the composite S2/S3 foliation in the Dalradian of Central Ox Mountains inlier (Flowerdew et al. 2000). Younger 40Ar_39Ar and Rb-Sr ages (429-410 Ma) have been strongly influenced by the intrusion of the Ox Mountains granodiorite. Similar Grampian (c. 470-460) ages have also been obtained from Dalradian rocks in the NE Ox Mountains (Fig. 1). Here Dalradian rocks are interleaved during D3 with the Slishwood Division, a unit of psammitic gneisses which has experienced late Precambrian granulite-facies metamorphism (Sanders et al. 1987). Three 40Ar-39Ar hornblende ages and six Rb-Sr muscovite ages defining the S3 foliation within the interleaved Dalradian range from 470-446 Ma. The older S3 mineral ages are indistinguishable from the older ages (472 ± 7 Ma, 467 ± 3 Ma) obtained from the composite S2/S3 fabric in the Central Ox Mountains inlier (Flowerdew et al. 2000).

Sampling In this study we present isotopic data from two Dalradian samples from two localities. DC-79 is a semipelite from South Achill (Fig. 3) in which muscovite defining both the local S2 and S3 foliations was dated in situ by the 40Ar-39Ar laserprobe spot fusion method. DC-03/02-1 is a semipelite from the Central Ox mountains inlier (Fig. 7) from which a bulk mineral separate of muscovite defining the main (composite S2/S3) foliation was analysed by the Rb-Sr method. The whole rock was also analysed.

REACTIVATION OF A CALEDONIAN SHEAR ZONE

347

Table 1. 40Ar-39Ar spot fusion data Spot

Laser power

36

Ar(a)

37

Ar(Ca)

38

Ar(Cl)

39

Sample DC-79, S2 and S3 ms. Dalradian, South Achill 2(S 2 ) Fusion 0.0002 0.0000 0.0001 4(S 2 ) Fusion 0.0002 0.0266 0.0000 6(S2) Fusion 0.0000 0.0299 0.0002 10 (S2) Fusion 0.0010 0.0208 0.0000 11 (S2) Fusion 0.0004 0.0308 0.0004 1(S3) Fusion 0.0001 0.0000 0.0000 3(S 3 ) Fusion 0.0000 0.0000 0.0001 5(S 3 ) Fusion 0.0001 0.0345 0.0001 7(S3) Fusion 0.0013 0.0511 0.0002 8(S3) Fusion 0.0002 0.0371 0.0000 9(S 3 ) Fusion 0.0003 0.0178 0.0000 Weighted average* (S2 spots): 451 ± 2 Ma (2a) Weighted average (S3 spots): 448 ± 3 Ma (2a) J value: 0.003987 ± 0.5% (la)

Ar(K)

40

Ar(r) Age (Ma) ±2o

(L69089511). 0.1262 9.0407 0.0320 2.2206 0.0209 1.5135 0.5635 40.1131 0.1893 13.5803 0.1492 10.5535 0.0972 6.8772 0.1325 9.3329 0.0683 4.7528 0.0891 6.2796 0.3099 22.0392

453.4 441.3 457.8 450.7 453.8 448.2 448.2 446.6 441.7 446.9 450.4

3.4 10.1 14.5 1.3 3.5 2.8 3.5 3.6 5.5 4.8 1.9

40

Ar(r) (%)

39

99.2 97.3 99.4 99.3 99.1 99.7 99.8 99.8 92.4 99.0 99.6

Ar(k) (%) 7.1 1.8 1.2 31.7 10.7 8.4 5.5 7.5 3.8 5.0 17.4

*Weighted averages calculated using ISOPLOT (Ludwig 1999) and use the 2a error associated with each analysis.

Analytical methods 40

39

Ar- Ar spot fusion analyses were carried out using the VULKAAN argon laserprobe (Wij brans et al 1995) at the Vrije Universiteit in Amsterdam. Samples were irradiated at the HPPIF facility in the high flux research reactor at Petten, Netherlands. Polished slices were interspersed between the Al tablets containing the flux monitor DRA-1 sanidine (24.99 ± 0.07 Ma; Wijbrans et al. 1995) prior to irradiation. Four flux monitors were used to construct a J-curve with a 0.5% error (lo). Samples were analysed within six months of irradiation to minimize the interference effects produced by radioactive decay after irradiation. The analytical procedure is described in detail by Wijbrans et al. (1995) and is outlined below. Samples were heated using a continuous 18 W argon ion laser (454.5-514.5 nm wavelength). For spot fusion experiments, several short laser pulses (0.1 s) excavated a pit approximately 30 microns in diameter, surrounded by a crater of melt material. The Ar released was cleaned with Fe-V-Zr getters (250 °C), prior to analysis on a MAP-215/50 mass spectrometer. Data reduction was carried out using in-house software, ArArCALC V20. Blank intensities were measured every 3-5 sample runs and mass fractionation was corrected for by regular measurement of shots of clean air argon. For Rb-Sr analyses, standard ion exchange methods were used for chemical separation of elements. Samples were loaded on tantalum filaments and were analysed on a semi-automated

single collector VG Micromass 30 mass spectrometer at the Department of Geology, University College Dublin. During the course of analysis, NBS SRM 987 gave 87Sr/86Sr ratios of 0.71027 ± 5 (n = 8, 2o) and NBS SRM 607 yielded 87Rb/86Sr ratios of 8.005 ± 13 (n = 7,2o). Sr blanks averaged 1.5 ng and are not significant. 2o analytical uncertainties of 1.5% for 87Rb/86Sr and tabulated values (Table 2) for 87Sr/86Sr ratios were used in age calculations which employed a value of 0.0142 Ga"1 for the 87Rb decay constant (Steiger and Jager 1977).

Constraining dextral shear in NW Mayo (South Achill) New S3 mica growth is not commonly associated with the D3 deformation event in South Achill, but locally S3 muscovite is found overgrowing a crenulated S2 fabric in the hinges of asymmetrical buckle folds (reverse-slip crenulations associated with D3 dextral shear). One polished slice from the hinge of an F3 fold was selected for in situ 40Ar-39Ar laserprobe spot fusion analyses (Table 1). Sample DC-79 is semipelitic, with an S2 foliation defined by muscovite and minor chlorite and epidote. Individual muscovite grains are typically around 500 um long and 50 um wide, but both the S2 and S3 fabrics are composed of seams of white mica which are typically several grains in width. Typically S3 seams are approximately 150 um across (Fig. 5e), whereas S2 lithons are larger and may be up to 500 um wide (Fig. 5e). Chlorite grains are

348

D. M. CHEW£r,4L.

typically similar in size to the muscovite grains (c. 500 urn long and 50 urn wide) and the largest epidote needles observed are 200 urn long. The S2 foliation is overgrown by MP2 plagioclase which is augened by the seams of S3 muscovite. The calcic component in some of the Ar analyses (Table 1) is probably derived from the epidote, as the MP2 plagioclase is essentially pure albite. Muscovite defining the S2 foliation is typically more celadonite-rich and paragonitepoor than the later (S3) fabric (Chew et al 2003). In situ 40Ar-39Ar laserprobe dating of the S3 muscovite seams yields a weighted mean age of 448 ± 3 Ma, whereas the older, crenulated S2 muscovite seams yield a weighted mean age of 451 ± 2 Ma (Table 1). The 40Ar-39Ar system is only reliably recording the youngest deformation fabrics present, as undeformed S2 muscovite from South Achill yields consistent c. 460 Ma Rb-Sr and 40Ar-39Ar ages (Chew et al. 2003). The 448 ± 3 Ma age for muscovite within the S3 crenulation-slip fabric is interpreted as a crystallization age based on the low-greenschist facies assemblages observed in the S3 crenulation seams as detailed above. The possibility of extraneous argon cannot be ruled out, particularly in a high-pressure, low temperature terrane such as South Achill. Excess argon (argon with 40Ar/36Ar ratios which differ from the modern atmospheric ratio of 295.5) has been documented in white mica from several high-pressure, low-temperature terranes (e.g. Arnaud & Kelley 1995; Sherlock & Kelley 2002). The presence of excess argon may be evaluated by using an inverse isochron correlation diagram (a plot of 36Ar/40Ar versus 39 Ar/40Ar) as the incorporation of excess argon will result in the intercept of the isochron on the ordinate axis deviating from the modern atmospheric 40Ar/36Ar ratio of 295.5. However, K-rich phases (such as white mica) can produce large quantities of radiogenic 40Ar, and hence the data often cluster close to the 39Ar/40Ar axis. The presence of excess argon is therefore difficult to assess, as the intercept with the 36Ar/40Ar axis is poorly constrained. This is the case with the South Achill Ar data. The presence of excess argon has also been documented in samples which yield intercepts within error of the modern 40Ar/36Ar atmospheric ratio on an inverse isochron correlation diagram (Sherlock & Arnaud 1999). Studies that have documented the presence of excess argon in white mica in high-temperature, low-pressure terranes instead are based on 40 Ar-39Ar phengite ages that are significantly older than either the corresponding Rb-Sr phengite ages or other geochronometers with

significantly higher closure temperatures (e.g. Arnaud & Kelley 1995; Sherlock & Arnaud 1999). This disparity is not observed in South Achill as both the Rb-Sr and 40Ar-39Ar ages for undeformed S2 muscovite ages cluster at c. 460 Ma (Chew et al. 2003) and are mutually within error. The S3 muscovite ages are temporally distinct, and are clearly post-Grampian (c. 470-460 Ma) in age.

Constraining sinistral shear in the Central Ox Mountains The Central Ox Mountains displays extensional crenulation cleavage development superimposed on what is believed to be a pre-existing Grampian (S2/S3) foliation. However, whereas dextral shear in South Achill was constrained by dating both the pre-existing foliation-slip surface and the later crenulation-slip surfaces within the same sample, this strategy has proved impossible in the Central Ox Mountains. Dalradian metasediments in the Central Ox Mountains display only limited growth of extremely fine-grained muscovite on extensional crenulation cleavage surfaces, which is not sufficient for isotopic dating. However, pegmatites and granite sheets associated with the Ox Mountains granodiorite are intruded into the Dalradian metasediments, and in common with the Ox Mountains granodiorite, the pegmatites and granite sheets were emplaced synkinematically with respect to sinistral deformation in the country rocks (McCaffrey 1992, 1994). Sinistral shearing in the Central Ox Mountains can thus be constrained by dating pegmatite crystallization. The age of the foliation surface that is affected by sinistral extensional crenulation cleavages in the Central Ox Mountains has until now remained uncertain, as previous geochronological studies in the inlier (e.g. Flowerdew et al. 2000) were undertaken on foliated samples that were not affected by later deformation. Whereas the reactivated foliation surface is believed to be the regional S2/S3 composite foliation based on detailed field mapping (Fig. 7), it may have developed contemporaneously with sinistral shear development in the Central Ox Mountains, as the timing relationship between mylonitic foliation development and cross-cutting shear bands can be often difficult to establish (e.g. Lister & Snoke 1984). The age of the pre-existing foliation has been constrained by a Rb-Sr muscovite-whole rock age from a semi-pelitic schist sample displaying a pervasive sinistral extensional crenulation cleavage (Fig. 5d). This age of 448 ± 9 Ma (Table 2) is

REACTIVATION OF A CALEDONIAN SHEAR ZONE

349

Table 2. Rb-Sr geochronology Sample

Locality and Irish National Grid Ref.

DC-03/02-131

Textural relationship

Mineral

Rb(ppm) Sr(ppm)

Lismoran Main foliation muscovite 338.77 126.60 (G323026) (composite S2/S3) whole rock 77.8571 141.07

87

Rb/86Sr

7.79 1.60

87

Sr/86Sr ± 2a

0.776157 ± 56 0.736629 ± 58

87

Sr/86Sr(i) Age ± 2a (Ma) 0.72641

448 ± 9

Table 3. Rb-Sr geochronology of Ox Mountains pegmatites from Flowerdew et al. (2000) Sample 37 38 39 40 37 40 41

Irish National Grid Ref.

Sample description

Minerals dated

Age ± 2o (Ma)

G270003 M167963 G198956 G242069 G270003 G242069 G198956

Dalradian-hosted pegmatite Dalradian-hosted pegmatite Dalradian-hosted pegmatite Ox Mountains granodiorite Dalradian-hosted pegmatite Ox Mountains granodiorite Dalradian-hosted pegmatite

coarse igneous muscovite-K feldspar coarse igneous muscovite-plagioclase coarse igneous muscovite-K feldspar coarse igneous muscovite-K feldspar recrystallized muscovite-K feldspar recrystallized muscovite-K feldspar recrystallized muscovite-plagioclase

392 ±6 400 ±6 402 ±6 400 ±6 381 ±6 384 ±6 384 ±5

slightly younger than the c. 470-460 Ma age estimates for the Grampian 83 foliation in the Central and NE Ox Mountains inliers (Flowerdew et al. 2000). However, it is broadly coincident with most of the mineral cooling age data from the Central and NE Ox Mountains inliers which cluster at or around 460^50 Ma (Flowerdew et al. 2000), and the reactivated foliation surface is thus thought to represent the regional composite Grampian S2/S3 foliation. The pre-existing foliation surface is markedly older than the age constraints for sinistral extensional crenulation cleavage development in the Central Ox Mountains inlier. Sinistral shearing is constrained by four Rb-Sr muscovite-feldspar ages from pegmatites which range from 402-392 Ma (Table 3) which are interpreted as recording igneous crystallization (Flowerdew et al. 2000). Late sinistral shearing has recrystallized coarse magmatic muscovite within both the Ox Mountains pegmatite suite and the Ox Mountains granodiorite, yielding three Rb-Sr muscovite-feldspar ages of between 385 and 381 Ma (Table 3; Flowerdew et al. 2000).

Discussion on the intrusion age of the Ox Mountains granodiorite Sinistral shearing along the FHCBL in the Central Ox Mountains is effectively constrained by the age of intrusion of the Ox Mountains granodiorite and its presumed coeval pegmatite suite. However, in contrast to the c. 400 Ma age

obtained from the Ox Mountains pegmatite suite (see above), the Ox Mountains granodiorite has yielded c. 480 Ma Rb-Sr whole rock isochrons (Pankhurst et al. 1976; Max et al. 1976). The old Rb-Sr ages suggest that granite emplacement and therefore sinistral strike-slip deformation occurred either before or very early in the Grampian orogeny. However several lines of evidence mitigate against a c. 480 Ma intrusion age. Most of the Ox Mountains granodiorite Rb-Sr whole rock data are characterized by low 87Rb/86Sr values, typically less than 2. Rb-Sr whole rock isochrons characterized by low 87Rb/86Sr values have also been obtained from other Caledonian granites, and these too yield c. 480 Ma intrusion ages even though the intrusion is demonstrably younger (c. 400 Ma) by independent evidence (Kennan 1997). Additionally, unpublished U-Pb multigrain zircon data suggest a c. 415 Ma intrusion age for the Ox Mountains granodiorite (MacDermot et al., 1996). This is consistent with the youngest mineral cooling ages (c. 410 Ma) obtained from the Central Ox Mountains inlier close to the Ox Mountains granodiorite which are likely to be due to thermal resetting, and a c. 400 Ma emplacement age for the pegmatite suite (Flowerdew et al. 2000); a c. 400 Ma intrusion age for the Ox Mountains granodiorite is more likely.

Earlier movements along the FHCBL This study documents two examples of postGrampian strike-slip movement along the

350

D. M. CHEW ETAL.

FHCBL. However, there is evidence for earlier stages of movement along the FHCBL during the Grampian orogeny, illustrating further how this important crustal-scale shear zone has had a long and complicated history of movement. In Tyrone, the D3 Omagh Thrust has translated inverted Dalradian rocks towards the ESE (Fig. 1) over Arenig-Llanvirn shales of the Tyrone volcanics (Alsop & Hutton 1993). In southern Donegal and the NE Ox Mountains inlier, Dalradian rocks were thrust to the SE over granulite-facies basement of the Slishwood Division (Fig. 1) along the Lough Derg Slide (Alsop 1991) and the North Ox Mountains Slide (Flowerdew 1998/9) respectively. Tectonic juxtaposition (D3) of the Dalradian and Slishwood Division is likely to have occurred between 470 and 460 Ma based on 40Ar-39Ar, Rb-Sr and Sm-Nd mineral ages (Flowerdew et al. 2000). Thus in Donegal, the NE Ox Mountains and Tyrone, the earliest constrained phase of movement (c. 470-460 Ma) along the Fair Head-Clew Bay Line involves the translation of the main Dalradian nappes over outboard terranes to the SE.

Conclusions The Fair Head-Clew Bay Line has been shown to have been reactivated several times. It was originally active as a ductile thrust during the (c. 470-460 Ma) Grampian orogeny, where Dalradian nappes adjacent to the Fair Head-Clew Bay Line were thrust over outboard terranes to the SE. The Dalradian metasediments have undergone two separate phases of post-Grampian strike-slip movement adjacent to the Fair Head-Clew Bay Line. These two phases of movement have produced reverse-slip and normal-slip crenulations which modified the earlier Grampian nappe fabrics, and tilted the initially recumbent Grampian nappes into a vertical orientation (Fig. 2b, c). The development of the reverse-slip and normal-slip crenulations produced by these two discrete phases of strike-slip movement has been constrained by isotopic dating. Dextral displacement along the Fair Head-Clew Bay Line in the NW Mayo inlier is constrained to c. 448 Ma, whereas sinistral displacement along the Fair Head-Clew Bay Line in the Ox Mountains inlier is constrained to c. 400 Ma based on previously published pegmatite intrusion ages. Major crustal-scale shear zones may therefore have a long and complicated history of movement, in which pre-existing planar anisotropies (e.g. foliations) act as slip surfaces during later

non-coaxial deformation. Careful analysis of the resulting crenulation morphologies combined with isotopic dating yields a more complete understanding of the reactivation history of major crustal-scale shear zones. D.M.C. gratefully acknowledges a Forbairt Basic Research Grant and a UCD Research Doctoral Scholarship. Barry Long is thanked for many fruitful discussions about Dalradian geology, and the Central Ox Mountains in particular. Sarah Sherlock and Grahame Oliver are thanked for their careful and constructive reviews, which significantly improved this paper.

References ALSOP, G.I. 1991. Gravitational collapse and extension along a mid-crustal detachment: the Lough Derg Slide, northwest Ireland. Geological Magazine, 128, 345-354. ALSOP, G.I. & HUTTON, D.H.W. 1993. Major southeastdirected Caledonian thrusting and folding in the Dalradian rocks of mid-Ulster: implications for Caledonian tectonics and mid-crustal shear zones. Geological Magazine, 130, 233-244. ALSOP, G.I. & JONES, C.S. 1991. A review and correlation of Dalradian stratigraphy in the Ox Mountains and southern Donegal, Ireland. Irish Journal of Earth Sciences, 11, 99-112. ARNAUD, N.O. & KELLEY, S.P. 1995. Evidence for excess argon during high pressure metamorphism in the Dora Maira Massif (western Alps, Italy), using an ultraviolet laser ablation microprobe 40 Ar-39Ar technique. Contributions to Mineralogy and Petrology, 121,1-11. CHEW, D.M. 2003. Structural and stratigraphic relationships across the continuation of the Highland Boundary Fault in western Ireland. Geological Magazine, 140, 73-85. CHEW, D.M., DALY, J.S., PAGE, L.M. & KENNEDY, MJ. 2003. Grampian orogenesis and the development of blueschist-facies metamorphism in western Ireland. Journal of the Geological Society, London, 160, 911-924. CLIFF, R.A. 1985. Isotopic dating in metamorphic belts. Journal of the Geological Society, London, 142, 97-110. DALY, IS. 1996. Pre-Caledonian history of the Annagh Gneiss Complex, north-western Ireland, and correlation with Laurentia-Baltica. Irish Journal of Earth Sciences, 15, 5-18. DENNIS, AJ. & SECOR, D.T. 1987. A model for the development of crenulations in shear zones with applications from the southern Appalachian Piedmont. Journal of Structural Geology, 9, 809-817. DENNIS, AJ. & SECOR, D.T. 1990. On resolving shear direction in foliated rocks deformed by simple shear. Geological Society of America Bulletin, 102,1257-1267. DEWEY, J.F. & SHACKLETON, R.M. 1984. A model for the evolution of the Grampian tract in the early

REACTIVATION OF A CALEDONIAN SHEAR ZONE Caledonides and Appalachians. Nature, 312, 115-121. DUNLAP, W.J. 1997. Neocrystallization or cooling? 40 Ar/39Ar ages of white micas from low-grade mylonites. Chemical Geology, 143,181-203. FLOWERDEW, MJ. 1998/9. Tonalite bodies and basement-cover relationships in the North-eastern Ox Mountains Inlier, north-western Ireland. Irish Journal of Earth Sciences, 17, 71-82. FLOWERDEW, M.J., DALY, IS., GUISE, P.O. & REX, D.C. 2000. Isotopic dating of overthrusting, collapse and related granitoid intrusion in the Grampian orogenic belt, northwestern Ireland. Geological Magazine, 137, 419^35. FREEMAN, S.R., INGER, S., BUTLER, R.W.H. & CLIFF, R.A. 1997. Dating deformation using Rb-Sr in white mica: greenschist facies deformation ages from the Entrelor shear zone, Italian Alps. Tectonics, 16, 57-76. FRIEDRICH, A.M., BOWRING, S.A., MARTIN, M.W. & HODGES, K.V. 19990. Short-lived continental magmatic arc at Connemara, western Ireland Caledonides: implications for the age of the Grampian orogeny. Geology, 27, 27-30. FRIEDRICH, A.M., HODGES, K.V., BOWRING, S.A. & MARTIN, M.W. I999b. Geochronological constraints on the magmatic, metamorphic and thermal evolution of the Connemara Caledonides, western Ireland. Journal of the Geological Society, London, 156,1217-1230. HARRIS, D.H.M. 1993. The Caledonian evolution of the Laurentian margin in western Ireland. Journal of the Geological Society, London, 150, 669-672. HARRIS, D.H.M. 1995. Caledonian transpressional terrane accretion along the Laurentian margin in Co. Mayo, Ireland. Journal of the Geological Society, London, 152, 797-806. HUTTON, D.H.W. 1987. Strike-slip terranes and a model for the evolution of the British and Irish Caledonides. Geological Magazine, 124, 405-425. HUTTON, D.H.W. & DEWEY, J.F. 1986. Palaeozoic terrane accretion in the Irish Caledonides. Tectonics, 5,1115-1124. JONES, C.S. 1989. The structure and kinematics of the Ox Mountains, western Ireland; a mid-crustal transcurrent shear-zone. Unpublished Ph.D. thesis, University of Durham. KENNAN, PS. 1997. Granite: a singular rock. In: Occasional papers in Irish science and technology, 15. Royal Dublin Society, 16. KENNEDY, M.J. 1980. Serpentinite-bearing melange in the Dalradian of County Mayo and its significance in the development of the Dalradian basin. Journal of Earth Sciences of the Royal Dublin Society, 3,117-126. LISTER, G.S. & SNOKE, A.W 1984. S-C mylonites. Journal of Structural Geology, 6, 617-638. LONG, C.B. & MAX, M.D. 1977. Metamorphic rocks in the SW Ox Mountains Inlier, Ireland; their structural compartmentation and place in the Caledonian orogen. Journal of the Geological Society, London, 133, 413-432. LUDWIG, K.R. 1999. Users Manual for Isoplot/Ex, Version 2.10: a geochronological toolkit for

351

Microsoft Excel. Berkeley Geochronology Center Special Publication, la, 49 pp. MACDERMOT, C.V., LONG, C.B. & HARNEY, S. 1996. Geology of Sligo-Leitrim: A geological description of Sligo, Leitrim, and adjoining parts of Cavan, Fermanagh, Mayo and Roscommon, to accompany the Bedrock Geology 1:100 000 Scale Map Series, Sheet 7, Sligo-Leitrim. Geological Survey of Ireland. MAX, M.D. & RIDDIHOUGH, R.P 1975. Continuation of the Highland Boundary fault in Ireland. Geology, 3, 206-210. MAX, M.D., LONG, C.B. & SONET, J. 1976. The geological age and setting of the Ox Mountains Granodiorite. Bulletin of the Geologial Survey of Ireland, 2, 27-5. MCCAFFREY, KJ.W. 1992. Igneous emplacement in a transpressive shear zone: Ox Mountains igneous complex. Journal of the Geological Society, London, 149, 221-235. MCCAFFREY, KJ.W. 1994. Magmatic and solid state deformation partitioning in the Ox Mountains granodiorite. Geological Magazine, 131, 639-652. MULLER, W, DALLMEYER, R.D., NEUBAUER, F. & THONI, M. 1999. Deformation-induced resetting of Rb/Sr and 40Ar/39Ar mineral systems in a lowgrade, polymetamorphic terrane (Eastern Alps, Austria). Journal of the Geological Society, London, 156, 261-278. PANKHURST, R.J., ANDREWS, J.R., PHILLIPS, W.E.A., SANDERS, I.S. & TAYLOR, WE.G 1976. Age and structural setting of the Slieve Gamph Igneous Complex, Co. Mayo, Ireland. Journal of the Geological Society, London, 132, 327-36. PLATT, J.P. & VISSERS, R.L. 1980. Extensional structures in anisotropic rocks. Journal of Structural Geology, 2, 397-410. RAMSAY, J.G. & GRAHAM, R.H. 1970. Strain variation in shear belts. Canadian Journal of Earth Sciences, 7,786-813. RYAN, P.D., SOPER, NX, SNYDER, D.B., ENGLAND, R.W. & HUTTON, D.H.W 1995. The Antrim-Galway Line: a resolution of the Highland Border fault enigma of the Caledonides of Britain and Ireland. Geological Magazine, 132,171-184. SANDERS, I.S., DALY, J.S. & DAVIES, G.R. 1987. Late Proterozoic high-pressure granulite facies metamorphism in the north-east Ox inlier, north-west Ireland. Journal of Metamorphic Geology, 5, 69-85. SANDERSON, D.J., ANDREWS, J.R., PHILLIPS, WE.A. & HUTTON, D.H.W 1980. Deformation studies in the Irish Caledonides. Journal of the Geological Society, London, 137, 289-302. SHERLOCK, S.C. & ARNAUD, NO. 1999. Flat plateau and impossible isochrons: Apparent 40Ar-39Ar geochronology in a high-pressure terrain. Geochimica et Cosmochimica Acta, 63, 2835-2838. SHERLOCK, S.C. & KELLEY, S.P. 2002. Excess argon evolution in HP-LT rocks: a UVLAMP study of phengite and K-free minerals, NW Turkey. Chemical Geology, 182, 619-636. STEIGER, R.H. & JAGER, E. 1977. Subcommission on geochronology: convention on the use of decay

352

D.M. CHEW#r,4L.

constants in geo- and cosmochronology. Earth and Planetary Science Letters, 36, 359-362. TAYLOR, W.E.G. 1969. The structural geology of the Dalradian rocks of Slieve Gamph, Cos. Mayo and Sligo, western Ireland. Geologische Rundschau, 57, 564-588. WEST, D.RW. & Lux, D.R. 1993. Dating mylonitc deformation by the 40Ar-39Ar method: an example from the Norumbega Fault Zone, Maine. Earth and Planetary Science Letters, 120,221-237. WIJBRANS, J.R. & MCDOUGALL, 1.1988. Metamorphic evolution of the Attic Cycladic Metamorphic Belt on Naxos (Cyclades, Greece) utilizing 40Ar/39Ar age spectrum measurements. Journal of Metamorphic Geology, 6, 571-594.

WIJBRANS, J.R., PRINGLE, M.S., KOPPERS, A.A.P. & SCHEEVERS, R. 1995. Argon geochronology of small samples using the Vulkaan argon laserprobe. Proceedings of the Royal Netherlands Academy of Arts and Sciences, 98,185-219. WILLIAMS, D.M., HARKIN, J., ARMSTRONG, H.A. & HIGGS, K.T. 1994. A late Caledonian melange in Ireland: implications for tectonic models. Journal of the Geological Society, London, 151, 307-314. WINCHESTER, J.A. 1992. Comment on 'Exotic metamorphic terranes in the Caledonides: Tectonic history of the Dalradian block, Scotland'. Geology, 20, 764.

Brittle-ductile shear zone evolution and fault initiation in limestones, Monte Cugnone (Lucania), southern Apennines, Italy S. MAZZOLI1, C. INVERNIZZI2, L. MARCHEGIANI2, L. MATTIONI2 & G. CELLO2 l Facoltd di Scienze Ambientali, Universitd di Urbino, Campus Scientifico Sogesta, 61029 Urbino (PU), Italy (e-mail:[email protected]) 2 Dipartimento di Scienze della Terra, Universitd di Camerino, Via Gentile III da Varano, 62032 Camerino (MC), Italy Abstract: The processes of brittle-ductile shear zone evolution and fault initiation by the coalescence of en echelon arrays of tensile cracks are quantitatively analysed in terms of displacement and temperature conditions at which they took place in very low-grade, well bedded micritic limestones from the southern Apennines, Italy. Three different types of structures are distinguished: (i) conjugate arrays of en echelon, calcite-filled tension gashes, showing extensional shear offsets; (ii) en echelon vein arrays showing incipient development of discontinuous shear-parallel fractures cutting through the tension gashes; and (iii) faulted vein arrays, in which vein array breaching by a continuous, discrete normal fault has occurred. Fluid inclusion microthermometry from vein calcite sampled from the different sets of structures (i) to (iii) above indicates that environmental conditions remained roughly constant during the different stages of vein array evolution and fault development, with average homogenization temperatures from primary fluid inclusions being in the range 130-140 °C. Our results show how displacement accumulation and shear strain essentially control vein array evolution by rotation of en echelon tension gashes, fracture linkage and, eventually, fault nucleation, at approximately constant temperature.

En echelon vein arrays, often organized in more or less well developed conjugate sets, have been intensely investigated by structural geologists during recent decades. The kinematic interpretation of these features has also been controversial (e.g. Roering 1968; Lajtai 1969; Ramsay & Graham 1970; Beach 1975; Durney & Ramsay 1983; Pollard et al 1982; Rothery 1988; Smith 1995,1996,1997), and an important step in their interpretation followed the application of the principles of continuum mechanics. The resulting model by Olson & Pollard (1991) suggested that selective vein growth forming en echelon arrays initially might be controlled by the mechanical interaction of neighbouring fractures. Once formed, macroscopic en echelon vein arrays might act as zones of weakness, localizing the later development of brittle-ductile shear zones (as defined by Ramsay & Huber 1987). The latter structures, in turn, might evolve to faults through fracturing between en echelon veins, with fault length increasing as more en echelon fractures link up (Roering 1968; Segall & Pollard 1983; Martel et al 1988). In recent years, a relevant contribution to the understanding of strike-slip fault nucleation by processes of this type came from the work of Peacock & Sanderson (1995), Willemse et al

(1997) and Kelly et al (1998). In all cases, the field examples were characterized by the occurrence of en echelon vein arrays in conjugate fault zones. The observed evolution along these zones from vein arrays to faults allowed the authors to implement a model of linkage in which pull-apart mechanisms play a primary role. This model can also explain the occurrence of vein arrays showing different structural styles in the same outcrop (Kelly et al 1998), which is observed within the area of the present study. Fault initiation by the coalescence of en echelon arrays of Mode I (tensile) cracks (Lawn & Wilshaw 1975) has been suggested by several authors (e.g. Knipe & White 1979; Pollard et al 1982; Etchecopar et al 1986; Cox & Scholz 19880, b\ Cowie & Scholz 1992). According to Scholz (1990), this process is likely to be important in the early stages of fault development. In this work we report arrays of en echelon tension gashes, occurring along shear zones of extensional type, which appear to have evolved locally into discrete normal faults. The studied structures are hosted in well-bedded, finegrained limestones exposed in a quarry in the southern Apennines in the Lucania province of Italy (Fig. 1). This outcrop displays several useful characteristics (Mazzoli & Di Bucci 2003;

From: ALSOP, G. L, HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) 2004. Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 353-373. 0305-8719/$15.00 © The Geological Society of London 2004.

354

S.MAZZOLI£T,4L.

Fig. 1. Location (a), geological setting (b)» and (c) cross-section (no vertical exaggeration) of the study area (modified after Mazzoli et al 2001).

SHEAR ZONE EVOLUTION AND FAULTING

Fig. 2): (i) numerous conjugate sets of vein arrays are well exposed; (ii) some vein arrays show incipient shear fracture development across the tensile cracks; (iii) some are faulted; (iv) in most instances the displacement can be accurately measured; and (v) the displacement associated with vein arrays and faults is quite low (never exceeding 1 m, and mostly below 20 cm). Arrays of fractures, minor faults and veins are commonly observed at the lateral and vertical tips of faults, where they are generally interpreted to record fault propagation rather than initiation/nucleation (e.g. McGrath & Davison 1995; Childs et al 19960; Marchal et al 2003). In these cases, vein arrays and shear zones are interpreted to represent 'process zones' or inelastic strains at the tip of faults that are out of the plane of observation (e.g. Childs et al. 19965). In the latter instance, differently developed structures could represent serial sections through similarly structured fault zones that are dominated by brittle deformation and strain localization (fault slip) over their central portions and brittle-ductile strain towards and around the fault tips. However, in our study area, not only brittle-ductile shear zones dominate the population of structures in terms of their frequency, but also the relatively few discrete brittle faults exposed in the quarry systematically overprint en echelon vein arrays (Fig. 2). Neither a single fault devoid of preexisting en echelon veins, nor vein arrays occurring solely at the tips of discrete fault planes have been observed. These features strongly suggest that the analysed shear zones are indeed precursors to brittle faulting, as opposed to being fault zone-tip related. In summary, evidence of the early stages of normal fault nucleation by the process of coalescence of en echelon vein arrays (e.g. Scholz 1990) is exposed in the studied outcrop. The thermal conditions governing these stages are also constrained by means of fluid inclusion microthermometry, permitting a comprehensive analysis of the early faulting process to be made.

Geological setting The work was carried out in Upper Triassic micritic limestones with chert levels (Calcari con Selce Fm) from the Mesozoic Lagonegro Basin succession (e.g. Scandone 1972) of the southern Apennines fold and thrust belt of peninsular Italy. The southern Apennines include mostly tectonic units derived from the telescoped Apulian (or Adriatic) continental palaeomargin (e.g.

355

Cello & Mazzoli 1999 and references therein) together with the remnants of a Cretaceous to Palaeogene accretionary complex (ophiolitebearing Liguride units). Structurally below the internal Liguride units, the thrust belt consists of allochthonous units derived from the deformation of both carbonate platform and pelagic basin successions (Apenninic Platform and Lagonegro Basin, respectively), of passive margin origin and Triassic to Palaeogene age, which are stratigraphically overlain by Neogene foredeep and wedge-top basin deposits (Miocene and Pliocene successions in Fig. la; Carbone et al 1991). These allochthonous units are completely detached from their original substratum and transported onto the foreland sequence of the Apulian Platform. Analysis of synorogenic deposits indicates that thrust accretion of the units derived from the deformation of the Apenninic Platform and Lagonegro Basin successions occurred mainly in Miocene times, while deeper thrusting involving the hinterland part of the Apulian Platform carbonates occurred mainly in Late Pliocene to Early Pleistocene times (e.g. Cello & Mazzoli 1999 and references therein). The late stages of thrusting were partially contemporaneous with kinematically compatible strike-slip faulting along roughly NNE-SSW trending, right-lateral, and WNW-ESE trending, left-lateral, structures (e.g. Monaco et al 1998). Active compression within the southern Apennines appears to have ceased by Middle Pleistocene times (e.g. Hyppoly te et al 1994). Subsequently, the geometry of the orogen has been modified mainly by NE-SW orientated extension (e.g. Cinque et al. 1993). The limestones analysed in this study are exposed in a quarry at Monte Cugnone, in the high Agri River Valley of Lucania (Fig. Ib). Structural analysis carried out by Mazzoli et al (2001) indicates that the Calcari con Selce Formation, a few hundreds metres thick, represented the mechanically dominant member during contraction and buckling of the sedimentary multilayer, which led to the formation of NW-SE to north-south trending, flexural-slip dominated folds. P-T conditions existing during the development of the analysed vein arrays were mostly controlled by the tectonic burial resulting from the previous contractional episodes. The Monte Cugnone regional structure consists of a faulted anticline of about 1 km wavelength, exposed in the footwall to a major thrust within the Lagonegro units (Marsico Nuovo Thrust; Fig. Ib, c). The area of our detailed study is located in the crestal region of the gently

356

S. MAZZOLI ET AL.

SHEAR ZONE EVOLUTION AND FAULTING

357

Fig. 2. Examples of different types of analysed structures, (a) Conjugate sets of vein arrays in the eastern quarry wall, (b) Line drawing from (a), (c) Intact, conjugate arrays of tensile cracks. Note deflection of the bedding (arrowed), (d) Vein array showing incipient development of discontinuous shear-parallel fractures (arrowed) cutting across en echelon tensile cracks, (e) Faulted shear zone, characterized by a continuous, discrete fault zone (arrowed) breaking through an original en echelon vein array, (f) En echelon vein array (E-W striking) cutting across the fold-thrust structure. Note how the vein array passes from thrust hanging wall to footwall without showing any displacement. Since angles in excess of 60° occur here between the quarry wall and the normal to the thrust transport direction, as well as between the strike of the vein array and the thrust transport direction, overprinting relationships are not just apparent, (g) Line drawing from (f).

358

S.MAZZOLI^r^L.

NW-plunging Monte Cugnone anticline, in an area characterized by subhorizontal to gently dipping bedding (Fig. Ic). The quarry is developed on three sides, one roughly trending WSW-ENE, the other two NNW-SSE. The latter two quarry sides are approximately perpendicular to the strike of the vein arrays, providing best exposures for their study (Fig. 2). In detail, outcrop surfaces are differently orientated and irregular, permitting correct measurement of veins and arrays as required for geometrical analysis (Smith 1995). The host rock consists of a strongly anisotropic (on metre scale) but rather homogeneous (on 10 m scale) multilayer, characterized by regular limestone beds (mostly 10 to 40 cm thick) containing chert layers, lenses and nodules. The portion of the Calcari con Selce Formation exposed in the quarry lacks in the marly/clayey intercalations that are typical of other parts of this formation (Scandone 1972). This rules out the possibility that the localization of brittle-ductile shear zones and discrete faults is controlled by competence contrasts between layers of differing composition.

intact arrays of en echelon, calcite-filled tension gashes (Fig. 2c); (ii) vein arrays showing discontinuous shear-parallel fractures cutting across the tensile cracks (Fig. 2d); and (iii) faulted vein arrays, characterized by a continuous, discrete fault zone breaking through, and evidently developed from, an original en echelon array of tensile cracks (Fig. 2e). These discrete fault surfaces are very likely to have formed by linkage of pre-existing shear fractures of the type shown in Fig. 2d. Therefore, a progressive evolution appears to occur from (i) to (ii), to (iii) above, with differently developed structures maintaining similar attitudes (Fig. 3). Most of the en echelon vein arrays are arranged in conjugate sets, generally striking NE-SW and steeply dipping (Fig. 3a). The veins belonging to the arrays are filled with calcite, and consist of a thick central portion that tapers off into narrow tails. They are mostly planar and only in scattered instances show sigmoidal or irregular shapes (probably due to rotation and/or folding as a result of shearing; Ramsay & Graham 1970). In profile, veins are generally tens of centimetres in length and have maximum aperture width of a few centimetres. They mostly strike NE-SW and show steep angles of Outcrop data dip (Fig. 3b). Most of the exposed en echelon vein arrays show Calcite shear fibres and striae from fully evidence of shear displacement, with exten- developed, discrete fault planes sometimes indisional offsets consistently defined by bedding cate oblique-slip with a right-lateral component (Fig. 2), irrespective of the different orientations of motion (Fig. 3c). However, structural features of the quarry faces. Kinematic analysis of en such as: overprinting relationships, invariably echelon vein arrays in conjugate shear zones, showing discrete faults developing from precarried out by means of the conjugate bisector existing en echelon vein arrays; fault attitude, method (e.g. Ramsay & Huber 1987), invariably consistently similar to that of pre-existing en show NE-SW orientated, subhorizontal shear echelon vein arrays; and the geometry of conjuzone intersections (representing the inter- gate faults, which maintain the same angular mediate axis of the related finite strain ellip- relationships (dihedral angles, intersection soid). Acute bisectors (i.e. shortening lines) as those of conjugate shear zones, all indidirections) of conjugate shear zone dihedral cate that faults originally formed as extensional angles are generally subvertical, whereas obtuse structures, and that a later, though minor, bisectors (i.e. extension directions) are subhori- oblique-slip reactivation has occurred. zontal and NW-SE orientated on average. Bedding-parallel shear veins are commonly These features clearly suggest a bulk strain observed. Shear fibres tend to be roughly perdominated by NW-SE horizontal extension, pendicular to the main fold axis; therefore, they compatible with purely dip-slip displacements can be best interpreted as a result of fold-related along conjugate shear zones. A few vein arrays flexural-slip processes. Early layer-parallel showing no detectable shear displacement also shortening (LPS) is documented by pressure occur, suggesting that these structures formed solution cleavage perpendicular to bedding and early, in response to bulk extension, and prob- by minor thrusting with associated outcrop-scale ably pre-dated shear strain localization. En folding. Cross-cutting relationships demonstrate echelon vein arrays possibly represented that LPS and flexural folding pre-dated en original zones of weakness that localized later echelon vein arrays (Fig. 2f, g). Therefore, the shear zone development, as suggested by, for structures analysed in this study can best be example, Roering (1968) and Olson & Pollard related to post-buckling extension, roughly par(1991). allel to the regional, NW-SE trending, fold axis. In outcrop, it was possible to distinguish: (i) Such a deformation seems to accommodate only

SHEAR ZONE EVOLUTION AND FAULTING

359

Fig. 3. Orientation data (lower hemisphere, equal area projection), (a) Poles to en echelon vein arrays (mean great circles for each of the two conjugate sets are shown, dipping 73° toward 310°N and 74° toward 129°N). (b) Poles to vein planes from en echelon arrays (mean great circles for the two dominant sets are shown, dipping 79° toward 305°N and 78° toward 127°N). (c) Discrete faults cutting through en echelon vein arrays (with slip vector determined from striae/shear fibres on fault plane; see text), (d) Poles to bedding (mean great circle shown, dipping 12° toward 264°N).

a relatively low amount of strain, at least within the studied outcrop (where bulk extension is
ship may be due either to the effect of pre-existing discontinuities (i.e. the stylolites) inhibiting the propagation of later extension fractures (e.g. Price & Cosgrove 1990) or, where bedding-parallel stylolites clearly cut across (otherwise continuous) extension veins, to stylolite reactivation due to their suitable orientation relative to vertical loading. In the latter instance, the related pressure-solution process may have provided a source of material for calcite precipitation within the veins. Five angular parameters can be used to describe the geometry of en echelon vein arrays (e.g. Srivastava 2000 and references therein): (i) the array dihedral angle 0, (ii) the average vein angle a, (iii) the en echelon vein array boundary angle ft, (iv) the inter-vein angle 0 and (v) the vein-array angle 8 (Fig. 4). Moreover, vein arrays can been characterized by average values of vein spacing, length and width. In this study, a total of 85 en echelon vein arrays, exposed over lengths of several metres to a few tens of metres, have been analysed (Table 1). Most of the measured vein arrays can be

S. MAZZOLI ETAL.

360

Fig. 4, Schematic diagram of a conjugate en echelon vein array showing geometric parameters discussed in the text. T: shear zone thickness; D: shear zone displacement; S0: bedding. Table 1. Structural data (displacement data for 54 features are from Mazzoli & Di Bucci 2003) Type

Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Fault Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Incipient fault

Shear zone attitude

Average vein attitude

dip direction angle of dip

dip direction angle of dip

298 284 311 130 125 152 168 145 281 316 299 324 156 124 140 152 321 302 128 331 149 121 156 286 168 127

86 81 71 76 78 62 83 56 80 66 79 65 69 78 71 70 69 69 71 77 78 67 59 76 66 68

122 100 149 304 291 330 355 217 118 136 310 127 161 298 311 336 131 129 138 141 133 304 297 111 332 304

77 80 76 84 80 87 63 76 82 81 74 80 79 79 81 84 82 71 68 75 74 77 76 71 70 74

Thickness Displacement (cm)

(cm)

4.9 4.8 12 3.6 3.2 3.7 3.4 8.3 5.3 4.2 2.7 9.6 1 7.2 6.2 5 13 6.7 2.6 15 4.1 4.3 13 9.5 4.3 8.8

0.1 1.1 1.2 1.5 1.5 1.6 2 2.2 2.2 2.2 2.2 2.4 2.5 3 3 3 3.2 3.3 3.5 3.7 4 4.1 4.4 4.5 4.8 5

j

0.02 0.23 0.09 0.42 0.48 0.44 0.61 0.27 0.43 0.53 0.81 0.24 2.65 0.42 0.49 0.61 0.24 0.5 1.35 0.24 0.97 0.94 0.35 0.47 1.11 0.57

SHEAR ZONE EVOLUTION AND FAULTING

361

Table 1. continued. Type

Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Incipient fault Shear zone Shear zone Fault Fault Shear zone Fault Shear zone Incipient fault Incipient fault Fault Shear zone Fault Fault Incipient fault Fault Fault Fault Fault Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Fault Fault Incipient fault Incipient fault Incipient fault Incipient fault Incipient fault Incipient fault Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Shear zone Incipient fault Shear zone Incipient fault Shear zone Shear zone Incipient fault

Shear zone attitude

Average vein attitude

dip direction angle of dip

dip direction angle of dip

121 144 311 148 351 312 334 300 318 302 345 336 129 125 298 151 132 168 301 306 319 114 333 319 152 175 136 316

134 310

74 71 77 62 77 71 71 76 62 74 72 75 78 84 64 75 71 79 71 79 77 83 60 76 78 69 69 85 65 58 75 72 87 70 88 86 77 63 78 60 55 70 60 70 71 82 65 60 65 75 73 73 74 63 83 75 72 68 65

318 126 299 164 172 131 144 126 154 121 149 160 302 295 129 300 306 322 142 137 139 304 158 144 318 310 301 143 320 150 306 300 130 145 120 316 310 140 320 152

321 318 138 140 312 308 304 192 158 128 138 304 170

87 89 62 76 76 71 53 75 71 66 76 72 76 42 71 73 79 75 67 64 61 81 81 65 64 81 85 63 85 80 85 72 70 72 80 75 83 80 78 86 70 61 70 83 85 75 70 70 65 80 84 62 60 57 53 72 82 60 85

Thickness Displacement

(cm)

(cm)

5.7 11 3.4 7.2 3.8 9.2 7.6 11 6.1 19 7.6 21 7 5.3 11 9.2 4.5 8.4 11 8.4 17 5.2 5 9 9.8 8.9 7.5 6.5

5 5.4 5.7 5.8 6 6.7 7.6 7.7 8.4 8.9 9.4 9.5 9.5 9.5 9.7 11 11 13 14 16 17 30 30 40 41 43 44 99

15.5 12.5 15 15 12 16 6 11 17 8 32.5 7 46 10 11 20 7

y

0.88 0.49 1.67 0.8 1.6 0.72 0.99 0.71 1.37 0.46 1.23 0.46 1.35 1.79 0.89 1.24 2.48 1.52 1.26 1.88 0.98 5.68 6 4.47 4.14 4.84 5.84 15.2

362

S.MAZZOLIETAL.

Fig. 5. En echelon vein arrays from Monte Cugnone plotted on the triangular graph of Srivastava (2000) using the three angular parameters <2>, <9, and 8.

classified geometrically, according to Beach (1975), as Type 1' (adopted as a strictly descriptive term, and not referring to the mode of formation of the vein arrays). They mostly range from the 'weakly convergent' to the 'strongly convergent' types of Smith (1996). For about half of the analysed vein arrays, the three angular parameters 0 (array dihedral angle), 3 (vein-array angle) and 0 (inter-vein angle) could be obtained and plotted on the diagram of Srivastava (2000). According to the classification by the latter author, our structures range from Type 1-2 to Type 1-3 (Fig. 5). Accurate measurements of shear zone displacement, using subhorizontal bedding (Fig. 3d) as a marker, were carried out by Mazzoli & Di Bucci (2003) from 54 structures belonging to the same data set of this paper (Table 1) and representing all different stages of the structural evolution outlined above. Shear zone displacement (D) has been measured, using the trace of offset bedding, in the XZ plane of the related finite strain ellipsoid and along the shear direction (assuming that the intersection of conjugate shear zones is parallel to the intermediate (Y) axis of the related finite strain ellipsoid; Ramsay & Huber 1987). Shear zone thickness (T) has been measured as the perpendicular distance between the two enveloping surfaces containing the tip lines of the tension gashes belonging to a single array (Fig. 4). The shear strain (7) can then

be obtained for each shear zone simply as (Fig. 4):

Equation (1) permits the determination of 7 also for vein arrays showing incipient development of shear-parallel fractures (Fig. 2d) or that have evolved completely into discrete faults (Fig. 2e). In the latter instances, the measured shear zone thickness (T) is that of the whole vein array from which the fault zone has developed, whereas the displacement (D) is the cumulative offset of bedding due to both brittle-ductile deformation and subsequent fault slip. By this method, the related shear strain (7) is determined as if all the deformation was of ductile nature, occurring across the whole thickness (T) of the shear zone. For faulted vein arrays, this method introduces an approximation (Mazzoli & Di Bucci 2003) which will be taken into account in the analysis of the data. Spacing distribution was also analysed by measuring the space between shear zones/faults in a 34 m long sample line performed along the eastern quarry wall (Fig. 6). This provided the opportunity to compare values of spacing for the different types of structures described above. Intact and incipiently faulted shear zones show a spacing ranging from 0.5-9.7 m, whereas faulted vein arrays are characterized by a more regular

SHEAR ZONE EVOLUTION AND FAULTING

363

Fig. 6. Sample line along the eastern quarry wall, showing spacing of different structures. Dashed lines are intact or incipiently faulted shear zones; continuous lines are faulted shear zones.

distribution, with an average value of spacing of 6.4 m. This spacing distribution can also be quantitatively assessed by means of the coefficient of variation (Cv), which is the standard deviation of the spaces divided by the mean spacing (Gillespie et al. 1993). The Cv is greater than one for structures that are clustered, less than one for anticlustered spaced structures, and equal to one and zero for distributions that are perfectly random and perfectly regular, respectively. Along the measured sample line, the coefficient for intact and incipiently faulted shear zones is 0.59, whereas the Cv for completely faulted ones is 0.12. The former value is indicative of a relatively anticlustered distribution of intact and incipiently faulted vein arrays, whereas the latter suggests a relatively regular frequency distribution of discrete faults breaking through pre-existing vein arrays.

Microstructures Microstructural analysis on the calcite-filled en echelon veins was performed in thin section by means of optical transmission microscopy. The sparry calcite filling the tension gashes consists mostly of euhedral crystals (smaller along the vein walls with respect to the vein centre), which may be indicative of rapid opening of the veins. Calcite fibres, where present, are orientated perpendicular to the vein walls and indicate a Mode I opening of the veins (Lawn & Wilshaw 1975). Vein calcite displays various degrees of twinning and undulose extinction: twin lamellae in the more intensely strained vein calcite consist of thick twins (> 1 urn, 'type IF in Burkhard 1993; Fig. 7a) generally organized in two, or rarely three, straight sets, and twins in twins ('type IIP in Burkhard 1993). Lobate crystal margins indicate that dynamic recrystallization by grain boundary migration also occurred locally (Fig. 7b). Transgranular fractures and microcracks cutting though thick twins (mainly occurring adjacent to impingement zones between grains) are frequent and marked by secondary fluid inclusion trails (Fig. 7c). The occurrence of undulose extinction and twinning is indicative of crystal plastic defor-

mation of the vein calcite, also testified by some grain boundary migration phenomena. However, the microstructure is characterized by a general lack of evidence for recovery (in the form of subgrain development or extensive dynamic recrystallization), hence suggesting that temperatures were not high enough for dislocation creep (e.g. Nicolas & Poirier 1976) to be important in these rocks. Low-temperature intracrystalline deformation dominated by glide processes alone (low-temperature plasticity; e.g. Frost & Ashby 1982) is likely to have led to strain hardening and to the widespread development of microfractures (Fig. 7c), thereby producing a switch to a dominantly brittle mechanical behaviour.

Fluid inclusion petrography and microthermometry Fluid inclusion microthermometry is the observation of phase changes in fluid inclusions under controlled conditions of heating and cooling. This non-destructive method of fluid inclusion studies was performed on vein calcite by means of an USGS gas flow heating-freezing stage (Goldstein & Reynolds 1994). Temperatures of phase changes at T < 0 °C are accurate to ± 0.1 °C; T > 50 °C are accurate to ± 1 °C. Microthermometry permits data for melting (Tm) and homogenization (Th) temperatures to be obtained for each inclusion. The temperature of homogenization (Th) is the point at which, in a two-phase inclusion (vapour V + liquid L), the vapour bubble disappears and the conditions of trapping of the fluid are somehow reproduced. The temperature of melting (Tm) for water-rich inclusions is referred to the melting of ice. Composition of fluid inclusions represented by the H2O-NaCl system are determined from temperatures of phase changes during microthermometric analysis. For salinities < 23.2% NaCl, the composition is determined from the ice-melting temperatures (Tm): the temperature at which the last of the ice melts defines the freezing point of the inclusion and this value is used to determine the salinity (Bodnar & Vityk 1994).

364

S. MAZZOLI ET AL.

Fig. 7. Examples of microstructures from vein calcite. (a) Two sets of thick twins, (b) Lobate grain boundaries (suggesting dynamic reerystallization by grain boundary migration processes) between calcite crystals, (c) Cracks with secondary inclusion trails (arrowed), (d) Iso-orientated primary inclusions (P) and straight microvein (MV) cutting through the crystals, (e) Primary fluid inclusion aligned along growth bands (arrowed), (f) Elongated two-phase primary inclusions (example arrowed), (g) Rounded two-phase secondary inclusions.

SHEAR ZONE EVOLUTION AND FAULTING

In diagenetic minerals, Th is a good approximation of the minimum trapping temperature of inclusions if a single phase of an immiscible fluid is entrapped (Goldstein & Reynolds 1994). For this work, samples were collected from: (i) calcite from straight en echelon veins belonging to intact en echelon arrays (sample SD2); (ii) calcite from en echelon veins belonging to incipiently faulted vein arrays (samples SD3, SA1A, SA3); (iii) calcite from dilational jogs and shear fibres associated with fully developed faults cutting across pre-existing vein arrays (samples SA1B, A2); and (iv) calcite from shear fibres indicating minor oblique-slip reactivation of fault planes (sample SD4). Double polished wafers 150 um thick were cut perpendicular or parallel to vein walls. Weakly deformed, clear crystals show numerous, small (2-10 \im) fluid inclusions of both primary and secondary origin (Fig. 7d-g). Following Goldstein & Reynolds (1994), a primary origin is interpreted for inclusions isoorientated in the grain-growth direction, aligned in growth bands or grouped into cloudy areas (Fig. 7 d-f). They show various shapes (round, elongated and negative crystal). Some primary inclusions suffered leakage, as shown by a variability of vapour percentage in the fluid inclusion assemblages and/or microcracking around the inclusions, especially along twin lamellae in intensely deformed crystals. On the other hand, well-preserved primary inclusions show a constant L/V ratio (about 10-15%). Very small primary inclusions are often one-phase liquid. In the samples from incipiently faulted vein arrays, the calcite fillings are frequently crosscut by microveins (MV; Fig. 7d) containing less deformed calcite crystals and both primary and secondary fluid inclusions. Secondary inclusions are abundant: they tend to be aligned along microfractures that cut across various grains. At least two sets of secondary inclusions can be distinguished: (i) inclusions aligned along healed microfractures (Fig. 7g); these generally show round or slightly elongate, two phase inclusions with a small vapour percentage (< 10%), but single phase inclusions are also observed; and (ii) iso-orientated, all-liquid inclusions with very elongate (acicular) and/or large-flat shapes, presumably associated with partially healed microfractures. Based on microthermometry results, no relevant differences in melting temperature (Tm) are identified for fluid composition between primary and secondary inclusions and between different samples (Fig. 8a). All the inclusions belong to the H2O-NaCl system. Eutectic temperature (Te), was rarely observed

365

(6 data) and ranges between -19 and -21 °C. This, together with Tm data, indicates a fluid composition consisting of low salinity water (1.74wt%NaCl system, following Bodnar's 1993 equation). Homogenization temperatures (7^) for primary and secondary inclusions are plotted together on the histograms of Fig. 8b-e. Primary inclusions record average Th of 130 °C, with scattered peak temperatures of up to 165 °C. Average homogenization temperatures remain roughly constant for primary inclusions from macroveins related to the different stages of vein array evolution, and also from microveins. In particular, microthermometric analysis on primary inclusions belonging to microveins gave results very close to the surrounding primary inclusions in the host vein. Sample SD4, which is from gently plunging shear fibres associated with strike-slip reactivation of a pre-existing normal fault, gave more disperse, but still quite consistent Th data (Fig. 8e). Secondary inclusions record Th values mostly ranging between 80-115 °C (Fig. 8b-e). A T^ITm graph also shows no evidence for a different fluid composition between primary and secondary inclusions and from structures belonging to different stages of vein array evolution (Fig. 8f).

Insights from fluid inclusion data The consistency of homogenization temperature data can be assessed by means of two criteria (Goldstein & Reynolds 1994): results from inclusions of variable size and shape, which tend to respond differently to thermal re-equilibration, are largely comparable; and a high percentage of data fall within a range of 10-15 °C. More than 70% of our Th data for primary inclusions cluster around the central value of 130 °C. Therefore, the examined fluid inclusion assemblages have presumably not undergone significant thermal re-equilibration. The results of our study suggest that, during the different stages of vein array evolution, the circulating fluids maintained an almost constant composition (i.e. very low salinity water) and a roughly constant temperature around 130 °C. Fluid composition remained approximately the same during the later entrapment of secondary inclusions, as temperature dropped down. Secondary fluid inclusion planes (healed microfractures) are hosted in calcite veins belonging to all the stages of shear zone evolution. Therefore, they are probably related to younger deformation episodes that took place during ongoing exhumation and cooling.

366

S. MAZZOLI ET AL.

Fig. 8. Fluid inclusion data, (a) Cumulative Tm data (P: primary inclusions; S: secondary inclusions), (b-e) Th data for the different stages of vein array evolution discussed in the text: (b) sample SD2; (c) samples SD3, SA1A, SA3; (d) samples SA1B, A2; (e) sample SD4. (f) ThITm diagram for all data (35 samples).

Primary fluid inclusions are well preserved in ealcite veins belonging to the different stages of the reconstructed vein array evolution. This is indicative of a gradual exhumation, without complete re-equilibration or decrepitation of the inclusions themselves. Th data suggests that the transition from brittle-ductile to fully brittle shear occurred at roughly constant temperature. As a result, we can rule out the possibility that changes in mechanical behaviour were controlled by changing burial and temperature conditions (e.g. during ongoing exhumation and

cooling), and are left with the alternative of focusing on the role of strain as an enhancing factor in the processes of shear zone evolution and normal fault nucleation. Kinematics Shear zone displacement data (Mazzoli & Di Bucci 2003) indicate that a critical value of about 9 cm can be identified for statistically significant onset of normal fault nucleation from brittle-ductile shear zones (Fig. 9a).

SHEAR ZONE EVOLUTION AND FAULTING

367

Fig. 9. Diagrams of displacement and shear strain for the 54 structures analysed by Mazzoli & Di Bucci (2003). (a) Population graph of cumulative number vs. displacement. Both single regression line (for data distribution in the range 3.0-40.2 cm) and two distinct regression lines (for data distribution in the ranges 3.0-9.4 cm and 9.7^0.2 cm; dashed) are shown, (b) Graph of shear strain vs. displacement.

368

S. MAZZOLI ETAL.

Fig. 10. Graph of vein angle (a) vs. zone boundary angle (j8). Different fields are defined by interpreting the plotted parameters as resulting from bulk strain geometry (modified after Rothery 1988; Kelly et al. 1998). Capital letters in parentheses (A-D) refer to models by Ramsay & Huber (1987, their fig. 26.42).

Displacement values in the range of 9-17 em appear to characterize a 'transition zone' to the development of discrete faults. Within this zone, different types of structures occur, ranging from unbroken shear zones to completely faulted ones. Faulted vein arrays occur where displacement values are in excess of 17 cm. Concerning the determination of possible critical values of shear strain for normal fault nucleation from shear zones, an approximation for the calculation of /from faulted vein arrays must be taken into account (Mazzoli & Di Bucci 2003). For accurate estimates on these structures, the following are needed: a complete partitioning of the total displacement into initial brittle-ductile and later fault slip components and an estimate of the thickness of the shear zone at the time of fault formation. Determination of these estimates is inherently difficult as these are finite strains. However, since displacements are small (few tens of centimetres at most; Table 1) it can be assumed that shear zone characteristics, specifically thickness upon which the shear strain is dependent, are close to those at the time of fault formation. Given this, the method adopted in the calculation of /is likely to lead to a progressively larger underestimation

of the shear strain as deformation localizes along a discrete fault zone (with the consequent narrowing of the actual slip zone thickness). This method is likely to represent a conservative approach for the determination of the cumulative shear strain in faulted vein arrays. The consequences of this approximation in the determination of /appear to be negligible in a diagram like that of Fig. 9b, where a possible underestimation of the shear strain for the larger faults (i.e. those with /> 2.7) would not substantially affect the main pattern. This suggests that, for /estimates above 2.7, relatively high values of shear strain tend to be taken up by throughgoing, discrete faults. Fig. 9b also emphasizes that vein array breaching by faulting starts to occur for shear strain values around 1.2. For shear strains in the range of 1.2 < /< 2.7, structures of different types (intact, incipiently faulted and faulted vein arrays) appear to define a transition zone to the development of discrete faults. Discussion Based on the results exposed above, it seems that shear strains below 7= 1.2 can be sustained

SHEAR ZONE EVOLUTION AND FAULTING

by brittle-ductile shear zones, at roughly constant temperature conditions around 130 °C, without significant development of shear parallel fractures breaching the tensile crack array. During these early stages of shear zone evolution, strain accumulation is likely to produce rotation of en echelon tension gashes (e.g. Ramsay & Huber 1987), and/or buckling of inter-vein columns of rock (Olson & Pollard 1991). Both of these processes should result in deformation and folding of en echelon veins (e.g. Ramsay & Graham 1970). However, only a few of our en echelon tension gashes show a sigmoidal shape; most of them are planar or nearly so. Therefore, it seems that shear strain accumulation, eventually leading to fault-slip, occurred in the shear zones without significant folding of en echelon veins. On the other hand, the occurrence of vein rotation during shear zone evolution might be suggested by a graph of vein angle (a) vs. zone boundary angle (/?) (Fig. 10). In a graph of this type (Rothery 1988; Kelly et al 1998), the central line (where a = (3) represents conditions of bulk stretching (corresponding to model A of Ramsay & Huber 1987, their fig. 26.42, with no shear zones forming). The two lines where a- ft = ± 45° represent conditions of simple shear (corresponding to model C of Ramsay & Huber 1987). These lines separate domains characterized by a component of extension (or positive dilation according to Ramsay & Huber 1987, their model B) normal to the shear zone walls from fields of contraction (or negative dilation according to Ramsay & Huber 1987, their model D) normal to the shear zone walls. In a different context, similar domains were defined by Kelly et al (1998) as fields of transtension and transpression, respectively (as these authors were actually dealing with strike-slip faults). The graph in Fig. 10 shows that most of the analysed vein arrays cluster in the domains of 'simple shear + bulk stretching'. In particular, the shear zones tend to plot closer to the central line of pure 'bulk stretching' than incipiently or completely faulted shear zones. The latter, in turn, tend to cluster closer to the lines of simple shear. These features suggest that the analysed shear zones were possibly characterized, during the early stages of their development, by a significant component of positive dilation (according to the model of Ramsay & Huber 1987) and/or of horizontal bulk stretching affecting the whole volume of host rock (as homogeneous strain is required to maintain strain compatibility; Ramsay & Huber 1987). This would have led to the formation of vein arrays characterized by an angle <5that, although likely to be comprised within a range of values, as suggested by Fig. 10, is mostly less than 45°.

369

The progressive rotation of tension gashes with increasing shear strain, leading to increasing vein-array ($) angles (Fig. 4), might have led to the clustering of more 'evolved' shear zones around the simple shear domains. If this is correct, en echelon vein arrays plotting in different areas Fig. 10 are not the result of strain geometry alone, i.e. simple shear with variable dilation/bulk stretching components normal to the shear zone walls, but also of progressive deformation involving rotation of the veins. As vein rotation appears to have occurred without general deformation of the tension gashes, which remain planar in most of the instances, it was probably accompanied by thickening of the shear zone, whose walls may have migrated toward the surrounding country rock during shear zone evolution (Fig. 11). The model in Fig. 11 implies no simple correlation between shear zone displacement and parameters such as shear zone thickness (7), average vein angle (a), and vein-array angle (S) at which fractures formed in different en echelon vein arrays. This is because of both the likely original variability of 8 and average vein length (and hence vein array thickness) within the shear zone population (Olson & Pollard 1991) and the non-linear function describing the rotation of a marker (e.g. vein) within a shear zone (Ramsay & Huber 1983). On the other hand, increasing shear zone thickness is also implied in the model of brittle-ductile shear zone evolution envisaged by Ramsay & Huber (1983, their fig. 2.11). In the latter model, involving progressive sigmoidal folding of en echelon tension gashes by increasing displacement, the fissure tips continue to propagate into the shear zone walls (with a propagation direction controlled by the incremental strains). As a consequence, 'the shear zone widens and the deformation front moves outward' (Ramsay & Huber 1983). For the structures analysed in this study, environmental (P-T) conditions of the deformation were not sufficient to produce fully ductile deformation and folding of early formed tension gashes and/or inter-vein rock columns. The resulting quasi-rigid block rotation shown in Fig. 11 could enhance strain localization and brittle fracturing along the edges of rotating blocks, thereby contributing to the development of the discontinuous shear surfaces (Fig. 2c) whose linkage may eventually produce throughgoing faults. This macroscopic process is compatible with the microstructural evidence of low-temperature plasticity (e.g. Frost & Ashby 1982) in the studied limestones. Brittle fracturing could be enhanced by the strain hardening intrinsic in intracrystalline deformation dominated by low-temperature mechanisms

370

S.MAZZOLIETAL.

(involving no recovery by thermally-activated processes; e.g. Nicolas & Poirier 1976). Thus in this study area, the progressive rotation of the veins inside the arrays led to shear stress localization, overlapping and, eventually, linkage among veins producing brittle shear (e.g. Olson & Pollard 1991). A possible model for the evolution of the studied en echelon vein arrays may have involved: (1)

(2)

(3)

the growth of initial flaws into microcracks that align parallel to the principal compressive stress and progressively become long enough to interact as en echelon arrays (Olson & Pollard 1991) showing a relatively anticlustered distribution (Fig- 6); successive strain localization along vein arrays producing shear deformation, vein rotation and, possibly, increase of shear zone thickness; and a final stage of transition from brittle-ductile to brittle shear.

The latter stage is marked at first by the development of discontinuous shear fracture segments and then by their progressive linkage, eventually producing discrete faults. As previously discussed, this process appears to be essentially controlled by the accumulation of displacement along the analysed vein arrays, with progressive strain localization occurring preferentially along regularly spaced faults (Fig. 6). Taking into account the likely role played by shear zone displacement (D) in the process of fault nucleation, the scaling properties of this parameter have also been investigated using the cumulative frequency distribution for D shown in Fig. 9(a). This shows a data point distribution which can be approximated by a power-law frequency distribution of the form:

Fig. 11. Schematic model for the rotation of tension gashes maintaining a roughly planar geometry and thickening of the shear zone.

where a is a constant, and the power-law exponent d is termed the fractal dimension of the population (e.g. Needham et al 1996; Yelding et al 1996). The distribution between about 3-40 cm defines an approximate straight line, which may suggest scale-invariance over more than one order of magnitude. Similar power-law relationships are well established for normal fault populations, which tend to be selfsimilar over a range of scales (e.g. Walsh & Watterson 1988; Knott et al 1996; Needham et al 1996; Yelding et al 1996 and references therein). Therefore, it seems that brittle-ductile shear

SHEAR ZONE EVOLUTION AND FAULTING

zones, dominated by viscous deformation processes accompanied by brittle fracturing, display a behaviour of the type generally shown by faults in the frictional regime. In Fig. 9a the transition from brittle-ductile shear zones to fully developed faults is not marked by a significant deviation in the straight line trend. However, a minor kink in the distribution occurs for displacements around 9 cm. Bearing in mind that the style of strain accumulation is different below (brittle-ductile) and above (brittle fault slip) D = 9 cm, the location of the kink in the distribution may not be coincidental, although no obvious explanation exists for it. On the other hand, when the data distributions either sides of the kink are analysed separately, the resulting straight line segments show very similar slopes (Fig. 9a). A marked deviation from the straight line trend occurs at the extreme right-hand side of the diagram. This 'tail' of the graph, consisting exclusively of faulted vein arrays, may be the result of undersampling of structures showing relatively large displacements (i.e. > 40 cm). Alternatively, maximum fault size could be constrained by mechanical layering. Once completely developed, through-going discrete faults may be able to take up relatively large displacements, leading to progressive strain localization and to the development, in a cumulative frequency distribution, of a non-power-law 'tail' containing the largest faults. Progressive strain localization onto larger faults, accompanied by the abandonment of smaller structures, can in fact produce a change in the active fault population from power-law to scale-bound (Walsh et al 2003). Clearly, the very few data points in the right hand 'tail' of the diagram in Fig. 9(a) are statistically not significant to discriminate between the possible different hypotheses. Conclusions This study emphasizes the fundamental role played by finite strain in the process of fault initiation by the coalescence of en echelon vein arrays in naturally deformed limestones. Fluid inclusion analysis suggests that the different sets of structures recognized in the field developed at approximately constant temperature. Critical values of displacement (D ~ 9 cm) and shear strain (7 ~ 1.2) were determined by Mazzoli & Di Bucci (2003) for the onset of normal fault nucleation from shear zones in the studied limestones. Structures displaying shear displacements below these critical values consist, for the vast majority (Fig. 9, Table 1), of brittle-ductile shear zones as defined by Ramsay

371

& Huber (1987). The deformation processes responsible for their formation are likely to be of dominant viscous type (i.e. intracrystalline deformation, solution mass transfer), accompanied by brittle fracturing (as indicated by the coeval formation of tension gashes). Such deformation processes are likely to become progressively less important as shear-parallel fractures, characterizing the incipient stages of fault nucleation, start to form. Eventually, once throughgoing faults are developed, deformation becomes dominantly frictional. In this context, during progressive strain localization leading to incipient fault nucleation (i.e. for 9 < D < 17 cm and 1.2 < / < 2.7; Fig. 9b), a transition from a dominant viscous to a frictional behaviour would occur. In the studied area, shear zone evolution may be reconstructed as follows: (i) strain localization occurs along en echelon crack arrays (e.g. Olson & Pollard 1991); (ii) en echelon veins, mostly formed originally at angles less than 45° with respect to shear zone walls, progressively rotate within the shear zones; this process is probably accompanied by thickening of the shear zones, mostly without significant folding of the tension gashes themselves; (iii) vein rotation and increasing shear zone displacement produce shear stress localization, leading to linkage of the veins by shear-parallel fractures and eventually producing brittle shear. In conclusion, our results, concerning normal fault initiation from en echelon vein arrays in limestones deformed at very low-grade conditions, should hopefully be applicable to the analysis of fault growth processes within the upper crust. Similar studies, carried out in rocks characterized by various rheological behaviours and/or deformed under different environmental conditions, may be useful to implement existing models of fault nucleation, thereby improving our comprehension of the processes of fault initiation and early development. This study greatly benefited from discussions with Daniela Di Bucci. Thorough and constructive reviews by W. Bailey and M. Barchi substantially helped to improve the paper. Financial support from the Italian MIUR Cofin 2002 (Resp. G. Cello, prot. 2002043912) and Universita di Urbino are gratefully acknowledged.

References BATHURST, R.G.C. 1971. Carbonate sediments and their diagenesis. Elsevier, Amsterdam. BEACH, A. 1975. The geometry of en echelon vein arrays. Tectonophysics, 28, 245-263. BODNAR, R.J. 1993. Revised equation and table for determining the freezing point depression of

372

S. MAZZOLI ETAL.

H^O-NaCl solutions. Geochimica et Cosmochimica Acta, 57, 683-684. BODNAR, R.J. & VITYK, M.O. 1994. Interpretation of microthermometric data for H2O-NaCl fluid inclusions. In: DE Vivo, B. & FREZZOTTI, M.L. (eds) Fluid inclusions in minerals: methods and applications. Virginia Tech, Blacksburg, 117-130. BURKHARD, M. 1993. Calcite twins, their geometry, appearance, and significance as stress-strain markers and indicators of tectonic regime: a review. Journal of Structural Geology, 15, 351-368. CARBONE, S., CATALANO, S., LAZZARI, S., LENTINI, F. & MONACO, C. 1991. Presentazione della carta geologica del Bacino del Fiume Agri (Basilicata). Memorie della Societd Geologica Italiana, 47, 129-143. CELLO, G. & MAZZOLI, S. 1999. Apennine tectonics in southern Italy: a review. Journal ofGeodynamics, 27,191-211. CHILDS, C., NICOL, A., WALSH, J.J. & WATTERSON, J. 1996a. Growth of vertically segmented normal faults. Journal of Structural Geology, 18, 1389-1397. CHILDS, C., WATTERSON, J. & WALSH, JJ. 19966. A model for the structure and development of fault zones. Journal of the Geological Society of London, 153, 337-340. 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. COWIE, PA. & SCHOLZ, C.H. 1992. Physical explanation for the displacement-length relationships of faults using a post-yield fracture mechanics model. Journal of Structural Geology, 14, 1133-1148. Cox, SJ.D. & SCHOLZ, C.H. 1988a. An experimental study of shear fracture in rocks: Mechanical observations. Journal of Geophysical Research, 93,3307-3320. Cox, SJ.D. & SCHOLZ, C.H. 19886. On the formation and growth of faults. Journal of Structural Geology, 10,413-430. DURNEY, D.W. & RAMSAY, J.G 1983. Incremental strains measured by syntectonic crystal growth. In: DE JONG, K.A. & SCHOLTEN, R. (eds) Gravity tectonics. Wiley, New York, 67-96. ETCHECOPAR, A., GRANIER, T. & LARROQUE, J.-M. 1986. Origin des fentes en echelon: propagation des failles. Comptes Rendues de I'Academie des Sciences, Paris, 302, 479-484. FROST, H.S. & ASHBY, M.F. 1982. Deformationmechanism maps. Pergamon Press, Oxford. GILLESPIE, P. A., HOWARD, C.B., WALSH, JJ. & WAITERSON, J. 1993. Measurement and characterisation of spatial distribution of fractures. Tectonophysics, 226,113-141. GOLDSTEIN, R.H. & REYNOLDS, TJ. 1994. Systematics of fluid inclusions in diagenetic minerals. SEPM Short Course 31, Tulsa, USA. HYPPOLYTE, J.-C, ANGELIER, J. & ROURE, F. 1994. A major geodynamic change revealed by Quater-

nary stress patterns in the Southern Apennines (Italy). Tectonophysics, 230,199-210. KELLY, P.G., SANDERSON, DJ. & PEACOCK, D.C.P. 1998. Linkage and evolution of conjugate strike-slip fault zones in limestones of Somerset and Northumbria. Journal of Structural Geology, 20, 1477-1493. KNIPE, RJ. & WHITE, S.H. 1979. Deformation in low grade shear zones in Old Red Sandstone, S.W. Wales. Journal of Structural Geology, 1, 53-66. KNOTT, S.D., BEACH, A., BROCKBANK, P.J., BROWN, J.L., MCCALLUM, IE. & WELBON,A.I. 1996. Spatial and mechanical controls on normal fault populations. Journal of Structural Geology, 18, 359-372. LAJTAI, E.Z. 1969. Mechanics of second order faults and tension gashes. Geological Society of America Bullettin, 80, 2253-2272. LAWN, B.R. & WiLSHAW,T.R. 1975. Fracture of brittle solids. Cambridge University Press, Cambridge. MARCHAL, D., GUIRAUD, M. & RIVES, T. 2003. Geometric and morphologic evolution of normal fault planes and traces from 2D to 4D. Journal of Structural Geology, 25,135-158. MARTEL, S.J., POLLARD, D.D. & SEGALL, P. 1988. Development of simple strike-slip fault zones, Mount Abbott Quadrangle, Sierra Nevada, California. Geological Society of America Bulletin, 100,1451-1465. MAZZOLI, S. & Di Bucci, D. 2003. Critical displacement for normal fault nucleation from en-echelon vein arrays in limestones: a case study from the southern Apennines (Italy). Journal of Structural Geology, 25,1011-1020. MAZZOLI, S., BARKHAM, S., CELLO, G, GAMBINI, R., MATTIONI, L., SHINER, P. & TONDI, E. 2001. Reconstruction of continental margin architecture deformed by the contraction of the Lagonegro Basin, southern Apennines, Italy. Journal of the Geological Society of London, 158, 309-319. McGRATH, A. & DAVISON, 1.1995. Damage zone geometry around fault tips. Journal of Structural Geology, 17,1011-1024. MONACO, C, TORTORICI, L. & PALTRINIERI, W 1998. Structural evolution of the Lucanian Apennines: Journal of Structural Geology, 20, 617-638. NEEDHAM,T., YELDING, G. & Fox, R. 1996. Fault population and prediction using examples from the offshore UK. Journal of Structural Geology, 18, 155-167. NICOLAS, A. & POIRIER, J.-P. 1976. Crystalline plasticity and solid state flow in metamorphic rocks. Wiley, New York. OLSON, J.E. & POLLARD, D.D. 1991. The initiation and growth of en echelon veins. Journal of Structural Geology, 13, 595-608. PEACOCK, D.C.P. & SANDERSON, DJ. 1995. Pull-aparts, shear fractures and pressure solution. Tectonophysics, 241,1-13. POLLARD, D.D., SEGALL, P. & DELANEY, P.T. 1982. Formation and interpretation of dilatant echelon cracks. Geological Society of America Bulletin, 93, 1291-1303. PRICE, NJ. & COSGROVE, J.W 1990. Analysis of geological structures. University Press, Cambridge.

SHEAR ZONE EVOLUTION AND FAULTING RAMSAY, J.G. & GRAHAM, R.H. 1970. Strain variations in shear belts. Canadian Journal of Earth Sciences, 7, 786-813. RAMSAY, J.G. & HUBER, M.I. 1983. The techniques of modern structural geology. Volume 1: Strain analysis. Academic Press, London. RAMSAY, J.G. & HUBER, M.I. 1987. The techniques of modern structural geology. Volume 2: Folds and fractures. Academic Press, London. ROERING, C. 1968. The geometrical significance of natural en echelon crack arrays. Tectonophysics, 5,107-123. ROTHERY, E. 1988. En echelon vein array development in extension and shear. Journal of Structural Geology, 10, 63-71. SCANDONE, P. 1972. Studi di geologia lucana: carta dei terreni della serie calcareo-silico-marnosa e note illustrative. Bollettino della Societd dei Naturalisti in Napoli, 81, 225-300. SCHOLZ, C.H. 1990. The mechanics of earthquakes and faulting. University Press, Cambridge. SEGALL, P. & POLLARD, D.D. 1983. Joint formation in granitic rocks of the Sierra Nevada. Geological Society of America Bullettin, 94, 454-462. SMITH, J.V. 1995. True and apparent geometric variability in en echelon vein arrays. Journal of Structural Geology, 17,1621-1626. SMITH, J.V. 1996. Geometry and kinematics of conver-

373

gent conjugate vein array systems. Journal of Structural Geology, 18,1291-1300. SMITH, J.V. 1997. Initiation of convergent extension fracture vein arrays by displacement of discontinuous fault segments. Journal of Structural Geology, 19,1369-1373. SRIVASTAVA, D.C. 2000. Geometrical classification of conjugate vein arrays. Journal of Structural Geology, 22, 713-722. WALSH, J.J. & WATTERSON, I 1988. Analysis of the relationship between displacements and dimensions of faults. Journal of Structural Geology, 10, 239-247. WALSH, J.J., CHILDS, C., IMBER, I, MANZOCCHI,T, WATTERSON, J. & NELL,P.A.R. 2003. Strain localisation and population changes during fault system growth within the Inner Moray Firth, Northern North Sea. Journal of Structural Geology, 25, 307-315. WILLEMSE, E.J.M., PEACOCK, D.C.P & AYDIN, A. 1997. Nucleation and growth of strike-slip faults in limestones from Somerset, UK. Journal of Structural Geology, 19,1461-1477. YELDING, G, NEEDHAM,T. & JONES, L. 1996. Sampling of fault population using sub-surface data: a review. Journal of Structural Geology, 18, 135-146.

This page intentionally left blank

Index Page numbers in italic refer to figures. Those in bold refer to entries in tables. alpine-type peridotite massifs 11-12 localized deformation 13 Anderson-Byerlee frictional fault mechanics 95 brittle-ductile shear zone evolution and fault initiation at Monte Cugnone, Italy 353-355, 368-372 cross-section 354 examples of structures analysed 356-357 fluid inclusion petrography and microthermometry 363-366 geological setting 354, 355-358 kinematics 366-368 microstructures 363,364 outcrop data 358-363 structural data 360-361 conjugate shearing domain (CSD) 219-221,220 continental crust, metamorphic signature of subduction in Corsica/northern Apennine orogen 321-322, 329-331 structural and metamorphic history of inner Tuscan metamorphic units 329 structural and metamorphic history of Tenda Massif deformation history 324-328,325, 326, 327 geological outline 323-324,324 metamorphic history 328-329,328,329 tectonic setting 322-323,322, 323 crenulation-slip development in NW Ireland 337-338, 350 evidence for strike-slip motion in Central Ox Mountains 344 extensional crenulation cleavages 344 high-strain zones 344 evidence for strike-slip motion in Mayo 339-341 asymmetrical buckle folds 341,342 crenulation-slip morphologies produced by oblique foliation-slip 341 extensional crenulation cleavages 341-342 orientation of D3 dextral shear zone 342-344, 343 predicted angular relationships 341 geological significane of the Fair Head-Clew Bay Line 338-339 isotopic dating of crenulation-slip surfaces 344-346 analytical methods 347 40 Ar/39Ar spot fusion data 347 constraining dextral shear in NW Mayo 347-348 constraining sinistral shear in Central Ox Mountains 348-349 intrusion age of Ox Mountains granodiorite 349 pre-existing age constraints on foliation-slip fabric 346 Rb-Sr geochronology 349 sampling 346 regional geology 338

dauphine twinning and misorientation 39, 54, 58-59 microstructural evolution 54-55 microstructural stability 56-57 misorientation angle distributions 57-58 study details crystallographic misorientation analysis 43-44, 44 grain boundary (mis)orientation analysis 44-45 relationship between crystal slip and boundary orientation 45 relationships between quartz crystal slip systems 46 relationships between specific quartz crystal slip systems 45 summary of Loch Torridon shear zone data 43 study results and interpretations boundary formation 52-54,53 dauphine twinning and twin boundaries 54 microstructure and LPO 47 misorientation analysis 47-52 petrofabric and misorientation analysis 51 SEM/EBSD analysis of dauphine twin microstructures 48-49 (sub)grain boundary formation 55-56 deformation in a complex crustal-scale shear zone 229-230, 246-247 Archaean granitic gneiss 230-231,233 Errabiddy Shear Zone 230,231 evolution of Capricorn Orogen 232 felsic gneiss and Erong Shear Area 234-237,237, 238, 239 kinematic evolution of Errabiddy Shear Zone 244 palaeoproterozoic metasedimentary rocks - Camel Hills 239 deformation in migmatized pelitic schist and gneiss 241-242,242 deformation in psammitic gneiss 239-241,240, 241 structural geometry within Errabiddy Shear Zone 243-244 summary 243 temporal and tectonic evolution of Errabiddy Shear Zone 244-246,245 ductile shearing 161-162,173-174 basement lithology on Sikinos 163-164,164 basement-cover contact on Sikinos 172 geology of cover sequence on Sikinos 162-163 maps 162,163 high-pressure metamorphic imprint in Cycladic basement 172-173 pressure-temperature conditions of metamorphism on Sikinos 170-172,171 feldspar porphyroclast populations, kinematics and strain 265-266 application to western Idaho shear zone 277-278, 279

376

INDEX

assumptions 279-281 feldspar shape preferred orietation data 281 field conclusions 284 implications for shear zone studies 284 location map 275 model conclusions 283-284 quantification of field data 281-282 study results 282-283,283 forward model of clast rotation 266-268 construction of fabric ellipsoid 269 three-dimensional description of clast orientation 268-270 model results 270 fabric ellipsoid versus finite strain ellipsoid 277 orientation of fabric ellipsoid to shear sense 276-277,277 populations of oblate clasts 276 populations of prolate clasts 272-276,273, 274, 275 rotation of single clasts 270 rotations of populations of clasts 271-272 single oblate clasts 270-271 single prolate clasts 270,271 flattening strain 252-253,253, 253 fluid-rock interactions in West Fissure Zone, Chile 141-142 comparison with San Andreas Fault, California 157-158 description of fault rocks 143-147,147 sampling profiles 146 fluid sources and fluid composition 156-157 geochemistry of fault rocks carbon and oxygen isotope relationships in calcite 148-149,149,150 fluid inclusions 147-148,148, 155-156 major elements 150-154,153,156,156 oxygen versus distance relationships in monzodiorite 154-155,154,155 trace elements 149-150,150-154,151,152,153, 156,156 geological setting 142-143 regional map 144-145 regional overview 142 stratigraphy 143 sampling and analytical methods 143 variations in fluid-rock interaction 157 Geographic Information Systems (GIS) applied to deformation patterns 73-76 aeromagnetic dataset 68-70,69-70 combined and integrated dataset combination of all available datasets 72-73, 73 combined directional structural data and shaded total magnetic field map 70-72, 71 vertical gradient of total magnetic field, foliation trends and metamorphic data combined 72, 73 database 65 directional structural dataset 66-68,67 fabric type dataset 68 lithological dataset 65 metamorphic dataset 65-66 proposed indentor model 74, 75 west Greenland case study 64-65,64,66 geostatistical analysis

kriging interpolation 305 variogram computation and interpretation 303-305,304 grain-size sensitive (GSS) flow 32,33, 34,34 granulites, instability and deformation localization in the lower crust 25-26, 35-36 Clarke Head megabreccia 26-27,26 deformation microstructures cherty ultramylonite 30,31 host mylonite 27,28 ultramylonite 27-29,29, 30 experimental procedures 27 interpretation of microstructures and deformation deformation environment 31-32 deformation mechanisms 32-33,33, 34 deformation partitioning and localization 34-35 mechanism transitions 33-34 high-pressure metamorphism 161-162,173-174 basement lithology on Sikinos 163-164,164 basement—cover contact on Sikinos 172 geology of cover sequence on Sikinos 162-163 maps 162,163 high-pressure metamorphic imprint in Cycladic basement 172-173 pressure-temperature conditions of metamorphism on Sikinos 170-172,171 hydrous fluid channelling 161-162,173-174 basement lithology on Sikinos 163-164,164 basement-cover contact on Sikinos 172 geology of cover sequence on Sikinos 162-163 maps 162,163 high-pressure metamorphic imprint in Cycladic basement 172-173 pressure-temperature conditions of metamorphism on Sikinos 170-172,171 indentor tectonics 73-76 aeromagnetic dataset 68-70,69-70 case study in west Greenland 64-65, 64,66 combined and integrated dataset combination of all available datasets 72-73, 73 combined directional structural data and shaded total magnetic field map 70-72, 71 vertical gradient of total magnetic field, foliation trends and metamorphic data combined 72, 73 directional structural dataset 66-68,67 fabric type dataset 68 lithological dataset 65 metamorphic dataset 65-66 proposed indentor model 74, 75 kriging interpolation 305 lattice preferred orientation (LPO) 39,54,58-59 dauphine twinning and (sub)grain boundary formation 55-56 dauphine twinning and microstructural evolution 54-55 dauphine twinning and microstructural stability 56-57 misorientation angle distributions 57-58 shear zone grain size reduction model 55 study details

INDEX crystallographie misorientation analysis 43-44, 44 grain boundary (mis)orientation analysis 44-45 relationship between crystal slip and boundary orientation 45 relationships between quartz crystal slip systems 46 relationships between specific quartz crystal slip systems 45 samples 39-41, 40, 42 SEM/EBSD technique 41-43, 42 summary of Loch Torridon shear zone data 43 study results and interpretations boundary formation 52-54,53 dauphine twinning and twin boundaries 54 microstructure and LPO 47 misorientation analysis 47-52 petrofabric and misorientation analysis 51 SEM/EBSD analysis of dauphine twin microstructures 48-49 low angle normal faults (LANFs) 95-97,105-109 active versus exhumed LANFs Altoberina Fault in Umbria region 97-102, 98 Zuccale Fault in Isle of Elba 102-105,105 regional setting of Northern Apennines 97 lower crust granulites, instability and deformation localization 25-26, 35-36 Clarke Head megabreccia 26-27,26 deformation microstructures cherty ultramylonite 30,31 host mylonite 27,28 ultramylonite 27-29,29, 30 experimental procedures 27 interpretation of microstructures and deformation deformation environment 31-32 deformation mechanisms 32-33,33, 34 deformation partitioning and localization 34-35 mechanism transitions 33-34 microstructure evolution during deformation lower crust granulites 33 mylonitic quartz simple shear zone 39 study details 39-46 study results and interpretation 47-54 misorientation analysis 39, 54, 58-59 crystallographie relationships 43-44 crystal slip systems and boundary orientations 45 quartz 44, 45,46 dauphine twinning and (sub)grain boundary formation 55-56 dauphine twinning and microstructural evolution 54-55 dauphine twinning and microstructural stability 56-57 grain boundary analysis 44-46 microstructure and LPO 47 misorientation angle distributions 57-58 SEM/EBSD technique 41-43,42 localized dauphine microstructures 48-49 shear zone grain size reduction model 55 study results 47-50 boundary formation 52-54,53 dauphine twinning and twin boundaries 54 localized dauphine microstructures 48-49

377

misorientation angle distributions 50 misorientation axis/angle pairs 50-52 petrofabric and misorientation analysis 51 study samples 39-41,40 Nabarro-Herring creep 81,81 ophiolite-type peridotite massifs 11-12 localized deformation 13 orthorhombic fabrics, development within a simple shear sinistral transpression zone 215-216 Arronches gneisses, structural analysis conjugate shearing domain (CSD) 219-221,220 grain-size reduction and deformation mechanisms in fabric formation 221 intermediate sinistral domain (ISD) 221 peralkaline gneisses 216-219 sinistral domain (SD) 221 Arronches Tectonic Unit regional setting 216 structure and metamorphism 216 study area 27 7, 218, 219 conjugate shear band formation 224-226,225 dynamic recrystallization and development of fabric and texture 221-224,224, 225 relative timing of orthorhombic and monoclinic fabric formation 224 partially molten rocks (PMR), application of twophase rheology 79-81, 91 experiments 89-90 development of instabilities 90-91 non-linear effects 90 implications for other two-phase systems 87-89 importance of shear deformation 89 main principles 81-82 schematic map of plastic deformation 81 stress versus strain diagram 82 Theological responses 84-87,86,87 rheology of two-phase materials 82-84,83, 84 pelitic rocks, shear deformation 113,121-124 composition of sheared clays 116,116 srnectic/illite transformation 116-117,117 geological framework of Scorciabuoi Fault (SBF) 113-115,774,775,71(5 grain size analysis 118-119,118 shear zone fabric 119, 719 fine scale analysis 119-121,122,123 mesoscale observations 119,120,121 peridotite mylonites 16,17-19 plane strain 252-253,253, 253 plastic deformation, application of two-phase rheology 79-81, 91 experiments 89-90 development of instabilities 90-91 non-linear effects 90 implications for other two-phase systems 87-89 importance of shear deformation 89 main principles 81-82 schematic map of plastic deformation 81 stress versus strain diagram 82 rheological responses 84-87,86,87 rheology of two-phase materials 82-84,83 plate convergence, shear and fluid flow 127

378

INDEX

comparisons and contrasts between the study sites 135 deformation structures with faults 136-137 fault thickness 136 fault-zone margins 137 hydrogeology 137 internal geometry of fault zones 137 lithological influence of propagation 136 summary of features 138 deformation features Barbados 129-130, 729 Costa Rica 130-131,130 Nankai 131,131 fluid transport Barbados 131-133,132,133,134 Costa Rica 133-134,135 Nankai 134-135,136 implications from study sites for other mega-shear zones 137-138 tectonic settings Barbados 128,128 Costa Rica 128-129,128 Nankai 128,129 rheology of two-phase materials 79-81, 91 experiments 89-90 development of instabilities 90-91 non-linear effects 90 extrapolating the end-members 83 implications for other two-phase systems 87-89 importance of shear deformation 89 main principles 81-82 schematic map of plastic deformation 81 stress versus strain diagram 82 rheological responses 84-87,86,87 thermodynamic considerations 83-84,83, 84 rigid percolation threshold (RPT) 80 shear zone folds 177-178,196-197 Caledonian Moine Nappe, Sutherland 179-180, 180,181 curvilinear fold patterns and evolution 189 fold evolution model 189-194 fold inheritance model 194-195,195,196 hybrid fold model 195-196 fold types 178 sheath folds 178-179 synshearing flow folds 179 Melness folds case study 181-185,183,185 topological relationships between sheath folds and synshearing folds 186-189,188,189, 190-191,192-193,194 transection relationships between sheath folds and synshearing folds 185-186,186,187 shear zones 1, 8 fault controls and shear zone development 4 anastomosis around low-strain augen 5 grain-scale controls 4 lithospheric-scale controls 4 network geometry-scale processes 4-5 histories 7-8 lithosphere deformation and rheology of shear zones 5-6,5

occurrence on different scales 1, 2 partitioning processes in shear zones 6-7, 7 strength, strain rate histories and fault rocks at depth 1-4 deformation regimes and typical fault rocks 3 schematic strength profile through crust and upper mantle 4 strain and deformation history in a syntectonic pluton, Roses granodiorite 307-308, 315-318 displacement versus width diagram 318 main lithological units 308 progressive development of structures in Roses granodiorite 308 elongated enclave of quartz diorite 313 geological setting 309 late brittle fractures 314 leucocratic dykes 313 magmatic fabric and enclaves 308-313 mesoscopic scale structures 311 pre-dyke finite strains 312 qualitative model and structural history 310 shear zones and associated mylonites 313-314 shear strain analysis 317 structural map and strain analysis 316 structure and strain profiles 314 deformation postdating dykes 315 deformation predating dykes 314-315 strain removal within Hercynian Shear Belt, methodology and tectonic implications 287, 287, 300 data processing cleavage trajectory model 292-293,293 domainal distribution of cleavage directions 294-295,294 geostatistical analysis of cleavage directions 291-292,292 geological setting 288 lithologies 288-289 structures 289 kinematic data cleavage and finite strain ellipsoid 289,290, 291 deformation regime 289-291,297 model validation and regional implications 296 at the boundaries 298-299,298 within restored area 296-298,297 restoration of eastern Central Brittany 295-296, 296 strike-slip deformation 250,251 tectonites relative softening fine grained 16-17 medium-to-coarse grained 16 structures and microstructures fine grained 15 medium-to-coarse grained 15 transpression terrane boundaries, geometric and kinematic analysis 201-202, 213 fault rocks 203 fault zone deformation mechanisms 211-212 fault zone kinematics 210-211 framework of the Minas fault system 202-203 location of the Minas fault system 202

INDEX strain partitioning and localization 212-213 structural elements and geometric relationships 204, 206, 207, 208, 209, 210 crenulation cleavage 208-209 faults 209-210 folds 203 foliations 203-205 lineations 205 S-C fabrics 208 shear bands 209 veins 205-208 transpressional high-strain zones, strain and vorticity analysis 249-250, 262 interpretation at study areas Brookneal high-strain zone (BHSZ) 258-259, 259 Spotsylvania high-strain zone (SHSZ) 259-260, 259, 260 strain compatibility 260-261 tectonic significance of Piedmont high-strain zones 261-262 kinematic deformation models 249 kinematic vorticity and vorticity analysis 251-253, 252, 253, 253 transpression and general shear 250-251,251

379

upper mantle shear zones 11-12,19-20, 21 features 12 implications for mantle strength 21 possible tectonite and mylonite shear zones 20 relative softening mechanisms 16 fine grained tectonites 16-17 mantle cross-section at Hilti, Oman 17 medium-to-coarse grained peridotite tectonites 16 olivine deformation mechanism map for Othris, Greece 17 peridotite mylonites 17-19 P-Tgridl9 SEM image of grain boundary alignments for Turon de Tecouere, France 17 SEM images of fine grain production 18 structures and microstructures 12-15 fine grained tectonites 15 localized deformation 13 medium-to-coarse grained tectonites 15 peridotite mylonites 16 photomicrographs 14 variogram computation and interpretation 303-305, 304 vorticity 251-253,252